Embed Size (px)
Citation preview
ED 203 767
TITLE.
INSTITUTION
REPORT NOPUB DATENOTEAVAILABLE FROM
FDRS PRICEDESCRIPTORS
IDENTIFIERS
DOCUMENT RESUME
HE 014 022
Peport of the Graduate Medical Education NationalAdvisory Committee to the Secretary, Department ofHealth and Human Services. Volume III. GeographicDistribution ,Technical Panel.Health Resources Administration (DHHS/PHS),Hyattsville Md. Office of. Graduate MedicalEducation.HPA-B1-653BO246p.: For related documents see HE 014 021-026.Superintendent of Documents, U.S. Government PrintingOffice, Washington, DC 20402 (Stock No.017-022-00726-6, $6.50).
MF01/PC10 Plus Postage.Advisory Committees: *Educational ,Demand: ForeignStudents: Futures (of Society) : 31c.36graphicDistribution: *Graduate Medical Education: *GraduateMedical Students: Labor Market: *Labor Reeds:*Physicians: Program Evaluation: Training MethodsManpower Research
ABSTRACTThe Graduate Medical Education National Advisory
Committee's (GMENAC) report on the present and future supply andrequirement's of physicians by specialty and geographic location isprovided. Part I presents a listing of Panel recommendations andoptions for addressing the problem, while Part II presents anintroduction to this report. Part III summarizes Charge 1--adescription of the underlying problem, a literature review ofphysician location factors, a representation of data by specialty bycounty, and recommendations for further efforts at data collection.Part IV summarizes Charaes 2 and 3, detailing empirical evidence ofvariations in per capita expenditures and rates of surgicalprocedures. It also reviews the literature on procedure ratevariation studies, providing implications of the variations for-manpower policy. Criteria for acceptable levels of equity and accessusing the indicators of physician to population ratios and traveltimes to service are presented-in 'Part V. Problems of datautilization, as well as recommenOtions for the level of geographicanalysis that should be employed,' are also expanded on in Part V.Part VI summarizes Charges 4, 5, and 6 linking the work of theGeographic Distribution Panel withthat of the Financing,Nonphysician Provider, and Educational Environment Panels, andpresents a taxonomy of "strategiesp\or programs that might beimplemented to address distributional and access problems. Strengthsand weaknesses of the past, present,' and potential programs arediscussed, as well as extensive documentation on each of themechanisms.,Among the 31 recommendations made by the Panel are:serious attention should be given to making available to physicianstheir utilization rate experiences relative to the norms of otherphysicians practicing in their immediate region, area, or in thenation: economic incentives and/or the provision of higher paymentlevels for service as an inducement for physicians to practice inunderserved areas, should be explored: and the effectiveness ofaovernment loan and scholarship programs should be cataloged andevaluated. A review of the literature on physician distribution isappended. (Author/LC)
Report of theGi'duate Medical Education National Advisory Committee
to the Secretary, Department of Health and Human Services
Volume IllGeographic Distribution
Technical Pane/rf,
N. '4U.S. DEPARTMENT OF EDUCATION
%,,,NATIONAL INSTITUTE OF EDUCATION
, EDUCATIONAL RESOURCES INFORMATION
..tr is....,
7 CENTER IERICIThis document has been reproduced asreceived from the person or organization
(
orrurnatrnq It.
'Ai',.../
Minor changes have been made to improve
***N-...
reproduction quality.
Points of view or 01,1,11011S Staled 111 this docu-
ment do not necessarily represent official NIE
position or policy.
1/ 1'4/7/ kf 7:N 7. (.0-t If .1, rqust,tei rivIL t s
vote,-11., cf ,,p1 tr. PAP, it,
The Report of the Graduate Medical Education National Advisory Committee tothe Secretary, Department of Health and Human Services, consists ofseven volumes:
Volume I GMENAC Sum Mary Report
Volume II Modeling, Research, and Data Technical Panel
Volume III Geographic Distribution Technical Panul
Volume IV Financing Technical Panel
Volume V Educational Environment Technical Panel
Volume VI Nonphysician Health Care Providers Technical Panel
Volume VII GMENAC Members' Commentaries and Appendix
For sale by the Superintendent of 1)ocurnents, U.S. Government Pr Intim: (Mire, Washington, D.C. 20402
Report of theGraduate Medical Education National Advisory Committee
9
to the Secretary, Department of Health and Human Services
Volume IIIGeographic DistributionTechnical Panel
0/2 02_ 2_.
U.S DEPARTMENT OFHEALTH AND HUMAN SERVICESPubic Health ServiceHealth Resources AdrncustrabonOffice of Graduate Madicai Education
OHMS Publicaton Na (HRA)
-2
September 30, 1980
The HonorableSecretaryDepartment ofWashington, D.
Patricia Roberts Harris
Health and Human ServicesC. 20201
Dear Madam Secretary:
The attached Report of the Graduate Medical Education National AdvisoryCommittee (GMENAC) is in fulfillment of the Committee's responsibilitiesunder th Charters of April 20, 1976, and March 6, 1980.
The charge of the Committee was to advise the Secretary on the number ofphysicians required in each specialty to bring supply and requirementsinto balance, methods to improve the geographic distribution ofphysicians, and mechanisms to finance graduate medical education.
GMENAC significantly advanced health manpower planning in direct andindirect ways.
GMENAC introduced new scientific methodology: Two new mathematicalmodels were developed to estimae physician supply and requirements.
GMENAC refined the data bases; figures for estimating the supply ofpractitioners in every specialty and subspecialty from thedistribution of first-year residency positions have been developed.
GMENAC integrated the estimates of supply and requirements forphysicians with nurse practitioners, physicianassistants, and nursemidwives.
GMENAC introduced new concepts to clarify assessment of thegeographic distribution of physicians and services; standards areproposed for desic.ating areas as adequately served or underservedbased on the unique habits of the people in the area.
GMENAC recommends that medical service revenues continue to providethe major source of funds to support graduate medical education.
GMENAC has initiated a collaboration between the private sector andthe Government; the unique expertise of each achieves a level ofcomprehensiveness in health manpower planning not previouslyexperienced.
September 30, 1980Secretary HarrisPage 2
GMENAC estimates a surplus of 70,000 physicians by 1990. Mostspecialties will have surpluses, text a feW will have shortages. Abalance by 1990 cannot be achieved. Until supply and requirementsreach a balance in the 1990s, GMENAC recommends that the surplus bepartially absorbed by expansion of residency training positions ingeneral/family practice, general pediatrics, and general, internalmedicine.
Recommendations are directed at achieving five manpower goals:
1. To achieve a balance between supply and requirements ofphysicians in 90s, while assuring that programs to increase therepresentation of minority groups in medicine are advanced byprograms to broaden the applicant pool with respect tosocio-economic status, age; sex, and race;
2. to integrate manpower planning of physicians and nonphysicianproviders when their services are needed, and to facilitate thefunction of nonphysician providers;
3. to achieve a better geographic distribution of physicians and toestablish improved mechanisms for assessing the adequacy ofhealth services in small areas;
4. to improve specialty and geographic distribution of physiciansthrough financing mechanisms for medical education, graduatemedical education, and practice, and
5.. to support research for the next phases of health manpowerplanning.
The. Committee unanimously recommends the immediate Establishment of asuccessor to GMENAC. Its establishment is essential to theimplementation of the manpower goals and recommendations in the Report.The full GMENAC methodology must be applied to the six specialties whichhave not been analyzed. The requirements estimates for each of thespecialties and subspecialties must be tested, monitored, and reassessedan a continuing basis. Important studies on financing, geography, andnonphysician providers should be undertaken.
6
September 30, 1980Secretary HarrisPage 3
The collaborative working relationship between the private sACtor and theGovernment facilitated a congruence of interest in planning and inimplementing improvements to best meet the needs of the Nation. Themomentum of this collaboration should be continued without interruption.
Respectfully submitted,
Alvin R. Tarlov, M.D.ChairmanGraduate Medical EducationNational Advisory Committee
For the Committee
Enclosure: Volumes IVII
GRADUATE MEDICAL EDUCATION NATIONAL ADVISORY COMMITTEE
GEOGRAPHIC DISTRIBUTION TECHNICAL PANEL
Conveners
ZUBKOFP, Michael, Ph.D.Professor and ChairmanDepartment of Family and
Community MedicineDartmouth Medical SchoolProfessor of EconomicsAmos Tuck School of Business
AdministrationDartmouth CollegeHanover, New Hampshire
ANDERHOLM, Lynn, M.S.Research AssociateDartmouth Medical School
and Staff ConsultantOffice of Graduate Medical
Education
BUTTER, Irene, Ph.D.Professor of Health Planning
and AdministrationUniversity of MichiganAnn Arbor, Michigan
GREENE, Richard J., M.D., Ph.D.Director, Health Services
Research and DevelopmentVeterans AdministrationWashington, D.C.
WENNBERG, John E., M.D.Professor of EpidemiologyDartmouth Medical SchoolHanover, New Hampshire
CARBECK, Robert B., M.D.Executive Vice PresidentCatherine McAuley HealthCenter
Ann Arbor, Michigan
MORGAN, Beverly C., M.D.Professor and ChairmanDepartment of PediatricsUniversity of WashingtonSeattle, Washington
Panelists
LAPENAS, Coralea, J.D.Instructor of Family and
Community MedicineDartmouth Medical SchoolHanover, New Hampshire
MODDERMAN, Melvin, M.H.A., M.B.A.Doctoral CandidateHealth Services AdministrationGeorge Washington UniversityWashington, D.C.
VIETS, Hilary P.,/M.P.H.Research AssociateDepartment of Health Planning
and AdministrationUniversity of MichiganAnn Arbor, Michigan
Staff
WESTCOTT, Mary L., M.A.Program AnalystOffice of Graduate Medical
Education-Health Resources AdministrationDepartment of Health and Human
ServicesHyattsville, Maryland
CONTENTS
PageLETTER OF TRANSMITTALROSTER OF GEOGRAPHIC DISTRIBUTION TECHNICAL PANEL ii
ROSTER OF MEMBERS iv
I. SUMMARY OF GEOGRAPHIC PANEL RECOMMENDATIONS 1
II. INTRODUCTION AND OVERVIEW OF CHARGESTO THE GEOGRAPHIC PANEL 6
III. PHYSICIAN DISTRIBUTION: AN ANALYSISOF CAUSES
DATA ON DISTRIBUTION OF PHYSICIANS IN THE U.S.LITERATURE REVIEW OF REASONS FOR VARIATIONS INDISTRIBUTION 42RECOMMENDATIONS 45
IV. VARIATIONS IN POPULATION-BASED USE RATES: IMPLICATIONSFOR MANPOWER POLICY, EMPIRICAL EVIDENCE OF VARIATIONS INUSE RATES AND IMPLICATIONS OF VARIATIONS ON ACCESS TO ANDDISTRIBUTION OF HEALTH MANPOWER 46RECOMMENDATIONS 56
V. CRITERIA FOR EQUITY AND ACCESS 57,
PHYSICIAN TO POPULATION RATIOS 64TIME TO SERVICE CONCEPT 70RECOMMENDATIONS 74
VI. TAXONOMY OF STRATEGIES TO IMPROVE THE GEOGRAPHICDISTRIBUTION OF PHYSICIAN SERVICES 7;
References 109Appendix - A Review of the Literature on Physician Distribution . 117
vi 9
TABLES
Table Page
1 Geographic Distribution Panel Final Charges 7
2 Selected Physician Distribution Statistics bySpecialty, 1975 10
3 Distribution of adult practitioner to populationratios for counties with less than 10,000 population . 14
4 1990 Estimated Manpower Requirements for SevenSurgical Specialties for Surgical Care 49
5 Summary Table of Significant Data Items 63
6 Range of Physician Specialty Requirements Expressedas Ratios of Physicians per 100,000 65
7 Summary Table: Range of Selected "RequirementsRatios" by Occupation
\
68
8 Range of Selected Requirements Ratios: FamilyPractice, General Surgery, Internal Medicine,Obstetrics/Gynecology 69C
9 Recommended Time-to-Service Standards by five majortypes of care and source 73
10 Practice Location Preference of Students and ResidentPreceptees by Location of Preceptor Practice 82
11 Distribution f Preceptors by Specialty According toRegion in w ich program was located 83
12 Distribution f Active Preceptors by Specialty, Formof Practice and Medical Shortage Designation ofPractice Si es 84
13 GP and Specie ist fees in California 104
14 GP and Speci list fees in Georgia 106
vii1j
FIGURESNumber
Histogram for U.S. of all counties: Total allphysicians of all counties
Page
1
202 Histogram for U.S. Adult Medical Care 212.1 Histogram of Californian Adult Medical Care 223 Histogram for U.S. of all counties: General Surgery 233.1 Histogram of California of all counties: General Surgery 243.2 Histogram of Alabama of all counties: General Surgery 253.3 Histogram of Colorado of all counties: General Surgery 264 Histogram for U.S. of all counties: Obstetrics 275 Histogram for U.S. of all counties: Pediatrics 285.1 Histogram of California of all counties: Pediatrics 295.2 Histogram of Arkansas of all counties: Pediatrics 306 Histogram for U.S. of, all counties:. Psychiatry 317 Histogram for U.S. of all counties: Orthopedic Surgery 328 Histogram for U.S. of all counties: Ophthalmology 339 Histogram for U.S. of all counties: Urology. 3410 Histogram for U.S. of all,counties: Otolaryngology 3511 Histogram for U.S. of all counties: Dermatology 3612 Histogram for U.S. of all counties: Neurosurgery 3712.1 Histogra of California of all counties: Neurosurgery 3812.2 Histogr of Arkansas of all counties: Neurosurgery 3913 Histogram for U.S. of all counties: Plastic Surgery 4014 Histogram for U.S. of all counties: Thoracic Surgery 41
(
viii
1
RECOMMENDATIONS OF THE
GEOGRAPHIC DISTRIBUTION TECHNICAL PANEL
I. Monitoring and Evaluating the Distribution
of Physicians and Services
II, Future Research
?. netermination of
medical market areas
3. Population based data
5. Dissemination of
utilization rates to
physicians
7. Outcome Studies
9, Minimum
acceptable ratios
11. Health manpower
shortage area
criteria
2. Data systems
4. Collection of
utilization rate
variation data
61 Dissemination
of aggregate
statistics
8. Comparison of
manpower estimates
ti
10. Minimum
acceptable
travel times
12, Factors affecting
location choice
13. Economic factors
14. Evaluation of programs
15. Selective admissions
16. Economic incentives
17, Demonstration
projects-reimbursement
III,Undergraduate/Graduate Medical
Medical Education Institutions
18. Role model development and preceptorships
19. Family practice training programs
20. Residency/education experiences
21. Nonphysician,proyider training programs
22. Decentralized medical tducation
23. Specialty s cieties/training programs
III-IV Action Oriented Strategies for Change
IV, Medical
Students/ Residents V. Practicing Physicians
24. National Health Service Corps 28. RR, RR and HURA evaluations
25, Loans and scholarOips 29. Group practices
26. Area Health Education Centers 30. Geographic shortage area differentials
27. Loan forgiveness 31. Reimbursement for primary care
(Federal and State)
I. SUMMARY OF GEOGRAPHIC PANEL RECOMMENDATIONS
RECOMMENDATION 1: The functional medical service areas, specialty byspecialty, are recommended as the geographic unit for assessingavailability of ph sician services. The Graduate Medical EducationNational Advisory Committee CTIGIZIDii.so recommends that physicianmarket areas by specialty be determined empirically based on patientorigin data derived from such information as discharge and claims data,until such time as total enumeration of physician services is possible,and that the resulting areas be compared to those previously determinedby specialty societies. The specialties of dermatology, obstetrics-gynecology, orthopedics, and neurosurgery have developed methods fordetermining the market areas for their respective specialties based onzip codes, economic service areas, and time-to-service concepts.(See pp. 18-19, Chapter III)
RECOMMENDATION 2: CMENAC supports the evaluation of alternative datasystems for monitoring of geographic distribution of providers. (See pp.
18-19, Chapter III)
RECOMMENDATION 3: GMENAC urges the use of small area populationbased data on the availability, requirements for and utilization rates of
hospital and physician services as a manpower planning tool. (See
Chapter 7V)
RECOMMENDATION 4: GMENAC urges that the ranges of variations in theutilization of specific procedures and services among service populationsand communities be collected and analyzed (includin: communities withdiffering financing and organizational arrangements for the delivery ofmedical care services). (See Chapter IV)
RECOMMENDATION 5: Serious attention should be given to makingavailable to physicians their utilization rate experiences relative tothe norms of other physicians practicing in their immediate area, region,or in the Nation. (See p. 53, Chapter IV)
RECOMMENDATION 6: Serious attention should be given to the voluntarycollection and dissemination' for analytical purposes of aggregatestatistics relative to utilization rates in various service areas.(See pp. 52-53, Chapter IV)
RECOMMENDATION 7: GMENAC encourages the support of efforts withinthe profession to assess the outcomes of common medical and surgicalpractices which exhibit high variation across communities as an importantstep for establishing the long-range requirements for suppliers ofmedical services in the United States. (See,pp. 51-53, Chapter IV)
1
RECOMMENDATION 8: Future health manpower planning groups shouldcompare manpower estimates (whether derived as a "need-based", "demand-based," or "requirements-based" model) against empirical estimatesselected from areas in the United States which exhibit high and lowutilization patterns. (See pp. 47-51 and pp. 53-56, Chapter IV),
RECOMMENDATION 9: GMENAC recommends that five basic types of healthcare services should be available within some minimum time/accessstandards: Adult medical care, child care, obstetrical services,surgical services, and emergency services. In order to monitor thegeographic distribution of tpiycis,Is GMENAC recommends that a minimumacceptable physician-to-population ratio for all areas in the U.S. beestablished. It is/recommended that 50 percent of the GMENAC ModelingPanel ratio of physician specialists per 100,000 for 1990 be establishedas the minimum acce table ratio for all areas. (See pp. 62-65, Chapter V)
RECOMMENDATION 10: GMENAC recommends maximum travel times of 30minutes for emergency medical care, 30 minutes for adult medical care, 30minutes for child medical care, 45 minutes for obstetrical care, and 90minutes for surgical care services for 95 percent of the population in1990, recognizing that unusual circumstances may arise which make thesetravel times impossible to achieve for all areas. (See pp. 65 and 71-76,Chapter V)
RECOMMENDATION 11: GMENAC recommends that the definition of healthmanpower shortage area include minimum physician/population ratios and aminimum travel time to service for general surgery, emergency medicalservices, and obstetrical services. (See pp. 73 and 76, Chapter V)
RECOMMENDATION 12: Incoul. . information exists on the direction ofcausation of many of the factof. ,.fecting physician location.Additional research is needed to z.tudy (1) how background factors such associodemographic factors affect specialty and location choices and theinteraction between specialty and location choices and (2) what factorsaffect permanent location choices in underserved/rural areas.(See pp. 43-46, Chapter IV)
RECOMMENDATION 13: Since the role of economic factors in locationchoice is not clear, attempts should be made to improve methodologies todetermine this role and to gather data on previously nonquantifiabletopics such as income as a motivating force in specialty or locationchoices. (See pp. 43-46, Chapter III)
RECOMMENDATION 14: Those strategies which GMENAC deemed mostpromising, such as preceptprships and tax incentives, and those which aremost amenable to evaluation efforts, should be evaluated morevigorously. (See pp. 77-110, Chapter VI)
\:
2
J4.
RECOMMENDATION 15: There is some evidence that selective admissions
Lolicies may improve the geographic distribution of physicians. Anationally mandated alteration in admissions policies is not recommendedat this time; further study into the location decisions of students with
particular ethnic or sociodemographic characteristics is rtcommended.(See pp. 77-78, Chapter VI)
RECOMMENDATION 16: Economic incentives (such as tax credits anddeductions) and/or the provision of higher payment levels for services asan inducement for physicians to practice in underserved areas should be
explored. (See pp. 78-80, Chapter VI)
RECOMMENDATION 17: Demonstration projects should be developed and
evaluated to determine the impact of differential rates of reimbursementfor technology-intensive versus time-intensive (counseling, patienteducation) services upon the geographic distribution of physicians andservices. (See pp. 80-81, Chapter VI)
RECOMMENDATION 18: It is recommended that practicing physicians andfaculty convey to students that the practice of medicine can be deliveredin a variety of geographic settings, including both rural and urban shor-tage areas. As a means of accomplishing this, urban and rural preceptor-slaps for medical students should be continued and expanded in schoolswith an interest in monitoring such programs. (See pp. 81-86, Chapter VI)
RECOMMENDATION 19: Given the geographic distribution patterns offamily practitioners, graduate medical education programs in familymedicine should continue to be supported as a strategy to increaseprimary care services in certain geographic areas of underservice. (See
pp. 87-88, Chapter VI)
RECOMMENDATION 20: Incentives should be created to broaden residencyeducation experiences to encompass training in underserved areas, pro-vided the appropriate resources are available and standards of educationof the relevant accrediting body are met. (See pp. 88-89, Chapter VI)
RECOMMENDATION 21: Data suggest that nonphysician health careproviders favorably affect the distribution of medical services by their
tendency to select shortage area locations more frequently than is the
case with physicians. It is recommended that nonphysician health careprovider training programs should continue to be supported for this
reason. (See pp. 89-91, Chapter VI)
RECOMMENDATION 22: Decentralized medical education programs such asWAMI (in Washington, Alaska, Montana, and Idaho) and WICHE (WesternInterstate Commission for Higher education) were'developed to coordinatemedical education and placement programs:in a relatively isolated andsparsel ovulated re ion. These t pes of programs have been effectiveand attention should be given to their replicablllty. See pp.
Chapter VI)
RECOMMENDATION 23: GMENAC encourages the medical profession, throughits training program directors and various specialty societies, in makingdecisions as to residency training programs, to consider, in addition tothe guality of residency programs, the aggregate number of programs,their size and the eo ra hic distribution of their :raduates to bettermeet national and regional needs. (See pp. 92-93, Chapter VI)
RECOMMENDATION 24: The National Health Service Corps (NHSC) and theNHSC Scholarshi. Program for increasin the availabilit of primary carephysician services in designated health manpower shortage areas impactfavorably on the geographic distribution of physicians; therefore, theNHSC and the NHSC Scholarship Program should continue to be supported.(See pp. 93-94, Chapter VI) .
RECOMMENDATION 25: Government sponsored loan and scholarshipprograms should be catalogued and evaluated to determine theireffectiveness in improving the geographic distribution of physicians.(See pp. 94-95, Chapter VI).
RECOMMENDATION 26: Despite limited evaluation, there is evidencethat several Area Health Education Center (AHEC) models are effective ininducing physicians to practice in underserved areas and/or to practiceprimary care. These types of AHECs should receive continued support.
\ (See pp. 95-97, Chapter VI)
RECOMMENDATION 27: Loan forgiveness programs modeled after thoseWhich have been successful should be used as a strategy for attractingphysicians into underserved areas. Mae pp. 97-99, Chapter VI)
RECOMMENDATION 28: Comprehensive evaluations of programs to recruitand retain providers in underserved areas (e.g., Rural Health Initiative,Rural Health Clinics, Health Underserved Rural Area Program) should beperformed after a reasonable period of time. Continued funding of theseprograms should be contingent upon a positive evaluation of theireffectiveness. (See pp. 99-100, Chapter VI)
RECOMMENDATION 29: Programs that foster or support group practicearrangements ;n rural areas, coupled with the appropriate communicationand transportation networks, should be developed or established on anexperimental basis as a means of attracting physicians to ruralcommun..ties. If these delivery modes prove to be successful indelivering care to underserved areas, start-up funding should beencouraged for new programs. (See pp. 100-103, Chapter VI)
RECOMMENDATION 30: Discontinuation of geographic differentials inpayment levels by third-party payors when this is in excess ofdifferences in costs of delivering those services as a means ofinfluencing geographic distribution should be the subject of futvreresearch. Present reimbursement systems (Federal, State and L, vete)
4
"/-4-
tend to sustain historical saifferences in fees and incomes amonggeographic areas and thus provide incentives for physicians to locate inhigh income communities within metropolitan areas. (See pp. 103105,Chapter VI)
RECOMMENDATION 31: GMENAC recommends that all physicians, both thosein,primaty care specialties and those in nonprimary care specialties, bereimbursed at the same payment level for the same primary care services.(See pp. 105-110, Chapter VI)
5
II. INTRODUCTION AND OVERVIEW OF CHARGES TO THE GEOGRAPHIC PANEL
The Graduate Medical Education National Advisory Committee (GMENAC)was created in April 1976 as an advisory body to the Secretary of Health,Education, and Welfare. The Charter of GMENAC mandated thatrecommendations be made by September 1980 on the present and futuresupply and requirements of physicians by specialty and geographiclocation. Despite a 31 percent increase in the aggregate supply ofactive physicians over the 1965-75 decade, the specialty and geographicdistribution of physicians and the assurance of equal access to healthservices by all segments of the population has caused continuingconcern. The overriding questions of how many specialists andgeneralists are needed for a given size population in a given geographicarea remain highly controversial. The Geographic Distribution TechnicalPanel of GMENAC was established to address these concerns.
The central goal of the Geographic Panel since its inception in March1979 was to develop options and recommendations to assist in reducing theunequal accessibility and availability of medical care services amongstcommunities. The Geographic Panel has looked at a wide range of issuesrelating to geographic considerations, including the relative importanceof reimbursement and life style, the different problems of rural andurban areas, the effect of geographic origin on students' decision tolocate, and procedure rate variations by small arc-is.
The GeographiApanel at its initial meeting proposed to report oneight topics, listed as "Charges" to the Panel (Table 1). The chargeswere approved in. May 1979 by the LIMENAC as a whole, and work proceeded onthe development of documentation in each area. Beginning with the firstcharge, which was aimed at identifying the underlying causes of thegeographic problem, the Panel described what currently existed in thearea of physician distribution, what should exist in terms of acceptablelevels of variations in manpower, and whf,, could be done to redress anyimbalances. The last charge (No. 8) presented a series of options and acompendium of recommendations on the opher areas (See V1).
The Interim Report of GMENAC described the distribution of totalphysicians on three levels of geography--the Health Services Area (OSA),'the State, and the county. Only data on selected States were presented.The Geographic Panel has expanded on this presentation, providing data on21 specialties by each county, RSA, and State in the United States forfull -time equivalent physicians. Also analyzed and debated were many ofthe problems of utilizing data on these levels of aggregation.
6
Table 1
GEOGRAPHIC DISTRIBUTION PANEL FINAL CHARGES
1. Identify the underlying causes of the unequal distribution(availability and utilization) of physicians in the United States.Describe and analyze data on rates of physicians per capita in 22specialties by the smallest geographic area possible (e.g., county,health service area, zip code).
2. Describe the variations in per capita expenditures for and rages ofmedical/surgical services among selected communities; relate thesevariations to local physician distribution and project theimplications for future manpower needs.
3. Given that local physician supply will always vary, what criteria canbe used to indicate acceptable levels of variations in localphysician supply, per capita rates of medical and surgical use, andper capita expenditures for medical services.
4. Analyze the potential impact and record to date of programs such asthe National Health Service Corps, Area Health Education Centers, and
Rural Health Clinics. in terms of eliminating gross distributionalproblems.
5. In-conjunction with the Educational Environment Panel, identify howthe sociodemographic characit4ristics and educational experiences of
students affect the decision of practice location.
6. In conjunction with the Financing Panel, determine if financialincentives for practitioner and training institutions might be usedto effect more equitably distributed medical services.
7. In conjunction with the Modeling Panel, identify distributivecharacteristics for each of the 22 specialties and describe theimpact of these characteristics on the recommendations derived fromeach specialty model.
8. Given the existing and projected number of providers by specialty,recommend how issues of geographic over- and undersupply could becorrected.
7
The organization of the Panel's remrt /is a follows: Part I presentsa listing of Panel recommendations and options for addressing theproblem, while Part II presents an Introduction to this report. Part IIIsummarizes Charge 17a description of the underlying problem,, aliterature review of physician location factors, a presentation of databy specialty by county, and recommendations for further efforts at datacollection.
Part IV summarizes Charges 2 and 3, detailing empirical evidence ofvariations in per capita expenditures and rates of surgical procedures.It also reviews the literature on procedure rate variation studies,providing implications of the variations for manpower policy. Criteriafor acceptable levels of equity and access using the indicators ofphysician to population ratios and travel times to service are presentedin Part V. Problems of data utilization, as well as recommendations forthe level of geographic analysis that should be employed, are alsoexpanded on in Part V. Part VI summarizes Charges 4, 5, and 6 linkingthe work of the Geogtaphic Distribution Panel with that of the Financing,Nonphysician Provider, and Educational Environment Panels, and presents ataxonomy of "strategies" or programs which might be implemented toaddress distributional and access problems. Strengths and weaknesses ofpast, present, and potential programs are discussed, as well as extensivedocumentation on each of the mechanisms.
The Appendix to this Report is a review of the literature onphysician distribution.
8
III. PHYSICIAN DISTRIBUTION: AN ANALYSIS OF ITS CAUSES
Under Charge No. 1, which was aimed at "identifying the causes of theunequal distribution of physicians" and describing and analyzing data on
physicians, the Geographic Panel addressed the question of geographicdisparity, provided an understanding of the nature, causes, and extent of
the problem, and identified reasons for physician location decisions.This charge was addressed in two parts: A presentation of data on thecurrent distribution of physicians and a literature review of reasons forvariations in the distribution.
DATA ON THE DISTRIBUTION OF PHYSICIANS IN THE U.S.
It was necessary to first investigate the current levels of physician
specialty distribution by the geographic area closest to the truephysician market area in order to determine where physicians by specialtywere located. This data can assist in an assessment of what actionsmould be necessary to address the problem of unequal distribution. The
data presented are for 1975, and do not indicate trends over time.
Data on Specialty Distribution
Data are presented in this section for the total number of non-Federal physicians in the U.S. and for 18 specific specialtyclassifications. The specialty classifications (listed in Table 2) aregrouped according to those of the Modeling Panel Delphi Groups ofGMENAC. Data are analyzed for the District of Columbia and for eachcounty in the U.S. in addition to State and national aggregates.Physician data were self-reported for 1c75 as recorded on the AmericanMedical Association (AMA) Physician master file tape and include onlyprivate practice physicians. County population estimates are those ofBureau of the Census Series P-26 for 1975 as recorded on the Bureau of
Health Profession's Area Resource File.
The figures do not necessarily represent the number of personsengaged in the practice of the specialty. A physician indicating more
than one practice specialty will be counted in more than one specialty
using the appropriate full-time equivalent (FTE) figures. The resident
population figures represent all residents of an area, including military
and institutional populations.
9
Table 2
SELECTED PHYSICIAN DISTRIBUTION STATISTICS BY SPECIALTY, 1975
Number of
Physicians
Specialty in U.S.
Physicians
per h, 000
Population
in U.S.
Number of
Counties
Without
Physician
Percent of
All Counties
(3084)
County Ratios per 100,000 Population
50th 90th 100th
Percentile Percentile Percentile
All Specialties 307,155 144.2 167 5.4% 52.8 122.8 1299.4Adult Medicine 109,615 51.4 180 5.8% 32.9 57.6 485.9
General Practice/
Family Practice 49,521 23.2 208 6.7% 25.2 47.4 136.4Internal Medicine 48,459 22.7 1633 53.02
, 0.0 15,7 380.8Cardiovascular
Diseases 6,381 3.0 2377. 77.1% 0.0 3,0 28.2Gastroenterology 1,945 .9 2730 88.5% 0.0 .5 25.0Pulmonary Disease 1,885 .9 2700 87.5% 0.0 .7 43.5 I
Allergy 1,424 .7 2703 87.6% 0.0 .6 55.61-.1
General Surgery 31,640 14.8 1202 39.0% 5.8 16.1 108.50 Obstetrics 21,177 9.9 1881 61.1% 0.0 10,1 55.9Pediatrics 20,399 9.5 1978 64.1% 0.0 8.2 86.0Psychiatry
Orthopedic
19,525, 9J... 2144 69.5% 0.0 6.8 104.2
Surgery 10,666 5.0 2218 71.9% 0.0 6.1 74.6Ophthalmology 8,952 4.2 2079 67.4% 0,0 5.2 53.0Urology 6,092 2.8 2256 73.2% 0.0 3.7 35.8Otolaryngology 4,791 2.2 2366 76.7% 0.0 2.7 58.8Dermatology 3,372 1,5 2504 81,22 0.0 1.1 52.9Neurosurgery 2,886 1.3 2675 86.7% 0.0 1,2 23.5Plastic Surgery 2,066 .9 2750 89.2% 0.0 .6 44.4Thoracic Surgery 21044 .9 2668 86.5% 0.0 .9 19.6
ri
The American Medical Association Physician ClassificationsGrouped to Correspond with GMENAC Delphi Panel is as follows:
GMENAC Grouping Corresponding AMA Specialties
Adult Medicine
Dermatology
General Surgery
Neurosurgery
Obstetrics/Gynecology
Ophthalmology
Orthopedic Surgery
Otolaryngology
PediatricsPediatric cardiologyPediatric allergy
Plastic Surgery
Psychiatry
Thoracic Surgery
Urology
General practice (includes familypractice)Internal MedicineCardiovascular diseaseGastroenterologyPulmonary diseaseAllergy
Dermatology
General Surgery
Neurosurgery
Obstetrics/Gynecology
Ophthalmology
Orthopedic Surgery
Otolaryngology
Pediatrics
Plastic Surgery
r* PsychiatryChild psychiatry
Thoracic Surgery
Urology
For each of the 14 histograms of all counties in the U.S. on pages21-42, the 3,084 counties of 50 States plus the District of Columbia aregrouped according to their physician /population ratios for each specialtyso that approximately 10 percent of the counties cluster into eachgroup.- (See Figures 1-14) The widths of the intervals determined foreach specialty were used to create histograms of counties in each State.As a consequence, comparisons of the-gistribution of county physician/population ratios within a State to the distribution of all counties inthe United States are facilitated.
11
The vertical axis illustrates the percent of all counties in theState or Nation, whichever, the case may be, whose physician/kopulationratios for the specialty in question fall within a particular interval.The exact percent is listed as the first data item in each decile groupalong the horizontal axis. The next two numbers under the percent figureare the width of the interval.
The fourth number indicates the number of counties whose ratios fall,'within the interval. It should be noted that the horizontal axis is notto scale and that the range is FTE physicians per 100,000 population.
Data are presented for full-time equivalent (FTE) physicians.Responses to the opportunity to list up to three specialties and thenumber of hours per week worked in each for the AMA survey are the basisof the FTE counts which consider both differences in hours worked andspecialty. The mean number of hours per week worked by all physicians inthe. U.S., approximately 60, are used to calculate FTEs for each specialtythat a physician lists.* Then'the number of FTEs in each specialty for acounty is summed and rounded to the nearest integer. State and nationaltotals are the summation of rounded county figures. It should be notedthat the AMA derived FTE counts do not differ substantially from simplehead counts of physicians by specialty.**
Histograms were created for all physicians and for 13 specialties ofphysicians organized according to the GMENAC Modeling Panel's DelphiGroups. County-specific histograms on active non-Federal physicians were,generated from the 1975 AMA master file of self-reported data. Anyphysician who indicated he or she was not retired, on sabbatical, etc.,was considered an active physician.
In 1975 there were 307,155 active, non - Federal physicians of allspecialties in the U.S. or 144.2 physicians per 100,000 population (SeeTable 2 - Selected Statistics by Specialty). Figure 1 illustrates thedisparate geographic distribution pattern of physicians of all specialtytypes. The median county ratio was 52.8 physicians per 100,000population indicating that half of the counties had ratios higher andhalf had ratios lower than 52.8. Ten percent of counties had ratiosbetween 122.8 and 129.4-physicians per 100,000 population. It should be
* While significantly higher and different. from the Mendenhall datapresented at the July 20, 1979 GMENAC meeting, these numbers should givefigures with correct relative orders of magnitude.
** Abt Associates' analysis of differences among head counts, distributedhead counts, and FTE figures by county using Chi-square analysis showsonly about 2 percent of U.S. counties (about 62) having significant (95percent confidence interval) differences among the counts. (1979)
12
rN
noted that the intervals necessary to cluster ten percent of U.S.counties' physician/population ratios grow increasingly, larger after the
50th percentile, indicating a broad right tail distribution.
The 109,615 FTE physicians practicing adult medicine are slightly
more than one third of all physicians and are distributed similarly to
physicians of all specialty types. Although there are 51.4 FTE adultmedicine physicians per 100,000 population in the Nation, more than 80percent of all U.S. counties have lower physician/population ratios. The
adult medicine histogram, Figure 2, shows a median county physician/population ratio of 32.9 FTE physicians. per 100,000. Further, 90 percent
of all counties had less than 57.6 FTE adult medicine practitioners per100,000. The highest county ratio was 458.9, and there were 35 counties
with more than 100 FTE adult medical practitioners per 100,000 population.
At the other extreme only 180 counties (5.3 percent) had no adult medicine
physician in 1975.
The distribution of FTE physicians practicing adult medicine within
individual States is quite striking. For example, 62.1 percent ofCalifornia counties (36) had 47.4 FTE adult medicine physicians per100,000 or higher while only 20.1 percent of all U.S. counties had ratios
in this range (Figure 2.1). At the other extreme, only 11.9 percent ofCalifornia counties (7) had county physician/population ratios less thanor equal to the national median of 32.9. Many other examples could be
given. The distribution of physicians within individual States does notusually reflect the distribution of physicians across the Nation.
Because county units have widely varying numbers of people and
physical geographies, one would expect the distribution of physicians tobe related to some degree to population concentration. It is intuitive
that the probability of a county having a physician of a particularspecialty type increases as the size of a county in terms of populationincreases. That physician-to-population ratios vary widely acrosscounties with populations of similar size is not always intuitive.Similarly, counties with comparable physician-to-population ratios oftenhave widely varying populations.
Analysis of physician-to-population ratios for adult medicine by sizeof county populations reveals varied distributions across counties with
particular ranges of population and across counties with similarphysician-to-population ratios. Of the 197 counties with no adultpract.tioner, 90 percent had county populations\of less than 10,000.Howev r, the 785 counties with populations under 10,000 displayed aremar able range of ratios. About 23 percent of these had adultpractitioner ratios between 10 and 29 physicians per 100,000 populationas Table 3 shows. At the other extreme almost 2 percent of thesecounties had ratios of more than 100 adult practitioners per 100,000population. This 2 percent of counties with populations under 10,000,however, represented 42 percent of all counties with physician-to-
population ratios of 100 or more adult practitioners per 100,000
13
population, and, counties of all population sizes examined wererepresented by this ratio. While it is quite true that the probabilityof a county having an adult practitioner increases with county size, thedistribution of the physicians among counties of similar size fi-Variedand no simple relationship exists. Similarly no simple relationshipexists for the distribution of counties by size within a range ofphysician-to-population ratios.
The varied relationships between county size and physician-to-population ratios found for the specialty of adult practitioner also holdfor pediatricians and plastic surgeons. Further, investigation of otherspecialties should yield similar conclusions about the distribution ofcounties by population size and physician-to-population ratio. Therelationships for the specialties of pediatrics and plastic surgery havebeen analyzed and are consistent with those for adult medicine.
Table 3
DISTRIBUTION OF ADULT PRACTITIONER TO POPULATION RATIOSFOR COUNTIES WITH LESS THAN 10,000 POPULATION
Practitioners to Population Ratios per 100,000 Population0 10-29 30-49 50-69 70-99 100+
Number ofCounties 179 224 207 113 48 14
% of countiest with populationsunder 10,000 22.8% 28.5% 26.4% 14.4% 6.1% 1.8%
% of allcounties witheach ratio 90.8% 21.4% 16.4% 29.1% 44.4% 42.4%
Another way that physician distribution data can be analyzed is by
health service area (HSA). The National Health Planning and ResourcesDevelopment Act of 1974 (P.L. 93-641) created approximately 200 local
planning agencies and health service areas. Guidelines for the creation
of health service areas showed a clear intention to preserve local
trading patterns by the requirement that Standard MetropolitanStatistical Areas (SMSAs) be preserved to the maximum extent possible.
Further stipulations on HSA population size and the designation of areasmay result in health service areas being more appropriate approximations
of market areas for some specialties of physicians than county areas.Data on 202 HSAs designated in 1975 show a substantially tighter range of
distribution - -from 24.5 FTE adult medicine physicians per 100,000 to
139.0 per 100,000. The HSA median ratio of 42.2 and 90th percentile of
63.6 are both higher than the Corresponding county ratios for the Nation.
The 31,640 FTE general' surgeons in the U.S. have a somewhat less
diffuse distribution than those physicians practicing adult medicine.Although there are 14.8 FTE general surgeons per 100,000 population, 39
peFcent (1,202) of all counties do not have a physician of this type
(Figure 3). The median county general surgeon/population ratio is 5.8,
and 10 percent of all counties have more than 16.6 FTE general surgeonsper 100,000 population.' The highest ratio is 108.5 per 100,000, and only30 counties have ratios higher than 30 FTE surgeons per 100,000.
Distribution of general surgeons within States again is marked.Whereas 39 percent of all U.S. counties do not have any general surgeons(Figure 3), 8.6 percent of California counties (Figure 3.1), 32.7 percentof Alabama counties (Figure 3.2), and 57.1 percent of Colorado counties
(Figure 3.3) have no general surgeon. At the other extreme some 20percent of U.S. counties have 12.4 or more FTE general surgeons per100,000 while California, Colorado, and Alabama have 37.9 percent, 20.6
percent and 10.5 percent of their respective counties' physician
population ratios in this range. Consequently, compared to the U.S.,
California counties tend to have a disproportionate share of highergeneral surgeon/population ratios, Alabama has a disproportionate shareof counties with ratios at or slightly above the national median, andColorado has a disproportionate share of counties with no general surgeon.
The range of ratios for HSA areas,',4.61 to 46.4 FTE general surgeonsper 100,000, is also smaller than that for counties, 0 to 108.5. Only
five of 202 health service areas have ratios less than the median countyratio for the U.S., 5.8. The median HSA ratio is 11.3 and the 90thpercentile is 19.76. In the highest group of HSA ratios, there were
19.76 and 25.0 FTE general surgeons per 100,000.
The histograms for obstetrics, pediatrics, psychiatry, orthopedicsurgery, ophthalmology, urology, otolaryngology, dermatology,
neurosurgery, plastic surgery, and thoracic surgery, Figures 4-14,
illustrate patterns of geographic distribution radically different from
those of all specialties, adult medicine, or general surgery. The number
15
0 1'1Not.]
and percent of counties without physicians of these specialties arestriking, from 64.1 percent of all counties without a pediatrician to89.2 percent of all counties without a plastic surgeon. As a consequence,the distribution of pediatricians, psychiatrists, orthopedic surgeons,ophthalmologists, urologists, otolaryngologists, dermatologists,neurosurgeons, plastic surgeons, and thoracic surgeons tends to beconcentrated in a small proportion of counties. The broad ranges of the10th deciles further emphasize the concentration of these specialties inrelatively few geographic areas. Significantly less than 10 percent ofall counties have pediatrician/population or psychiatrist/populationratios which equal or exceed the national ratios of 9.5 and 9.1 per100,000 respectively.
Once again many California counties appear to have a disproportionateshare of physicians by specialty, illustrating their uneven distributionacross the Nation and within States. The distribution of pediatricianswithin California (Figure 5.1) shows a marked cluster of counties withthe highest pediatrician/population ratios whereas the opposite tends tohold for the distribution of pediatricians in Arkansas (Figure 5.2). Thesame pattern holds true for neurosurgery. While only 13.2 percent of allU.S. counties had a neurosurgeon, 56.9 percent of California counties(33) had a physician of this specialty (Figure 12.1). The correspondingpercent of Arkansas counties is but 5.3 percent or four counties (Figure12.2).
At the HSA level the distribution of the preceeding specialties ofphysicians becomes much less diverse as the area used to calculate thephysician/population ratios grows in both size and population. Whereas30.5 percent of all counties in the U.S. had a psychiatrist, all healthservice areas had a physician of this specialty. While only 10.8 percentof all counties have a plastic surgeon, 89.2 percent of all HSAs had aplastic surgeon. Interestingly, the range of HSA physician/populationratios for psychiatrists is from .66 FTE psychiatrists per 100,000 to45.39 while the ranges for neurosurgeons is 0 to 5.36 and for plasticsurgeons from 0 to 3.09. The median ratios are 5.04, 1.07, and .68respectively for psychiatry, neurosurgery, and plastic surgery. Thecorresponding 90th percentiles are 13.66, 2.11, and 1.53 respectively.
Caveats--Physician/Population Data
In the.utilization and analysis of these data it is necessary tounderstand caveats surrounding the use of physician / population ratios.Perhaps the most basic caveat concerns whether or not the numerator anddenominator used are appropriate. Ideally the figures used would bethose for market areas; the numerator would represent those physiciansactually serving the residents represented in the denominator. Thecounty ratios used here are the best current approximations of physicianmarket areas available nationwide. It should be noted that physicians
16
30
represented in the numerator may serve more than the total number of-persons represented in the denominator, and all persons in a denominatorbecause of their residence within a particular county may not receivecare from numerator physicians. Market areas differ in size andpopulation among specialties such that a county may be too small (orlarge) to represent service areas for some specialties accurately.
Consequently, physician/population ratios premised on county datamust be interpreted cautiously because the true market areas may be
different. The following section will describe in more detail theshortcomings of physician/population data and the alternative of marketarea data.
Data on Market Areas
Because market area data are important for the realistic analysis andresolution of manpower problems, it seems imperative to the GeographicDistribution Panel that physician market areas be determined. The
weaknesses of past measures of distribution based on inappropriatelydefined areas are well known. The shortcomings of physician/populationratios which render them unsatisfactory as measures of distribution aresummarized here, as well as in Part V of this report.
Availability of physician services is determined not only by thesize of the manpower pool residing in an area but equallyimportantly by the composition and characteristics of the poolsuch as professional, activity, that is, patient care versusnonpatient caret specialty, type of delivery mode, age, andproductivity.
Service requirements are determined not only by population size,but also by its make-up: Age, sex, race,, income, education, andwillingness to travel, or proximity to another large community.
There is a difficulty with deriving physiCian/population ratioswith numerators and denominators that are confined to persons whoeither deliver or consume health services exclusively in thedesignated geographical unit of observation.
Physician/population ratios contain numerators and denominatorswhich are not homogeneous across geographic units of observation.
Thus, the findings of geographic distribution studies using theseratios have frequently been misrepresented and/or misinterpreted. Such
research can describe the geographic unevenness of health manpowerdistribution in the sense that the physician distribution may depart fromdirect proportionality to the population distribution. But without astandard of adequacy that takes into account effective demand and
17
requirements of the population at risk as well as the actual andpotential productivity of the relevant manpower pool, unevenness of amanpower distribution should not ipso facto be translated intomaldistribution.
Some of the basic flaws of physician/population ratios could beameliorated through the use of data at the market level,where thenumerator would represent those physicians actually serving the residentsrepresented in the denominator. However, until the U.S. has a system toidentify all, ambulatory as well as inpatient visits and salientdescriptive statistics of each visit, a second beat approach to thedefinition of service areas must be followed.
Use of claims data from Medicare, Medicaid, Blue Cross -Blue Shield,and other commercial insurers could approximate all ambulatory visits tophysicians and could allow an empirical estimation of service areas.Physicians could be allocated to small geographic areas according to eacharea's residents' relative proportion of visits. The physicians servingthe residents of each small area can then be determined, and small areascan be clustered to form the service area of a physician or group ofphysicians. While the use of claims data does not take into accountself-pay clients' visits, only through empirical investigation of serviceareas for specific physicians of specific specialty types can one beginto determine generalizable characteristics of service areas.
It is imperative that the distribution of physicians by specialty bemonitored at the market level since it is dynamic and changing ?Ver timeas new physicians enter markets and older physicians change market areasor leave markets altogether. Consequently, a system of angoing datacollection and analysis must be put into place to assess changingdistributional patterns, the success of programs to alter distributionalpatterns, and the attainment of societal goals for access to physicianservices,
/
At the minimum the system should monitor the numhe/ r of physicians ofeach specialty type in each market area, their characteristics, andcharacteristics of area residents. At a minimum,/ outputs of the systemwould be physician/population data at the market level and sunmary tablesand histograms based on market areas to display/the ranges Ofdistribution.
Summary
From the preceding tables, figures, and analysis, it is obvious thatthe distribution of all physicians or a particular specialty of physicianis not uniform. As a consequence the needs and demands for care whichresidents of different areas have may not be met uniformly. GMENAC inits efforts to quantify the optimal number of physicians by specialty in
18
4?
the U.S., has considered that the':c is some minimal physician/populationratio for each specialty that should exist in all areas. (See Chapter
V) This ratio will be affected not only by the aggregate number ofspecialists in the U.S., but also by the existing distributional patternsand future location tendencies of each specialty. The GMENAC ModelingPanel has sought to adjust its aggregate physician requirements to take
these factors into account.
The following summarizes the findings of the Geographic Panel on thedistribution of physicians and data to describe them:
1. There is a disparate geographic distribution of physicians byspecialty in the U.S. as measured by county physician/populationratios. In 1975 less than 10 percent of all counties hadphysician/population ratios which were as high or higher than thenational ratio of 144 non-Federal physicians of all specialtytypes per 100,000 population. There were 167 counties (5.4percent) without a physician. The highest physician/populationratio was 1,299.4 per 100,000 population.
2. The geographic disparity is greater for some specialties, e.g.,plastic surgery or neurosurgery than for others, e.g., adultmedicine or general surgery.
3. Good data do not exist nationally on physician market areas by
specialty.
RECOMMENDATIONS
RECOMMENDATION ,1:'1: The functional medical service areas, specialty by-
specialty? ate recommended as the geographic unit for assessingavailability of physician services. It is also recommended thatphysidian market areas by specialty be determined empirically basedon patient origin data derived from such information as discharge and
claims data until such time as total enumeration of physicianservices is possible, an t at the resulting areas be compared tothose previously determined by specialty societies.
The specialties of dermatology, obstetrics-gynecology, orthopedics,and neurosurgery have developed methods for determining the market
areas for their respective specialties based on zip codes, economic
service areas, and time-to-service concepts.
RECOMMENDATION 2: GMENAC supports the evaluation of alternative data
systems for the monitoring of the geographic distribution ofproviders.
19
% OF
UNIVERSE
40
35
30 -
25
20 -
N0
15 -
figure 1PANEL 1 FRAME 15
HISTOGRAM OF U.S. OF ALL COUNTIES
TOTAL ALL PHYSICIANS
10 -
0 -:....10.0%:....10.1%: 98%;...,10.1%:,....9.9%:....10.0%:....10.0%:,...10.0%:..10,070.,10.1%:RANGE X 0.00/ 17.29/ 27.86/ 36.44/ 44.50/ 52.37/ 61.35/ 73.73/ 91.00/ 122,85/
17.24 ' 27.78 36,42 44.44 52.33 61.32 73.72 90.91 -122.55 1299.441 OF COUNTIES 308 313 303 311 305 308 308 309 307 312
DECILE 1 2 .3 4 5 6 7 8 910o '
u0.) %1`iX PHYSICIANS PER 100,000 POPULATION
N
% OF
UNIVERSE
40 -
35 -
30
25 -
20 -
15 -
ID
5 -:
Figure 2
PANEL 1 FRAME 2
HISTOGRAM FOR U.S. OF ALL COUNTIES
I I I I I I II
ADULT MEDICAL CARE
II I I 1,1 I I I
./
-:....10.0%:....10.3%: .....
12.92/ 20.45/
20.41 25.00
318 297
2 3
RANGE m 0.00/
12.90
I OF COUNTIES 309
DECILE 1
It PHYSICIANS PER 100,000 POPULATION
36
.
25.08/ 29.17/ 32.89/ 36.86/ 41.40/ 47,39/ 57.61/
29.13 32.88 36.85 41.38 47.30 57.55 485.88
311 305 308 309 307 308 312
4 5 6 7 8 9 ID
Figure 2.1
X OF
UNIVERSE HISTOGRAM OF CALIFORNIA
40 -
35 -
30 -
a
25 -
20 -
OF ALL COUNTIES
ADULT MEDICAL CARE
N
15
10
5
0
10
-
-
7
:f11111111:
.011111111:
-: 1 7%: 3 4"/: 0
:111111111:111111111:
0%: 3 4%; 3 4%;
PANEL I FRAME 58
;111111111:
:
:1111111111
..... 5.2%1..12.1%; 8 6%:....29.3%1....32.8%;
41.40/ 47.39/ 57.61/
47.30 57.55 485.88
005 017 019
RANGE 11 0.00/ 12.92/ 20.45/ 25.08/ 29.17/ 32.89/ 36.86/12.90 , 20.41 25.00 29.13 32.88 36.85 41.38
0 OF COUNTIES 001 ' 002 000 002 002 003 007
DECILE I 2 3 5 6
j Lit PHYSICIANS PER 100,000 POPULATION
a 9 10
OF
UNIVERSE
40' -
35
30
25
20
15 -
10
5 -:
PANEL 1 FRAME 6
4
Figure 3
HISTOGRAM. FOR U.S. OF ALL COUNTIES
tJ
GENERAL SURGERY
, .
0 -:,..710.0%:....10.0%:,..19.0%: ..... 1.0Z:....10.0%:....10.1%: ..... 9.9%:....10,1%:....10,0%:...,10.0%:
RANGE M 0.00/ °4 0.00/ 0.00/
0.00 0.00 0.00
J 0E-41UN1IES 308 ..308- 586
DECILE--------- 1 '2 3
$ PHYSICIANS PER 100,000 POPULATION
0.93/ 2.56/ 5.79/ 7,71/ 9.86/ 12436/ 16.17/
2.55 5.78 7.70 9,85 12.35 16.13 108.47
030 308 310 306 312 307 309
4 5 6 7 8 9 10
Figure 3,2
% OF
UNIVERSE HISTOGRAM OF ALABAMA OF ALL COUNTIES
40 - 'GENERAL SURGERY
35 -
30 -
25 -
:111111111:
15 -
10 .:111111111:111111111:
rt
: :111111111:
0 "......10.(1%: .1. o10.47#1,...11.9%: .....
RANGE X 0.00/ 0.00/ 0.00/
0.00 0,03 0.00
0 OF COUNTIES 007 007 008 001 015 011 008 003 005 002
PANEL 1 FRAME
011111111;
: 1111 111 11:
1.5%:.,..22,4%:,...16.4%:.,,,11.9%:,....4.5%: 7 5Z: 3 0%1
0,93/ 2.56/ 5.79/ 7.71/ 9.86/ 12.36/ 16,17/
2.55 5.78 7.70 9.85 12.35 16.13 108.47
DECILE 1 2 3 4 5 6 7 11 9. 10
ts X PHYSICIANS PER 100,000 POPULATION
N
Figure 3.3
% OF
UNIVERSE HISTOGRAM OF COLORADO
40 -
35 -
30 -
25 -
'20
15 -
10
OF ALL COUNTIES
GENERAL SURGERY
PANEL 1 FRAME 76
.:111,11111:111111111:111111111:
:\
5 h.:
:111111111: :111111111:
: : \\0 -; 9 5%; 9\5%: .38.1%: .. 0.0%: 3 2%: ..... 4.8%;,....9.5%; .....4,8%;..11.1%; ..... 9,5%;
RANGE X 0.00/ 0.00/ 0.00/ 0,93/ 2.56/ 5.79/ 7.71/ 9.86/ 12.36/ 16.17/
0.00 x,0.00 0.00 2,55 5.78 7.70 9.85 12.35 16.13 108.47
# OF COUNTIES 006 006 024 000 002 003 0061 003 007. 006
DECILE 1 2 3 4 5 6 7 a 9 10
X PHYSICIANS PER 100,000 POPULATION
Al 4.c
15
UNIVERSE
40 -
35 -
30 -
25
20 -
15
Figure 4
HISTOGRAM FOR U.S. OF ALL COUNTIES
OBSTETRICS
10 -
PANEL 1 FRAME 4
0
RANGE N
B OF COUNTIES
DECILE
0.00/
0.00
308
1
0.00/
0.00
bi 308
2
0.00/
0.00
308
3
0.00/
0.00
308
4
0.00/
0.00
308
5
0.00/
0.00
341
6
8 94/0....10.070....10.0%:....10.17.:
1.08/ 4.51/ 6.94/ 10.02/
4.55 6.93 10.00 55.87
275 309 309, 310
7 8' 9 10
11"'
X PHYSICIANS PER 100,000 POPULATION
PANEL 1 FRAME 5
% OF
UNIVERSE
Figure 5
HISTOGRAM FOR U.S. OF ALL COUNTIES
40 - PEDIATRICS
35 -
30
-
25
20
15
10 - . . .
5-: r
-1
:
1
.
i:
0 -:.,-10.0%:-.10.0%:,,10.0%:,,,.10.0%:,,10.0%)..14.2%: . - .. 5.8%:.,..10.0%:,..10.0%:,,10.1%:
RANGE N 0.00/ 1,0.00/ 0.00/ 0.00/ 0.00/ 0.00/
0.00 0.00 0.00 0.00 0.00 0.00
1 OF COUNTIES 308 308 308 308 308 438
DECILE 1 2 3 4 5 6
N PHYSICIANS PER 100,000 POPULATION
43
0.57/ 3.04/ 5.30/ 6,21/
3.03 5.29 8.20 86.01
178 308 309 /511
7 8 9 10
0:1
Figure 5.1
% OF
UNIVERSE HISTOGRAM OF CALIFORNIA OF ALL COUNTIES
40 - PEDIATRICS
35 -
30
25 -
20
IN
15 -
IQ .:/1/1/1111:111,111,1: 11111111LIIIII11111IIIIIIIIIIIIII11111
5 -:
0 -:.,..10,3%:..,.10.31 :....13.87,: 0 OY. 0 0% 0 0%:
RANGE W 0.00/ .0.00/ 0,00/ 0,00/ 0,00/ 0.00/
0.00 0.00 0.00 0.00 0.01," 0.00
OF COUNTIES 006 006 008 000 000 000
PANEL 1 FRAME 61
5 2%:,,..19.0%:....17.2%:...,24.1%:
0.57/ 3.04/ 5.30/ 8.21/
3.03 5.29 8.20. 86.01
003 011 010 014
DECILE 1 2 3 4. 5 6 7 a 9 10
W PHYSICIANS PER 100,000 POPULATION
W ,
t.;
Figure 3.1% OF
UNIVERSE HISTOGRAM OF CALIFORNIA
40 -
35 -
30 -416
25 -
20
15 -
10
.011111111:
5
OF ALL COUNTIES
GENERAL SURGERY
8 6%: 0 0% 0 0% 0 0%.RANGE X 0.00/ 0.00/ 0.00/ 0,93/
0.00 I. 0.00 0.00 2.55
OF COUNTIES 005 000 000 000
DECILE 1 2 3 4
PHYSICIANS PER 100,000 POPULATION
:111111111:
PANEL 1 FRAME 0.
;111111111:
;111111111:
:111111111:
6 9%:....20.7%:...10.3%:....15.5%:....27.6%:....10.3%:
2.56/ ,5.79/
5.78 7.70
7.71/ 9.86/ 12.36/ 16.17/
9.85 . 12.35 16.13 108.47
004 012 006 009 016 006
5 6 7 a 9 10
53
Figure 5.2
% OF
UNIVERSE HISTOGRAM OF ARKANSAS OF ALL COUNTIES
40 - PEDIATRICS
35 -
30
25 -
20 -
15 -
.:11,11,111:111111111:111!Atill:111111110:111111111:111111111
PANEL 1 FRAME 47
:111111111:111111111:
:1111/1111:
-:..
0 -:..10.7%:,...10.7%:.,..10,7%:....10.7%:,...10.7%:....28.0%: 2 7%;
RANGE m 0.00/ 0.00/ 0.00/ 0.00/ 0.00/ 0,00/ 0.57/
0.00 '0,00 0.00 0,00 0.00 0.00 3.03
OF COUNTIES 008 008 008 008 008 021 002
6'7%4
3.04/
5.29
005
:111111111:
6 7%: ..... 2.7%:
5.30/ 8.21/
8.20 86,01
005 002
mc.
hjk.
1,EfILE 1 2 3 4 5 6 7 a0
9 10
Jk)M PHYSICIANS PER 100,000 POPULATION
% OF
UNIVERSE
40 -
35
30 -
sol
25 -
to
20 -
lo
15 -
Figure 6
HISTOGRAM FOR U.S. OF ALL COUNTIES
PSYCHIATRY
10 -
5 -:
PANEL t FRAME
- :. 0 .
..
.
.
. . . . .
0 -:....10.0%:....10.0%:...\.10.0%:....10.0%:.,..10.0%:....19.67 0.4%:....10.0%:....10.07.:....10.1%:
RANGE 11 0.00/ .* 0.00/ '0.00/ 0.00/ 0.00/ 0.00/ 0.71/ 0.94/ 3.22/ 6.76/
0.00 0.00 0.00 0.00 0.00 0.00 0.93 3.19 6.75 104.22
0 OF COUNTIES 308 308 308 308 308 604 012 309 307 312
DECILE 1 2 \ 3 4 5 6 7 a 4 10
M PHYSICIANS PER 100,000 POPULATION
PANEL 1 FRAME A
OF
UNIVERSE
Figure 7
HISTOGRAM FOR U.S. OF ALL COUNTIES
40 - ORTHOPEDIC SURGERY
E. I
35
30
25
20
15 -
10 -
5 -:
-1
0 -: .....
, RANGE X 0.00/ 0.00/ 0.00/ 0.00/ 0.00/ 0.00/ 0.00/ 0.64/ 3.35/ 6.13/
0.00 0.00 0.00 0.00 0.00 0.00 0.00 3.38 6.12 14.58
I OF COUNTIES 308 308 308 308 308 308 370 24) 308 311
DECILE 1 2 3 4 5 6 1 4 9 10
x PHYSICIANS PER 100,000 POPULATION
PANEL 1 FRAME 10
% OF
UNIVERSE
40
35 -
30 -
25
20 -
15 -
Figure 8
HISTOGRAM FOR U.S. OF ALL COUNTIES
Ophthalmology
10 -
5
-: :ItIIIIIII:.: . .
0 -:,...10.0%:....10.0%:,...10.0%:...,10,02:..10.0%:,...17.5%: ..... 2.5%:....10.0%:-.10.0%:....10.1%:
RANGE N 0.00/ ', 0.00/ 0.00/
0.00 0.00 0.00
I OF COUNTIES 308 308 308
DECILE 1 2 3
N PHYSICIANS PER 100,000 POPULATION
0,00/ 0,00/ ' 0.00/ 0,54/ 1.69/ 3,43/ 5.20/
0.00 0.00 0,00 1.68 3.42 5,19 52,98
308 308 539 077 308 308 312
4 5 6 7ll .9 10
% OF
UNIVERSE
40 -
35 -
Figure 9
HISTOGRAM FOR U,S. OF ALL COUNTIES
UROLOGY
PANEL 1 FRAME 11
30
25
20
15
10
5
0
de
-
-
la
la
-
la
-:
-:,...10.0%:....10.0%:....10.0%:....10.0%:....10.0%:,...10.0%:....13.2%: 6 9%: ..... 9.8%:.,..10,1%:
RANGE * 0.00/ 0.00/ 0.00/ 0.00/ 0.00/ 0.00/ 0.00/
0.00 '0.00 0.00 0.00 0.00 0.00 0.00
4 OF COUNTIES 308 308 308 308 308 308 408
DECILE 1 2 3 4 5 6 7
X PHYSICIANS PER 100,000 POPULATION
0.47/
2.23
213
A
2.24/ 3.69/
3.66 35.76
303 312
9 10
Figure 10
% OF
UNIVERSE HISTOGRAM FOR U.S OF ALL COUNTIES
40 - OTOLARYNGOLOGY
35 -
30 -
25 -
20 -
15
10 -
5 -;
0
RANGE K 0,00/
0.00
' 0,00/
0.00
0.00/
0.00
0.00/
0.00
0,00/
0.00
0 OF COUNTIES 308 308 308 308 308
DECILE I 2 3 4 5
PHYSICIANS PER 100,000 POPULATION
PANEL 1 FRAME 13
.....
0.00/ 0.00/ 0.44/ 1.32/ 2.74/
0.00 0.00 1.31 2.73 58.82
308 518 100 306 312
6 7 8 9 10
Or,
\,
PANEL 1 FRAME 3
Figure 11
% OF
UNIVERSE HISTOGRAM FOR U.S. OF ALL COUNTIES
40 - DERMATOLOGY
35
30 -
25
IN
20 -
15 -
1 1 1 1 1 1 1 1 1 :
10 -
:111111111:
5
0 ..... 8.7%:..10.1%:
RANGE M 0.00/ 0.00/ 0.00/ 0.00/ 0.00/ 0.00/ 0.00/ 0.00/ 0.34/ 1.67/
0.00 ' 0.00 0.00 0.00 0.00 0.00 0.00 0.00 1.66 52.94
1b0,5 COUNTIES 308 308 308 308 308 308 308 348 269 311
DECILE ti 2 3 4 5 6 7 8 9 10
)1 PHYSICIANS PER 100,000 POPULATION
Figure 12
'4 OF
UNIVERSE HISTOGRAM FOR U.S. OF ALL COUNTIES
40 - NeuroSurg ry
35 -
30
25
20
c.4
15 -
10 -
5 7:
0
PANEL 1 FRAME 12
RANGE m 0.00/ 0.00/ 0.00/ 0.00/ 0,00/ 0,00/ 0.00/ 0.00/ 0.15/ 1,15/
0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 1.13 23.531 OF COUNTIES 308 308 308 308 308 308 308 519 097 312
DECILE 1 2 3 4 5 6 7 8 9 10
PHYSICIANS PER 100,000 POPULATION
PANEL 1 FRAME 68
Figure 12,1
% OF
UNIVERSE HISTOGRAM OF CALIFORNIA
40 -
35 -
30 -
25
20 -
m
15 -
OF ALL COUNTIES
Neurosurgery
10 .:11111111v:IIItillit:mit1111; II111111111III111111111111111111111111111111111111
:1111111111:
0 0 0% 0 0% 0 0% 0 0%: 8 6%;.,..48,3%:
RANGE X 0.00/ 0.00/ 0.00/ 0.00/ 0.00/ 0.00/ 0.00/ 0.00/ 0,15/ 1.15/
0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 1.13 23.53
OF COUNTIES 006 006 006 007 000 000 000 ,000 005 028
DECILE 1 2 3 4 5 6 1 8 9 10
X PHYSICIANS PER 100,000 POPULATION
Figure 12,2
OF
UNIVERSE HISTOGRAM OF ARKANSAS
40 -
35 -
30
25
20
kip
15
OF ALL COUNTIES
Neurosurgery
10 .:1111111,1111,111111011,111,1:11,11111,011,11111:1111,1111:111,11111:111111111:
5 -:
RANGE X
PANEL 1 FRAME 54
:1,1111111:
0 Or,:
0.00/ 0.00/ 0.00/ 0.00/ 0.00/ 0.00/ 0.00/ 0.00/ 0.15/0.40 ' 0.00 0.00 0.00 0.00 0.00 0.00 0.00 1.13
5 3%:
1.15/
23.53# OF COUNTIES 008 008 008 008 008 008 008 015 000 004
DECILE 1 2 .3 4 5. 6 7 8 9 10
x PHYSICIANS PER 100,000 POPULATION
73
Figure 13
% OF
UNIVERSE HISTOGRAM FOR U.S. OF ALL COUNTIES
40 - PLASTIC SURGERY
35 -
30
25
C20
gn
15 -
, 10 -
PANEL 1 FRAME 14
0 10.0 "/.:,...19.3%......0.)%:....10.1;
RANGE W 0.00/ ' 0.00/ 0.00/ 0.00/ 0.00/ 0.00/ 0.00/ 0.00/ 0.17/ 0.57/
0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0:56 44.44
I OF COUNTIES 308 308 308 308 308 308 308 59 022 312
DECILE 1 2 3 4 5 6 7 4 9 10
X PHP,ICIAN5 PER 100,000 POPULATION
% OF
UNIVERSE
Figure 14
HISTOGRAM FOR U.S. OF ALL COUNTIES
40 - THORACIC SURGERY .
35 -
1.
30 -
25
20 -
15 -
10
5 -:
PANEL 1 FRAME
0 -:....10.0%:-.10.0%:....10.0%:....10.0%:....10.0%:....10.0%:..10.0%:....16.6%: ..... 3.4%:....10.1%:
RANGE X 0.00/ . 0.00/ 0.00/ 0.00/ 0.00/ 0.00/ 0.00/ 0.00/ 0.10/ 0.90/
0.00 '0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.89 19.61
1 OF COUNTIES 308 308 308 308 308 308 308 512 104 312
DECILE 1 2 3 4 /05, 6 7 8 9 10
x PHYSICIANS PER 100,000 POPULATION
LITERATURE REVIEW OF REASONS FOR VARIATIONS IN DISTRIBUTION
The second part of Charge 1, provided in this section, is a synopsisof the current literature on physician distribution. An extensiveappendix, Appendix 4, contains a review of 90 factors found to be relatedto physician location including influences on and objectives of physicians.,
as well as environmental, practice-related and demand-related decisionfactors. Many of these factors are found to be closely interrelated, aswell as overlapping. There are conceptual differences found in thestudies, and it is difficult to separate each factor into a discreteinfluence.
The Problem
The Secretary of The Department of Health, Education, and Welfare, inan address to the Association of American Medical Colleges in October1978, stated that the geographic maldistribution of physicians is one oftwo basic concerns in national health policy (Califano, 1978). Morerecently, Dr. David Rogers, President of the Robert Wood JohnsonFoundation, reported that while there will not be an overall shortage ofphysicians providing general medical care in the 1990s, seriousinequities still exist in the availability of care (Rogers, 1979). Otherauthors cite similar findings (Bobula and Goodman, 1979).
Influences: Background Factors
Several categories of influences have been reported to affectlocation decisions. These include the geographic background of thephysician, such as place of residency, or some combination of geographicorigin, medical school, internship, and residency location. A recentstudy found that National Health Service Corps (NHSC) physicians whoeventually locate in rural areas and those who locate in urban areas arenot differentiable in terms of background (Wilson, 1979). The studyfound, however, that physicians who entered practice in non-NHSC ruralareas usually attended medical schools located in a relatively smallcommunity, had residency experience in a nonhospital setting and/or in arural area, and had a rural background (Wilson, 1979). Even though aphysician may deLide to practice in a rural community, this is not
42
sufficient to compel the continuation of practice in that community(Parker and Sorenson, 1979). The research results are not definitive onlong-term retention in rural or shortage areas.
Objectives
Physicians' choices are not only affected by their sociodemographiccharacteristics and origin, but individually they have a number ofpersonal objectives regarding their life-styles. Both the physician andspouse have a set of preferences with respect to careers and place ofresidence, which includes the role of other individuals. Thesepreferences or goals can be primarily family-oriented, economic,prestige-oriented, or professional.
Professional goals/preferences can include:
Accessible hoppital facilitiesAvoidance of long hours of practice/excessive workloadContact with colleagues/professional interaction
Personal goals can include:
Spouse/family/friends--influence of/desire to locate nearTime for leisure/family pursuitsClimate or geographic area preferencesCommunity/school preference
The Government Accounting Office (GAO) found rural physiciansgenerally more influenced by personal and community factors and urbanphysicians more influenced by factors related to the practice ofmedicine. Hassinger (1979) found that when rural physicianscharacterized the salient differences between urban and rural practice,they emphasized the quality of patient-physician interaction, while urbanphysicians cited lack of facilities and support personnel. Theseconclusions are borne out by Wilson's data which indicated that studentsfrom rural areas, who are not interested in research or teaching, aremore likely to locate in rural areas.
Decision Factors
The decision factors which a physician considers in practice locationchoices can be numerous, but basically involve: (1) the environment ofthe area; (2) the type of practice he/she will be able to pursue; and (3)the perception of demand and need for services in an area (Kane et al,1979). Although these may each impinge on location choice, they are notnecessarily decisive in all cases. It is clear that the communities with
43
the fewest resources will have the greatest difficulty attractingproviders. Surveys of physicians indicate that community characteristicssuch as cultural and recreational opportunities, climate, the quality ofthe educational systems, and the wealth of an area may play a predominantrole in the decision making process. In general, climate and geographicpreferences appear to be strong motivating factors, as are preferencesfor leisure activities. However, it should be noted that these factorsare not very amenable to policy manipulation. It should also be notedthat anecdotal information from physicians about an area (need, demand,income) can play an important, although undocumented, role in the \
decision-making process.
Some studies strongly suggest that one of the factors necessary forretaining physicians in rural areas is good professional support.Professional considerations that often figure prominently in aphysician's choice of practice location include the availability ofclinical and/or hospital facilities, the opportunity for contact withother physicians, and the possibility of continuing education.
Considerable attention has been de, ,Led to the'influence of economic
factors on physician distribution. Research to date has not shown astrong correlation between economic facI'ors and specialty or locationchoice. Both fees and incomes vary coisiderably by urban-rural area,region, and State. Except for Canadian studies, research relatingphysician location, or geographic distribution, to income differences hasnot shown a strong relationship. Tiffs may be due to: (1) Social taboosagainst citing income as a motivation factor; (2) the lack of gold data;or (3) the intercorrelation of the and inability to separateout factors. It is possible that physicians do take fee levels intoaccount in making location decisions (Institute. of Medicine, 1979),the available evidence presents no firm conclusions.
A number of factors are important in affecting the demand (orperceived need) for health services: These include population size, age(population 65 and over), income of the population; education (affectingutilization), physician income, and the existence of Federal programs.Most research has dealt with population size, per capita income, andphysician income. Results show the migration of physicians towell-populated areas with high per capita incomes.
Summary
There are a number of factors in the literature relating to locationchoice, and it is difficult to weigh the importance of each. Therelative importance of professional, sociodemographic, lifestyle,personal, community, and demand-related factors has not been determined,so that it is difficult to sort out where policy applications should bedirected.
44
)
Each of the studies is conceptualized differently. Some are surveysof an individual physician's performance; others are market studies, orregression'analyses of sociodemographic data. No two location models aremethodologically comparable. Differences in specification, estimationprocedures, and, data contribute to differences in the results obtained.
RECOMMENDATIONS
RECOMMENDATION 12: Incomplete information exists on the direction ofcausation of many of the factors affecting physicians location.Additional research is recommended to study: (1) How backgroundfactors such as sociodemographic factors affect specialty andlocation choices and the interaction between specialty and locationchoices, and (2) what factors affect permanent location choices inunderserved/rural areas.
RECOMMENDATION 13: Since the role of economic factors in locationchoice is not clear, attempts should be made to improve methodologiesto determine this role and to gather data on _previouslynonquantifiable topics such as income as a motivating force inspecialty or location choices.
RECOMMENDATION 14: Those strategies which,GMENAC has deemed mostpromising, such as preceptorships and ,tax incentives, and those whichare most amenable to evaluation efforts, should be evaluated morevigorously.
45
IV. VARIATIONS IN POPULATION-BASED. USE RATESAND EXPENDITURES: IMPLICATIONS FOR MANPOWER POLICY
Charge #2--Describe the variations in per capita expenditures for andrates of medical/surgical services among selected communities; relatethese variations to local physician distribution and project theimplications for future manpower needs.
Charge #3--Given that local physician supply will always vary, whatcriteria can be used to indicate acceptable levels of variations in localphysician supply, per capita rates of medical and surgical use, and percapita expenditures for medical services?
Since a major goal of physician manpower policy is to equalize thedistribution'of medical services on a geographic basis throughout thecountry, it is extremely important to examine the relationship betweenlocal service use rates and physician supply. Current manpower policy,which uses physician-to-population ratios as the primary measure ofsupply adequacy, assumes a simple relationship between local physicianavailability and the amount and kind of medical services consumedlocally. A number of studies provide evidence that the physiciansupply/service distribution relationship is a complex onE, but one thatcan and should be factored into policy decisions on the geographicdistribution of,physicians. (See Wennberg et. al., in press.)
A review of the literature revealed that variations in the amount andkinds of medical services used by residents of different geographic areasin this country and Canada are dramatic. In general, the smaller thegeographic units that are being compared, the greater the variations intheir resident's use rates for medical services. Utilization variationsdid not reflect statistical artifacts nor differences among populationgroups studied. While most of the studies of variations in specificprocedure rates concentrate on surgical services, there is evidence thatvariations in nonsurgical service rates are at least as great as thosefor surgical procedures. In most studies there is a correlation betweenthe rates of use of specific procedures and the local availability ofmedical specialists who perform these procedures.
However, these correlations do not allow accurate prediction ofspecific procedures rates based on the local distributionkof physicianspecialists. Even in communities whose residents experience the sameoverall rate of surgery, specific procedure rates vary dramatically.
46
QLs ti
spec. ing accevtable levQ1c of variation involves choosing amongranges possible ways of practicing medicine. In making the choice,the objective ihformation necessary to specify an optimal pattern ofconsumption of medical cars' i, not available. The choice must be madeunder uncertainty because i%fcirmation on the outcomes of alternative waysof treating a specific condition is commonly not available. Consequently,an optimum treatment configuration cannot be specified and the physicianmanpower "needed" cannot be ifecified in terms of outcome objectives.Furthermore, under circumstances when r-ormation on outcome isavailable; there is some evid hat patient choice of treatment maydiffer substantially from th as physicians make for their patients(McNeil et al., 1978). Unde ;ircumstance, "need" assessment basedon physician preference woul, iy produce a different specificationfor physician supply than would "need" assessment based, on patientpreference.
ESsence of an optimal solution to the question of acceptablecr' poses a significant challenge to currentlmethods for estimatingmanp *.,1 requirements based on professionally defined needs. Thesemethods assume that current national average utilization rates representa first approximation for projecting manpower requirements. However, thenational average rates result from the weighted average of per capitarates in local markets whose rates vary extensively. The crucialquestion is: "Which rate is right?" Unfortunately no clear answer isavailable.
In assessing the various patterns of medical practice, it isrecommended that ,particular attention be paid to those costs and benefitsthat can be clearly identified. The empirical ('-,ta indicate that currentsupplies of specialty physician manpower in lc rate feeforserviceareas in New England (and in Health Maintenance Organizations (HMOs))meet the demand fc, low variation services and that the surplus labor isinvested in high variation workloads, as will be described. Higher ratesof use of higher variation services implies greater per capita costs,and, particularly for invasive care such as elective surgery, higherrates of mortality and morbidity. Whether the risk is worth the benefitremains moot because the longer range benefits of alternative approaches,are not clear. Drs. Wennberg, Lapenas, 'and Greene argue that identifiedrisks and costs should be given more weight than uncertain Venefits andthat specialty manpower requirements based on a choice to mimize costsand morbidity and mortality are reasonable and consistent with what cancurrently be known about the public welfare.
As an example of the extent to which national requirements forphysician manpower will, vary depending on the level of use of serviceswhich is adopted as the normative rate, the authors estimated the number'sof physicians required in 1990 to accomplish surgical workloads underalternative estimates of the preferred workload-. The estimates arebased on the extrapolaeon from low and high surgery rates among the tenhospital service areas with the largest populations in Rhode Island,
47
Maine, and Vermont: the extrapolation is to the United States populationin 1990, using the same estimate for 1990 as was used in the GMENACmethodology.
The results of the estimation are summarized in Table 4. If the lowrate of use of the operations is the norm for the planning of surgicalmanpower, the number of surgeons required for the seven surgicalspecialties* listed is 17,762. This number is 41% fewer than the numberestimated as needed by the GMENAC panels. If the high rates were takenas the target for projecting manpower requirements, then 37,824 surgeonswould be needed to meet the required workload. This estimate is 26percent higher than the estimate made by the GMENAC modeling panel.
Table 4 gives the ratio of the estimates for each of the listedspecialties to ths GMENAC requirement. For most specialties, the rangeof rates of use of services produce manpower need estimates that bracketthe GMENAC estimate. However, for ophthalmology services and particularly for plastic surgery, the rates of use of service in most areas inthe New England study are substantially below the rate on which theGMENAC need estimate is based. The implication is that for each of thelisted specialties, there are a number of areas in New England for whichGMENAC estimates represent an ove.9stimation of the number of neededphysicians.
Details concerning the methodology used for making the estimatestogether with tables gi,-ring the range of utilization and manpowerrequirements for specific procedures are given in the complete paper, tobe published in 1980 as a background paper of the Geographic Panel.
Necessary vs. High Variation Care
Drs. Wennberg, Lapenas, and Greene reviewed the el,idemiologic datacomparing the distribution of services among relatively homogeneouspopulations served by different medical care organizations, includingfeeforservice, HMO, and the National Health Service in the UnitedKingdom., The analysis demonstrates that some condition/procedurecombinations--for example, operations for cancer, hernias, certain typesif neurosurgery, hospitalization for hip fractures--showlittlevariation between populations, and the pattern of distribution of theseservices is compatible with a model in which "need" is met more or lessuniformly throughout all markets. By, contrast, most conditions orcondition/procedure combinations are hospitalized at strikingly differentratesNand the distribution suggests that there is considerable discrlionwith regard to the decision concerning the "need" for the procedure.
* As listed in Battelle Document dated February, 1980 and in theestimates for gynecologists from the second draft entitled "OB/GYN DelphiPanel Responses for Manpower Requirements."
48
SPECIALTY
Table 4
1990 U.S. MANPOWER REQUIREMENTS FOR SEVEN SURGICAL SPECIALTIES(SURGICAL CAREXOMPONENT ONLY)
COMPARISON OF REQUIREMENTS BASED ON GMENAC'S ESTIMATES AND ONNEW ENGLAND'S a OBSERVED UTILIZATION EXPERIENCES IN LOW AND
HIGH USE AREAS
BASED ON BASED ON NEW ENGLANDa
GMENAC'S ESTIMATE UTILIZATION EXPERIENCES
General Surgery 8,422
Low
Ratio
to
GMENAC
6,019 .71
Obstetrics/Gynecology 11,334 5,705 .50
Ophthalmology 840 394
Orthopedic Surgery 4,126 2,290 .56
Otolaryngology 1,532 1,126 .73
Plastic Surgery 1,548 60211
.39
Urology 2,156 1,6261
.75
TOTAL PHYSICIANS 29,958 17,752 .59
Ratioto
Hi,-- GMENAC
11 1.31
15,116 1.34
872 1.j4
4.921 1.19
2,449 1..60
93:. 0.61
2,521 1.17
37,124 1.26
a) For th,, 10 largest hospital service areas in :he 3-state zrea of Maine, node D.A.endand Vermont, rates for surgical procedures by specialty for a 3year period wereobtained. Populations in the 10 -.2as ranged from 62,5-: to 17!.:396. For each
procedure category, rates in the high ar low area were used to estimate the manpowerrequirements, using the assumptions concecning productivity made by f:MENAC'sRec:irements Model. Detailed computations are available in GMENAC Paper. "Var ationsin Population-Based Use Rates and Expenditures: Implications for Manpower Polity" byJohn Wennberg, Coralea Lapenas, Richard Greene, and Michael Zubkoff.
49
The consistency of the indicators of variation among different setsof populations, together with the fact that patterns of variation betweenprocedures can be so different, suggests that the relative ranking of aprocedure along the high to low variation scale fixes that procedurealong a continuum descriptive of discretion in clinical decision making.Along the continuum, low variation proc ures are those for which (1) thecondition to be treated is well defined as to its presence or non-presence and (2) there is consensus concerning the value of the procedure(as distinct from an alternative .innroach) in treating the condition.
By contrast, high variation pr,:-dures aL,:: (1) those in which"need" is less clear, certainly not a dichotomous (yes or no) situation,and/or (2) those about which people do not agree concerning theirefficacy.
The argument is made that at current levels of supply, variations arerestricted to those conditions for which there is little agreement onwhat treatment should be used, given the general relationship betweenspecialty manpower supply and specialty-specific workload rates.Planning decisions to specify requirements for specialists in effect aredecisions on what is the appropriate level of services for high variationprocedures.
Varidtions in Rate of Surgery Among Non-Fee-For-Service Systems
Within western nations, procedure-specilic-pAfterns of variationappear to be similar even though the way health services are organizedand financed may differ. In the United States, patterns of procedurerate variation among HMOs appear to be similar to those seen underfee-for-service systems. Luft, who has studied surgical services amongHMOs, found the_overall surgery rates to be consistently lower than underfee-for-service (LLft, 1978). However, the speci.7ic proceduresdemonstrated considerable variation among t' HMOs studied.
On the Causes for Variation in Rate of Services
Drs. Wennberg, Greene, and their colleagues reviewed the crucialimportance of the physician as a producer of variations in rates of,service use. In addition to the aggregate supply, the evidence points tothe importance of differences in physician decision-making which areattributed to uncertainty concerning the outcome of various methods fortreating a particular condition (e.g., surgical or nonsurgical treatmentof gallstones) and to differences in values physicians attribute tovarious modalities of treatment. They amplify the two points made
50
earlier regarding: (1) Professional uncertainty, and (2) failure of thephysician agency role to guarantee that patient rather than physicianpreferences dominate the decision to uae a oarticulaormodality of care.
Professional Uncertainty as a Source of Variation
The interpretation that variations reflect differences inprofessional judgments made under uncertainty concerning the existence ofdisease or the value of specific treatments is supported by the medicalliterature. The extensive literature concerning the variability ofphysicians as observer and interpreters of medical evidence has beenreviewed by Koran (1975).
The uncertainty concerning diagnostic or therapeutic decisions thatexist in the med* -al literature are reflected in the decisions that aremeasured by the 8mall area techniques reported here.
The Flawed Agency Role as a Source of Variation
The importance of physician choice in influencing utilization is notrestricted to cases of uncertainty about diagnosis or long-rangebenefits. It Involves the values he or she assumes to be those of thepatient.
The divergence between physician values and patient values--and thefact that the former can dominate.decision making--is shown by examiningthe choices patients would make if they were better informed about thealternatives when the outcomes are known or where there is the risk ofdeath when using alternative choices. Here the choice is not underuncertainty but there is the existence of objective information whoseutilie.es (e.g. values) are being expressed in the decision process. Aparticularly impressive example of the fact that the agency choicesphysicians make on behalf of their patients can be radically different:from the choices patients would make had they been provided more -.-
information is given by LcNeil and her colleagues (McNeil, et al.,1978). When they asked patients who had already undergone an operationto remove a. cancerous lung to a, ess their preferences for greatercertainty for short-term survival vs. greater probability for survival acfive years, most patients preferred greater immediate certainty.
Consequently, these patients would have been better served byradiation tl-wx1: than by surgery for their cancerous condition. Theresults thus indicate that patients' preferences for treatment , Idality
can be m( -e conservative than their physicians and McNeil concludes thatpatients appear to be more adverse to
51
risk than their physicians. The
lluststudy
i tes the differences that can exist between physician andPatient utlxratties /rid how, with more information, patients may makechoices more consistent with their own values and more divergent fromthose of their physician-agents.
ConO/usions n"nncernin Preferred Criteria
HMO experience) the United Kingdom experience, and the low ratefee_for-servicemarket areas seem to provide compelling examples of
"Peri ences i-'c4 which Porta ations receive care with which they are more orleas satisfied and which involve lower costs and less morbidity andmortality' Given`:tven the highet costs and greater numbers of certain deathsassociated with greater use of high variation services, the manpowerrequirement decision for specialists should be made to minimize thechance of significant Type B errors, which are having too much manpowerwho create istrogenic disease ant unnecessary costs'by intervening with
unwanted ineffective"fective" medical care. A Type A error would behaving insufficient man to meet true need. The data summarized bywennhorg, Lan
-:11:1s1161.1t10::rne support the notion that there is nofor
evidence unde rutilization.for low variation hospitalservices, and tl,lt much of the current workload is inveted in highvariation services.
Serious attLation should be given to making available to physiciansutilization rate experiences relative to other physicians practicing inan area
Experience among 13 Vermont Hospital Service Areas and in theProvince of Saskatchewan suggests that feedback of population-based dataon ino idence of _rocedures may be a valuable tool for peer review and canaffect average P
rates of Procedures in areas (Wennberg et al., 1977; Dycket 1977).
The Sini ficane of Professional Uncertainty: Some Current Public Policy
Dilemmas
Surprisingly, little uisagreement exists (among western nations) onthe social
objectives.J tives of a public health policy. Across a wide spectrum
of polit"regul!t--
economies) People appear to agree on the basic values thatto charactershould the ter af their national health care systems. A
cons ensus health tiexists that care is a right of citizenship that should
not depend cn an individual's social or economic circumstance or ongeographic location. This right to health care includes the right toequitable use of ncurinc6 , technologies that improve the length and'quality" c.).life and "caring servicesP which without pretense of cure,
make disabltltie more endurable and death more tolerable.
52
A parallel consensus exists that the "free market" is an inefficientmeans for achieving agreed-upon social goals. As a result, nations ofdivergent political ideologies all intervene directly into health caremarkets. Although their methods differ, the strategies of intervention--those regulatory efforts designed to assure that the social goals are infact realized--are necessarily' based on a model of how the private marketworks, particularly the causes of and remedies for its failures. Healthinterventions range from the org,-aization of suppliers into a NationalHealth Service, as has occurred in England, to the regulation of specific,selected elements of the market through rate setting by regulatory andplanning agencies, as typified by the United States. Public
interventions, at least in western nations, primarily address distributionof resources, access to services, and freedom from untoward financial lossof individuals who use services.
With the qualified exception of certain drug. outcome studies, theissue of the effectiveness of allocated services with regard to their"curing" or their "caring" expectations has not been of direct concern topublic policy. Nor has the issue of how resources are allocated amongcompeting priorities (for example between "curing" and "caring" effortsor between alternative approaches to either) been of direct concern. Theissue of how and when resources designated for use in the health sectorare allocated to optimize individual welfare.is a responsibility tacitlydelegated to the medical profession, to be resolved through theprofession's traditions whicl, emphasize science and ethics as a sufficientmeans for resolving questions or fact and value. This process assumeshealth resource allocation based on professional consensus concerning thevalue of technology and the nature of patient need or utility.
Much e: the current policy is thus base( .H.1 the belief that ar
underlying professional consensus exists on the optimum methods foxallocating medical technology. For mple, such programs as theProfessional 'Indards Review Organization (PSRO) assume that meaningful"onsensus exists on which standards of care can be developed that willpromote the public health by iusuing that only necessary hospitalationsand procedures are undertaken. Similarly, the development of nationalstandards for health resources--physicians per capita an. facilities percapita--rests on the assumption that there is a determined r,..tationshipbetween resource input and health care outcomes. Needs-based planningsuch as the GMENAC approach t, manpower planning is an example. However,
if the existing evidence is usually insufficient to settle controversieson the value of common medical practices, consensus standards representweighted averages of those selected to establish them and cannot serve asa basis for the rational allocation of medical care te.:Imology. Indeed,
without more direct information on the outcomes.of alternative approaches,rational allocation is not possible if the criterion for rationality isthe maximization of health through the efficient use of health care.
53
In a similar vein, unless the choices made in rusource alloationrepresent the informed wishes of the patientrather _ban the physician- -then the same criterion for rationality cannot be met. An importantimplication of the evidence cited previously is that tl,e limits oninformed decision making are more severe than generally realized,affecting physicians as well as their patients. This is a problem evenwhen outcomes are reasonably well-known: McNeil shows that the valuesreflected by the physician's choice of treatment do not reflect thevalues of the patient; had the patient known more, he/..ea would havechosen otherwise. And when the outcomes of alternative treatments arenot well understood as is often the case--informed decision making orgiving informed consent is difficult indeed.
While most regulatory strategies assume so, it is not clear thathealth services provide what patients want. At the level at which healthcare is used in U.S. markets, marginal case selection is discretionary inthe very fundamental sense that more care may not necessarily be betteror that more intensive care would necessarily be selected were patientsfully informed concerning the possible outcomes. For this reason,current regulatory decisions concerned with the equitable distribution ofresources or an efficient use of resources need to be re-examined.
The dilemma of regulating "equity" without regard to effectiveness isillustrated by the patterns of income transfer between communities thatoccur because of taxation or insurance premium policies. At the level oflocal markets the premiums charged or taxes levied are not closely relatedto population ,consumption rates so that low consumption populationscommonly pay the same amount as high consumption populations. The cross-population subsidies persist over time, (unrelated ?o differences inpopulation need) as a function of the distribution of services. If theservice paid for by the income transfer were an unequivocal hood, thenthe interpretation would he easy. But a dilemma is clear: Populationsthat use more health services may not be a well off as those using less;it is not clear how the pri, .iples of "equity" can be used to interpretthe observed inequalities in per capita expenditures.
Similar unce:tainti., exist concerning the efficient production ofservices. Because the outcome value of many common procedures such astonsillectomies, hysterectomies, or cesarean section is in disput,traditional ways ' evaluating efficiencies based on relative amounts ofresources used to produce a treatment are subject to criticism. It isnot realistic to evaluate technical etiiciency without regard to theefficacy of the product produced. Clearly, it is absurd for regulatoryagencies to award institutions for efficiency if they are producing a mixof cases that decrease population welfare; yet this may be precisely whatis happening.
Manpower requirements planning, undertaken using either econometricmodels (e.g. the Bureau of Labor Statistics or the Bureau ot. HealthProfessional models) or needs-based models (e.g. the adjusted needs-
54
based GMENAC approach) begin with the assumption that current utilizationrepresents an important first approximation of the optimum situation. In
the case of econometric approaches where utilization equals demand, theproblem of projection involves estimating the influence such variables asfuture income will have on consumer demand for care. For needs-basedplanning, the issue is to what degree current utilization meets "real"need, by which is meaut the need that physicians identify afterevaluation of patient morbidity. The implication of the previousevidence for the econometric approach is that current utilization doesnot reflect what would occur if'informed consent were the basis forclinical decision making; therefore, current utilization cannot beinterpreted to represent an optimum allocation based on consumer marketchoice. The implication for needs-based planning is that informed choiceamong treatments based on expected outcomes--even if those outcomes arebased on physician, rather than patient values--cannot be systematicallyapplied across the range of treatments currently employed in the commonpractices of medicine.
Primary Care Physicians Manpower Requirements
Most of the Panel's discussions were concerned with the neel forspecialists, a n,-spective due largely to the fact that the empiricalanalyses on whir' the main arguments are based concern the use ofhospitals. Sonic information is available on the implications of variouschoices for primary physicians, including the possible colkseqwic-r ofdifferences in distribution between the alternative models for primarycare delivery, namely the per capita numbers of internists and f nily
practitioners.
Briefly, Vermont data show that population areas with more internistsper capita have greater numbers of diagnostic ests, more re-visits percapita and may experience greater difficulty in r:hieving access tophysicians for' new episodes of illness than do areas with greater numbersof general practitioners per capita. The experiences seem to beexplained by the fact that internists are more accustomed to detailedworkups, are more likely to see patients for follow-up and have fullschedules which are not designed for drop-in patients. The contrastingdata for general practitioners are based largely on the "old" generalpractitioner (GP) model rather than family practice and the workhfadprofiles of this hc%rly emerging profeSsional subgroup have not been
studied. If they emulate the GP, then greater emphasis on the familypraCtitioner model should improve access to car., and minimize re- visits.Based on existing information, the impact of either, approach onpopulation health cannot be assessed.
What criteria should be used to develop estimates of primaryphysician manpower? Under the assumption that they do not undertakeinvasive technology, the arguments developed for making value judgments
55
concerning specialists supply requirements do not apply to primaryphysicians, leading to the conclusion that the targeted proportion ofprimary physicians should be considerably larger than the 55 percent ofall physicians that now stands as a national average. This reason isprimarily a "negative" one--because the number of specialists should bereduced, and the existing supply must be shifted into primary careavenues. In the longer run, better estimates of the work of primaryphysicians should be made based on a more systematic analysis of rates ofinitial contacts and revisit rates.
RECOMMENDATIONS .
RECOMMENDATION 3: GMENAC urges the use of small area populationbased data on the availability_, requirements for and utilizationrates of hospital and physician services as a manpower planning tool.
RECOMMENDATION 4: GMENAC urges that the ranges of variations in theutilization of specific procedures and services among communities andservice populations be clearly delineated (including communities withdiffering financing and organizational arrangements for the deliveryof medical services).
RECOMMENDATION 5: Serious attention should be given to maki,available to physicians their utilization rate experiences ativeto the norms of other physicians practicing in their immediate area,region, or in the Nation.
,,MMENDATION 6: Serious attention should 1,- given to the voluntarycollection and dissemination of aggregate statistics for analyticalpurposes relative to communitywide utilization rates in variousservice areas.
RECOMMENDATION 7: GMENAC encourages the support of efforts to assessthe outcomes of common medical and surgical practices which exhibithigh variation across communities, as an important step forestablishing the long-range requirements for suppliers of medicalservices in the United States.
RECOMMENDATION 8: Future health. manpower planning'groups shouldcompare manpower estimates (whether derived as a "needs-based,""demAnd-based" or 'requirements-based" model) against,empirical
estimates selected from areas in the United States which exhibit highand low ntili7atin patterns.
56
V. CRITERIA FOR EQUITY AND ACCESS: PHYSICIAN-TO-POPULATION RATIOS AND TRAVEL/TIME DISTANCE TO SERVICE
IN DUCTION
Because of the complexity and scope of physician distribution issues,it has not been possible to measure accurately the geographicmaldistribution of physician services. Numerous studies have beenconcerned »iLh clocnmenting the extent of variation in the distribution o.physicians and with the characteristics of this variation. Less
attention has been devoted' to the development of criteria and standardsfor defining an optimal or adequate distribution of physician services.While it is acknowledged thatIthe curreat distribution of physician's bygeographic area has resulted in problems of accessibility for somepopulation groups, for example those in some rural areas, no agreementexists concerning the number of physicians which would be required tomeet the health care needs of these populations. Further, no consensusexists concerning criteria for determining whether an area has a"shortage" or a "surplus" of available health manpower.
In the development of criteria to determine an equitable distributionof physicians by specialty type, several factors emerge from a host ofpossible criteria as key factors to consider. These factors, which arefound frequently on criteria lists, include relative population need,resource availability, and resource accessibility. Relative populationneed is often measured through adjustments for differing sociodemographiccharacteristics and disease incidence from some standard population.Resource availability is sometimes measured through resource-to- popula-tion ratios, and geographic accessibility through time-to-service.
Although numerous estimates of ratios of physicians by special_y-to-population exist, relatively few time-to-service requirements have beenreported. None are defensible in terms of rigorous methodologicaldocumentation, and all suffer from a basic ad hoc, unscientific approachto setting requirements for meeting essential medical needs. Estimatesof maximum "acceptable" distances-to-services or travel times-to-servicesare often based on surveys of what is acceptable to are residents or onconsensus estimates of experts or representative groups of areacitizens. The standards consequently have little scientific basis interms of, the maximum time to care to prevent ser4u: adverseconsequences. It is therefore not possible to staunchly defend or attacka 30 minute driving time, fo' instance, as opposed to a 2(; linute or a 45
minute driving time.
57
Physician-to-Population Ratios--Caveats
Physician-to-Population ratios have frequently been used as a proxymeasure of the relative availability of physician services. The Ilasis ofthe method is the selected ratio of manpower to population. This ratiois multiplied by the population in the area being modeled to provide anestimate of the number of providers "needed" or "required" in the modeledarea. The validity of the estimate depend:. upon the appropriateness ofthe ratio selected to the target population. In the ratio method it isnecessary to assume that the manpower count in the numerator and thepopulation count in the denominator operate in the area being modeled inthe same way as in the area frOm which the ratio is selected. (Forexample, the same State ratio is not equivalent to the same countyratio.) The method generally assumes that population size explainsmanpower requirements and consequently ignores important influences whichdo not operate through population size.
Implicit in ri..8 approach iE the assumption that evenness of therati=, represents a desired distribution. Consequently, variability inphys.. ,n population ratios across geographic units of observation hasbeen viewed as evidence of maldistribution and the range of variation inthe ratioshas been used as an indicator of the extent ofmaldistribution. But if physiciaris were to be equally distributed, somephysicians, for example those in aetropolitan areas and in areas with anoider population, would face a larger demand for services than those inrural and more youthfully Populated areas. Apart from possibleinteractions between the supply and\demand for services, a so- calledequal distribution of physicians could lead to excess capacity in somerural areas and unde rcapacity in some metropolitan: areas, and thus-to-amaldistribution of physician servires.-
In order to circumvent soz of these pitfals, the numerator anddenominator may be defined and adjusted to relate specifically to theproblem and_geographic area under study. Many different assumptions anddefinitions can underlie the determination of the number of a specificspecialty of physicians.
The number of physicians, (the numerator) might be adjusted to reflect:
Federal physiciansForeign medical graduatesActivity statusFull-time equivalency
Availability of nurse practitioners and physician's assistantsand task delegationMix of physicians' specialtiesMedical care organization and practice settingPhysician case loadProductivity
58
The population figure might be adjusted to reflect:
Age, race, and sex structure of populationIncome and ability to payDisease incidenceMorbidity and presence or prevalence of chronic conditionsMigrant or seasonal populationsPopulation densityArea geography
As a result ratios may differ widely and cannot always be properlycompared.
While the methodology of physician-to- population ratios is slightlymore sophisticated than that of either time -to- service or distance-to-service, the resultant area ratios may not be appropriate for modelingmanpower requirements in an area. Differences between the area selectedfor the rati,, and the area modeled may exist, and the differences forwhich explicit allowances have not been made may compromise the validityof the resultant ratio in the aree modeled. Since ratios may be premisedupon wid, differing assumptions and definitions, it is not possible tounequivocally argue that one ratio is superior to another for general usethrougl ,t the country as a standard.
Geogr. : units of Analysis
An important factor impacting on physician -to- population criteria forassessing equitable distribution of services is the geographical areauN;.:1 consideration. Since the unit of geographic analysisgraphic ana is is notspE itied in optimal physician -to- population ratios, meaningful --,
e4 :ulation of the ratios and assessment of the distribution of servicesand manpower on a small area basis is dependent upon the extent to whichthe unit of analysis is an appropriate unit to study.
Several types of geographic units of analysis are possible. Theyinclude:
StateHealth systems agency service areaCountyZip codeHospital service areaPhysician service areaTrade or market areas for commercial products
Historically the availability of basic statistics on geopoliticalunits has been the primary criterion for selecting a unit ofmeasurement. The widespread use of county data obviously has to do with
59
the relative ease of obtaining information on physicians' practicecharacteristics and socioeconomic and demographic descriptors of thepopulation. The stability of county boundaries themselves alsocontributes to ease in analysis. However, consumers seeking physicians'services are not confined to the county, SMSA, or even State of residencein their search for sources of care. Contiguous areas may provide neededresources. In multi-county, multi-State metropolitan areas, there is somuch commuting of this nature that State and county lines probably dividerather than encircle health market areas.
Consequently the unit chosen for analysis probably will be grouped orclustered for study. The extent to which medical care services are bothavailable and accessible to the population of a specifiedNgeographiclocale will depend upon many other factors in addition to naturalgeographic boundaries. Population density and terrain may influence theapplication of an optimal or required physician to population ratiothrough modifibation based on time- or distance-to-service. Sparselypopulated areas may require higher physician-to-population ratios becauseof unacceptably long driving times as may areas with rugged ormountainous terrains.
Data to support the delineation of health service market areas havenot been readily available, however, despite real market areas being thepreferred unit of geographic analysis. ,Given the complex task ofdefining these areas and the fact that the necessary data are neitheravailable nor accessible on a uniform or a routine basis for almost allareas of the U.S., reliance on administratively designated, albeitarbitrary, geographic boundaries provide an inexpensive, interimalternative t6 determine service areas. Nonetheless, boundaries which donot reflect market areas are inadequate units of analysis if therelationships among the requirements for physicians' services, the supplyof these services, and their utilization are to be elucidated.
The American Academy of Dermatology (AAD) has examined alternativemethods of analyzing the geographic distribution of dermatologists. Theydeveloped a framework for studying distribution using Zip Code SectionalAreas._ The Academy believes these areas to be conceptually superior toStates, counties, or federally designated Health Service Areas asgeographic units of analysis. Areas formulated to facilitate thecollection of census data, those designed for planning hospitalfacilities, or those defined for political purposes, are very o'cteninappropriate for studying the accessibility of physician care. Inaddition, it became apparent to the AAD that a single unit was inadequateto deal with all types and levels of physicians' services, as eachpossesses a unique catchment or service area (Ramsay, et al., 1977).
Zip Code Sectional Areas approximate economic trading areas, in mostcases, and thus generally conform to the transportation and communicationpatterns of each place. It appeared reasonable to the AAD thatgeographic patterns of health service utilization largely conform to
60
those for goods and services in general. At the very least, Zip CodeAreas could be effectively served by a physicist. specialist at a centrallocation, regardless of existing traffic patterns. Two other specialties- -ophthalmology and orthopedic surgery--have also used this approach.
Another approach utilizes the central place theory to compare theeconomic system and the health care system within a specific geographicsetting and demonstrates that consumer patterns for health care servicesapproximate those for goods and services in general (Leyes, 1974). Itwas concluded that the development of an economic system is a necessacondition for the development of the health system. Using a theoryrelating the spatial results of supply and demand decisions, Leyesdemonstrated that there is a "hierarchical demand structure" such thatthe amount of and frequency with which consumers purchase goods variesaccording to the type of goods. Over 500 communities were grouped intoseven categories or economic and health-related service areas. It wasrecommended that these areas be used rather than political oradministrative units.
Utilization of physician-to-populatioh ratios along with drivingtimes is another means of analyzing the distribution of servicep whileconcomitantly accounting for access and availability. Consequently it isindeed unfortunate that defensible physician-to-population ratios andmaximum acceptable driving time criteria do notexist. NeverthelessGMENAC addressed the questions of standards for assessing distribution,fully cognizant that ultimately HSAs and local and State planners willknow best their individual needs, factoring in considerations ofdifferences in areas. While the time to service and physician topopulation standards presented in Table 5 are no more scientific orjustifiable than those presented in the following literature review, theyare the best possible estimates which the Geographic Distribution Panel'could identify for consideration and debate.
Summary Table of Significant Data Items
Table 5 presents the ranges of information available on physician-to-population ratios and times-to-service as well as first cut recommenda-tions of the Geographic Distribution Panel. Two basic premises underliethe development of Table 5.
First, several assumptions were made in defining time-to-servicestandards. Large variations in the scope of public transportationsystems, terrain, climate, and traffic and road conditions wererecognized. Mileage figures corresponding to travel times were notestimated for the preceding reasons. A broad range of modes oftransportation were acknowledged. Also it was understood that accessproblems are considerably reduced for those with private transportation,and that the poor, the elderly, and the handicapped are disadvantagedwhen public transportation must be relied upon.
61
Second,'the reality that physicians of a defined specialty alsoperform services outside the unique scope of their specialty wasacknowledged. (See Aiken, 1979; Mendenhall, 1978). Because of thepossibility of substituting different specialists' services in thetreatment of some conditions, five major groupings of physicians arepresented. They are: Adult Medical Care Services, Child Care Services,Obstetrical Services, General Surgical Services, and Emergency MedicalServices. GMENAC felt that these basic types of health care servicesshould be available and accessible to residents of all areas of the U.S.,and that minimum ratios of physicians-to-population in these five basicgroupings, as well as maximum times-to-service, for 95 percent of thepopulation should be given.
Table 5 is formatted as follows:
Specialty - =The first column provides a way of organizing thespecialties under the five major types of care which the Geographic.Distribution Panel deemed essential. These types of care seemedintuitively to hover the range of basic services which should be provided.
U.S. Ratio per 100,000--As described in detail in Part III of thisreport, these data are 1975 FTE figures per 100,000 total populationobtained from special computer runs. The figures marked (*) were notavailable on the special computer runs and were obtained from head countnumbers of physicians by specialty in the U.S.
Range of Physician Specialty Requirements Expressed as Ratios ofPhysicians per 100,000--According to a review of over 200 articlesconducted by Applied Management Sciences and reported in Review of HealthManpower Population Requirements Standards, (1976), these are the highestand lowest ratios of physician specialty requirements found in a reviewof the literature. The ratios represent a broad spectrum of studies fromthose attempting to provide sufficient resources to treat all medicalneeds to those reporting observed staffing patterns in HMOs.
GMENAC Modeling Panel Adjusted Needs-Based Ratio of Physicians bySpecialty per 100,000--These are estimates of the Delphi groups, and havebeen approved by GMENAC. These are final estimates.
Recommended Geographic Area of Analysis (Real Market)--As discussedin Part III and recommendation i., it is absolutely necessary thatphysician services be analyzed at the medical market level on empiricallydefined medical market service areas. (See Wennberg et al., 1980; Leyes,1974; Ramsay, 1976.)
Minimum Acceptable Ratio for All Areas--The figures representsone-half of the ratio ofspecialists-to-population in the U.S. which theGMENAC Modeling Panel deemed acceptable for 1990. For example, if theModeling Panel recommends 8.4 pediatricians per 100,000 for 1990, theGeographic inr.l would recommend 4.2 per 100,000 as the minimum
acceptable ratio.
62
Table 5
SUMMARY TABLE OF SIGNIFICANT DATA ITEMS
CMENAC Modeling Panel
1975 Range of Physician Adjusted Needs Based Recommended
U.S. Specialty Require- Range of Ratios of Geographic Interim Minimum Range of Recommended
Ratio ments Expressed as Physicians by Specialty Area of Geographic Acceptable Times-to Times-to
Per Ratios of Physician per 100,000 Analysis Area of Ratio for Service/ Service/
100,000 per 100,000 Pop. for 1990 (Real Market) Analysis All Areas I/ Manpower Manpower
Emergency Medical
Services 1,4-4.7 2/ N, A, 5.3-5.7
EMS
Region N.A. 2.7-2.9
5-30
Min. 30 Min.
Obstetrical Services Hospital HSA 30 Min, -
(08 /GYN) 19.4* 2.6-12.2 18.4-20.0* Service Service 9.2-10.0* 1 Hour 45 Min.
Area Area
Child Care Sva/ Physician HSA 30 Min. -
Pediatric Medical 32,9*** 5,5-23.2 47.1-51,2** Service Service 23,6-25,6** 1 Hour 30 Min.
Care Area Area
Adult Medical Care 23.2 (FP/GP) 50 -133.0 Physician
(FP/GP/Internal 16,7-30.8 (IM) Service County 13.4-15.4 (IM) 20 Min. -
Medicine) 22,7 (IM) 3.3-50.0 ',.2-35.7 (FP/GP) Area 16.6 -11.9 (FP/
GP) 1 Hour 30 Min.
Surgical Care Hospital HSA
General Surgery 14,8 1,7-12.4 9.4- 9,8 Service Service 30 Min, -
Area Area 4,7 -4.9 2 Hours 90 Min.
I) OnPhalf of GMENAC Modeling Panel needs-based estimate.
2/ Range of ratios from AMA, DOS and Am. College of Emerg. Phys, (1978-80)
* Per 100,000 women of all ages
** Per 1001000 children under 11
*** Per 100,000 children under 16
Range of Times-to-Service--Thesefigures represent the range of timesreflected in the literature for each of the five major types of care.
Some types such as adult medical care are documented more fully thanothers. The five types of care were arranged so that the type requiringthe shortest time-to-service was 'listed first.
Recommended Times-to-Service--The Geographic Distribution Panelrecommended these as maximum travel times for 95 percent of thepopulation, recognizing that unusual circumstances may arise which makethese travel times impossible to achieve for all areas. Therecommendations are based on an extensive review of the literature,including consensus judgments of experts and health officials. TheGeographic Panel members then reached their own consensus on these data.
PHYSICIAN-TO-POPULATION RATIOS
The literature on physician-to-population ratios is rich in numbersand types of studies reported. The findings of two major papers whichsummarize ratios found in the literature are presented. The first is aGMENAC Staff Paper on physician manpower requirements which examinedphysician-to-population ratios for 19 specialties (1978). The paperreports wide ranges for individual specialties, with internal medicineexhibiting almost a 20-fold variation compared to neurosurgery whichexhibits less than a 2-fold variation. Table 6 summarizes these findings.
Applied Management Sciences under contract with the Bureau of HealthProfessions examined ratios for 25 physician specialties. This secondmajor study compared the wide range of ratios by grouping the ratiosfound in the survey of 200 plus articles in the literature according tothe basis for the recommended estimate. Four groupings or clusters ofratios were developed. The first related to treatment of all medicalneeds, as determined by disease prevalence and morbidity data. The otherthree categories deal with requirements for analysis of the demand forcare, as derived from the opinions of health professionals; "effortanalysis" calculations of service requirements and manpower productivity;and observed staffing patterns in prepaid group practice (HMO) settings.
Furthermore, the factors each ratio takes into account are importantfor proper interpretation and are listed for each study reported.
Morbidity/Demand-Related Factors.
Prevalence of isease conditionsConsideratioL, of selected disease conditions.Backlog of untreated conditionsRequirements for preventive care
64
1
Table 6
RANGE OF PHYSICIAN SPECIALTY REQUIREMENTSEXPRESSED AS RATIOS OF PHYSICIANS PER 100,00 POPULATION
SPECIALTY LOW HIGH
Primary CareGeneral and Family Practice 11.5 55.0Internal Medicine 5.0 96.0Pediatrics 4.0 37.0
Medical SpecialtiesAllergy 2.0 4.0Cardiology 1.0 6.0Dermatology 2.0 6.0
Surgical SpecialtiesGeneral Surgery 8.0 15.2Neurosurgery 0.9 1.4Ophthalmology 0.5 5.8Orthopedic Surgery 1.0 6.3Otolaryngology 1.c 23.0Plastic Surgery O. 2.2Thoracic Surgery 0.5 1.1Urology 1.0 5.0
Other SpecialtiesAnesthsiology 2.0 12.0Neurology 1.0 5.0Pathology 1.0 6.9Psychiatry 2.0 54.4Radiology 3.0 13.0
Source:
From:
Review of Health Manpower Population Projections, DHEWPublication No. (HRA) 77-22, October 1976.
Physician Manpower Requirements, DHEW Publication No. (HRA)78-10.
65
1'
-- Quality of care-- Changing definitions of health-- Utilization rates (visits per week or year)
Supply/Productivity-Related Factors
-- Time required to produce services or visits-- Case loads per health professional-- Technological advances
Task delegation
Delivery System-Related Factors
-- OrganizationSetting
-- Role definition of health professionals (e.g., primary andspecialty care functions)
Other Factors
-- Patient subpopulations-- Reimbursement mechanism-- Ability to pay-- Urban/rural location
Type of Evidence on Factors
-- Observations of actual conditions-- Test or survey results-- Single source judgement
Table 7 summarizes the range of requirements ratios for 21 physicianspecialties and displays the wide ranges that exist for individualspecialties. Psychiatry,_for example,-exhibits.a 26-fold variation whileneurosurgery exhibits a two-fold variation. Ratios of primary carephysicians show less than a three-fold variation and pediatrics less than'a five-fold variation. In almost every case the professional judgementratios were higher than the needs-based or demand/productivity-based orHMO-based ratios.
The four specialties of general/family practice, general surgery,internal medicine, and obstetrics-gynecology were omitted from thesummary table and are presented in Table 8. General/Family Practiceratios of population to physician range from 2000 persons perpractitioner under a professional judgement-based ratio to 30,000 personsper practitioner in an HMO setting. General surgeon ratios vary almostsix-fold from 8,056 to 58,824 persons per general surgeon. The range of
66
internal medicine ratios is 1,042 persons per internist under a needs-L.ased model to 20,000 persons per internist under a professionaljudgement based ratio. Finally, obstetrics-gynecology ratios range from8,190 to 38,402 persons per obstetrics-gynecology under different HMOratios.
From the host of possible ratios, several federally adopted ratiosexist to determine areas with a defined shortage of physician services.In other words, these are accepted minimum ratios to trigger Governmentaction and do not necessarily reflect the lowest "acceptable" physicianto- population ratio. Further, the ratios adopted for Federal use wereadapted from ratios in the literature, and carry with them many of thecaveats detailed earlier.
In the proposed National Guidelines for Health Planning (P.L. 93-641,1979) is the following goal statement for primary care: "The supply ofprimary care health personnel in a community should be no less than theequivalent of one physician per 2,000 population. This ratio can beachieved by fostering the use of nurse practitioners and physician'sassistants." This sets forth a goal of a "minimum acceptable" ratio forprimary care health personnel. A primary care physician for the purposesof this goal has been defined to include general practitioners, familypractitioners, internists, obstetrician-gynecologists, and pediatricians.
In response to the Health Professions Educational Assistance Act of1976 (P.L. 94-484), criteria for the designation of Health ManpowerShortage Areas were estabJshed. Either geographic areas or populationgroups could be designated shortage areas for several types of healthmanpower classifications. Central to the criteria is the population-to-manpower ratio. For designation, the area must have a population-to-primary care physician ratio of at least 3,500 population per physician.The population count may be adjusted to reflect the area's age-sexcomposition and transient, seasonal tourist, and migrant populations.The physician count of non-Federal physicians practicing general orfamily practice, general internal medicine, general pediatrics, andobstetrics-gynecology may be adjusted for area interns and residents,foreign medical graduates, full-time equivalents of semiretiredphysicians, restricted practices, and the contribution of nursepractitioners and physician's assistants. In addition the unusually highneeds for primary care services, the capacity of existing providers, andthe resources of contiguous areas are considered.
The physician-to-population ratio has been a commonly accepted methodof defining physician manpower shortage areas since the passage of theHealth Professions Educational Assistance Act of 1965 (P.L. 89-290). Theratio has been progressively refined since 1965 to reflect moreaccurately the physician resources available to the population.Unfortunately the federally adopted criteria and standards cannot respondto the question of "what should be," and they present only a step towarddefining a minimum acceptable accessibility to services.
67
Table 7
SUMMARY TABLE; RANGE OF SELECTED "REQUIREMENTS RATIOS" BY OCCUPATION
Professional Judgement
Need Based RatioBased Ratio
Demand /Productivity
Based Ratio1010 Based RatioPop. per
Professional
Professional per Pop, per Professional per
100 000 Po . Professional 100,000 Pop.
Pop, per Professional per
Professional 100,000 Pop.
Pop. per Professional per
Professional 100,000 Pop.
ALL gDs 596 167.8 534.724 138.1-187.3 547-989 101,1-182.8 651-1,493 67.0-153.6Primary Care 752 -1,585 63.1-33.0 1,264 79.1 2,000 50.0 1,177-1,739 57.5-85,0Allergy25,000 4.0
40,000-250,000 0.4-2.5Anesthesiology8,457 - 14,000 7.1-11,8 10,764-17,544 5,7-9.3 25,641-55,556 1.8-3.9Cardiology
25,000-771000 1,3-4.0Dermatology
40,000-50,000 2.0-2,532,258. 333,333 0.3-3,1Gastroenterology
50,000 2.5Neurology
60,000-77,000 1.3-1.7107,000-125,000 0.8-0,9Neurosurgery
77,000-125,000 0.8-1.3 162,000 0.6 100,000-135,000 0.7 -1.0Ophthalmology20,000 5.0
24,400- 125,000 0.8-4,1Orthopedic Surgery20,000-351000 2.9-5.0 28,900 3.4 271800-76,900 1.3-3.6Otolaryngology25,000-331000 3.0-4.0 70,400 1.4 25,600-62,T1 1.6-3.9Pathology17,400- 33,000 3.0-5.7
19,700-111,111 0.9-5,1Pediatrics10,000 10.0
4,300-18,200 5.5-23.3Pediatric Surgery
780,227 0.1Plastic Surgery50,000-1501000 0.7-2.0
Psychiatry4,800-14,300 7.0-20.8
12,500-38,500 0.8-7.7Pulmonary Disease100,000 1.0
Radiology7,400-20,000 5.0-13.5
14,400-100,000 1.0-6.9Thoracic Surgery100,000-200,000 0.5-1.0-
Urology21,300 - 50,000 2,0-4.7
251600-90,900 1.1-3.9
From Review of Health ManpowerPopulation Requirements Standards,
DHEW Publication No. (HRA) 77-22
Table 8
RANGE OF SELECTED "REQUIREMENTS RATIOS"GENERAL/FAMILY PRACTICE, GENERAL SURGERY, INTERNAL MEDICINE, OBSTETRICS-GYNECOLOGY
General/FamilyPractice
General Surgery
Obstetrics-Gynecology
Internal Medicine
Needs BasedRatio
2,123
1,042-14,347
22,568
PopulationProfessionalJudgmentBased Ratio
2,000 4,000
9,436 - 10,000
3,846 - 20,000
10,000 - 11,000
Per PhysicianDemand/ProductivityBased Ratio
11,650
HMO Based Ratio
2,300 - 30,303
8,056 - 58,824
.2,192 - 14,286
8,190 - 38,46
From: Review of Health Manpower Requirements Standards, DHEW Publication No. (HRA) 77-22
69
TIME-TO-SERVICE STANDARDS
Health care is'said to be a right rather than a-privilege. As aconsequence geographic barriers to entry into the health care system aredeemed not acceptable and a geographic distribution of physician servicesconsonant with equity of access for the population is encouraged. Thegeographic accessibility of resources within an area depends on the tirand distance travelled to care. Particularly in rural areas, much efiucthas been directed toward reducing geographic and tempc.ral barriersseparating residents from facilities. Further, time and distance factorsfor urban and rural residents have acquired an even greater significanceas out-of-pocket costs are reduced through the growth of third-partypayments (Shannon et al., 1979).
No definitive studies have found to what extent lengthy travel times'impact upon decisions to forego needed available medical care (Weiss,1971; Cibcco and Altman, 1954; Bosanac, 1976; Parkin, 1979). However,geographically "inaccessible" populations have been characterized ashaving sociodemographic attributes associated with high medical needs(Bosanac, 1976; Aday and Anderson, 1975; DREW, 1979; Acton, 1973).Residents of rural and inner city areas have less access to physiciansthan residents of suburban or metropolitan areas_ excluding the innercity.
Early studies (Jehlik and McNamara, 1952) of geographic accessibilitystressed distance (mileage) rather than time, whereas more recent studieshave taken 4.nto account differences between rural and urban areas byplacing greater emphasis on travel time (SRI, 1968). Studies relating totravel time are divergent in method, although they usually deal withpatient preferences (Shannon, 1979; Hershey, 1975; Kane, 1969). Somestudies are large scale surveys, (NCHS, 1979; Aday and Anderson, 1975)and others are ideals proposed by planning groups (Wisconsin, 1972;Northern Virginia, 1980; Pennsylvania, 1975; Midland, 1978; SoutheastNebraska, 1978).
In the analysis of times-to-service, factors which should be takeninto account include mode of transportation (e.g. bus or car), roadsurface, speed, terrain of area, climate and population preferences. Itis generally recognized that travel times and distances increase withincreased specialization of care, although different groups definespecialty care differently. While consumer studies or surveys usuallyfocus only on sociodemographic-characteristics of the respondents, Kanelooked at how realistic the desire' and, perceptions of consumers were inrelation to travel patterns (Kane, 1969). He found that their desireswere generally congruent with the general trade and travel patterns ofthe community. As they described the distances and their-generalexpectations regarding willingness to travel to care, the responses
70
formed a continuum in which distance varied with the specialization ofthe service. The consumers, recognized that they would have to travelfarther to get many of the more sophisticated services, while desiringthe security of quick access to more basic services.
Use of time guidelines appears to vary widely, from a recognition oftravel time as an important planning consideration to actual employmentof standards (See Table 9). Based on a consensus of the literature,health planning guidelines, and health officials, a 30 minute travel timeto nonemergent care has become a standard in health planning. Primarycare manpower shortage area criteria suggest that utilization of medicalservices is seriously affected by travel times greater than 30 minutes(DREW, 1978). A nt..ber of States have recommended that a largeproportion of the population be within 30 minutes to nonemergency care.Public hearings in Pennsylvania produced standards for community healthinstitutions which indicated that "a maximum travel time of 30 minutes(nonemergency) was appropriate." (Pennsylvania, 1975) A Wisconsin HealthTask Force stated that "a broad range of health services should beavailable for all citizens as soon as practicable, within a one-waytravel time of not more than 30 minutes." (Wisconsin, 1972). TheKentucky State Department of Health also recommended a 30 minute traveltime standard for hospital care (Bosanac, 1976).
The Federal 30 minute standard is utilized as a mechanism to define a"rational service area." A shortage area must be "a portion of a county,or an area made up of more than one county, whose population, because oftopography, market or transportation patterns, distinctive populationcharacteristics or other factors, has limited access to contiguous arearesources" as measured by a travel time greater than 30 minutes.Distances used to correspond to 30 minutes travel are: (1) 20 milesunder normal conditions with primary roads available, (2) 15 miles inmountainous terrain or in areas with only secondary roads and (3) 25miles on Interstate highways or in flat terrain. It is suggested thatinformation on the public transportation system be used to determinetravel times in urban areas (DREW,. 1978).
Data on time-to-service for emergency care are more variable anddetailed and many standards have been proposed. The Texas RegionalMedical Program recommended minimal time-to-service requirements for 14medical ailments, most of which were emergency (Permian Basin, 1975).Using the Delphi method, members of the medical profession made judgmentsranging from a minimal requirement of less than 15 minutes for a traumainvolving head injury to less than 30 minutes for stroke. Alsorecommended were services required for, treatment of the ailment rangingfrom a phar c..y to an intensive care unit. Most of the morbiditiesrep:Fixed a hospital with specialized facilities and emergency capability.This study reiterated the viewpoint represented by Dr. Richard Wilburthat fine-tuning in cases of emergency treatment beyond 15 minutes is notproductive; when dealing with the Emergency Medical System (EMS), thetype of condition should be specified (Wilbur, 1980). Dr. Wilbur also
71
1
stated that except for acute emergencies, the length of time taken toobtain service is usually less important than the quality of theservice. The quality is dependent on'a number of factors beyond thepresence of a qualified specialist.
A review of a number of health systems plans indicates that most seemto analyze travel times to emergency care as well as set atts.nablegoals. However, the range of times and goals is rather broad, dependingupon the metropolitan or nonmetropolitan nature of the area. In NorthernIndiana, for example, a goal of the health systems agency (HSA) is that"EMS services should be available within a travel time of 60 minutes"while response times for emergency vehicles and personnel ranged from 3-4minutes in urban areas to 20-30 minutes in other rural areas. Many HSAplans recommend a 10minute response time for EMS in urban areas and a.20minute response time for rural areas.
Obstetrical standards for hospitals are rarely described in theliterature although traveltime and distance standards to short stayhospitals are frequently cited. The National Advisory Commission onHealth Manpower reported that in 1967 "simple physical distance was not amajor barrier for either urban or rural residents. Hospital facilitieswere within a 25 mile distance of all but two percent of the population,and only .01 percent had to travel more than 50 miles. In the urbanareas, none of the population was more than 10 miles from a hospital."(National Advisory Commission on Health Manpower, 1967) In Kansas, shortstay hospitals are accessible to 90 percent of residents within 30minutes from home or work. Other HSA plans cite similar 30minutestandards to hospital care.
The Health Systems Plan for Southeast Nebraska recommends that "asource of inpatient obstetric services should be available within 30minutes travel time." The Health Planning Council of the Midlandsrecommends .a 30minute travel time for 80 percent of the population forobstetric services. The Committee on Perinatal Health recommendsservices and facilities for caesarean section capability within 30minutes at any hour.
Even less information on traveltime standards is available onpediatric services. The Omaha HSA recommended that services should beaccessible within 1 hour for 80 percent of the residents of an area. Noother information could be found. In addition, very few articles orhealth system plans provide information on traveltime to surgery.\-?TheOmaha Health Council recommends 2 hours for 80 percent of area residentsto neurosurgery services, open heart surgery, organ transplant, andorthopedic surgery. However) it also suggests that anesthesiologyservices, general surgery, and diagnostic radiology be available to 80percent of the population within 30 minutes. In addition a study iscurrently being undertaken by Dr. Clark Watts at the University ofMissouri Medical Center which will indicate that "for good medical carethe maximum a patient should be from a neurosurgeon is 100 miles--30
72
Emergency Care
Time
30 min Rural
10 min Urban
10 min Urban
20 min Rural
(response time)
20 min
10 min Urban
20 min Rural
15-30 min
Trauma-head-15
Trauma-Mult-15
Gunshot -15
Poisoning -15
Myo Infarct-I5
Burn -30
CYN Hemorr -30
Asthma -30
Stroke -30
30 min
(802)
Source
EMSS Act of
1973
Northern
Indiana HSP**
1977
Southeast
Nebraska HSP**
1978
Eastern
Connecticut HSP**
Permian
Basin
Texas
1975
Midlands
Omaha HSP**
1978
60 min Northern
(EMS Services) Indiana HSP**
1977
Table 9
RECOMMENDED TIME-TO-SERVICE STANDARDS BY
FIVE MAJOR TYPES OF CARE AND SOURCE*
OB/GYN
Time Source Time
Child Care
Time Source
1 hr Omaha HSP**
Adult Services Surgical Services
Source Time Source
30 min Southeast
Nebraska
1978
30 min 1/ CNN, 1977***
30 min 2/ Omaha HSP**
30 min 3/ No, VA, HSP**
1979
30 min 4/ No. VA HSP**
'45 1979
1 hr
30 min Florida HSP**
(90 X Peak 1978
capacity
for hosp)
1/ Cesarean Section
2/ 80% of population
3/ Hospital Services
4/ 2000 delivery hosp 30
1000 delivery hosp 45
500 delivery hosp 1 hr
20 min Urban
40 min Rural
30 min
No Indiana HSP** 50 min (Radi-
1977 ology)
Southeast
Southeast
Nebraska
HSP**
Nebraska HSP** 2 hrs OmahaI/
30 min 2/ Midlands
30 HEW HSP**
1978
2 hrs U of Mo
30 Wisconsin HSP** Med Cntr
1972 1980
30 PA 1972 2 hrs Northern
Virginia.30 E. Conn HSP** HSP**
1979
30 KY 1972
30 No, VA HSP**
30 (FP) Omaha HSP**
30 (IM)
30 N. Central
GA 1978 HSP**
*There is a paucity of references by aetailed specialty groupings, and they will not be included here.
** HSP Health Systems Plan
*ftCommittee on Neonatal Health, March of Dimes.
11.
I/ 80% of area residents
Neurosurgery, Open
i Heart, Organ,
) Orthopedic
2/ Anesthesiology, general
surgery and dia'gnoayic
Radiology
minutes by helicopter, 2 hours or less by ambulance. Every communitythat had a neurosurgeon was identified, and a radius of 100 miles waschosen. It was found that there is essentially no place in the country(excluding four areas) that was not covered." (Watts and Adelstein, 1980)
What are the linkages between physician-population ratios anddistance as measures of access? Sloan's investigations looked atvariations in the physician population ratio and compared them toobserved travel differentials (Sloan, 1977). He found that communitiesin which the patient's time commitment was the greatest tended to be themetropolitan areas. Further variations in the physician population ratioexplained only a small proportion of travel (and waiting) timedifferentials. Yet Bosanac concluded that various travel time standardscan assist'in refining standard measures of physician availability(Bosanac, 1976).
Health Manpower Shortage Area criteria attempt to utilize both thephysrcian-population ratio as well as the time-to-service concept indefining underserved areas for specific services. These criteria couldbe expanded to include general surgery,, emergency services, obstetricalservices, and pediatric services and standards could be developed likethose which have already been developed for shortage area criteria indental manpower, psychiatry, vision care, podiatry, pharmacy, andveterinary medicine.
RECOMMENDATIONS
RECOMMENDATION 9: GMENAC recommended that five basic types of healthcare services should be available within some minimum time/accessstandards: adult medical care, child care, obstetrical services,surgical services, and emergency services. In. order to monitor thegeographic distribution of physicians, GMENAC recommends that aminimum acceptable physician-to-population ratio for all areas in the
be 'established. It was recommended that 50 percent of theGMENAC Modeling Panel ratio of physicians specialists per 100,000 for1990 be established as the minimum acceptable ratio for all areas.
RECOMMENDATION 10: GMENAC recommends maximum travel times of 30minutes for emergency medical care, 30 minutes for adult med'r.c1caret 30 minutes for child medical care, 45 minutes forcare, and 90 minutes for surgical care services for 95 percent of tiepopulation in 1990, recognizing that unusual circumstances may arisewai:h make these travel times impc Bible to achieve for all areas.
RECOMMENDATION GMENAC recommends that the definition of healthmanpower shortage area include minimum physician/population ratiosand minimum travel 'times to service for general surgery, emergencymedical services, and obstetrical services.
74
1"
VI. TAXONOMY OF STRATEGIES TO IMPROVE THEGEOGRAPHIC DISTRIBUTION OF PHYSICIAN SERVICES
This report began by summarizing the recommendations of theGeographic Distribution Technical Panel (Part I). Part II presented thecharges of the Panel as well as its history. Charge No. 1, covered inPart III, was a presentation of data on the geographic distribution ofphysicians by county accompanied by a literature review on physicianlocation choice, and reasons for geographic variation. A discussion ofthe implications of procedure rate variations on manpower policy (ChargeNo. 3) followed a review of the evidence ua variations in the rates bY,small area (Charge No. 2) and was presented as Part IV. Part V addressedissues of equity, access and the proper and adequate distribution ofphysicians and reviewed the literature on physician-population ratios andtimes to service. Part VI will describe the "Taxonomy of Strategies toImprove the Geographic Distribution of Physician Services," theculmination of all of the work of the Panel, including Charges 4-8.
The attached Taxonomy presents a set of action orientedrecommendations from the Geographic Distribution Technical Panel. Therecommendations are strategies which can be directed at one of threelevels: undergraduate and/or graduate medical educational institutions,medical students and/or residents, and practicing physicians.
For each recommendation, the format described below was followed: first,a statement of the recommendation followed by a brief statementexplaining it in more detail and then the advantages and disadvantages orpros and cons of each recommendation are presented. A complete list ofcitations, grouped with each recommendation, appears at the end of thesection.
UNDERGRADUATE/GRADUATE MEDICAL EDUCATIONAL INSTITUTIONS (INCLUDINGTEACHING HOSPITALS)
RECOMMENDATION 15: There is some evidence that selective admissionspolicies may improve the geographic distribution of physicians. Anationally mandated alteration in admissions policies is notrecommended at this time; further study into the location decisionsof students with particular ethnic or sociodemiographic characteris-tic,is recommended. Selective admissions policies have been used toincrease the likelihood that medical students will choose practicewithin a State or in an underserved area of .a State by grantingpreferential admissions treatment to in-State residents or applicants
75
with particular backgrounds or personal characteristics. Selectiveadmissions policies may be used in conjunction with loans, servicecommitments, and the like. At the present time, outcomes of theseprograms do not warrant a nationally mandated policy.
Pros
Cons
There has been adequate time to assess the results of selectiveadmissions programs (this strategy has been used since 1948);however, data are conflicting and additional evaluation isneeded. Some programs have been successful in adding to the'numbers of practitioners in rural areas of particular 'States fromthe pool of students selected on a preferential basis.
This strategy is theoretically an attractive policy device.
These policies have practical problems. Causal relationshipsbetween individual traits and physicians.' career and locationdecisions have not been clearly established.
There is a, high drop-out rate of pre' -entiel admissions programparticipants.
-- These programs are susceptible to capricious management.
The costs per retained physician are high.
Questions as to the legality of preferential admissions policieshave been raised.
Sources
Hadley, V. New York: (ed). Medical education financing policy analysisand options for the 1980s. New York: Prodist, 1980.
Stefanu, C. et al. Selection of Primary Care as a Medical Career:Demographic and Psychosocial Correlates. S Med J 73:924-927, July 1980.
Cuca, J. 1978 U.S. Medical School Graduates: Practice SettingPreferences,. Other Career Plans, and Personal CharacteristiCs. J MedEdtic 55:465-68, May 1980.
RECOMMENDATION 16: Economic incentives (such as tax credits anddeductions) and/or the provision of higher payment levels forservices as an inducement for physicians to establish practice inunderserved areas, 'should be explored.
76
11;
Pros
Cons
While a major theme in specialty and location research is thatphysicians are unresponsive to differences in income opportunities,this assertion remains largely unproven. This strategy is similar toother financial incentives such as bonuses, direct grants,scholarships, and loan forgiveness which change the net present valueof the physician's income stream. Use of this mechanism wouldincrease net present values associated with practicing in underservedareas relative to those generated in adequately served areas. Incometax incentives could be provided at the Federal or State level.(Hadley, p. 195). This mechanism might be structured through taxcredits, e.g., freeing a certain portion of the physician's incomefrom the income tax, analogous to the tax ::redits allowed during 1978for investment in insulation of residential homes for the purpose ofenergy conservation. Alternatively, the tax rate applied to aportion of a physician's' income might be lowered. Reimbursementschemes might be changed to reflect higher payments for provision ofservices in underserved areas.
The pool of physicians who might be affected by a tax incentiveprogram is much larger than the number of medical studentseligible under loan or scholarship programs (Hadley, 1980,pp. 195-96).
The impact of a tax benefit program can be achieved more rapidlythan lo-a or scholarship programs (Hadley, 1980, p. 195-96).
Thess mechanisms relate directly to the locationphysicians rather than influencing the amount orservices a physician would deliver.
See pp. 103-105, Recommendation 30.
decision oftypes of
-- There is no precedent in the use of these mechanisms.
It is not certain how large a tax advantage would be needed toattract physicians into underserved areas (Hadley, 1980,pp. 195-97).
See pp. 103-105, Recommendation 30.
Sources
Eichenholz, J. Reducing the supply of physicians in overserved areas(unpublished). June 13, 1978. Presented at Institute of Medicine,Health Manpower Policy Meeting, Washington, D.C. June 21, 1978.
77
Eichenhollz mentions another type of tax incentive (disincentive),namely, encouraging graduated state licensure taxes and graduatedtaxes on issuance of narcotics numbers to all physicians.
Hadley, J. "State and local financing options," in Medical EducationFinancing, New York: Prodist, 1980.
Ernst, R.L. and Yett, D. "Econometric studies of the geographicdistribution of physicians," in Physician Specialty and LocationChoices Final Report on Grant #3330 from the Robert Wood JohnsonFoundation, Human Resources Research Center, Accepted forPublication, Lexington, M.A., D.C. Heath & Co., May 29, 1979.
RECOMMENDATION 17: Demonstration projects should be developed andevaluated to determine the impact of differential rates ofreimbursement for technology-intensive versus time-intensive(counseling, patient education) services upon the geographicdistribution of physicians and services.
Research evidence indicates that differential economic valuation ofmedical care services implicit in reimbursement policies may havesignificant implications for the equitable distribution of healthmanpower. Presently there are financial incentives to performin-office medical procedures and laboratory tests, which are oftenprovided more frequently in large medical centers in urban areas.Even within primary care fields, substantial financial incentivesexist to perform procedures and tests rather than personal services.
Pros
For example, using a standard conversion of the 1974 edition ofthe California Relative Value Studies (CRVS), a physician cancharge $48.50 for a periodic annual physical examination, whichis commonly scheduled for 45 minutes. Yet, performance andinterpretation of an electrocardiogram, which usually requires afew minutes of the physician's time, carries a $28.00 charge. AsPetersdorf emphasizes, current national health manpower policiesemphasizing generalists rather than specialists are out of phasewith current reimbursement mechanisms which preferentially rewardspecialty care (Schroeder and Showstack, 1978, p. 297).
A recent report suggests, for example that an internist canincrease net income up to threefold by changing the service mixfrom one emphasizing consultative service to one emphasizingtechnological procedures and tests (Schroeder and Showstack,1978, p. 294),.
78
Another illustration is the fact that in some hospitals anall-night vigil for the treatment of diabetic, acidosis or hepaticcoma is rewarded with a standard $8.00 fee for professionalcare. In contrast, a 10-minute endoscopy in the same institutionearns $100.00 (Seldin, 1976, p. 2'7).
Relative valuation of services in benefit schedules shouldencourage physicians, or at least not discourage them, to providethose services deemed socially needed and beneficial, i.e., timeintensive primary care services such as patient histories,physical exams, patient education, and counseling.
Comprehensiveness of care, including health education andpreventive measures, is an attribute essential to the practice ofgood primary care. The provision of a broad range of services,including services for basic medical problems, psychosocialproblems, and health education, distinguishes the primary carepractitioner from the secondary care practitioner and thereferral specialist.
Cons
There are two major argumen*:s against offering third-partypayment to physicians for providing health education andcounseling services. First, it is difficult to assess theefficacy of many such services; second, it is possible that totalhealth care expenditures may rise in the short run as a result ofthis, change in reimbursement. The probability of an immediateincrease in expenditures for health care must be weighed againstthe probability of future savings, both economic and in terms ofhuman suffering.
Sources
Schroeder, S.A. and Showstack, J. Financial incentives to performmedical procedures and laboratory tests: illustrative models ofoffice practice. Med. Care, 16:289-98, April 1978.
Seldin, D. W. Specialization as scientific advancement and over-specialization as social distortion. Clin. Res. 24:245-48, October,1976.
RECOMMENDATION 18: It is recommended that practicing physicians andfaculty Convey to students that the practice of medicine can be 4delivered in a variety of geographic settings, including both ruraland urban shortage areas. As a means of accomplishing this, urbanand rural preceptorships for medical students should be continued andexpanded schools with an interest in monitoring such programs.
79.
1-I- V
Role model development is aimed at changing the educationalenvironment to stress the positive aspects of primary care practiceand practice in underserved areas. Federal, State and privatepreceptorship programs are intended to influence physicians toestablish practices in urban and rural underserved areas.
Pros
Cons
Analysis of the locational decisions of 486 physicians and theirspouses reveals that physicians now in practice in shortage areaslink their choice to faculty role models in their residencyprograms and report that practice in shortage areas was valuedhighly by their faculty (Wilson, 1979).
A 1978 assessment of preceptorship programs by Applied ManagementSciences (1978) reported that preceptor location seemed to have asignificant impact on the practice location of preceptees (Referto Table 10).
Programs sponsored by family medicine departments have been morelikely than those in other specialties to emphasize primary careand rural service in their preceptorships, and to have preceptorslocated in underserved areas.
Programs located in the South have had a significantly higherpercentage of preceptors located in definitely underservedareas. The distribution of preceptors by specialty also differedaccording to region (Applied Management Sciences, 1978; refer toTable 11).
Little significant evaluative work has been performed oneducational innovations such as the influence of role models.
There is contradictory evidence as to the efficacy of ruralpreceptorship programs in'influencing physicians to practice inrural areas. The results of a study conducted by Steinwald andSteinwald (1975) indicate that the overall impact ofpreceptorship programs and other rural training programs is mostpronounced with respect to urbanreared physicians in nonprimarycare specialties.
A good proportion of preceptorship programs (39.1 percent) arenot located in medically underserved areas. Only 16 percent arein definitely underserved areas; 80 percent are not in criticalshortage areas (Applied Management Sciences, 1978; refer to Table12).
80
Programs do not have uniform goals (whether they are trainingprograms, recruitment programs, or both), and vary as to lengthand timing of preceptorship, size of program, and voluntary ormandatory character of preceptorship experience (GAO, 1978, p.73).
-- Additional data are needed. Estimation of program influence isdifficult because programs are relatively new and long term dataare not available. Cost effectiveness needs to be determined.
Sources
U.S. General Accounting Office, Progress and Problems in Improving theAvailability of Primary Care Providers in Underserved Areas. Reportto the Congress by the Comptroller General of the United States.August 22, 1978. Chapter 6, Remote Site Training Programs.
Wilson, S.R. An analytical study of physicians' career decisionsregarding geographic location. Draft Digest of the Final Report,American Institutes for Research Contract No. HRA-231-77-0088,Prepared for DHEW 1979.
Applied Management Sciences. An assessment of the influence of medicalpreceptorships and other factors on the career choice and medicaleducation of physicians. Unpublished final report. DHEW ContractHRA-231-76-0040. Executive,Summary, March 31, 1978.
Barish, A.M. The influence of primary care preceptorships andother factors on physicians' career choices. Public Health Rep., 94,36-47, January-February, 1979,,
Sax, E. Review of Incentive Programs Promoting Practice in UnderservedAreas. New York: National Health Council, February 1976, pp. 8-15.
Steinwald, B. and Steinwald, C. The effect of preceptorship andrural training programs on physicians' practice location decisions.Med. Care 13:219-229, March 1975.
81
co
Table 10
PRACTICE LOCATION PREFERENCE OF STUDENTS AND RESIDENT
PRECEPTEES BY LOCATION OF PRECEPTOR PRACTICE
Practice Location of Preceptor
Preferred
Practice
Location
Students Residents
Inner-city Rural Other urban Total Inner-city Ruril Other-urban Total
Inner-city 8(28.6)* 5(17.8) 15(53.6) ':(,'6.5)** 1(11.1) 2(22.2) 6(66.7) 9( 4.8)
Low income (44.4)** ( 2.5) ( 7.1) (14.3) (1.9)( 5.4)
Small town/ 7( 3.2) 130(59.6) 81(37.2) 218(50.6) 4( 4.9) 38(46.9) 39(48.1) 81(43.3)
rural area (38.9) (64.4) (38.4) (57.1) p5 (55.9) (34.8)
Other urban 3( 1.6) 67(36.2) 115(62.2) 185(42.9) 2( 2.1) 28(28.9) 67(69.1) 97(51.9)
(16.7) (33.1) (54.5) (28.6) (41.2) (59.8)
TOTAL 18( 4.2)* 202(46.9) 211(49.0) ' 431(100.0) 7( 3.7) 68(36.4) 11 59.9) 187(100.0)
* Row percent.
** Column percent.
Applied Management Sciences, An Assessment of the Influence of Medical Preceztorshipe and Other Factors on the Career Choice and MedicalEducation of Physicians. Unpublished final report of Contract RRA- 231 -76 -0040. Executive Summary, March 31, 1978.
Table 11
DISTRIBUTION OF PRECEPTORS BY SPECIALTY ACCORDING TO REGION
IN WHICH PROGRAM WAS LOCATED
Preceptor
Specialty
Proportion of Preceptors in Each Specialty
Northeast
North
Central South West
Family medicine* 39.7 65.7 73.7 58.0
Pediatrica* 26.4 7.2 7.4 13.2
Internal medicine 24.1 14.0 9.3 13.8
Other specialties* 9.4 12.8 9.5 15,0
Unknown 0,4 0.6 0,1 0.0
* Significant at .05 level
Applied Management Sciences, An Assessment of the Influence of Medical Preceptorships and Other Factors on
the Career Choice and Medical Education of Physicians. Unpublished final report of Contract HRA-231-76-0040.
Executive Summary, March 31, 1978.
Table 12
DISTRIBUTION OF ACTIVE PRECEPTORS BY SPECIALTY, FORM OFPRACTICE, AND MEDICAL SHORTAGE DESIGNATION OF PRACTICE SITE
Preceptor Characteristics Number Percent
SpecialtyFamily medicine 2095 62.4Pediatrics 392 11.7Internal medicine 472 14.0Other 386 11.5Unknown 14 0.4
Total 3359 100.0
Medically Underserved AreaDefinitely underserved 536 16.0Perhaps underserved 1509 44.9Definitely not underserved 1246 37.1Unknown 68 2.0
Total 3359 100.0
Critical Shortage AreaDefinitely underserved 154 4.6Perhaps underserved 449 13.4Definitely not underserved 2428 72.3Unknown 328 9.7
Total 3359 100.0
Applied Management Scien'ces,An Assessment of the Influence ofMedical Preceptorships and Other Factors on the Career Choiceand Medical Education of Physicians. Unpublished final report ofContract HRA- 231 -76- 0040.. Executive Summary, March 31, 1978.
RECOMMENDATION 19: Given the geographic distributional patterns offamily practitioners? graduate medical education programs in familymedicine should continue to be supported as a strategy to increaseprimitry care services and their availability in certain geographicareas of underservice.
Federal support to hospitals and medical schools to develop, expand,and improve residency training programs in family medicine should beprovided because data presently available show that graduates offamily practice residency programs, more than physicians in any otherspecialty, tend to locate their practices in underserved areas. Ifgraduates of residency programs in internal medicine and pediatricsmake location choices similar to those of family practice physicians,then the recommendation should be broadened to include support forresidency programs in these two primary care specialties.
Pros
Cons
Unlike physicians in general, family practitioners areconcentrated in rural areas. In 1976, 54.9 percent were locatedin cities with populations of 30,000 or less, and 11.1 percent incities with populations between 2,000 and 5,000. (Institute ofMedicine, 1978, p. 52).
Establishment of departments of family medicine reflectsincreasing emphasis on primary care in medical school curricula.
Family practitioners are usually trained in ambulatory caresettings. It is possible that these physicians are less likelyto require a hospital in close proximity as a prerequisite forpractice.
-- Since these programs are relatively new, results are justbeginning to be discerned.
Sources
Institute of Medicine, A Manpower Policy for Primary Health Care.National Academy of Sciences: Washington, D.C., May 1978.
Geyman, J.P. Fam4ly practice in evolution: progress, problems, andprojections. New Engl. Med. 298:593-601, March 16, 1978.
U.S. General Accounting Office, Progress and problems in improving theavailability of primary care services in underserved areas. Reportto the Congress by the Comptroller General of the United States.
85
1 ti
Public Law 94-484-, the Health Professions Educational Assistance Act of1976. October 12, 1976.
Pros
RECOMMENDATION 20: Incentives should be created to broaden residencyeducational experiences to encompass training in underserved areas,provided the appropriate resources are available and standards ofeducation of the relevant accrediting body are met.
Because of research findings which suggest that the frequency andrecency of a medical school graduate's contact with a specificgeographic area influence the probability of practice in the area,residency experience in a geographic area has been considered as amechanism for equalizing the geographic distribution of physicians.It was one of the strategies considered in the debates about Federalhealth manpower legislation in the early 1970s. The absence ofclear-cut goals for geographic redistribution, uncertainty aboutlinkage of location of residency to location of practice, and lack ofmethods and methodology for redistribution of residencies on ageographic basis led to the formation of GMENAC.
A variety of studies have lent support to the hypothesis thatpolicies designed to attract physicians to a geographic areathrough residency training have a higher payoff in terms ofphysician retention than any other single medical training"event" (Statewide House Officers Training System, 1979, p.6).
Satellite training as part of a residency program is particularlysuitable for attracting physicians in primary care specialties tounderserved areas.
Both in rural and inner city areas, house officer trainingprograms may be established in independent community centers andsatellite facilities under the direct control of hospitals.Change in the structure and process of medical education appearsto be a necessary component of any attempt to increase thenumbers of physicians in these areas (Wilson, 1979).
Besides preparing physicians for practice in noninstitutional,low technology settings, this strategy would increase the supplyof services in underserved communities.
Satellite centers which have a graduate medical educationtraining component provide practicing physicians with anopportunity to establish collegial relationships with assignedhouse officers and program faculty.
86
Cons
The development of residency programs may not be feasible inarea's where the population base,is inadequate; it may bedifficult to attract faculty and costly to refurbish or build newfacilities.
Data on local needs and resources must be developed to facilitatethe planning process for extension of residency programs.
Correlations between residency location and practice locationwhich have been observed in the literature have used States asthe unit of geographic analysis. This literature does not speakto the linkage between residency location and practice locationwithin a given State (Statewide House Officers Training System,Inc., 1979, p. 10).
Sources
Statewide House Officers Training System, Inc., Alternative strate-gies for addressing the issues of physician specialty and geo-graphic distribution through graduate medical education, May 16,1979. Appendix H to Policy Proposal for Graduate Medical Educationin Michigan, September 15, 1979.
California Office of Statewide Health Planning and Development, Graduatemedical education in California: A position paper. Health andWelfare Agency, State of California, December 1978.
Wilson, S.R. An analytical study of physicians' career decisionsregarding geographic location. Draft Digest of the Final Report,American Institutes for Research, Contract No. HRA-231-77=0088prepared for DHEW, NCHSR, 1979.
Recommendation 21: Data suggest that nonphysician health careproviders favorably affect the distribution of medical services bytheir tendency to select shortage area locations more frequently thinis the case with physicians. It is recommended that nonphysicianhealth care provider training programs should continue to besupported f r this reason.
Because these ograms train health care providers to deliverservices traditionally provided by physicians only and improve theaccess to those services without compromising quality of care andoften with considerable cost savings, continuation is appropriate.
87
I Or;tiv
Pros
Cons
-- Distributional patterns exhibited by nonphysician health careproviders are favorable towards improving access to primary careservices in underserved areas. Physicians' assistants and MEDEXtend to locate in nonmetropolitan areas more than physicians,while nurse practitioners work more frequently in inner citycommunities (Congressional Budget Office, 1979, p. 1; Weston,1980; Sultz, 1978).
Nonphysician health care providers cost between $10,000 and$12,000 to educate versus approximately $60,000 for a physician(Congressional Budget Office, 1979, p. 26).
Under assumptions of national health insurance, changes inreimbursement policies, and increased support of healthmaintenance organizations, the potential of this strategy forcost effective provision of primary health care services is good(Congressional Budget Office 1979, xii-xiii).
-- It is difficult to estimate the effect of locationalcharacteristics of nonphysician health care providers because ofthe relatively small number of this type of provider in practice(Morris, 1977, p. 1049).
-- Continuation of programs which would increase the supply ofnonphysician health care providers may be contrary to costcontainment efforts.
Training of nonphysician health care providers can be supportedonly when coordinated with other programs to assure the effectiveutilization of these personnel (refer to Section III, Practicingphysicians).
Support of the training programs should be made contingent onstrengthening their placement mechanisms to channel graduatesinto underserved areas.
-- Data are not available to measure retention of nonphysicianproviders in underserved areas.
Sources
Morris, S.B. The distribution of physician extenders. Med. Care15:1045-57, December 1977.
88
Sultz, H.A. et al. Longitudinal study of nurse practitioners. Phases I,II, III. Bureau of Health Manpower Division of Nursing DHEW Publ.No. (HRA) 76-43, 78-92, 80-2.
Weston, J.L. Distribution of Nurse Practitioners and PhysiciansAssistants: Implications of Legal Constraints and Reimbursements,Pub. Health Rep 95:253-58, May-June, 1980.
Congressional Budget Office, Physician extenders: their current andfuture role in medical care delivery. background paper, Washington,D.C., April 1979.
Scheffler, R.M.; Yoder, S.G.; Weisfeld, N. and Ruby, G. Physicians andnew health practitioners: Issues for the 1980's. Inquiry16:195-229, Fall 1979.
RECOMMENDATION 22: Decentralized medical education programs such asWAMI (in Washington, Alaska, Montana, and tdaho) and WICHE (WesternInterstate Commission for Higher Education) were developed tocoordinate medical education and placement programs in a relativelyisolated and sparsely populated region. These types of programs havebeen effective and attention should be given to their replicability.Decentralized medical education is aimed at changing the educationalenvironment,,to stress the positive aspects of primary care practice.Programs such as WAMI and WICHE provide'a coordinated approach forincreasing the number of primary care physicians trained and forredressing geographic maldistribution problems in Statesparticipating in these programs.
Pros
Results of a study of the WAMI program indicate that residentswho completed community clinic rotations have increasingly chosenpractice in small towns (Schwarz, 1979, pp.388-389; GAO, 1978, p.82).
Cons
-- It may not be possible to isolate the "pure" effects of a changededucational environment on the career decisions of medical schoolgraduates.
While programs such as WAMI and WICHE ave been shown to have amodest effect on physicians' prac,Ice locations, it is unlikelythat these programs will have a maj influence on the geographicdistribution of physicians nationwide.
89
%.1
Sources
Schwarz, M.R. The WAMI program: A progress report (MedicalEducation) West. J. of Med. 130:384-390, April, l979.'
RECOMMENDATION 23: GMENAC encourages the medical profession,' throughits training program directors and various specialty societies, inmaking decisions as to residency training programs to consider, inaddition to the duality of residency programs, the aregate numberof programs, their size, and the geographic distribution of theirgraduates to better meet national and regional needs. Trainingprogram directors can assist in redistributing numbers of individualsamong specialties and geographic areas according to goals establishedfor meeting national and regional needs. Clearly, successes reportedin some specialties indicate that program directors and specialtysocieties can influence the specialty distribution of physicians.
Pros
Cons
As a substitute for or in conjunction with public sectorinitiatives to solve geographic maldistribution problems, theprofession can exercise leadership by influencing the specialtyand geographic distribution of practicing physicians.
This strategy can be carried out by the voluntary sector.
Currently, most specialty societies are engaged in studies of theadequacy of numbers of physicians in their specialties. Thesedata could assist in the overall determination of relativesurpluses or inadequacies among specialties (GAO, 1978, p. 32).The Placement Bureau of the American Academy of Dermatologistshas printed and distributed a map of the distribution ofdermatologists by zipcode area to all training programs in orderto encourage trainees to consider practicing in relativelyunderserved areas.
While considerable agreement exists on the need for additionalprimary care physicians, opinions differ as to what constitutes asufficient supply in other specialties (GAO, 1978, p. 22). Theproblem of defining an adequate supply is complicated by the factthat certain specialists, other than primary care specialists,provide substantial amounts of primary care (GAO, 1978, pp. 10,21; Aiken et al., 1979). Furthermore, there appears to be somequestion about the extent to which specialists and subspecialistsshould be relied upon to provide primary care.
90
Sources
Aiken, L.H. et al. The contribution of specialists to the delivery ofprimary care: a new perspective. New Engl. J. of Med.300:1363-1370, June 14, 1979.
U.S. General Accounting Office, Are enough physicians of the right typestrained in the United States? Report to the Congress by theComptroller General of the United States. May 16, 1978.
American Medical Association, Profile of Medical Practice 1978. Chicago:American Medical Assocation, Center for Health Services Research andDevelopment, 1978.
MEDICAL STUDENTS AND/OR RESIDENTS
Pros
Cons
RECOMMENDATION 24: The National Health ServiTA Corps (NHSC) and theNHSC Scholarship Program for increasing the availability of primarycare physician services an designated health manpower shortage areasimpact favorably on the geographic distribution of physicians;therefore, NHSC and the NHSC Scholarship Program should continue tobe supported.
The objectives of the NHSC werf to improve the availability of healthservices in underserved areas and to encourage the permanentsettlement of health professionals in the areas in which they wereplaced.
The NHSC service and scholarship program have definitely improvedgeographic maldistribution problems by increasing access toprimary care medical services in health manpower shortage areas.
The potential of the NHSC program for increasing the productivityof NHSC physicians in existing sites and expansion of the programinto a variety of underserved areas (e.g., inner cities) mayenhance its future success (GAO, 1978, p. 40).
Long-term retention rates have been poor, but appear to beimproving.
Adequate data are not yet available to estimate to what extentthe goal of increasing per 2nt settlement (beyond the 2-yearinitial service period) h oeen achieved (Modderman, p. 26).
91
1 00%..." 4,
The assignment of a minimum of two physicians to underservedareas appears uneconomical, although necessary. (GAO, 1978, p.39). The trade-off between productivity and retention should beconsidered.
Data which would allow the comparison of costs and benefits ofthis strategy with others are not available and should bedeveloped.
The strategy has so far not been successful in recruitingphysicians to more remote, less populated areas (Modderman, 1979,p. 17-19).
Sources
Kane, R., Olsen, D., Wright, D., Kasteller, J., and Swoboda, J. Changesin utilization patterns in a national health service corpscommunity. Med. Care 16:828-36, October 1978.
Michelsen, E. and Cronquist, N. Scholarships now for service later:Meeting the health manpower needs of medically underserved areas.Public Health Rep, 94:11-15, January-February 1979.
Modderman, M.E. The potential impact and the record to date of thenational health service corps (NHCS), area health education centers(AHEC) and rural health clinics in terms of eliminating grossphysician distribution problems. Prepared for the GeographicDistribution Technical Panel, GMENAC, 1979.
U.S. General Accounting Office, Progress and problems in improving theavailability of primary care providers in underserved areas. Reportto the Congress by the Comptroller General of the United States.August 22, 1978. Chapter 3, National Health Service Corps.
Sax, E. Review of incentive programs promoting practice in underservedareas. National Health Council, February 1976, pp. 22-23.
Recommendation 25: Governmentally sponsored loan and scholarshipprograms should becatalogued and evaluated to determine theireffectiveness in improving the geographic distribution of physicians.
Too little is known about the success of loan and scholarshipprograms and the effectiveness with which they fulfill theirpurposes. For this reason it is appropriate that an organizationwithin the Federal government (HHS, GAO etc.) undertake a carefulevaluation of State as well as Federal loan and scholarship programs,particularly in light of the skyrocketing costs of medical educationand the uncertainty about the large indebtedness of graduatingphysicians.
92
Pros
Financial support for the high costs of medical education is animportant part of most programs designed to attract healthpersonnel to underserved areas; however, success of theseprograms has been insufficiently documented, and a comparison ofthe effectiveness of both public and private programs is in order.
Cons
Evaluation efforts must be based on adequate data and criteriafor evaluation. These have not been available to date forevaluations of loan programs.
Sources
Sax., E. Review of incentive programs promoting practice in underservedareas. New York: National Health Council, February 1976, pp. 2-5.
Hadley, J. (ed). Medical Education Financing Policy Analysis and Optionsfor the 1980s. New York: Prodist, 1980.
RECOMMENDATION 26: Despite limited evaluation, there is evidencethat several Area Health Education Center (AHEC) models are effectivein inducing physicians to practice in underserved areas and/or topractice primary care. These types of AHECs should receive continuedsupport.
To provide a regionalized organizational framework for coordinatinghealth manpower education and health care delivery, individualeducational and service institutions are linked to form an AHECsystem. The AHEC concept was originally proposed by the CarnegieCommission in 1970, and is designed to provide a basis forinstitutional change by offering (1) new affiliation arrangements forincreasing clinical training op ortunities without additional medicalschool construction (2) tr ing opportunities located away from thehealth sciences cent o he medical school, thereby increasing thedelivery of health services throughout the AHEC region and addingstimulation to the profesp;onal environment for practicingphysicians, and (3) training opportunities for allied health manpower.
Pros
These programs have long-term potential for significant impact onthe distribution of health personnel in underserved areas.Efforts to incfease the attractiveness of these areas for medicalpractice focus on the development of professional support systems.
93
The activities conducted by AHECs are varied. The greatestpotential appears to be realized when AHECs are components oflarger systematic efforts to improve health manpower distributionGAO, 1978, p. 70).
An analysis of the effects of AHECs between 1972 and 1976 showsthat AHEC programs have had a significant impact on increasingthe growth of physician/population ratios in AHEC target countiesin comparison with other nontarget counties. The effect to dateis believed to result from the attraction of practicingphysicians and clinical faculty to AHEC counties, since fewindividuals trained in AHEC-related programs had establishedpractices in the four years covered by the analysis (Modderman,p. 40-41).
-- Evidence suggests a high level, of program acceptance among themedical community.
Cons
Programs vary as to level of development and because of therecency of their inception are not amenable to evaluation. Fewindividuals trained in AHEC-related programs have so farestablished practice (Modderman, p. 31).
Available financial data alce not sufficiently detailed"to permitcomparison of the benefits of specific program elements withtheir costs.
AHECs lack a strong national strategic basis. Overlap with NHSCand rural health initiatives indicates a need to document commonelements, operations, outputs, and measures of effectiveness ofall strategies.
Systematic data are not-aVailable on parallel programs (calledAHECs al o)-whiCh have not received- Federal funding.
Sources
U.S. Department of Health, Education, and Welfare, AHECs: A Response toGeographic and Specialty haldistribution. Prepared under ContractNo. 231-76-1109 for the Bureau of Health' Manpower, Health ResourcesAdministration, Washington, D.C. DREW Publication No. (RRA) 77-61,1976.,
U.S. Department of Health, Education and Welfare. Report of theSecretary: An Assessment of the National Area Health EducationCenter Program prepared by the Health Resources Administration,Bureau of Health Manpower, Washington, D.C.: November 9, 1979.
94
135
Modderman, M.E. The potential impact and the record to date of thenational health services corps (NHSC), area health education centers(AHEC), and rural health clinics in terms of eliminating grossphysician distribution problems. Prepared for the GeographicDistribution Technical Panel, Graduate Medical Education NationalAdvisory Committee, 1979.
U.S. General Accounting Office. Progress and problems in improv\ing theavailability of primary care providers in underserved areas.\ Reportto the Congress by the Comptroller General of the United Staes.August 22, 1976. Chapter 5, Area Health Education Centers.
Sax, E. Review of incentive programs promoting practice in underservedareas. New York: National Health Council, February 1976, pp. 20-21.
RECOMMENDATION 27: Loan forgiveness programs modelled after thosewhich have been successful should be used as a strategy forattracting physicians into underserved areas.
The goal of loan forgiveness programs is to attract physicians intounderserved areas while making medical education financiallyaccessible to students who could not otherwise afford it.
Pros--
There are indications of success on Federal and State levels ofloan forgiveness_ programs with some instances of respectableretention rates beyond the period of contractual service.Examples of successful programs are those conducted in Illinoisand Kentucky (the Illinois Medical Student Loan Fund and theRural Kentucky Medical Scholarship Fund). A commoncharacteristic of both of these programs is a high degree ofinvolvement in identifying, screening, and placing loanrecipients (Hadley, 1980, p. 186; Sax, 1976, pp. 2-5).
It is generally agreed that the incorporation of more stringentpenalty clauses into loan forgiveness agreements might reduce theincidence of "buying out" (Hadley, 1980, pp. 181-186). Withattractive rates of loan forgiveness for each year of contractualservice and high penalties attached to breach of contract, loanforgiveness programs may become increasingly promising. InMassachusetts, students who refuse service must pay back part ofthe difference between what they paid for their education andwhat it cost the State to educate them (Hadley, 1980, p. 187).Federal regulations have increased the penalty for failure toperform obligated service under the NHSC scholarship assistance,plus interest at the maximum prevailing rate. The sum is payablein one year ("Fact Sheet," 1979, p. 15).
95
iTti
Cons
On the other hand, "buying out" is not a totally negativefeature. Although loan forgiveness is designed to assist inalleviating physician shortages, it is, unrealistic to expect allapplicants for forgiveness programs, many of whom are first-yearstudents, to make irrevocable commitments as to where they willpractice. Since buyouts 4o repay their loans, they provide fundswhich may be used by future applicants who decide to utilize theservice payback option (Hadley, 1980, p. 186).
It is not known whether those who enter loan forgiveness programswould have settled in shortage areas without the inducement ofsuch programs. Hadley (1980, p. 181) found it unlikely that manystudents who did not already have a strong interest in practicingin a shortage area would choose the loan forgiveness option.
Data which would allow one to calculate the cost -for each year ofphysician service in an underserved area obtained through loanforgiveness programs are not available.. In some States, theamount of maximum annual loans available was rather low, as apercentage of a student's total costs, and served as adisincentive to service paybacks. In 19 of 27 programs Hadleyreviewed, the loan forgiveness buyout provision was too easy; itwas simply repayment of the amount borrowed plus interest of nogreater than 10 percent (pp. 181-86)
-- The percentage of physicians receiving loan forgiveness bypracticing in shortage areas and the proportion retained beyondthe period of contractual obligation also impact on the costs ofthese programs. Mason's analysis of State loan forgivenessprograms (1971) indicates that payback and buyout provisions weregenerally liberal enough to discourage service paybacks, unlessstudents were already predisposed toward rural practice (inHadley, p. 181). The majority of States were fortunate if 60percent of the borrowing physicians followed through on practicein rural areas (Mason, 1971, p. 576).
Sources
CONSAD Research Corporation. An evaluation of the effectiveness of loanforgiveness as an incentive for health practitioners to locate inmedically underserved areas. A report prepared for the Department ofHealth, Education and Welfare, Office of the Secretary, Contract No.HEW-OS-73-68, Washington, D.C., 1973.
U.S. Dept. of Health, Education and Welfare, Fact Sheet: aealthProfessions Educational Assistance Act of 1976 (P. L. 94-484, asamended). February 1, 1979.
96
Hadley, J. (ed) State and local financing options, in Medical EducationFinancing, New York: Prodist, 1980.
Mason, H.R. Effettiveness of student aid programs tied to aservice commitment. J. of Med. Educ., 46:575-83, July 1971.
PRACTICING PHYSICIANS
Pros
RECOMMENDATION 28: Comprehensive evaluations of programs to recruitand retain providers in underserved areas (e.g., Rural HealthInitiative, Rural Health Clinics, Health Underserved Rural AreaProgram) should be performed after a reasonable period of time.Continued funding of these programs should be contingent upon apositive evaluation of their effectiveness.
P. L. 95-210 (the Rural Health Clinic Services Act of 1977) providesfinancial support for facilities using physician extenders to provideprimary care in rural, medically underserved areas. Provider-basedor independent facilities which meet specified criteria may becertified as rural health clinics and may receive Medicare andMedicaid reimbursement for specified ambulatory services.Reimbursement is provided for professional services furnished bynurse practitioners, physician assistants, nurse midwives, andspecialized nurse practitioners. The Rural Health Initiative hasbeen deveicTed over several years since the enactment of PL 91-419,the Rurai evelopment Act of 1972.
Reimbursement of services provided by nonphysician health careproviders may assist in the recruitment of physicians tocommunities which have not been able to attract or retain doctors.
Under Federal regulations, services can be provided by physicianassistants and nurse practitioners whether or not a physician isphysically present at the time the service is provided (GAO,
p. 106).
L. 95-210 also provides funding for demonstration projects foriysician- directed clinics in urban medically underserved areasRural Health Clinic Services Act, Dec. 13, 1977, Section 3).
-- The Rural Health Initiative is designed to integrate variouselements from multiple programs.
97
Cons
Reimbursement provisions under the Rural Health Clinic ServicesAct (P.L. 94-210) can be beneficial only in those States whereregulations governing the practice cud supervision of physicianextenders are consistent with conditions stipulated in the act.
Very little evaluative information on the effects of rural healthclinics on the availability or use of services in rural areas isas yet available. Data associated with program elements of theRural Health Initiative have not been identified. Data on otheraspects and strategies of the rural health initiative should becollected and examined.
Sources
Holland, C.D., and Durmaskin, B.T., Progress in development of ruralprimary care clinics in West Virginia. Public Health Rep. 94:369-71,JulyAugust 1979.
Modderman, M.E. The potential impact and the record to date of thenational health service corps (NHSC, Area Health Education Centers(AHEC), and Rural Health Clinics in terms of eliminating grossphysician distribution problems. Prepared for the GeographicDistribution Technical Partel, Graduate Medical Education NationalAdvisory Committee, pp. 441-47.
Public Law 94-210, Rural Health Clinic Services Medicare/MedicaidAmendments. December 13, 1977.
U.S. Dept. of Health, Education, and Welfare, Health Care FinancingAdministration. Rural Health Clinics Regulations, containingregulations as of July 1, 1978, on certification and Medicare andMedicaid coverage and reimbursement of rural health clinics underPublic Law 94-210, enacted December 13, 1977.
Wallen, J. The geographic distribution of nonphysician healthcare providers. Background Paper prepared for the GeographicDistribution Technical Panel, July 1979.
RECOMMENDATION 29: Programs that foster or support group practicearrangements in rural areas, coupled with the appropriatecommunication and transportation networks, should be developed orestablished on an experimental basis as a means of attractingphysicians to rural communities. If these delivery modes prove to besuccessful in delivering care to underserved areas, additionalstartup funding should be encouraged for new programs. , As anorganizational framework for the delivery of medical care, group
98
practice may be effective in redressing the geographic imbalance ofphysician services and especially primary care services. Whetherthis strategy succeeds on a long-term basis depends upon whethergroup practices can be located in areas of greatest need, and whetherphysicians can be successfully retained by rural communities usingthis approach.
Pros
-- Group medical practice emphasizes services provided by a group ofprimary care physicians, or by a primary care physician workingin a team with nurse practitioners, physician assistants, andnurse midwives. Group practice guarantees access to a broadrange of medical services.
-- Studies of recent medical graduates in primary care specialtiesgive a high ranking to "the opportunity to join a desirablepartnership or group practice" as a factor affecting locationdecisions (Cooper et al., 1975, p. 20). Of a group of physicianswho left primary care practice for other medical careers,overwork was given as the chief reason for leaving.
-- Practice in a group setting has been shown to alleviatephysicians' concerns about professional isolation, because ofgreater ease of collegial interaction. Research findings supportthe view that professional isolation is a strong deterrent topractice in rural areas (Steinwald and Steinwald, 1974, p. 16).
-- Group practice compares favorably with the option of solopractice in rural areas because of greater access to capitalequipment, ancillary personnel and other facilities needed toprovide thorough patient care. Primary care physicians choosingthe multispecialty group practice delivery mode cited thefollowing features as important in their decision: (1) freedomfrom the business aspects of medical practice, (2) predictableworking hours, and (3) immediate access to other physicians forconsultation and referrals (Madison, 1973, p. 761).
-- Potential long-term payoff for the redistribution of physiciansro areas where they are most needed may offset the difficultiesand costs encountered in establishing new delivery modes(Madison, 1973, p. 760). It has been shown that it is relativelyeasy for small communities to attract physicians when grouppractices already exist.
99
3
Cons'
Group practice does not easily overcome the perceiveddisadvantages of medical practice in rural areas; physicians seemto be attracted to nonmetropolitan areas which have relativelymore of the characteristics of urban environments (Cotterill andEisenberg, 1979, p. 150). Additional study is needed to clearlyidentify those characteristics of group practice that cansuccessfully offset the disadvantages of rural medical practice.
Investigators do not agree to what extent rural areas can supportincreases in group practices (Steinwald and Steinwald, 1974, p.15). While group practice may be a "compensating factor" inphysicians' choices to locate in rural areas, American MedicalAssociation data indicate that this may not be the case in ruralcounties with a population of less than 10,000 (Langwell andBudde, 1978, pp. 104-106). In these counties, it may not befeasible to establish HMOs since they require a certainenrollment to break even.
Sources
Cooper, J.K., Heald, K., Samuels, M., and Coleman, S. Rural or urbanpractice: factors influencing the location decision of primary carephysicians. Inquiry 12:18-25, March 1975.
Cotterill, P.G. and Eisenberg, B.S. Improving access to medical carein underserved areas: The role of group practice. Inquiry16:141-53, Summer 1979.
Davis, K. and Marshall, R. Section 1502(1): Primary health care servicesfor medically underserved populations, in Papers on the NationalHealth Guidelines: The Priorities of Section 1502. U.S. Departmentof Health, Education, and Welfare, January 1977.
Eisenberg, B.S. and Cantwell, J.R. Policies to influence the spatialdistribution of physicians: A conceptual review of selected programsand empirical evidence. Med Care 14:455-68, June 1976.
Langwell, K.M. and Budde, N. Urban-Rural differences in generaland family practices: An examination of location choice incentives,in Reference data on socioeconomic issues of health, 1978. Chicago:American Medical Association, Center for Health Services Research andDevelopment, 1978, pp. 101-119.
Madison, D.L., Recruiting physicians for rural practice. HealthSery Rep, 88:758-762, October 1973.
100
Perkoff, G.T., An effect of organization of medical care uponhealth manpower distribution. Med Care 16:628-40, August 1978.
Steinwald, B. and Steinwald, C. The effect of preceptorship andrural training programs on physicians' practice location decisions.Chicago: American Medical Association, Center for Health ServicesResearch and Development, November 1973 (revised April 1974).
Pros
Cons
Recommendation 30: Discontinuation of geographic differentials inby
differences in costs of delivering those services as a means ofinfluencing geographic distribution should be the subject of futureresearch/ Present reimbursement systems (Federal, State and private)tend to ustain historical differences in fees and incomes amonggeograph c areas and thus provide incentives for physicians to locatein high 'ncome communities within metropolitan areas.
Under the UCR method (usual, customary, and reasonable charges) largediffereices in payment exist for the same procedure. To illustrate,the ma imum Medicare would, pay in 1977 for a lens extraction in.Buffal , New York, was $561, compared to $957 in the Bronx, $1,085 inManhattan, and $480 in Arkansas. Thus, to the extent that Medicarereinforces existing physician fee differences between metropolitanand nonmetropolitan areas, the Medicare program provides incentivesfor doctors to locate in high-charge areas, potentially increasingthe problem of geographic maldistribution of physicians (Showstack etal., 1979, p. 237). The establishment of uniform reimbursementlevels within a geographic area, such as a State, would appear toeliminate the financial disadvantage to practice in underserved areas.
The present reimbursement system does not appear to be neutralvis a vis the geographic distribution of physicians; it can berevised to create incentives (or at least, to removedisincentives) for physicians to locate in underserved areas.
Reimbursement methods can be redesigned so as to bring thedistribution of services more in accord with social objectives.
Implications of the preceding may be counter to cost containment,at least in the short run.
Policies aimed at equalizing fees across areas in and ofthemselves may not provide sufficient inducement to redistributephysicians and possibly may decrease work effort.
101
-- The impact of discontinuation of fee differentials has not beendetermined.
Sources
Institute of Medicine, A manpower policy for primary health care.National Academy of Sciences, Washington, D.C., May 1978.
Burney, I.L., Schieber, G.J., Blaxall, M., and Gabel, J.R. Geographicvariation in physicians' fees (Special Communication). JAMA,240:1368-71, September 22, 1978.
"To study geographic differences in physician feesrecognized by the Medicare and Medicaid programs, weanalyzed physician reimbursement rates at the national,State, and county levels. The results indicate thatnationally, Medicaid specialist fees are 77 percent ofMedicare specialist fees. Medicare specialist fees inmetropolitan areas are 23 percent higher than those innonmetroplitan areas, but there are no differencesunder Medicaid. State Medicare specialist fees variedfrom 73 percent to 132 percent of the national Medicareaverage, while Medicaid specialist fees ranged from 49percent to 179 percent of the national Medicaidaverage: State Medicaid fees for specialists rangedfrom 39 percent to 100 percent of Medicare specialistfees. These results indicate that under nationalhealth insurance, fees set at national or statewidelevels could have notable effects on physicianremuneration in some localities."
Cantwell, J.R. Implications of reimbursement policies for the locationof physicians." Agriculture Econ Res, 31:25-35, April 1979, p. 30.
"If prevailing charge levels are established at higherlevels in urbail than in rural areas, we can expectavailability of physicians' services in rural areas tobe affected adversely. One policy which meritsconsideration is to actively use the reimbursementsystem to encourage physicians to locate in areaslacking them. A first step in implementing this policywithin the Medicare program would be for the carrier toset limits on prevailing charges which would make themuniform throughout a State."
Gabel, J.R. and Redisch, M.A. Alternative physician payment methods:incentives, efficiency, and national health insurance. Health andSociety, 57:38-59, 1979, pp. 50-54.
102
113
Showstack, J.A., Blumberg, B.D., Schwartz, J., Sch oeder, S.A.Fee-for-service physician payment: Analysis of current methods andtheir development. Inquiry. 16:230-59, Fall 19
RECOMMENDATION 31: CMENAC recommends that all physicians, both thosein primary care specialties and those in non-primary are special-ties, be reimbursed at the same payment level for the came primarycare services.
To increase the availability and quality of primary care 'aervices,changes in the structure of reimbursement methods are required.Physician manpower policy has emphasized the production of moreprimary care physicians treating populations in underserved areas,while current reimbursement policies tend to show fiscal rewards fornonprimary care physicians in adequately served areas (Hadley, 1978;Cantwell, 1977) Cantwell found that general practitioners' mean feeswere significantly lower than specialists' mean fees (Cantwell,1977). The Institute of Medicine estimates that specialists arereimbursed an estimated 7 percent more for the same procedure thanare general practitioners (Institute of Medicine, 1976, p. 335).Available evidence is that Medicare "allowed" charges have increasedmore rapidly for surgeons than for primary care physicians (Instituteof Medicine, 1976, p. 345). Existing physician maldistributionproblems may be exacerbated by the UCR method of reimbursement(Showstack et al.,,1979, p. 237). See Tables 13 and 14.
Pros
This strategy lessens the financial disincentive to enter primarycare specialties by equalizing third-party payments to allphysicians for the same primary care services, and allows equalpayment for identical services of acceptable quality (Instituteof Medicine, 1978, p. 49).
This mechanism also lessens the incentive of nonprimary carephysicians to deliver primary care services. This may prove tobe advantageous because nonprimary care physicians may not be aswell trained in the primary care role as in the specialty role,and there may be a tendency to use specialty skills andprocedures when these are not needed (Institute of Medicine,1978, p. 49).
It is possible that the cost of primary care services may bereduced through this change in reimbursement methods.
103
Table 13GP AND SPECIALIST FEES IN CALIFORNIA
Procedure(1969 CRVS)
SpecialtyDesignation
Mean S.D. C.V.
Medical VisitInitial OfficeVisit (90000) GP $ 17.18, $ 5.27 .31
(115)Specialist 25.86 12.69 .49
(270)
Follow-up OfficeVisit (90040) GP 11.99 3.03 .25
(118)Specialist 13.88 3.84 .28
(284)
Periodic Examination(90088) 33.56 15.30 .46
(113)Specialist 30.81 14.79 .48
(196)
Follow-up HospitalVisit (90240) GP 19.41 11.79 .61
(108)Specialist 21.23 12.05 .57
(249)
SurgicalAppendectom (44950) GP 397.31
(29)139.75 .35
Specialist 418.93 79.11 .19(45)
Tonsillectomy(42840) GP 166.87 44.46 .27(38)
Specialist j 179.24 42.72 .24(31)
Hernia Repair GP 390.61 136.45 .35(49505) (28)
Specialist 407.81 77.97 .19(42)
104
1L)
Table 13 (contip:?...::
GP AND SPECIALIST FEES IN CALIFORNIA
Procedure Specialty Mean S.D. C.V.(1969 CRVS) Designation
Distal Radial GP 148.54 60.81 .41Fracture (25605) (35)
Specialist 182.36 61.61 .34(28)
Obstetrical Care(59400) GP 356.04 82.88 .23
Specialist 481.19 121.02 .25(42)
Radiologicalx-ray of Forearm GP 18.15 4.92 .27(73090) (38)
Specialist (61)
LaboratoryRed Blood CellCount (85020) GP 5.60 2.28
(56)Specialist 6.28 2.65 .42
(73)
Urinanalysis (81000) GP 4.05 1.15 .28(91)
Specialist 4.89 2.27 .46(134)
Source: Tenth Period Survey of Physicians (Fall 1975), Center forHealth Services Research and Development, American MedicalAssociation.
In James R. Cantwell, "Physician Fee Variation and Reimbursement inCalifornia and Georgia." Working Paper No. 7803, delivered at theMedical Economics Session, Western Economic Association 1977 Meetings,Anaheim, California, June 22, 1977.
105
1 44
Table 14GP AND SPECIALIST FEES IN GEORGIA
Procedure(1969 CRVS)
SpecialtyDesignation
Mean S.D. C.V.
Medical VisitInitial OfficeVisit (90000) GP $ 14.92 $ 7.82 .52
(18)Specialist 22.41 16.66 .74
(48)
Follow-up OfficeVisit (90040) GP 9.44 2.37 .25
(16)Specialist 29.77 16.07 .54
(40)
Follow-up HospitalVisit (90240) 1 GP 11.42 5.38 .47
(13)Specialist 16.89 8.50 .50
(40)
Surgical
Appendectomy (44950) GP
Specialist 274.17 69.38 .25(12)
Tonsillectomy (42940) GP
Specialist 156.11 33.15 .21(9)
Hernia Repair (49505) GP
Specialist 261.50 31.80 .12(10)
Continued
Table 14 (continued)GP AND SPECIALIST FEES IN GEORGIA
Procedure(1969 CRVS)
SpecialtyDesignation
Mean S.D. C.V.
Distal RadialFracture (25605) GP 84.58 24.00 .28
(6)
Specialist
Obstetrical Care(59400) GP -
Specialist 335.83 92.76 .28
(6)
RadiologicalX-ray of Forearm(73090) GP 11.94 2.24 .19
(8)
Specialist 13.81 3.91 .28
(9)
LaboratoryRed Blood CellCount (85020) GP 5.55 2.39 .43
(10)Specialist 3.77 1.95 .52
(13)
Urinalysis (81000) GP 3.36 1.01 .30(14)
Specialist 3.57 0.93 .26
(22)
- Sample size less than 5.
Source: Tenth Periodic Survey of Physicians (Fall 1975), Center forHealth Services Research and Development, American Medical(Association.
In James R. Cantwell, "Physician Fee Variation and Reimbursement inCalifornia and Georgia." Working Paper No. 7803, delivered at theMedical Economics Session, Western Economic Association 1977 Meetings,Anaheim, California,June 22, 1977.
107\
.4 c)1 L)
Cons
-- Implementation of this strategy could potentially decrease theamount of primary care services available in those areas andsettings where specialists participate substantially in thedelivery of primary care-(Aiken et al., 1979).
Sources
Aiken, L.H., et al. "The contribution of specialists to the deliveryof primary.care: A new perspective. New Engl J Med 300:1363-70,June 14, 1979.
Institute of Medicine, A Manpower Policy for Primary Health Care.National Academy of Sciences, Washington, D.C., May 1978.
Cantwell, J.R. Physician Fee Variation and Reimbursement in Californiaand Georgia. Working Paper No. 7803, delivered at the MedicalEconomics Session, Western Economic Association 1977 Meetings,Anaheim, California, June 22, 1977.
Institute of Medicine, Medicare-Medicaid Reimbursement Policies. SocialSecurity Studies Final Report. Washington, D.C., National Academy ofSciences, March 1976. Resource Paper #10, Physician choice ofspecialty and geographic location: Analysis of third-party paym ntlevels, pp. 327-359.
108
REFERENCES
Abt Associates, Inc., Analysis and Assessment of Full-time EquivalentPhysician Data Base Produced by the AMA and NCHS. Final Report.Greber, F. (Principal Investigator). Prepared for the Office ofGraduate Medical Education. HRA Contract No. HRA 232-78-0127,Hyattsville, Md.,1979.
Acton, J. P. Demand for Health Care Among the Urban Poor, with SpecialEmphasis on the Role of Time. Prepared for the U.S. Office ofEconomic Opportunity and the New York City Health ServicesAdministration. The New York City Rand Institute, R-1151-0E0/NYCNew York, Apr. 1973.
Aday, L. and Anderson, R. Development of Indices of Access to MedicalCare, Health Administration Press, Ann Arbor, 1975.
Aiken, L.H., et al. The contribution of specialists to the delivery ofprimary care. N Engl J Med 300:213-70, June 14, 1979.
American Medical Association, Center for Health Services Research andDevelopment. Profile of Medical Practice, 1978, The Association.Chicago.
Applied Management Sciences, Review of Health Manpower PopulationRequirements Standards, Prepared under Contract No. 230-0079 forBureau of Health Manpower, DHEW Publication No. (HRA) 77 -22.Bethesda. Md. 1976.
Applied Management Sciences;Preceptorships and OtherEducation of Physicians,final report of ContractMarch 31, 1978.
An Assessment of the Influence of MedicalFactors on Career Choice and MedicalApplied Management Sciences, unpublishedHRA-231-76-0004n. Executive Summary,
Barish, A.M. The influence of primary care preceptorships and otherfactors on physicians' career choices. Public Health Rep 94:36-47,Jan-Feb 1979.
Board of Directors, Health Systems Agency of Northern Virginia, HealthSystems Plan, Health Systems Agency of Northern Virginia, FallsChurch, Va. Oct, 19/9.
Bobula, J.D. and Goodman, J.J. The Supply and Distribution of Physicians:A Look at the Issues. American Medical Association, Chicago: 1979.
109
Bosanac, E.M., Parkinson, R. and Hall D. Geographic access to hospitalcare: A 30-minute travel time standard. Med Care 14:616-624, July,1976.
Burney, I.L., et al., Geographic variation in physicians fees. (specialcommunication) JAMA 240:1368-71, September 22, 1978.
Califano, J.A., Jr. Remarks of the Secretary, USDHEW, before the NationalAcademy of Sciences. Washington, D.C. Oct 26, 1978.
California Office of Statewide Health Planning and Development. GraduateMedical Education in California: A Position Paper. Health andWelfare Agency, State of California, Sacramento, December, 1978.
Cantwell, J.R. Implications of reimbursement policies for the location ofphysicians. Agric Econ Res 31:25-35, April 1979.
Cantwell, J.R. Physician Fee Variation and Reimbursement in Californiaand Georgia. Working paper number 7803, Delivered at MedicalEconomics Association 1977 Meeting, Anaheim, California, June 22,1977.
C..E. Pagan Associates, Inc. AHECs: A Response to Geographic & SpecialtyMaldistribution. Contract No. 231 -76 -0019 Modification for U.S.Department of Health, Education, and Welfare, Health ResourcesAdministration. DHEW Publication No. (HRA) 77-61. Hyattsville, Md.,1976.
Ciocco, A. and Altman, I. Medical Service Areas and Distances Traveledfor Physician Care in Western Pennylvania, U.S. Public HealthMonograph #19, Washington, D.C. 1954.
Committee on Perinatal Health, Toward Improving the Outcome of Pregnancy,National Foundation March of Dimes, New York: 1977.
Commonwealth of Pennsylvania Department`of Health. 1975 Public Hearingson Critical Health Issues Towards Development of a StateComprehensive Health Plan, State Advisory Council on ComprehensiveHealth Planning, Harrisburg, Pa., 1975.
Congressional Budget Office, physician Extenders: Their Current andFuture Role in Medical Care Delivery, Background Paper. GovtPrinting Office, Washington, D.C., Apr 1979.
CONSAD Research Corporation, An Evaluation of the Effectiveness of LoanForgiveness as an Incentive for Health Practitioners to Locate inMedically Underserved Areas. A report prepared for the Department ofHealth, Education and Welfare, Office of the Secretkry, Contract No.HEW-05-73-68, Washington, D.C., 1973.
110
Cotterill, P.G. and Eisenberg, B.S. Improving access to medical care inunderserved areas: The role of group practice. Inquiry 16:141-53,Summer, 1979.
Cooper, J.K., Heald, K., Samuels, M., and Coleman, S. Rural or urbanpractice: Factors influencing the location decision of primary carephysicians. Inquiry, 12:18-25, Mar 1975.
Davis, K. and Marshall, R. "Section 1502(1): Primary health careservices for medically underserved populations" in Papers on thenational health_videlines: the priorities of section 1502, US Deptof Health, Education and Welfare, Washington, D.C., Jan 1977.
Dyck, F.J., et al. Effect of surveillance on the number ofhysterectomies in the province of Saskatchewan. N Engl J Med296:1326-28, June 9, 1977.
Eichenholz, J. Reducing the Supply of Physicians in Overserved Areas(unpublished), June 13, 1978, presented at Institute of Medicine.Washington, D.C., Health Manpower Policy Meeting, June 21, 1978.
Eisenberg, B.S. and Cantwell, J.R. Policies to influence the spatialdistribution of physicians. Med Care 14:455-68, Jurie 1976.
Ernst, R.L. and Yett, D.E. "Econometric studies of the geographic anddistribution of physicians," in Physician Specialty and LocationChoices. Final Report of Grant #3330 From the Robert Wood JohnsonFoundation, Human Resources Research Center, Acccepted forpublication, D.C. Heath & Company, May 29, 1979.
Gabel, J.R. and Redisch, M.A. Alternative physician payment methods:Incentives, efficiency and national health insurance. Health andSociety 57:38-59, 1979.
Geyman, J.P. Family practice in evolution: Progress, problems andprojections. New Engl J Med 298:593-601, Mar 16, 1978.
Hadley, J. (ed). Medical Education Financing, Policy Analysis andOptions for the 1980s. Prodist, New York, 1980.
*j.
Hassinger, E.W., et. al. A Restudy of Physicians in Twenty Rural MissouriCounties. U.S. Dept. of Health, Education and Welfare, HRA, Officeof Graduate Medical Education, DHEW Pub. No. (HRA) 79-634,Hyattsville, MD. 1979.
Health Planning Council of the Midlands. Health Systems Plan 1978-83.Health Planning Council, Omaha, Jan 1978.
Hershey, J., Gianabis, J., Luft, H. Making sense out of utilizationdata. Med Care 13:838-854, Oct 1975.
111
Holland, C.D. and B.T. Durmaskin, Progress in development of rural primarycare clinics in West Virginia. Public Health Rep 94:369-71,July-Aug 1979.
Institute of Medicine, A Manpower Policy for Primary Health Care NationalAcademy of Science: Washington, D.C., May 1978.
Institute of Medicine, Medicare-Medicaid Reimbursement Policies. SocialSecurity Studies Final Report, National Academy of Sciences,Washington, D.C. Mar 1976.
Kane, R., et al., Changes in utilization patterns in a national healthservice corps community. Med Care 16:828-836, Oct 1978.
Kane, R.L. Determination of health care priorities and expectations amongrural consumers. Health Sery Res 4:142-51, Summer, 1969.
Koran, L. M. The reliability of clinical methods, data and judgments.N Engl J Med 293:642-646, Sep 25, 1975.,
Krasner, M., Ramsay, D., and Weary, P. A Framework for PhysicianDistribution Analysis Based on Zipcode Areas as Applied toDermatologists, American Academy of Dermatology Manpower Study,Contract No. 231-75-0021, prepared for Division of Medicine, HealthResources Administration, Department of Health, Education andWelfare, Hyattsville, Md., 1977.
Langwell, K. and Budde, N. Urban-Rural differences in general and familypractice: An examination cf location choice incentives, in ReferenceData on Socioeconomic Issues of Health, 1978. American MedicalAssociation) Center for Health Services Research and Development, TheAssociation, Chicago, 1978.
Leyes, J.M. The Delineation of Economic and Health Service Areas and theLocation of Health Manpower Education programs U.S. DREW, HealthResource Administration. Contract No. NIH 72-4083, Bethesda; Md.,Feb 1974.
Luft, H.S. How do health maintenance organizations achieve their"savings?" N Engl J Med 298:1336-43, June 15, 1978.
Madison, D.L. Recruiting physicians for rural practice. Health SeryRep 88:758-762, Oct 1973.
Mason, H. R.) Effectiveness of student aid programs tied to a servicecommitment. J Med Educ 46:575-83, July 1971.
McNeil, B.J., Weichselbarm, R., and-Parker, S.G. Fallacy of the five yearsurvival in lung cancer. N Engl J Med 299:1397-1401, Dec 21, 1978.
112 (..
Mendenhall, R.C., Girard, R.A., Abrahamson S. A national study of medicaland surgical specialties: I. Background, p rpose, and methodology.JAMA 240:848-852, Sep 1, 1978.
Mendenhall, R.C., et al. A national study of medical and surgicalspecialties: II. Description of the survey instrument. JAMA240:1160-1168, Sep 8, 1978.
Michelsen, E. and Cronquist, N. Scholarships now for service later:Meeting the health manpower needs of medically underserved areas.Public Health Rep 94:11-15, Jan-Feb 1979.
Modderman, M.E. The.Potential Impact and Record to Date of the NationalHealth Service Corps (NHSC), Area Health Education Centers (AHEC),and Rural Health Clinics in terms of Eliminating Gross PhysicianDistribution Problems. Prepared for the Geograrhic DistributionPanel, GMENAC, Feb 1980.
Morris, S.B, The distribution of physician extenders (Communication).Med Ca-e 15:1045-57, Dec 1977.
National Advisory Commission on Health Manpower, ("Miller Report")Report of the Commission, Volume I. Washington, D.C.: U.S.Government Printing Office, (of 2 vols) 19S7,
Parker, R.C. and Sorensen, A.A. Physician attitudes toward ruralpractice: Answers to questions that were not asked. Forum on Med411 -416, June 1979.
Parkin, Distance as an influence on demand in general practice.Epid and Community Health 33:96-99, 1979.
Perkoff, G.G. An effect of organixation of medical care upon healthmanpower distribution. Med Care 16:626-40, Aug 1978.
Permian Rasir Regional Planning Commission, How to Use the Time to Sr:rviceConcept in planning Health Care Resources. Permian Basin, MidlandAir Terminal. Division of Regional Medical Programs, HealthResources Administration, DHEW. Gran: #75-109C, (Project Director,Steven Oliver): Sep 1975.
Prue, D.M. et al. An analysis of distance variables that affect aftercareattendance. Community Ment Health J 15:149-154, Summer 1979.
Public Law 95-210, 95th r,ongress. Social Security Act, ap:endments. Toamend Titles 7111 and XIX, to provide payment for rural health clinicservices, and for other purposes. Dec 13, 1977.
Public Law 94-484, 94th Congress. Health Professions EducationalAssistance Act of 1976, 42 USC, Oct 12, 1976.
113
Richardson, J.D. and Scutchfield, F.D. Priorities in health care: Theconsumer's viewpoint in an Appalachian community. Am J Public Health63:79, Jan 1973.
Sax, E. Review of Incentive Programs Promoting Practice in UnderservedAreas. National Health Council, New York: Feb 1976.
Scheffler, R.M., et al., Physicians and new health practitioners: Issuesfor the 1980's. Inquiry 16:195-229, Fall, 1979.
Schroeder, S.A. and Showstack, J. Financial incentives to perform medicalprocedures and laboratory tests: Illustrative models of officepractice. Med Care 16:289-298, Apr 1978.
Schwarz, M. R., The WAMI program: A progress report. West J Med130:384-390, Apr 1979.
Seldin, D.W., Specialization as scientific advancement andoverspecialization as social distortion. Clin Res 24:245-48, Oct1976.
Shannon, G., Lovett, J., and Bashshur, R. Travel for primary care:Expectation and performance in a rural setting J Community Health15:113-125, Winter, 1979
Shannon, G.W., Bashur, R.L., and Metzner, C.A. The concept of distanceas a factor in accessibility of health care Med Care Review 26:143,1969.
Showstack, et al., Fee-for-service physician payment: Analysis ofcurrent methods and their development. Inquiry 1000-59, Fall 1979.
Sloan, F., Access to r,dical care and the local supply of physicians MedCare 15:338-346, Apr 1977.
Southeast Nebraska Health Systems Agency, Health Systems Plan 1978-83,Health Systems Agency, Lincoln, Neb, 1978.
Statewide House Officers Training System, Inc. "Alternative Strategiesfor Addressing the Issues of Physician Specialty and GeographicDistribution through Graduate Medical Education," Appendix H toPolicy Proposal for Graduate Medical Education in Michigan, Sep 15,1979.
Steinwald B. and Steinwald, C. The Effect of Preceptorship and RuralTraining Programs on Physicians' Practice Location Pecisions.American Medical Association, Center for Health Services Research andDevelopment. The Association, Chicago, Nov 1973 (revised Apr 1974).
114
Steinwald, B. and Steinwald, C. the effect of preceptorship and ruraltraining programs on physicians' practice location decisions. MedCare 13:219-229, Mar 1975
SultZ, H.A. Longitudinal Study of Nurse Practitioners, Phases I, II,III. Bureau of. Health Manpower, Division of Nursing. DREWPublication Nos. Phase I, (HRA) 76-43 (Mar 1976, Phase II, (BRA)78-92 (Sep 1978), Phase III, (HRA) 80-2 (May 1980). Hyattsville, Md.
U.S. Department of Health, Education, and Welfare, "Fact Sheet," HealthProfessions Educational Assistance Act of 1976 (P.L. 94-484, asamended), Feb 1, 1979.
Federal Register, Health Manpower Shortage Areas, Criteria forDesignation, Tuesday, Jan 10, 1978, Part II. pp. no 1586-1596.
Health Resources Administration, 42 CFR Part 121, Sub-Part B, DraftNational Guidelines for Health Planning, Notice of ProposedRulemaking.
Health Care Financing Administration, Rural Health Clinics Regulations,Washington, D.C., enacted Dec 13, 1977.
National Center for Health Statistics Medical Care of Acute Conditions,U.S., 1973-74. DREW Publication No. (PHS) 79-1557, Hyattsville, Md..Feb 1979.
Methodological Approaches for Determining Health Manpower Supply andRequirements. Health Planning Methods and Technology Series, 2.DREW Publication No. (HRA) 76-14512, Washington, D.C., 1976.
Health Resources Administration. Physician Manpower Requirements--StaffPaper #1, Prepared for the Graduate Medical Education NationalAdvisory Committee (GMENAC). DREW Pub. No. (HRA) 78-10.Hyattsville, Md., 1978.
U.S. General Accounting Office, Progress and Problems in Improving theAvailability of Primary Care Providers in Underserved Areas Reportto the Congress by the Comptroller General, Washington, D.C. Aug 22,1978.
115
U.S. General Accounting Office, Are Enough Physicians of the Right TypesTrained in the United States? Report to. the Congress by theComptroller General (HRD-77-92). Washington, D.C. May 16, 1978.
Wallen, J., The Geographic Distribution of Nonphysician Health CareProviders. Background paper prepared for the Geographic DistributionTechnical Panel, Hyattsville, Md., July, 1979.
Watts, C. and Adelstein, W. A Model for the Evaluation of NeurosurgicalManpower Needs Draft Report, University of Missouri-Columbia, May,1980.
Weiss, J., Greenlick, M.R., and Jones, J. Determinants of medical careutilization: The impact of spatial factors. Inquiry: 8:50-57, Dec1971..
Wennberg, J. et al., Changes in tonsillectomy rates associated withfeedback and review. Pediatrics 59:821-826, June, 1977.
Wennberg, J., Barnes, B., and Zubkoff, M. Professional uncertainty andthe problem of supplier-induced demand. Soc Sci Med, in press.
Wennberg, J., Lapenas, C. and Greene, R., Zubkoff, M. Variations inPopulation-Based Use Rates and Expenditures: Implications forManpower Policy. Background paper prepared for the GeographicDistribution Technical Panel, Hyattsville, Md., 1980.
Weston, J.L. Distribution of nurse practitioners and physiciansassistants: Implications of legal constraints and reimbursements.Pub Health Rep 95:253-58. May-June, 1980.
Wilbur, R. Personal communication. Council of Medical SpecialtySocieties, Apr 7, 1980.
Wilson, S. R. An Analytical Study of Physicians' Career DecisionsRegarding Geographic Location. Draft digest of the final report,American Institutes for Research, Contract No. HRA-231-77-0088,prepared for National Center for Health Services Research, Office ofthe Assistant Secretary for Health, U.S. Department of Health,Education, and Welfare, Hyattsville, Md., 1979.
Wisconsin Governor's Health Planning and Policy Task Force. The FinalReport, Madison, Wisconsin, 1972.
116
APPENDIX
LITERATURE REVIEW OF SELECTED FACTORS AFFECTING
LOCATION DECISIONS OF PHYSICIANS
This review was developed for use by the Geographic DistributionTechncal Panel in its deliberations. The appendix was not voted on byGMENAC as a whole.
This review was prepared by Mary L. Westcott, Program Analyst, Officeof Graduate Medical Education, Health Resources Administration.
117
I.
Table of Contents
A.
Introduction
Page
122
123
126
INFLUENCES
Family
1. Spouse 1262. Father's occupation 127
B. Social128
1. Friends 1282. Other physicians 1283. Community recruitment 129
C. Professional? 130
1. Specialization 1302. Medical school 131
a. Ownership 131b. Quality 132c. Socialization: faculty J32
3. Placement service 133
D. Financial Incentives 134
1. Loaf, forgiveness 1342. Scholarships 135
E. General I.ackground and SociodemographicCharacteristics 136
1. Geographic origin 1362. Geographic: medical school 1373. Geographic: residency 1384. Geographic: first practice 139
118
Table of Contents (Continued)
Page
5. Specialty, age, first practice 1406. Race, physician 1407. Academic ability 1408. Sex 1419. Time of location decision 141
II. OBJECTIVES 143
A. Income Maximization 143
B. Family - oriented Goals 144
C. Social Prestige 145
1. Friends 1452. Place in community 146
D. Professional Interaction 146
E. Practice 147
1. Work-leisure 1472. Patient case mix 149
a. Interesting 150b. -Quality can, 150c. "Challenge" independence 150
3. Physician/patient relationship 150
F. Personal (Lifestyle) 150
1. Geographic area preference 1502. Desire to be of help 151
III. DECISION FACTORS 153
A. Environment 153
1. Culture 1532. Recreation 1543. Housing 1544. Community security 1545. Transportation 1556. Schools 1557. Pollution 156
119
Table of Contents. (Continued)
Page
8. Shopping 1569. Climate 157
10. Population growth (change) 15811. Economic growth 15912. Urbanization population density 159
a. Proportion population in agriculture 159b. Proportion population professional 160
13. Racial mix 16114. Wealth of area 16215. Size of community 16316. Terrain 16417. Religion 16418. Population mobility 16419. Relocation and search cost 16520. Welfare support 16521. Zoning 16522. Public services 165
B. Practice Related 166
1. Availability, quality andquantity of hospital services 166
2. Availability of ancillary services(lab, pharmacy, x-ray, office) 168
3. Availability of personnel 1694. Availability of medical school 1695. Availability of consultations 1716. Emergency medical services 1727. Licensure 1728. Group practice possibility 1739. Contact with other professionals 175
10. Continuing education possibilities 17611. Costs of practice 17712. Best available position
Quick practice start 17813. Work with established physician 178
C. Demand Related 179
1. Population characteristics 179
a. Size 179b. Age 180c. Sex 182
120
Table of Contents (Continued)
Page
d. Race 182e. Income 184f. Education 185g. Morbidity and mortality 186h. Health insurance coverage 187i. Per capita health care expenditures 188j. Visits (hospital, MD) 188
2. Physician income in area 1893. Price of area services 1904. Perceived excess demand 1905. Existence of federal programs
Medicare, Medicaid, Research 191
References 193
Additional Recommended References 202
121
Introduction
The Graduate Medical Education National Advisory Cc-mittee (GMENAC)was chartered by the Secretary of the Department of Hea" h, Education,and Welfare on April 20, 1976 to make recommendations on the present andfuture supply and requirements for physicians and their specialty andgeographic distribution. The Geographic Distribution Technical Panel ofGMENAC was established in 1979 to address concerns related to access andequity in the current availability of medical care services. An-exploration of the underlying causes of .the geographic problem, as wellas reasons for physician location choices was undertaken.
The results of this exploration are presented in the following pagesas a comprehensive up-to-date literature review encompassing some 90factors and 160 bibliographic references. The review is the most currentand comprehensive available to date. (see Institute of Medicine, 1976;Cantwell, 1975 and Long, 1975). A number of previous attempts have beenmade to organize the literature on physician location factors (Cooper etal., 1972; Cantwell, 1975; McFarland, 1972, and Crane and Reynolds,1974). The following schema is derived largely from Crane and Reynolds,with slight modifications.
The classification scheme is divided into three major sections- -Influences, Objectives, and Decision Factors. Influences arerepresentative of social and personal forces which can shape and directthe objectives a physician has for his/her lifestyle and practice.Objectives suggest general and specific criteria which a physician mayuse in choosing a practice location. Decision Factors provideinformatip on environmental, practice-related, and demand-reloteddecision factors which can describe, explain, or predict phys'%irinlocation (Crane and Reynolds, 1974).
The .task of clarifying and identifying definitions of variables hasnot been attempted. Many of these variables are overlapping and muchwork needs to be done on conceptualizing the variables. "Community size"may refer to a demand-related decision factor or an'influence, and may bedefined operationally in a number of ways (small, medium, or largecommunities; rural-urban, b; population). There is the additionalproblem that each factor cannot always be separated_ into a iiscreteinfluence.
122
1; .1t.;
StThe following is a summary of the major findings of this literature
review. It should be noted at the start that research on physicianlocation is generally divided into efforts to explain the distribution ofcurrently practicing physicians and efforts to explain the locationalbehavior of new physicians. In terms of policy implications, newphysicians Eh..e highly mobile, and action to influence their locationalbehavior is considered far more feasible from a political standpoint thanmeasures to change established physicians' locations. It should also bepointed out that no consensus exists on the most salient factorsinfluencing distribution, and thus on the direction for public policy.
Influences: Background Factors
Several categories of influences have been reported to affectlocation decisions. These include the geographic background of thephysician, such as place of residency, or some combination of,geographicorigin, medical chool, internship and residency location. Airecentstudy found that National Health Servic(L\ Corps (MSC) physicians whoeventually locate in rural'areas and those. who locate in urban areas arenot differentiable in terms of background (Wilson, 1979). The studyfound, however, that physicians who entered practice in non-NationalHealth Service Corps (NHSC) rural areas usually attended medical schoolslocated in a relatively small community, had residency experience in anonhospitil setting and/or in a rural area, and had a rural background(Wilson, 1979). Although type of background appears to be:an importantvariable in determining whether a physician decides to practice in arural community, another study reported that it was not sufficient tocompel them to continue practice (Parker and Sorenson, 1979). Theresearch results are not yet in on long-term retention in rural orshortage areas.
Objectives
Physicians' choices are not only affected by their sociodemographiccharacteristics and origin, but individually they ha'e a number ofpersonal objectives regarding their life-styles. Both the physician andspouse have a set of preferences with respect to careers and place ofresidence, which includes the role of other individuals. Thesepreferences or goals can be family-oriented, economic, prestigeorientedor professional; For example, personal preferences about work. content;) the degree of professional stimulation in the work environment are
1-.or determinants of career choices. The Government Accountir OUice.1nd rural physicians generally more influenced by personal a,.)
community factors, and urban physicians more influenced by faciorsrelated to the practice of medicine. Hassinger (1979) found that when
rural physicians characterize the salient differences between urban andrural practice, they emphasize the quality of patient-physicianinteraction, while urban physicians cite lack of facilities and supportpersonnel. These conclusions are borne out by Wilson's data whichindicate that students from rural areas, who are not interested inresearch or teaching, are more likely to locate in rural areas.
Decision Factors
The decision factors which a physician considers in practice-locationchoices can be numerous, but basically involve: (1) The environment ofthe area; (2) the type of practice he/she will be able to pursue, and (3)the'perception of demand and need for services in an area (Kane et al.,t979). Although these may each impinge on location choice, they are notnecessarily decisive in all cases. It is clear that the communities withthe fewest resources will have the greatest difficulty attractingproviders. Surveys of physicians indicate that community characteristicssuch as cultural and recreational opportunities, climate, the quality ofthe educational systems and the wealth of an area may play a predominantrole in the decision-making process. However, it should be noted thatthese factors are not very amenable to policy manipulation. It shouldalso be noted that anecdotal information from physicians about an area(need, demand, income) can play an important, although undocumented, rolein the decision-making process.
Some studies strongly suggest that one of the factors necessary forretaining physicians in rural areas is good professional support.Professional considerations that often figure prominently in aphysician's Choice of practice location include the availability ofclinical and/or hospital facilities, the opportunity for contact withother physicians, and the possibility of continuing education. Again, itis difficult to assess the relative importance of thefactors--professional, sociodemographic, personal--that influence therecruitment and retention of physicians in rural areas.
Considerable attention has'also been devoted to the influence ofeconomic factors on physician distribution. Research to date has notshown a strong correlation between economic factors and specialty orlocation choice. Both fees and incomes vary considerably by urban-ruralarea, region and State. Except for Canadian studies, research relatingphysician location, or geographic distribution, to income differences hasnot shown a strong relationship. This may be due to: (1) Social taboosagainst citing income as a motivation factor; (2) the lack of good data,or (3) the intercorrelation of the variables and inability to separateout factors. It is possible that physicians do take fee levels intoaccount in making location decisions (Institute of Medicine, 1979), butthe available evidence presents no firm conclusions.
124
I
A number of factors are important in affecting the demand (or:perceived need) for health services: These include population size, age(population 65+), income of the population, education (affectingutilization), physician income and the existence of Federal programs.Most research has dealt with population size, per capita income, andphysician income. Results show the migration of physicians, towell-populated areas with high per capita incomes.
Summary
Many of the factors analyzed here are presented at the individuallevel and others at the community level. A regression analysis ofcommunity factors affecting physician distribution differs markedly froma physician survey of reasons for practice locations. This complicatesthe development of any kind of a coherent model and makes it diffkcill-t-t6sort out the importance of each factor reviewed here.
- _
The question that must be answered prior to development of policyinterventions to correct the distribution problem is: What is therelative importance of each of the professional, sociodemographic,lifestyle, personal, community and demand-related factors in influencingrecruitment and retention of health providers in 'needy' areas? Studiesthat analyze demand Land community characteristics fail to take account ofindividual differences--studies that only look at individuals'preferences and abilities fail to look at the context in which theindividual operates. A merging of these two separate spheres must takeplace before a complete analysis of physician distribution can occur.Perhaps the next decade of researchers addressing this problem can cometo grips-with this issr-.
125
GMENAC Geographic Distribution Panel
Summary Table:
Selected Factors Affecting Location Decisions of Physicians
VARIABLES MAJOR FINDINGS
I. INFLUENCES
A. Family
1. Spouse
126
Several studies have shownthat family ties and wives'preferences affect location choice(Hassinger, 1963; Bible, 1970;Champion, 1971). Rural and smalltown physicians and their wivestend to come from small communities(Hassinger, 1963). Taylor, Dickmanand Kane found that families inwhich both partners were from thesame background had strikingpreferences for living in an areasimilar in size-to their own(1973). Twenty-six percent ofphysicians in a later study saidtheir spouses had connections inthe area or were attracted to it(Parker and Sorensen, 1978).
Physicians in Wilson'snon-National Health Service Corpssample who were in rural areaswere more likely to have spousesraised in a rural area (1979).Also found was the fact that ruralretention is better for thosephysicians whose spouses did nothave any graduate education. Anearlier study by Coope: et al. hadreached similar conclusions, butstressed the fact that if a wifefelt that opportunity for her own
VkaABLES MAJOR FINDINGS
. INFLUENCES (cont'd)
A. Family (cont'd)
1. Spouse (cont'd)
2. Father'sOccupation
127
career was important, the physicianwas more likely to choose an urbanlocation (1975). Hassinger statedthat in general the locationalbackground of the spouse reinforcesthe tendency for physiciangofrural origin to practice in rurallocalities and physicians of urbanorigin to practice in urban areas(1979).
However, Parker and Tuxillplaced influence of spouse low ona list of factors which deterredphysicians from locating in asmall community (1967). Generally,the spouses' role in locationdecisionmaking did not appear tobe highly influential (Wilson,1979).
Schaupp studied migration ofWest Virginia medical graduatesand found that those remaining inWest Virginia had fathers who wereblue collar workers and weremarried to natives of the State(1969). Champion argued thatmetropolitan physicians as a grouphave higher social origins thannon-metropolitan physicians ;0971).Cordes found that 17 percend ofthe doctors he studied had fatherswho were either physicians,,dentists, or pharmacists--adisproportionately large number ofthe physicians came from familybackgrounds characterized byrelatively high socioeconomicstatus (1978). 'Hassingerconcluded that fathers' occupationis not an important predictivevariable for understandingeventual practice location (1979).
VARIABLES MAJOR FINDINGS
B. Social
1. Friends
2. Other Physicians
128
When Bible asked what factorsinfluenced physicians to choosetheir present location, family andfriends were cited as a majorinfluence (1970). When asked howthey decided upon their presentpractice location, more than halfindicated friends (Bible, 1970).The GAO reported that one of theinfluences found to be mostimportant in location choice was"personal influences" (GAO, 1978).The small community physician inone study felt his colleagues'decisions were influenced by others(such as professors, relatives,etc.) whereas their influence wasminor in his own decisionmaking(Parker and Tuxill, 1967). Aboutnine percent of Cooper's surveyphysicians reported that theinfluence of family and friendsranked among the top threelocation factors (Cooper et al.,1975).
On the other hand, Parker andTuxill described the influence ofparents or relatives as not parti-cularly strong. Others also feltthat the desire to locate nearfamily or friends seemed to influ-ence only a small percentage ofpractitioners (Peterson, 1968;Charles, 1971; Heald, 1974).
Eight percent of physicians inone study said that associationwith older physicians influencedtheir decision in finding theircurrent practice location (Bible,1970). Only two percent of/the
VARIABLES MAJOR FINDINGS
I. INFLUENCES (Cont'd) 1965 graduates of medical schoolsin another survey found the advice
B. Social (Cont'd) of an older physicians as rankingamong the top three reasons for
2. Other Physicians (Cont'd) locating where they did (Cooper etal., 1975). About 16 percent ofphysicians interviewed by Parkerand Sorensen cited the influenceof teachers in medical school orlater training as a factor whichinfluenced them to settle in asmall community (1978).
3. Community Recruitment
129
A survey of 1971 medicalschool graduates shoWed that ruralphysicians were generally moreinfluenced by community recruitmentefforts than were urban physicians(GAO, 1978). Only two percent of1965 medical school graduates foundorganized efforts of communityrecruitment an influential factor(Cooper et al., 1975). Thelocation decision for 29 percentof physicians surveyed by Cordesincluded the fact that they wererecruited by individuals withinthe community (1978). Communityrecruitment did not affect theprobability of rural choice,according to Coleman, but didattract some primary care groups ifthere was an interest in medicalneed, availability of loans andthe prospect of influence in thecommunity (1976). Wendling demon-strated that local communityefforts to recruit physicians wasa significant variable in thewestern regions of the U.S.explaining physician location(1978).
1
VARIABLES MAJOR FINDINGS
C. Professional
1. Specialization AMA data on the distributionof office-based physicians in 1975showed that 22% of all general andfamily practitioners lived incounties of less than 50,000persons as opposed to 5% ofmedical specialists and 7% ofsurgeons (Goodman and Mason, 1976).
It appears that generalpractitioners are attracted tosmall communities (Presser, 1975)and tend to come- from small towns(Coker, 1960; Paiva and Haley,1971, and Kritzer and Zimet,1967). However, Rushing found theattraction of larger communitiesfor general and narrow specialistsis stronger than the attraction ofsmall and rural communities forgeneralists (1975). Coker con-cluded, on the other hand, that a"greater proportion of studentswho prefer to practice in a ruralsetting intend to enter generalpractice, while students whoprefer large cities intend tospecialize (1960). Cullison etal. demonstrated that hometownsize and specialty choice may beinterrelated determinants ofpractice county size (1976).Their data showed that the graduateentering family medicine was threetimes more likely to select non-metropolitan practice as a physi-cian entering the other primarycare specialties of internalmedicine, pediatrics, andobstetrics/gynecology (Cullison etal., 1976). In Hassingers' study,
-11 specialization was characteristic
130
VARIABLES MAJOR FINDINGS
I. INFLUENCES (coned)
C. Professional
1. Specialization
2. Medical School:a. Ownership
131
of the type of practice which ledto urban locations (1979). At thenational level, Wendling et al.found practice specialty to herelated to location choice. Ageneral pract..tioner was morelikely to select a shortage countyin the North Central Region ascompared to the West (1978). Itseems clear that increasedspecialization involves a greaterreliance on other physicians andhospital-based care, both of whichtend to be found in large urbancenters (Werner, 1978).
No significant research hasbeen performed on the relationshipbetween practice location and thetotality of medical school charac-terisitics. One study showed thatState-owned medical schools retainabout 50 percent of their graduateswhile private schools retain 40percent (Yett, 1973). Heldbelieved it is important to differ-entiate between graduates ofmedical schools that are known tobe very selective of applicants(private schools) and those fromthe less selective schools. Hefound that graduates from theprivate schools showed a greatertendency to enter into practiceoutside their home State.
This pattern did not prevailin New York and California, where,the majority of graduates fromboth private and public medicalschools tended to remain in thosestates (1973). Graduates ofsouthern medical schools were alsoless likely to leave their State
VARIABLES MAJOR FINDINGS
I. INFLUENCES (cont'd) to practice. However, data on1,850 graduates in 28 medical
C. Specialization (cont'd schools collected by the AAMCfound the type of medical school
2. Medical School (cont'd) attended was not a usefulpredictor of location choice.
a. Ownership (cont'd) Even when characteristics of themedical school were entered intothe analysis, their contributionto the prediction of practicelocation was minor (Erdmann etal., 1979).
b. Quality Breisch (1970) found thatquality of medical school educa-tion and location of medicalschools accounted for almost 50percent of the variation inlocation of practice. The higherthe quality of the medical school'straining program 'and the more urbanits location, the more likely grad-uates would seek urban practicelocation. More recent studiesindicate that type of medicalschool and characteristics of themedical school are not usefulpredictors of location choice(Erdmann, 1979).
c. Socialization: Faculty Some argue that medicalschools turn out specialists wholearn that it is preferable todeal with only one type of patient(urban middle class). Studentsare selected who seem most likelyto mirror the attitudes and orien-tations of medical faculty members(Bynum, 1972). Even for studentsfrom rural States, "the cumulativeeffect of current medical educationis to dissipate their interest inbecoming generalists" (Taylor,1973). Sixteen percent of
132
VARIABLES MAJOR FINDINGS
I. INFLUENCES (Coned)
C. Professional
2. Medical School:c. Socialization:
Faculty
3. Placement Services
133
physicians cited the influence ofteachers in medical school or latertraining as a factor which influ-enced them personally to settle ina small community (Parker andSorensen, 1978). The respondentsof all previous surveys by Parkerand Sorensen generally felt thatmedical school faculty membersdisplayed a negative attitudetoward rural practice (1979). Asurvey of residents by theInstitute of Medicine showedfaculty influences to be a majorfactor in their choice ofspecialty (1979).
Wilson concluded thatphysicians now in practice inshortage areas tend to see thefaculties of their residencyprograms as having valued certainshortage area careers more highlythan does the average faculty(1979). "That family practice andrural medicine are so often viewedas complementary underscores theneed for positive role models andexperiences in rural familymedicine." (Taylor, 1973).Cullison felt that very few deter-minants of practice location couldbe influenced by the medicalschool, but the medical schoolcould select students with ruralbackgrounds, and through itscurriculum, influenCe specialtyselection of its graduates (1976).
In finding a location topractice, physicians used theassistance of State and AMAphysicians placement services(Bible, 1970).
VARIABLES MAJOR FINDINGS
D. Financial Incentives
1. Loan Forgiveness
134
Mason found thait only one Stateloan forgivenegs program had beenvery successf while in themajority of States about 60 percentof the borrpwing physicians did notcomplete their obligation (1971).In all of/the State programs, one-third chqe to buy out of theirobligati,on to practice in a smallcommunity. The CONSAD ResearchCorporation found that the HealthProfessions Education AssistanceAct (P.L. 94-484) loan forgivenessprogram had little success andthat higher service repaymentrates could be effected from morecareful selection of loanrecipients (1973). Althoughmedical students were given $155million in scholarships and lowinterest loans, only 86 physicianshad portions of their loanscancelled in return forestablishing practices indesignated shortage areas (Parkerand Sorensen, 1978).
The Human Resources ResearchCenter found that loan programsare attractive to substantialnumberstof medical students, butsubstantial numbers of physiciansbuy out of their service oblige''(Yett, 197n. Disproportionatelymore loan recipients who fulfilledtheir service obligations comefrom rural backgrounds. Little isknown of the extent to which loanforgiveness programs inducephysicians to settle permanentlyin medically underserved areas.
VARIABLES MAJOR FINDINGS
I. INFLUENCES (Cont'd) Sinclair Coleman reported thatpayment of a forgiveness loan
D. Financial Incentives (Cont'd) affects how rural the practicelocation is among rural general
1. Loan Forgiveness (Cont'd) practitioneis (GPs) though not the-GP's probability of selecting a'-rural practice. Loan forgivenesswas not found to be related tolocation type for nonprimary carephysicians. (Coleman, 1976;Cooper et al., 1975).
Recent surveys have shown thatthe loan forgiveness programexerted little influence on even-tual practice location (GAO, 1177;Wendling, 1978). However, intheir recent study of the effectsof debt on medical students'career preferences, Mantovani etal. (1976) did report that studentswith large debts exhibited muchstronger interests in practicingin physician-shortage areas thanthose --with no debt, regardless of
other background traits (sex, race,marital status, size of home town,and parental income). Althoughthey did not indicate the sourceof debt, it is possible that manyhigh-debt students were motivatedby the opportunity to obtain loan-forgiveness by serving in shortageareas.
2. Scholarships
135
Evaluations of such scholarshipprograms as the National HealthService Corps have not beencompleted and trends in retentionrates have been difficult toassess. The program has not beenable to staff some sites in moreremote and isolated communities.
VARIABLES MAJOR FINDINGS
E. General Backgroundand ,Sociodemographic
Characteristics
1. Geographic Origin The AAMC Longitudinal Study ofMedical School Graduates ifound thatonce specialty difeccences aretaken into acc-,unt-, only backgroundcharacteristics, particularly thesize of commimicy lived in most oflife, are significantly related tolocatiln choice (1979). Longreports several studies which havedetermined that physicians aremore likely to locate theirpractice in communities whose sizeresembles those in which theylived prior to attending medicalschool (1975). Hassinger alsoreports that rural as well asurban primary care physicians werelikely to return to practice inplaces similar in size to wherethey were born (1979). Theprobability of choosing a givenpractice location increases withthe amount of prior contact,including place of birth (Yett,1973). Hassinger found thatselection of n practice site isbased on preferences developed inthe socialization of early yearsand altered through training andcareer experiences (1979).Weiskotten's study of U.S. medical.school graduates for the period1915-1955 concluded that all priorlocation factors influenced aphysician's location choice (1960).
136
VARIABLES FfNDINGS
I. INFLUENCES (Cont'd) A and Sloan found thatgeneral practitioners (GPs) were
E. General Background substantially more likely thanand Loc.() :Dgraphic other physicians to practice inCharm tics (Cont'd) their States of birth (1974).
About 15 percent of all physicians3. Geo' r% is Origin (Cont'd) entered practice in their Btates
of birth; for GPs, internists, andpediatricians the figure was 27'percent, 16 percent and 10 percent.In cases where the physician didnot practice in his home State orcommunity, attitudinal evidenceindicates that preferences forcommunity size are probablyestablished by early types of areacontact. A number or recentstudies confirm th, -ignificanceof place of rearing in predictingpractice location (Cooper, 1975;Cullison, 1976; Cordes, 1978;Erdmann, 1979; Wilson, 1979;Wendling, 1978).
2. Geographic:Medical School
137
Place of birth, however, wasnot significantly related tochoice of practice location ineither Held's (1973) or Weber's(1972) analysis.
Weiskotten indicated that morethan half of the graduates in hisstudy were practicing medicine inthe State in which they residedprior to attending medical school(1960). Held reexam'ned the dataand reported that many of thephysicians P-G' went to medicalschool and t residency in thatState (1973). Breisch supportedthe hypothesis that physicianstended toilocate in areas similarin size t6 those of the medicalschool they attended, but the
1'7r')4,..)
VARIABLES MAJOR F's1DINGS
I. INFLUENCES (Cont'd)
E. General Backgroundand SociodemographicCharacteristics ( Cont'd)
2. Geographic:Medical School (Cont'd)
3. Geographic: Residency
study was not controlled forbackground of students (1970).Existing evidence indicates thatattending a medical school in aState without previous or laterState contact had a slight impacton the propensity of graduates topractice there. (Yett, 1973).Wilson found that physicians inthe nn-Corps sample who are inrural areas tended to have gine tomedical school in a relativelysmall community (1979).
Weiskotten (1960), Fahs andPeterson (1968), and Held (1973)have reported tha,t the State inwhich residency training was takenwas more significant in predictinglocation choice than the State inwhich the medical school waslocal: d. However, Fein and Webermaintained that 'Pl-yszians taketheir residency training in theState in which they plan to prac-tice, in order to gain helpfulcont-cts (1971). Hadley disputedWeiskotten, reporting that hismethod involved double countingsince the four contact events werenot mutually exclusive, and that asingle factor, site of residencytraining, has a small effect(1975). However, in combinationwith the caber three factors (bornor lived in area, medical school,internship and residency location),the influence of residency trainingsites was considerable.
Results of a recent study inTexas indicate that the selectionof a residency i, closely associ-ated with a choice of permanent
138
1 7
VARIABLES MAJOR FINDINGS
I. INFLUENCES (Cont'd)
E. General Backgroundand SociodemographicCharacteristics
3. Geographic: Resi ncy
4. Geographic:First Practice
139
location for practice, both ofin-State medical school graduatesand out-of-State graduates(Stefano, 1979). The dominantposition of the residency siteamong prior location factors wIsalso supported by Petersdor;:, whoreported that 53 percent ofspecialists in theWashington - Alaska Montana -Idahoregion had obtailied trainingin the region (1975). However,Hassinger in Missouri foundlocation of residency traini.ng,n3t,to be an important inf.' ence(1979). Yett and Sloan stiggPstedthat the more contact events r.physician has in aState the greater the probabilityof his locating ..i,: Lhetat,,:, andthat the lore tecen.i'. cOnttsappear t: have a grent'e- impactthan the earlier one,=:(174).
Fei*and We:)erinquirr' intrsanothe'r aspec_ of thf!
between residency training andphysician's cblice, and concludedthat physicians take their resi-den tr,n4ning in State:: 4herethey int d to establish theirpractice, and do not choo'epractice location subsequent toresidency training (1971).However, this conclusion wasdisputed by Yett and Sloan(1974).c- Hadley questions bothsets of findings and concludes thccausality issue has yet .:o be
resolved (1975'.
VARIABLES .70R Fi'mINGS
5, Specialty, Age,First Practice
6. Race of, Physician
7. Acad.(' is Ability
140
Harvey (1973) found thatspecialists who had left generalpractice were more likely thangeneral practitioners to establishtheir first practice in a town ot49,999 or smaller. Physicians inthe former group were more likelyto encounter diffii ,lty with lacko: .1.,onsulttInts and lack of
hospital facilities.
Tabulations of data on theState level location of physiciansin 1967 indicated that location ofbla.f. physicians were noX.substan-tially different from those ofwhite physicians (Yett, 1973). It
is commonly believed that blackphysicians are disproportionatelylikely to setrle in inner cityneighborhood: 4 to serve blackpatients, and .stentive' evidencehas been adv.? by Woolridge(1976), Guzick and Jahiel (1'976),and Breiger (1979).
In a study by Lloyd.et al. (1978)72 percent of black Howard gradu-ates had black patients, but onlya fourth were patients who werejudged to be "well-to-d:." Thereis also some indication that physi-cians practice in areas where thereare people of their own , 7hnicgroup (Liebersen, 1958) and thatpreferences for c les own ethnicgroup maximizer. physicians' income(Elesh and Schnilaert, 1972).
Few studies have relatedacademic ability to eventual choiceof practi e location. However,CONSAD found that the medicalstudent who enters general practice
VARIABLES MAJOR FINDINGS.
I. INFLUENCES (Cont'd)
E. General Backgroundand SociodemographCharacteristics (Cont'd)
7. Academic (Cont'd)
8. Sex
9. Time of LocationDecision
141
has low MCAT and medical schoolperformance scores (CONSAD, 1973).A number of other studies havereported that graduates with thelowest MCAT scores or classstandings are those most likely toenter unspecialized internships(Peterson, P963: Perlst,adt, 1972).
Nothing substantive is knownof the practice location behaviorof women, but it is suggested thata third or more of all women physi-cians select practice 1pcations inconcert with the decision of theirphysician husbands.
McGrath and Zimet (1977) founda preference for family practiceand internal medicine among womenin two western medical schools.According to some, the impact onlocation decisions of thesechanging preferences is that morewomen may be expected to settle inrural areas. (Lavin et al., 1980).
Some evidence indicates thatspecialty choice generally precedeslocatior -..hoice. General loctiondecisions are not made until aftergraduation from medical school byas many as 75 to 85 ercent of allphysicians (Parker and Tuxill,1967; Heald et al. 1974; Hassin5;er,1963) Physicians broug '7. up inrural areas tend to make earlierlocation choices than those fromurban areas. Results from Heald's1965 survey of medical sch-olgraduates concluded that 73 per-cent of all physicians and 68 per-cent of primary care physicians:,7ho responded to the survey
VARIABLES MAJOR FINDINGS
I. INFLUENCES (Cont'd)
E. General Backgroundand SociodemographicCharacteristics (Cont'd)
9. Time of LocationDecision (Cont'd)
142
determined their preference forcommunity size only after medicalschool graduation and physicianswho decided to settle in a ruralarea decided on a practicelocation sooner than those whodecided to settle in an urbanlocation (1974). Wilson's -ecentstudy of NHSC physicians as wellas non-corps physiciansdemonstrated that eventualgeographic region and practicecommunity size was decided earlierthan the eventual specialty,particularly for non-primary carephysicians. Most GPs and FPstenIpd to decide on practiceLocution and specialty a: aimutthe same time, and generallyearlier than those in otherspecinities :Wilson, 1979).
In another study of twoentering medi..:al school classes,99 percent of the students knewthe size of community in whichthey planned to practice, but only84 percent had a specialty choiceplanned. (Haug et al , 1980).Only 37 percent of the latterconsidered their specialty choicea definite decision.
1
VARIABLES MAJOR FINDINGS
II. OBJECTIVES
A. Income Maximization
143
Sloan found that the averagenet income received by physiciansin 1960 had a positive and statis-tically significant effect on aphysicians' decision to locate,and later research has borne outthis/finding (1968). Fahs andPeteyson (1968) and Ball andLISon (1968) also cite income ashaving high drawing power for allphysicians. Parker, Rix andTuxill found "anticipated smallincome" as a r, a why physiciansdid not practice in small commu-nities (-967). Parker andSorensen's respondents reportedthat the area in which theypracti,:ed yielded too limited anincome- 33 percent cited this as areason for leaving (1978). -ordesfound that economic reasons sucl-,as thr_ orportunity to acquire aFinancially sound practice werementioned 3,, almost 40% of doctorsas influencing locational decisions(1978). However, Benham (1968),Rimlinger and Steele (1963) andMarshall (1971) argue that incomeis not aiJ important as ;:riginallythought. Held found expectedearnings to have a small effect onmigration of physicians, and atendency for migrants to be drawntoward nonpecuniary benefits and
concentrations of physi-ts (1973). Rural physician's
Licomes are apparcncly not systema-Ccally 1Gwer f:hose of tJrban
physicians.
ti1
VARTA,BLES MAJOR FINDINGS
II, (IBJECTIVES (Cont'd) Hadley also found financialvariables were not significant in
A. Income Maximization (Cont'd) a physician's location choice.The physician's relative mean netincome was found to be an insigni-ficant factor, which may have beendue to the lack of exact physi-cian productivity data (1975).Wilson demonstrated that ruralphysicians tend to have a somewhatlower concern than other special-ists about having a relatively highaad/or increasing income (1979).Lave, Lave, and Leinhar't feltthat physician incomes are"currently so high that the lureof still higher income influel.-2sfew" (1975). 7 lysicians, theyrgue, are more concerned with
cultural, environmental, andprofessional amenities.
B. rnmil-Oriented Goals
144
Thus, the relative importanceof income has mixed and inconclu-sive results in all of the studies.Limitation:. in the economic litera-ture stem from lack of data, lackoc consideration of important non-economic variables, and studydesigns inhibiting focus on thisvariable. Most studies foundincome to have a positive andgenerally statistically signifi-cant impact on the number of atleast some types of physicians ina geographic area. Howeve, themagnitude of this relationshipvaries considerably (Hadley, 1980).
DeVise idenciAied four typesof life-style goals motivatingphysicians to locate in certainStates and urban centers. Family-oriented goals stressed the best
VARIABLES MAJOR FINDINGS
II. OBJECTIVES environment for bringing up chil-dren and satisfying the social
B. Family-Oriented Goals (Cont'd) daytime needs of the mother (1973).
C. Social Prestige
1. Friends
145
Peterson surveyed physicianswho listed lifestyle indicators asbeing of major importance by bothurban and rural physicians (1968).Bible found that respondents wholiked rural practice and livingdid so because of the feeling thatrural people were Eriendty anddependable, which resulted inclose personal ties (1970).Parker, Rix, and Tuxhill reportedthat of those physicians practicingin 'mall towns, pe sonal lifestylepreferences rather than profes-sional iss0,26 predominated (1969).
__Crawford and McCormack
reported, that of physicians wholeft primary practice in a small7ommunity, the social life andcompanions for their children weredescribed as adequate or betterthan average (1971). Bible citt.td
deficiencies in social opportuni-'ties as a major problem for physi-cians locating in rural areas, andthe development of close andlasting friendships as a desirablefactor (1970). Harvey reportedthat five percent of generalpractitioners and 11 percent ofspecialists' wives were dissat-isfied with the kind of social andfamily life possible when thehusband was in rural practice(1973).
VARIABLES MAJOR FINDINGS
2. Place in Community
D. Professional Interaction
Bible cited several desirablefactors in community life includ-ing "opportunity for communityleadership (1970)." The prospectof being more influential incommunity affairs was cited byphysicians as being relativelyimportant in location choice,particularly for older physiciansand urban physicians. It is deba.able whether this factor wasactually important in decision-making at the time, or if subse-quent success provided a rational-ization (Marker and -uxill, 1q76).Coleman described situations inwhich community recruitment wassuccessful for non-GPs and foundthat the prospect of influence inthe community was a related factor(1976). However, the opportunityto take an active part in communityaffairs seemed to be a minor inf'l-ence in another study (Heald etal., 1974).
DeVise identified professionalinteraction as one of the fourgoals which motivates a physician'in terms of location choice (1973).He hypothesized that this goal,stressing professional and socialinteraction among people sharingcommon skills, values and inte-rests, favors intrastate locationsaccessible to concentrations ofspecialists' offirs downtown andin large university medicalcenters. Parker and Tuxill citedlack of free and informalcommunication with medical peers(professional isolation) as adeterrent to rura. practice(1967). He -'1d et al. found that
146
VARIABLES MAJOR FINDINGS
II. OfJECTIVES (Cont'd)
D. ProfessionalInteraction (Cont'd)
E. Practice
1. Work-Lei3ure
147
among the factors cited mostfrequently by recent graduateswere opportunity for regularcontact with a medical school ormedical center and with otherphysicians (1974). Bible surveyedphysicians in nonmetropolitan areasand found that opportunity forprofessional growth was considereda problem but this varied with thesize of the community (1970). TheGAO study demonstrated that of thephysicians they surveyed, five ofthe seven top factors related toclinical support, contact withother physicians and continuingeducation (1977). An opposingviewpoint was proposed by Ducker,who in a more recent'study, foundthat for 46 physicians practicingin nonurban areaq, professionalisolation was no a major problem(1977).
Established rural physiciansare more likely than urban pbysi-cians to complain of excessiveworkloads, irregular hours, thedemands of being continu,usly oncall, and the availabilir:v of1;'*le leisure time. Parker al5.:,,!nsen cite too large a workloadand inability to get adequate timeoff as reasons physidians do notpraCtice in small communities(1978). Hassinger reported that aconfining work situation (inabilityto get away from patients) was adisadvantage reported by a subsT.na-tial proportion of rural M.D.s andD.O.s (1979). Crrlwford andMcCormack (1971) found that thegreatest complaint offered byprimary practitioners in rural
VARIABLES MAJOR FINDINGS
II. OBJECTIVES (Cont'd)
E. Pract (Cont'd)
1. Work-Leisure (Cont'd)
148
areas was overwork. At leasttwo-thirds gave this as a majorreason for leaving practice.
The vice-president of theSears-Roebuck Foundation statedth , insome cases, a physicianwho came to a town that was.doctor-short "wouldn't have enough workduring the days to keep busy, andthen he'd spend half the nighttaking care of emergencies."(Lavin, 1972). Riley et al.(1 Bible, 1970, and Taylor etal. (1973) reported that physiciansbelieve long hours and disruptionsof family life are drawbacks ofrural practice. Hambleton foundthat at the State level specialistsappear to be\most concerned aboutthe use of their leisure time. Atthe county level the costs of usingleisure time were still dominantin the decision of a specialist tolocate. Although the stror4estforce was the availability of a
general hospital, the next mostimportant was locally developedrecreational sites (1971).
Harvey disputed these argu-ments, claiming that physiciansleave general practice for tworeasons: the kinds of valueorientations they hold and medicalcompetence (1973). According toShannon and Dever, the desire forleisure is apparently not a strongmotivating...force in physician'schoice of location..(1974). Recentresearch indicates that ruralphysicians have a lesser aversionthan urban physicians long hoursand seeing large numbers of
VARIABLES MAJOR FINDINGS
II. OBJECTIVES (Cont'd)
E. Practice ( Cont'd"
1. Work-Lf gitre (Cont'd)
2. PPL,...r1,- ':age Mix
a. Int,?restingb of Carec. Calienge,
Independence
149
patients each day (Ducker, 1977;Wilson, 1979). Parker and Tuxill(1967) found that 85% of the urbanrespondents claimed that longhours were an important factor inother physicians location deci-sions, but less than 30 percentwere influenced themselves.
Parker and Tuxill (1967) foundthat one of the factors whichdeterred a significant number ofphysicians from locating in asmall community was the notionthat medical care might not be ofthe quality found in large medicalcenters; one factor likely toinfluence physicians to locate in asmall community was "challenge ofpracticing in a rural area lackingcomplete facilities." Crawford andMcCormack (1971) found that'no onedescribed boredom or lack of achallenge as a reason for leavingprimary practice, but the physi-cians were concerned about thequality of care they were able togive.
Physicians who report thatprofessional considerations predom-inate cite the ability to have avaried practice as associated withurban practice (Lave et al., 1975).Harvey found that important factorsinvolved in the decision to leavegeneral practice are "spectrum ofproblems faced in general practicetoo broad" and "unreasonable expec-tations on the part of patients."(1973). Lave, Lave, and Leinhardtreport that physicians seem moreconcerned with factors such asinteresting case variety which are
MICROCOPY RESOLUTION TEST CHARTNATIONAL BUREAU OF STANDARDS
STANDARD REFERENCE MATERIAL 1010a(ANSI and ISO TEST CHART Mi. 2)
VARIABLES MAJOR FINDINGS
II. OBJECTIVES (Cont'd)
E. Practice (Cont'd)
2. Patient Case Mix
a. Interestingb.L Quality of Carec. Challenge,
Independence
3. Physician-PatientRelationship
F. Personal (Lifestyle)
1. Geographic AreaPreference
usually associated with large popu-lations (1975). Wilson (1979) andHassinger (1979) found that ruralpractitioners clearly value theattributes that typify generalmedical. practice including varietyin procedures andechniques thatthey are called 'pporo employ andclose interaction with patients,while metropolitan physicians tendto emphasize the technical side ofmedicine.
Physicians dissatisfied withrural practice cited "the attitudeof inhabitants of small communitiestoward physicians bent on perfor-ming complete and extensivestudies" as affecting their abilityto feel comfortable. (Parker andSorensen, 1979). Also cited wasdistrust of the abilities of thedoctor by many small communities.Physicians happy with rural livingpreferred the closer ties topatients, their activities andproblems (1979). For rural physi-cians in Hassinger's study, theadvantage most often cited was thequality of the doctor-patient
/ relationship (1979).
There is a large body of datasuggesting that lifestyle prefer-ence is an important criterion inpractice location (Bible, 1970;Cooper et al., 1972; Taylor, 1973;Coleman, 1976; Hassinger, 1979).
150
Seventy-four primaryphysicians in group practice werequeried as to why they decided towork in their present organization.The majority said the community and
1 "
VARIABLES MAJOR FINDINGS
II. OBJECTIVES (Cont'd) the general geographic location ofthe group was unimportant to their
F. Personal (Lifestyle) (Cont'd) decision (Madison, 1971). Parkerand Tuxill found that one of the
1. Geographic Area three most important factors forPreference (Cont'd) locating in a small community was
the idea of small community living(Parker and. Tuxill, 1967).
2. Desire to beof help
151
A later study by Parker andSorensen found that a predisposition to small community living isessential for physicians to berecruited for rural practice(1978). The most frequently citedreason for physicians choosing topractice in their present communitywere "personal preference for thegeneral area and/or this type ofcommunity" in the Cordes study(1978). The advantage mostcommonly reported by rural physicians was their general preferencefor rural areas, and they oftencited their rural background as abasis (Hassinger, 1979).
Bible found that one of thethree most frequently mentionedinfluences on physicians wasgeographic preference (Bible,1970). A GAO report concludedthat rural physicians generallywere more influenced by preferenceto live in a rural or a particulargeographic area. (GAO, 1978).
One factor that influenced73 percent of physicians in Parkerand Sorensen's study was "need forphysicians in your community"(1978). Concern with an area'smedical needs was also found byWendling et al. to be an
19°
VARIABLES MAJOR FINDINGS
II. OBJECTIVES (Cont'd)- explanatory variable at thenational leve land in the regions
F. Personal (Lifestyle) (Cont'd) they studied (1978). GAO foundrural physicians more influenced
2. Desire to be by the high medical need in anof help*(Cont'd) area (GAO, 1978).
152
VARIABLES MAJOR FINDINGS
III.DECISION FACTORS
A. Environment
1. Culture
2
153
Virtually all conceivableaspects of.living conditions havebeen reported as of moderate tomajor importance in locationdecisions.
The availability of "cultural-resources" is reported by mostinvestigators to be a highly signi-ficant determinant of locationchoices, but the types of resourceshave not been identified and themeaning of the"term is not precise(Yett, 1973). Park and Tuxillfound that approxim4,-.7 80 per-p.ent,of physicians ratea "possiblescarcity of cultural events andplaces of entertainment" as factorsdeterring them from practicing ina small community (1967). Schaupprated "acceptable social andcultural environment" as one ofthe three most important factorsin location choice (1969).Crawford and McCormack found thatcommunities with a profile of defi-ciencies in physician numbers haveinadequate cultural resources(Crawford and McCormack, 1971;DREW, 1974). Cooper et al. notedthat urban physicians feel thatopportunities for social life andcultural advantages are importantfactors (1975). Hassinger reportedthat lack of. cultural activitiesranked highest among disadvantagesreported by rural physicians.
19 ;
VARIABLES MAJOR FINDINGS
III.DECISION FACTORS (Cont'd)
A. Environment (Cont'd)1. Culture (Cont'd)
2. Recreation
3. Housing:
4. Community Security
154
Although it is Yett's subjectiveimpression, this disadvantage wasa cliche among rural doctors anddid not appear to be serious (1979)
The extent of a community'srecreational facilities somehowappears to affect the geographicdistribution of specialists, butnot of GPs (Martin, 1968;HarOleton, 1971). Hambleton pin-pointed certain indicators ofleisure time activities--publicbeaches/campsites; tctnnia courtsand recreational spending--asexertinia strong influence on theState.location of specialists, butless so on the county. Biblefound that avocational opportuni-ties such as hunting or fishingwere considered desirable attrac-tions of a small community (1970).Hassinger reported that the disad-vantage "too far from recreation".(water sports and skiing) wascited almost exclusively bymetropolitan doctors (1979).
McFarland listed housing as afactor in location decisions whichis an inefficient policy variable(not subject to manipulation)(1972). Shannon and Dever statethat the most important considera-tions for site selection seem tobe the availability of housing andequipment (affecting costs ofsettlement) as well as otherprofessional considerations (1974).
McFarland also listed communitysecurity as a factor in locationdecisions which is not subject toa great deal of policy manipulation
1 t
VARIABLES MAJOR FINDINGS
III.DECISION FACTORS(Cont'd) (1972). Wilson found rural physi-cians had a prefervcP for living
A. Environment_(Cont' in the same community as one's
4. CommunitySecurity:
5. Transportation:
. Schools:
patients and being safe and secure(Cont'd) (1971).
155
McFarland'also cited intra-regional transport as a policyvariable not subject to a greatdeal of manipulation (1972).Thirty -two' percent of physicianscited "distance to hospital" as aliability in isolated ruralcounties as opposed to 12 percentin metropolitan areas (Bible,1970). Physicians listed lesstraffic and confusion as assets ofrural practice (Bible, 1970).Sixteen percent of respondents toParker and Sorensen's study foundthe distance from their office tothe hospital they used was toogreat, and this was the reason fortheir leaving a rural community(1978). Wilson demonstrated thatrural physicians preferred ruralpractice because they avoided a
lengthy commute to the office(1979).
Bible found that one of thefactors pertaining to communityliving which was of concern wasthe limited availability of educa-tional facilities (Bible, 1970).It was reported that the qualityand accessibility to primary andsecondary education is an importantconsideration in location plans(Charles, 1971; Long, 1975; Yett;1973). Crawford and McCormackstudied physician satisfactionwith aspects of rural life andfound that schools were described
1 9c
VARIABLES MAJOR FINDINGS
III.DECISION FACTORS (Cont'd)
A. Environment (Cont'd)
6. Schools (Cont'd)
7. Pollution:
8. Shopping:
156
as adequate or better. (1971)Schaupp found that of the threemost important factors in locationchoice, availability of good educa-tion facilities for children rankedfirst (1969). In a study comparingAppalachian rural physicians withthose practicing in Knoxville,Champion reported that urban educa-tional facilities were importantfor physicians with school agechildren (1971). The data suggestthat perceptions of characteristicsof schools may be influential; thatis, desirable characteristics neednot attract physicians but undesir-able ones may keep them away.
In sum, the qua/'ty of commu-nity life, includin, ducationalopportunities, appear to influencethe physician in his choice oflocation (DREW, 1974).
McFarland lists pollution asan environmental variable whichaffects location decisions butallows for little chance of policycorrection. (McFarland, 1972).
Also listed under environ-mental, quality of life, nonprofes-sional attractions of communitylife is access to shopping(McFarland, 1972). Crawford andMcCormack found that shoppingfacilities appeared to be adequatefor rural practitioners in hisstudy (1971). Dewey studied datafor 1950, 1960 and 1970 andobserved that suburban shoppingcenters and office buildings hadhigh attractive potential forphysicians (1973). The general
VARIABLES MAJOR FINDINGS
III.DECISION FACTORS (Cont'd)A. Environment (Cont'd)
Shopping (Cont'd)
9 Climate;
157
pattern of migration over recentdecades has been from colder towarmer sections of the country.
The climate of an area, itsgeographic features, or both, aregenerally ranked among the two orthree most influential of allprofessional and personal factors(Fein, 1956; Charles, 1971, Role-ston, 1973; Heald et al., 1974;Steinwald and Steinwald, 1974,Longnecker, 197 o)l- In one small-sample study, chaupp (1969) foundclimate prefeences to be an unim-portant decision factor to gradu-ates of West Virginia, but inanother study Martin et al.,(1968) reported that they ,stronglyaffected the decisions of Kansasgraduates to leave or remain inthe State. Physician-populationratios are reported to be signifi-cantly correlated with climate(Yett and Sloan, 1974; Martin,1968; Coleman, 1976). DeVisereports that an examination of theten States with the "highest physi-cian-population ratios confirmsthe strong attraction that Eastand West coast centers of glamourand good climate exert on physi-cians." (1973) Long states thatthe best established results arethat "on the average, physiciansare attracted to regions havingwarm winter climates." (1975)
A 1972 survey of 1,965 medicalschool graduates found that life-style factors such as climate,were among the major ones listedin location decisions (Steinwaldand Steinwald, 1974). The
VARIABLES 7 MAJOR FINDINGS
III.DECISION FACTORS (Cont'd) Illinois Board of Education foundthat climate was reported as being
A. Environment (Cont'd) the prime reason why physiciansleft the State (IOM, 1976).
9. Climate (Cont'd) Wilson found physicians who werein undesirable National HealthService Corps sites were retainedin that site if they were tolerantof adverse climatic conditions(1979).
fa
10. Population Growth:Change
Unfortunately, no substantive .
attempts have been made to determine what types of climates orgeographic characteristics physi-cians prefer. This limits thevalue of attitudinal evidence corspecifying and testing hypothesesconcerning the effects of climateand geography on location behavior.
Factors relating to communityliving which were of considerableconcern to respondent's of Bible'ssurvey included lack of a growingand thriving community (1970)New, physicians have been reportedto be attracted to areas character-izediby recent population growth(Rimlinger and Steele, 1976).Using State data, Fein and Weberfound that population growth wasthe most important single factoraccounting for the variation innumber of new physicians locatingin each State (1971). Rimlingerand Steele found, however, that anincrease in population is onlyeffective in attracting physiciansto urban, not rural areas.Mountin, Pennell and Brockettfound that population growth mustbe accompanied by high income(1945).
158
-1 I
VARIABLES MAJOR FINDINGS
11. Economi Growth: Smith supported the generalhypothesis that physicians preferto locate in stable economic areas(1961). Sloan corroborated thisby finding that physicians tend tolocate in areas subject to lesscyclical variations in income(1968). Fahs and Peterson foundthat 64 percent of "no-physician"towns had a decline in economic
12. Urbanization:
Population Densit
growth rates (1968).
Parker, Rix and Tuxill foundthat areas with a relatively highpercentage of the population
a. Proportion ofPopulation -Agriculture
159
engaged in agriculture and with alow percentage of high school grad-uates are disadvangated in attemptsto attract and retain physicians(1969). Several studies have notedthat urbanization also had signifi-cant effects on physician location,although the separate influences ofper capita income and urbanizationare difficult to determine (Steeleand Rimlinger, 1963; Benham, 1968;Hambleton, 1971). Joroff andNavarro (1971) and Scheffler (1971)explained large differences inphysician distribution by localdifferences in population (as wellas affluence and medical schoolproduction). Benh4ff found thatphysicians are more responsive intheir location choices to the sizeof the population than to its percapita income (1968). Kane (1979)stated thats less remote ruralarea having a higher populationdensity is less like_y to be defi-cient in health personnel. Biblecites Rix's emphasis on the accele-rating influence of urbanizationduring the past few decades in
VARIABLES MAJOR FINDINGS
III.DECISION FACTORS (Cont'd)
A. Environment (Cont'd)
attracting young physicians tolocate in larger centers ofpopulation. (Bible, 1970)
12. Urbanization: Marden concluded that popula-tion size is significantly and
Population Den ity(Cont'd) positively associated with thedistribution of, private practi-
a. Proportion of tioners (1966). Evashwick foundPopulation - population density to be salientAgriculture (Cont'd) in explaining physician locafion
in 1960 and 1970 (1976).
Yett and Sloan demons'_r.tedthat the .percent of the StaLe'spopulation residing in StandardMetropolitan Statistical Areasproved to be significant in one oftwo cases relating tit,specialists,and in the case of general practi-tioners the data indicated thatthey are attracted to less denselypopulated areas. (1974) Wilsonfound critical values on certaincommunity characteristics, suchthat above the critical value thecharacteristic becomes important.Having i8 county-wide populationdensity of 10 persons per squaremile or less was one such charac-teristic (Wilson, 1979).
b. Proportion According to the Humanof Population Resources Research Center, thereProfessional are some indications that, on the
average, GPs choose relativelypoor and elderly populations,while non-GPs locate in areas ofrelatively high educationalattainment (Yett. 1973).
160
.
Since urbanization as anindependent variable suggests no
VARIABLES MAJOR FINDINGS
III.DECISION FACTORS (Cont'd)
A: Environment (Cont'd)
12. Urbanization:Population Den.(Cont'd)b. Proportion of Pop.
Professional(Cont'd)
13. Racial Mix
161
policy initiatives, the effect ofpreference for urban areas byphysicians, should be examined inthe context of urban-ruraldistribution, and in the light ofcharacteristics of urban-ruralareas that attract or repelphysicians.
Cooper, Heald, and Samuelsmaintain that the influence offamily and friends as it appearsin, survey responses may be a 1
surrogate for ethnic or racialconcerns of physicians (1972).Lieberson concluded that the desireto maximize income appears to be aconsideration for physicians tolocate among their own ethnicgroups (1958). Elesh andSchollaert found that race producesan 18 percent difference, in thenumber of physicians between blackand white census tracts (1972).Physicians "do avoid practice inblack areas, and the avoiders arechiefly general practitioners."(Elesh and Schollaert).
Researchers do not agree onthe importance of race as a
determinant of intraurban practicelocations. Hambleton concludedthat more physicians! offices werelocated in postal zones with a
large black population (Hambletdn,1971). Elesh and Schollaert foundthat physicians did not locatetheir offices in areas with alarge black population. However,in a study of census tracts inBrooklyn, New York, race was notfound to be an important variable "(May, 1970). Wilson demonstrated
2')r)
VARIABLES MAJOR FINDINGS
III.DECISION FACTORS ( Cont'd)
A. Environment (Cont'd)
13. Racial Mix ( Cont'd)
14. Wealth of area(see also IIA-Income(Maximazation, IIIC lePer Capita Income,and III C2 PhysicianIncome in Area)
that having a county-wide non-white population greater than 20percent of the total was related'to retention in National HealthService Corps sites (1979).Non-white physicians were somewhatmore likely to be retained in thesite community under thesecircumstances.
A number of reports haveconcluded that physicians tend tolocate in high income areas, butthe relative importance of incomevaries (Rimlinger and Steele, 1965;Sloan, 1970; Fein and Weber, 1971).
According to several studies,high income areas offer a widevariety of attractive featuresthat may operate independently toattract physicians (Long, 1975).
Benham et al., conclude thateffective ,demand for physicians'services within a State depends litper capita income after populatiCin(1968). Hambleton's analysis showsthat per capita income was the mostimportant influence in the choiceof practice location of generalpractitioners, but had littleinfluence on a specialist'sdecision (1971). Ball and Wilson,on the other hand, found that"effective buying income" was amore relevant factor in the loca-tion of specialists than in thelocation of general practitioners(1.968).
162
2 "
VARIABLES MAJOR FINDINGS
IIi.DECISION FACTORS (Cont'd)
A. Environment ( Cont'd)
14. Wealth of, area(see also IIA-Income(Maximization, IIIC lePer Capita Income,and III C2 PhysicianIncome in Area) Cont'd)
15. Size of Community
163
Cooper cites the economy ofthe community as an impOrtantdetermining factor in locationchoice (1972). "Towns with lowmedian income or an otherwise pooreconomy are in an unfavorableposition for attracting orretaining physicians." Forty-twopercent of physicians wereinfluenced to settle in a smallcommunity because it was a"prosperous pait of the country"(Parker and Sorensen, 1978). 7Wilson found a county-wide medianfamily income of $8,500 or lessnegatively related to retention ina National Health Service Corpssite (1979).
In summary, area wealth affectsthe location of physicians eitherbecause high income induces rela-tively high demand for physiciansservices, thus creating attractivepractice opportunities, or becausehigh income areas are otherwiseattractive to physicians becauseof their professional, social andcultural environments (Yett, 1973).
When Benham examined theabsolute number of physiciansacross geographic areas, populationsize emerged as the single mostimportant determinant of the physi-cian supply (1,968). In a Kansasstudy, population size of countieswas also reported to be tie mostimportant factor in determiningphysician location (Marshall,1971). Fahs and Peterson found a
clear relationship between a town'spopulation and its physicianstatus. In the States and towns hestudied, 187 had no physician, butof those only four had a population
20
VARIABLES MAJOR FINDINGS
III.DECISION FACTORS (Coned) greater than 1,000 (1968). Joroffand Navarro, on the other hand,
A. Environment (Coned) found population not an importa-tdeterminant of variation in pijsi-
15. Size-of Community(Coned) cian-population ratios (1971).Marden's analysis of 204 SMSAs end165 cities found population sizeto have a strong effect, but otherfactors were not held constant(1966).
16. Terrain
II
17. Religion
18. Population Mobility
164
Thirty-five percent of respon-dents cited climate or geographicfeatures of an area as the firstranked factor in location decisions(Cooper et al., 1975).
Although some research hasbeen done on religious backgroundby specialty, the relationshipbetween religious background andlocation choice has not beenstudied (Yett, 1973). Theret areindications that non-religious andJewish physicians are dispropor-tionately likely to choose.psychiatry, academic medicine orinternal medicine (Kritzer anuZimet, 1967). These choices maydictate location choices. Parkerand Sorensen recorded comments byphysicians regarding religionincluding the following: "thereis . . . the realistic notion thatin this still religion-consciouscountry minority groups may findlimited acceptance in smallercommunities as well as limitedreligious facilities" (1978)..
Areas with low populationmobility are disadvantaged inattempts to attract and, retainphysicians (Parker, Rix andTuxill, 1969).
021)?
VARIABLES MAJOR FINDINGS
19. Relocation andSearch Costs
20. Welfare Support
21. Zoning
22. Public Services
165
It Xs assumed that if aphysician chooses a more populousarea, he can rely more on informa-tion supplied by pharmacists,bankers, his medical school andother doctOrs, reducing the costsof his search. One of the mostimportant considerations for siteselection seems to be the avail-ability of housing and equipment,which affect the costs of settle-ment (Shannon and Dever, 1974).
The availability of good socialservices, welfare, or home careservices was found to be a nonsig-nificant factor in Cooper et al.'s,reg-ession analyses of primary care
_ ,-ns in urban areas (1975).How'. for rural areas, goodsocial service or welfare agencieswere much more important.
Kaplan and Leinhardt noted thatcensus tracts with small amounts ofcommercial zoning do not seem toattract physicians. It is onlywhen a tract has a major amount ofcommercial activity that largeconcentrations of physicians exist(1973). Elesh and Schollaertstudied census tracts in Chicagoand found that percent commercialand in the tract had a positiveand significant coefficient, in theall physicians, general practi-tioner and specialist equaiions(1972).
McFarland lists the provisionof public services as an/environ-mental variable affecting locationchoices which is an inefficientpolicy variable (1972). Sloanfound that physician-populatio
VARIABLES MAJOR FINDINGS
III.DECISION FACTORS (Cont'd)A. Environment (Cont'd)
22. Public Services (Cont'd)
B. Practice Related
1. Availability, Qualityand Quantity ofHospital Services
166
ratios were significantlycorrelated with State and localpublic service expenditures(Sloan, 1968).
A large number of studies havedealt with the question of theavailability and proximity ofhospital services as a majordeterminant in physician location.(Parker and Tuxill, 1967; Bible,1970; Charles, 1971; Heald et al.,1974) The availability of hospi-tals is said to attract physicianssince nearly all physicians needhospital privileges in order topractice. Explanatory variableshave included the number ofhospital beds, the number of bedsper capita, the number of hospitalsin the area, and the percentageincrease in the number of hospitalbeds (IOM, 1975).
One study supported by theHuman Resources Research Centerfound a weak relationship betweenthe total number of physicians, ortotal physician-population ratios,and hospital capacity. They founda strong positive relationshipbetween the number of specialistsand hospital capacity. No causeand effect relationship betweenhospital capacity and physicianlocation was demonstrated. Theyconcluded that the role of thehospital in location decisions isexceedingly complex (Yett, 1973).
A number of studies indicatethat the hospital bed rate is animportant predictor of the loca-tion of some specialties, althoughnot of general practitioners.
2') ,
VARIABLES MAJOR FINDINGS
III.DECISION FACTORS (Cont'd) (Joroff and Navarro, 1971, Reskinand Campbell, 1974; Yett and
B. Practice Related ( Cont'd) Sloan,1974). Marden, on the otherhand, found the hospital bed rate
1. Availability, Quality had little influence on theand Quantity of location of specialists,Hospital Services ( Cont'd) especially in small cities; he
reported more GPs in thosemetropolitan areas having fewerhospital facilities (1966).
-167
Other studies indicate that theexistence of hospital facilities ina rural area will tend to drawphysicians to that area (Williamsand Uzzell, 1960; Royce, 1972).Bible's survey of physiciansrevealed that the "best openingwhen ready to practice" includedavailability or medical facilities.About half the respondents gavethis as the most important factorin decision location, in additionto geographic preference (1970).Parker and Sorensen found that 82percent of physicians named "agood community hospital" as afactor that influenced a physicianto settle in a small community411978).
Champion and Olsen, in asurvey of Appalachian physicians,found that the most frequentlycited disadvantage of medicalpractice was lack of facilitiessuch as hospitals and equipment(1971). Shannon and Dever cite astudy which shows a strong nega-tive aasc ation between GPs andhospitals: A decrease of 5.4 forevery new hospital per 100 squaremiles (Hambleton, 1971). In urbanareas Kaplan and Leinhardt foundthat if 100 hospital beds are
2i-1n
VARIABLES MAJOR FINDINGS
III.DECISION FACTORS (Cont'd)
B. Practice Related (Cont'd)
1. Availability, Qualityand Quantity of HospitalServices (Cont'd)
2. Availability ofAncillary Services(Laboratory, pharmacy,x-ray, office)
168
added to census tract it would,on the av age, increase physicianoffices in that tract 2.6 andincrease offices 1.2 inneighboring tracts (1973).
Yett summarized the results ofseveral studies, indicating, thatat the State and county levels,the number and number per caPitaof nongeneral practitioners wasvery positively associated with\the area's hospital capacity,whereas the number of generalpractitioners was not. The impli-cation is that hospitals attractreferral specialists, but exert amuch weaker influence on mostprimary care practitioners. Dataalso indicate that internistsproduce substantial percentages oftheir total patient visits inhospitals, and, among the primarycare specialists, internists areprobably relatively more attractedto those areas having the mosthospital capacity. In cities,evidence suggests all physiciansare attracted by the proximity ofhospitals (1973).
\About 90 percent of large andsmall\ community physicians surveyedsaid that the notion that smallercommu4ities lack complete facili-ties Mas an important deterrent tolocating there (Parker, Rix, and,Tuxill, 1976). Champion and Olsenreported that when asked to iden-tify deterrents to Appalachianpractice, physicians mostfrequently cited the inadequacy ofgeneral medical\Tparaphernalia"
VARIABLES MAJOR FINDINGS
III.DECISION FACTORS ( Cont'd)
B. Practice Related (Cont'd)
2. Availability ofAncillary Services(Laboratory, pharmacy,x-ray, office) ( Cont'd)
3. Availability ofPersonnel
4. Availability of' Medical Schools
and medical facilities. Lack ofmedical facilities was also citedin Bible's survey '(1970).
On the other hand, Ducker, ina more recent study, found that 84percent of physicians sampled inrural areas expressed satisfactionwith tteir office facilities(1977).
Attitudinal and interviewstudies indicat.1 the availabilityof clinical support personnel mayaffect physicians' locationchoices, although it is probablymore important for non-generalthan general practitioners. TheInstitute of MOicine stated thatthe availabili: of supportinghealth personnel may influence thelocation decisions of physicians,but investigations into thisrelationship are tOo "sparse andresearch results too ambiguous towarrant reporting" '(I0M, 1975).Physicians were reported not to besatisfied with their currentpractice in a rural area becauseof a shortage of physicians andallied health 'personnel (Bible,1970). McFarland lists "alliedhfialth personnel" as a professionalrelationship factor which is apotential policy variableaffecting location decisions.Lack of support personnel wasalmost exclusively reported byrural ph3rsician,in assinger'sstudy (1979). '12\
A substantially higher per 1
capita concentration of physi-cians, especially specialists, isfound in areas having medical
169'
VARIABLES MAJOR FINDINGS
III.DECISION FACTORS (Cont'd) training centers (Coleman, 1976).this concentration may be
B. Practice Related (Cont'd) attributed to the attraction oftechnical knowledge available
4. Availability of there, it is difficult to separateMedical Schools (Cont'd) this effect from that of urban
amenities (Shannon and Dever,1974). Joroff and Navarro reportmore conclusivo results: The:number of ical schools was thesingle mo important predictor ofthe location of eight specialties(1966). Scheffler also found thisto be a factor (1972). Parker andSorensen found a medical centernearby to be a factor influencingphysicians to settle in a smallcommunity (1978). Heald, Cooper,and Coleman (1974) also concludedthat among the most importantfactors in choice of location wereopportunity for regular personalcontact with a medical school ormedical center.
According to Ball and Wilson(1968) and Yett and Sloan (1970),the existence of medical schoolsin an area increases the physiciansupply because it creates opportu-nities for practice, particularlyfor specialists. The existence ofa medical school is thought tocreate an environment that makesthe area attractive for medicalpractice. Medical schools are thefocus of professional interaction,provide opportunities for ,contin-uing education; and otherwiseserve as a magnet for physicians(Long, 1975). Thus, nearness to amedical school has been reportedto be a moderately influentialfactor, but less important toprimary care physicians than to
170
VARIABLES MAJOR FINDINGS
5. Availability ofConsultations
171
referral specialists. (Fein,1956; Charles, 1971).
Bible's survey of nonmetro-politan physicians indicated thatthey were concerned with'thelimited availability of medicalfacilities and consulting services(Bible, 1970). Attitudinalfindings indicate that some physi-cians seek areas where they haveopportunities for contact withprofessional colleagues (Yett,1973). Although there are competi-tive forces leading physicians tochoose areas with low concentra-tions of established physicians,there are other forces which leadthem to prefer areas allowing amplecontact with other physicians(Yett, 1973). Heald, Cooper, andColeman cited opportunity forregular personal contact withother physicians as an importantfactor for the 1,161 physiciansthey studied (1974). A recentstudy by Ducker reportf2d that, only37.5 percent of physicians sampledwere satisfied with their oppor-tunities for professional contactsin a rural practice (1977). Asurvey by Parker and Sorensen in1978 revealed that 49 percent ofphysicians who left rural practicedid so because they were unable torely on the physicians in thatmedical community for much profes-sional support (Parker andSorensen, 1978).
In the QAO study physitiansindicated that if access to contin-uing education and consultingphysicians had been available, theywould not have rejected shortage
VARIABLES MAJOR FINDINGS
III.DECISION FACTORS (Cont'd)
B. Practice Related (Cont'd)
5. Availability ofConsultations (Cont'd)
6. Emergency MedicalServices
7. Licensure
172
locations (1977). Parker andSorensen's report on personalcomments as a result of four'surveys of physicians providesthis summary: "It is apparentthat new physicians are dependenton colleagues, a good communityhospital, consultants, and a
medical center in a nearby city."On the other hand, recent researchalso suggests that ruralphysicians tend to adapt to a
limited scope of supportingservices and facilities. Thisinvolves a shift in attitudes andvalue2 (Ducker, 1977). Wilsonalso reported that rural physi-cians seem to have a somewhatlower concern than other specia-lists with consultations (Wilson,1979).
Bible noted that one of fivefactors in which there was a signi-ficant difference between therating of''respondents in the motepopulated rural counties and theisolated rural counties was"facilities for handlingemergencies." (1970).
The hypothesis thai physicianswill be attracted to States withlow failure rates on medical licen-sure examinations was supp ted byBenham (1968). Scheffler found asignificant negative coefficientin four years and a significantpositive coefficient in one yearout of twelve in his physiciansper capita equation. It.wassuggested that a positive coeffi-cient may indicate that thelicensure variable measures quality(1971). On the other hand, several
2 1 '>
VARIABLES MAJOR FINDINGS
III.DECISION FACTORS (Cont'd)
B. Practice Related (Cont'd)
7. Licensure,(Coned)
8. Group PracticePossibility
studies found the failure rate onthe State licensure examination ispredicted to negatively affect aState's supply of physicians(Rimlinger and Steele, 1963;Sloan, 1968; Yett and Sloan,1970). It should be pointed outthat the effect of licensurerestrictions should be strongeston the locational behavior of newphysicians rather than on thetotal physician upply in a State(Long, 1 . There is indirectevidence that State board policiestoward foreign medical graduatesare not uniform, and that dispari-ties in licensure policies partlyaccount for the maldistribution offoreign graduates at the Statelevel (Yett, 1973).
Group practice opportunitiescan attract physicians to ruralareas (Cantwell, 1975; Charles1971; Heald, 1974). Over time,non - metropolitan areas that have arelatively high percent ofphysicians in group practice dobetter in maintaining and increa-sing their supply of physicians(Evashwick, 1976). Cooper et al.,in a mail survey of all 1965medical school graduates, foundthat 75 percent of new physiciansprefer group practice (1972).Other researchers found theavailability of group practice tobe an important policy variable(Bible, 1970; Steinwald andSteinwald, 1975; and McFarland,1972). Steinwald and Steinwaldfound that, overall, 57 percent of.all graduates indicated thatopportunity to join a desirablegroup practice or partnership was
173
2 1
VARIABLES MAJOR FINDINGS
III.DECISION FACTORS (Cont'd)
B. Practice Related (Cont'd).
8. Group PracticePossibility (Cont'd)
of prime importance in locationdecisions (1975). There is someevidence that multi-specialtygroups in rural areas have thepotential to act as medical"magnets" for doctors who desirefewer management duties andearlier economic returns inaddition to colleague support(Kralewski and Luke, 1979).
Since most group practices arein urban areas, it would seem thaturban group practice would beeasier to join. There are likelyto be more nonindigent patientswho have readier access totransportation (Cooper el1972). Since rural communitiesare typically too small to supportgroup practices, the implicationis that preference for grouppractice is leading them to avoidrural locations. Rural physicianshave often been solo practitioners,and they commonly report that ruralpractice would be more attractiveif coverage were available. Thedisadvantage of solo practice was
0 cited by 15 percent of the ruralM.D.s in Hassinger's study (1979).
174
Thus, there are some indica-tions that established grouppractices are attractive tophysicians not otherwise`discouraged from rural practice,but the evidence is mainlyqualitative (Yett, 1973) and oftencontradictory (Coleman, 1976).For example, Coleman found grouppractice not to be related to theurban-rural choice for any of thethree specialty groups. Interestin group practice opportunities
VARIABLES MAJOR FINDINGS,
III.DECISION FACTORS (Cont'd)
B. Practice Related (Cont'd)
8. Group PracticePossibility (Cont'd)
9. Contact with OtherProfessionals
was more strongly related to adesire for amenities than toprofessional interests. Wilsonfound rural physicians preferredsolo or small group practice(1979). Cotterill and Eisenbergconcluded that general factors'.:which affect location decisions ofall physicians also affect thedecisions of physicians to joingroups (1979). They also demon-strated that group practice ingeneral does not overcome theinhibiting effects of isolationnor any of the perceived disad-vantages of medical practice inrural areas. The feasibility ofsubsidizing groups as a device forattracting physicians into ruralcommunities is not yet fullyestablished.
About 43 percent of physicianssampled by Parker and Sorensenfound that they were not able toobtain the professional andeducational opportunities theydesired in a small community(1978). Thirty-nine percent saidthAiere too few peers to relateto, cially and intellectually.
About 40 percent of largecommunity physicians interviewedby Parker and Tuxill cited thepossible scarcity of nonmedical1.ntellectual companionship as adeterrent to small community/practice (1967). Over 60 percentof small and large community
'physicians felt it to be an/important, deterrent in general.Bible observed that liabi4ties inrural practice weretlimited socialactivities and opportunities for
175
2 I
VARIABLES MAJOR FINDINGS
III.DECISION FACTORS (Cont'd)
B. Practice Related (Cont'd)
9. Contact with OtherherProfessionals (Cont'd)
10. Continuing EducationPossibilities
A
17G
professional growth (1970).McFarland noted that professionalrelationships were*an importantaspect of a satisfying professionallife (1972). Cooper et al., founda positive relationship betweenthe educational level of thecommunity and number ofspecialists with the ability toattract physicians (Ccu,r, 1972).
It should be noted thatprofessional factors are usuallylinked together as a set: Medicalschool contact, the presence ofphysician specialists, and contin-uing medical education are part ofwhat many physicians consid r aprofessionally desirable env ron-ment (Cooper et al., 1977).Several studies have cited accessto continuing education as animportant consideration inlocation choice (Bible, 1970;Castleton, 1970; Heald, 1974; GAO,1977; Wendling, 1978). About 48percent of physicians in Ducker'srecent sample reported they weresatisfied at some level with theopportunities for continuingeducation available to them.Rural M.D.s in Hassinger's studywere less satisfied with opportuni-ties for continuing education thanj-5metropolitan physicians (1979).Respondents in the Taylor et al.,survey of medical students indi-cated that regardless of wherethey planned to locate theirpractice, they needed opportuni-ties to keep current (Taylor,Dickman and Kane, 1973). A Randstudy indicated that for twonon-GP specialty groups, access tocontinuing education influenced
2 I1 '."
VARIABLES MAJOR FINDINGS
III.DECISION FACTORS (Cont'd)
B. Practice Related (Cont'd)
10. Continuing EducationPossibilities (Cont'd)
11. Costs of Practice
177
location choice toward an urbanlocation. Thus physicians who areunconcerned with urban "amenities"may be recruitable to rural areasby an improvement in the profes-sional environment (Coleman, 1976).
McFarland lists opportunityfor continuing education as apotential policy-variableaffecting location decisions(1972). Parker and Sorensenpropose that "more adequateprograms of continuing educationcould do an enormous amount tolessen the ebb of physicians fromrural America" (1978). For thepracticing rural physicians,bringing them to the medicalcenters for continuing educationmay be less appropriate thanbringing the medical centers tothe physicians through continuingmedical education (Parker andSorensen, 1978). Despitecontinuing education, ruralphysicians are often reluctant touse new procedures and approachesbecause they have little oppor-tunity to practice those skills(Owens et al., 1979).
McFarland lists costs ofpractice as an economic factoraffecting geographic location(1972). Rated very low by Healdas an influential factor affectinglocation choice was "availabilityof loans for beginning practice"(1.4 of physicians percent sampledrated it 1, 2 or 3) (Heald et al1974). Hambleton included coststo the physician of deliveringservices such as rent, educationand health personnel and found
2' c)-A. Li
VARIABLES MAJOR FINDINGS
III.DECISION FACTORS (Cont'd) that at the postal zone level,rents and other costs of estab-
B. Practice Related (Cont'd) lishing a practice were one of twofactors which seemed to have the
11. Cost, of Practice (Cont'd) most impact on practice location(Hambleton, 1971). Feldsteinanalyzed costs per unit of serviceprovided, in addition to theavailability of interesting cases,and found that demographic charac-teristics of an area may be suffi-ciently influential to induce a
physician to locate where he other-wise might not satisfy simpleprofit maximization objectizy(Feldstein, 1970).
12. Best AvailablePosition, QuickPractice Start
13. Work with EstablishedPhysician
178
The reasons most commonlymentioned for choosing a practicelocation in a 1971 survey ofphysicians were best opening whenready to practice and geographicpreference (Bible, 1970).Forty-six percent of respondentswho settled in a rural communityin New York did so because "a goodpractice could be built upquickly" (Parker and Sorensen,1978). An earlier study of bothlarge- and small-community physi-cians revealed that one of thethree most important factorsinfluencing location decisions isthe "prospect of building a busypractice earlier" (Parker andTuxill, 1967).
Parker and Tuxill found thatabout 40 percent of both rural andurban physicians rated "opportu-nity to practice with an estab-lished physician" as an importantfactor likely to influence small-town location choice. Largecommunity physicians found it not
211i41- tar
VARIABLES MAJOR FINDINGS
III.DECISION FACTORS (Cont'd)
B. Practice Related (Cont'd)
13. Work with EstablishedPhysician (Cont'd)
C. Demand Related1. Population
Characteristicsa. Size of population
179
to be a factor deterring thempersonally from rural practice(1967). Twenty-four percent ofParker and Sorensen's physicianswere influenced to settle in asmall community because "an olderestablished physician needed apartner" (1978).
According to Benham, when theabsolute number of physiciansacross geographic areas isexamined, population size emergesas the single most importantdeterminant of the physiciansupply (Benham et al., 1968).Marden has also shown a strongrelationship between populationsize and number of physicians/specialists (1900). Fahs andPeterson found population veryimportant in his study of upperMidwest States -- there appears tobe a threshold population neces-sary to support a physician (1968).Elesh and Schollaert argue thatgiven a community's ability topay, cultural predisposition, andneed, physicians will distributethemselves with regard to popula-tion composition. As populationsincrease in size, the magnitudesof the demand factors also increaseand they can support morephysicians (1972).
New physicians are attracted toareas characterized by recentpopulation growth, especially ifit is caused by migration(Rimlinger and Steele, 1963; Feinand Weber, 1971). Although popu-lation growth is similar to popula-tion size, an increase in popula-tion only attracts physicians to
VARIABLES MAJOR FINDINGS
III.DECISION FACTORS (Cont'd)
C. Demand Related (Cont'd)
1, PopulationCharacteristicsa. Size of population
(Cont'd)
b. Age of Population
180
urban areas (Rimlinger and Steele,1963), and population size ofurban areas is positivelycorrelated with the location ofvarious groups of specialists butnegatively correlated with GPs(Yett and Sloan, 1974). Althoughpopulation size and growth areimportant determinants of numbersof physicians, it is analysis ofthe variation in the number ofphysicians relative to population(per capita) that is not affectedby size and growth (Long, 1975).While it is the single mostimportant determinant, there is a
significant amount of variationleft unexplained by populationvariables per se (Long, 1975).
Some studies have shown thatpercentage of the population 65years and older is a significantfactor in the location choice ofgeneral practitioners, whereas ahigh level of education in thecommunity favorably affectsspecialist's locations (Marden,1966; Joroff and Navarro, 1971).Coleman found percent over 65 tobe related to the physician/popula-tion ratio for both primary andnon-primary care physicians(1971). Concentration of thepopulation in very young and oldage groups positively affectsmorbidity, and thus the agecomposition of the populationaffects demand for physician'sservices and through demand,affects supply, (Long, 1975).Evashwick found .a positive rela-tionship between the physician/population ratio in 1960 and thepercent of the population over
VARIABLES MAJOR FINDINGS
III.DECISION FACTORS (Cont'd) age 64. This could be interpretedas an index of the amount of
C. Demand Related (Cont'd) demand for medical services or an1. Population index of the economic growth stage
Characteristics (Cont'd) of an area. However, she foundthat by 1970 the relationship had
b. Age of Population disappeared, indicating that(Cont'd) physicians had left, retired or
died (1976). At the postal zonelevel, specialists seem to heassociated with the percentage ofelderly population; a onepercentage point increase in theproportion of the population overage sixty-five results in a gainof 24.5 specialists per 100,000.This could be due to the tendencyof specialists to practice in thecommercial centers of cities . . .
"where many elderly reside"(Hambleton, 1971).
Elesh and Schollaert foundthat percent of the population 25years + had a positive and signi-ficant coefficient in the allphysician, GP and specialistsequations, (1972); however, thevariable chosen is not veryspecific to young or old age(Cantwell, 1975). Yett (1973)concludes that "there are someindications that, on the average,general practitioners chooselocations where the population ispoor and elderly. Disparitiesexist among the findings, and theconclusions are best regarded ashypothesee." (1973). Guzick andJahiel found that only elderlygeneral practitioners show signi-ficant estimates with respect tothe population over age 65 (1976).Since GPs tend to be older thanspecialists, it is possible that
181
2.)0
VARIABLES MAJOR FINDINGS
III.DECISION FACTORS (Cont'd)
C. Demand Related ( Cont'd)1. Pop. Char. ( Cont'd)
b. age of pop. ( Cont'd)a
c. Sex of population
d. Race of Population
e.i
182
the difference in elasticity oflocation estimates with respect toage 65 + is more tied to physicianAge than specialty status (Guzick,1976).
Reskin and Campbell looked atseveral variables, including per-cent of the population female, andfound all to be correlated withthe residuals from the number ofphysicians on population regres-sion equation (1974f. Cantwellsuggests that other covariantfactors were not held constant(Cantwell, 1975). The proportionof the population of women 15-44should be associated with thedistribution of physicians becausedue to childbearing, these womenhave special medical needs.(Reskin and Campbell, 1974).Marshall et al. also looked at thepercent of women 15-44 but did notfind this factor a determinant ofphysician location (1971).McFarland lists sex as a demanddeterminant not subject to policymanipulation (1972).
Elesh and Schollaert foundpercent black in census tracts inChicago to be negative in thegeneral practitioner (GP) andspecialist equations but signifi-cant in the GP (1972). Most black,inner-city areas are characterizedby low physician/population ratiosand, over time, a gradual decreasein the number of GPs accompaniedby an increase in population
(Roemer, 1966; McMillan et al.,1970; Robertsbn, 1970). Kaplanand Leinhardt, on the other hand,confirmed findings that physicians
VARIABLES MAJOR FINDINGS
III.DECISION FACTORS (Cont'd)
C. Demand Related (Cont'd)
1. PopulationCharacteristics (Cont'd)
d. Race of Pop. (Cont'd)
183
do not consider race as an impor-tant determinant in their decisionto locate their offices within acity (1973). Joroff and Navarroalso found race not to be a signi-ficant explanatory variable (1971).Marden had similar findings but hisdata suggested that the distribu-tion of GPs is affected by age andrace characteristics of the popula-tion (1966). Guzick and Jahielfound that office location of non-white and/or non-American physi-cians tends to be associated withareas having a population that isethnically similar to them (1976).
Hambleton found that a onepercentage point increase in thenon-white population increases thespecialists' population ratio by3.2. This "could be due to thetendency of specialists topractice in the commerdial centersof cities near black neighbor-hoods" (Hambleton, 1971). It
should also be noted that not onlydo racial groups vary in averagepersonal income, and as a conse-quence in their demand for medicalcare, but racial groups are charac-terized by distinctive tastes thataffect their demand for medicalcare (Marden, 1966). It appearsthat low levels of effectivedemand for physicians' services incertain neighborhoods, rather thanracial composition as such, maydiscourage physicians from settlingin them (Yett, 073). Patientsprefer to be treated by physiciansof the same ethnic origin, andphysicians locate in response tothe income opportunities createdby these demands (Yett, 1973).
2 .)
VARIABLES MAJOR FINDINGS
e. Income (Per capita)
Additional research is needed toclarify why physicians tend not tosettle in ghetto neighborhoods.
Fuchs and Kramer estimatedphysician distribution for 33States, in 1955 and found percapita income one of fourimportant variables (1972).However, the relevant variableswere so closely associated thatthey could not separate out thevarious hypotheses. Positiverelationships were found betweenthe number of physicians and percapita income at the State andcounty level (Weiskotten, 1960;Rimlinger and Steele, 1963; Cooperet al., 1972). In studying censustracts in Chicago, Elesh andSchollaert found that physicianslocated their offices in high-income areas with the implicationthat both poor and middle-incomegroups had difficulty obtainingaccess to physicians (1972).Sloan found the income of apopulation to have a positiveimpact on the supply of doctors(1968). Monmonier also concludedthat the dominant factor in thespatial distribution of physiciansis high per capita income (as wellas urbanization) (1972). Parker,Rix, and Tuxill too, found thatareas with low median income (aswell as other sociodemographiccharacteristics) are disadvantagedin terms of attracting andretaining physicians. Dewey andPetto (no date), Roemer (1966) andKaplan and Leinhardt (1973) also
/observed low-income areas withfewer physicians, while Guzick andJahiel in a later study found
184
291--
VARIABLES MAJOR FINDINGS
III.DECIBION FACTORS (Cont'd) positive elasticities of locationwith respect to area income (1976).
C. Demand Related (Cont'd) At the State level, Fein andWeber found insignificant correla-
1. Population tions between change in. per capitaCharacteristics (Cont'd) income and physician-population
ratios (1967). Hambleton demon-e. Income (Per capita) strated a positive relationship
(Cont'd) between income per capita andspecialists/GPs per 100,000.Rimlinger and Steele found high-income areas had larger numbers ofphysicians; this was caused by atendency. for patients to visitphysicians more often, to whomthey were able to charge higherfees (1965). Hambleton demon-strated that per capita income wasa minor factor in determining thelocation of specialists among theStates and within a city; however,it was related to location at thecounty level. High income areasoffer a wide variety of attractivefeatures that operate independentlyto attract physicians (Long, 1975).Confirmation is provided by anumber of studies that physiciansare indeed concentrated in areasof high per capita income, eitherbecause earnings opportunities areconcentrated there or for otherreasons. (See Coleman, 1976.)
f. Education ofPopulation
Areas with low median schoolyears completed and low percentageof high school graduates are disad-vantaged in attempts to attractand retain physicians (Parker,Rix, and Tuxill, 1969). Joroffand Navarro found median years ofeducation to be the most importantsingle variable in predicting thelocation of five specialties
185
200
VARIABLESMAJOR FINDINGS
III.DECISION FACTORS (Cont'd)
C. Demand Related (Cont'd)
1. PopulationCharacteristics (Cont'd)
f. Education ofPopulation (Cont'd)
g. Morbidity andMortality
186
(1971). Marden had similarresults, but found educationallevel of little importance to GPs(1966). Reskin and Campbellanalyzed 22 SMSAs in 1966 andfound education levels positivelyrelated for all specialties except
obstetrics-gynecology, and highlycorrelated with number of GPs(1974). Cooper, Heald, andSamuels found educational levelscorrelated with number ofspecialists (1972). Utilizationof medical services is higheramong those with relatively high
educational attainment, and as aresult physicians are concentratedin areas characterized by highattainment (Long, 1975). Sinceuse of specialists implies greatersophistication and knowledge aboutmedical care, specialists can beexpected to locate more respon-sively to a population's educa-tional level,than general practi-tioners (Elesh and Schollaert,1972).
Community health levels should berelated to physician distribution(Reskin and Campbell, 1970 .
Employing the concept of demand,we would expectkhat an unhealthypopulation would attract medicalpersonnel, yielding a positiverelationship between physiciandistribution and level of morbidity. On the other hand, a strongmedical presence should serve toreduce the extent of sicknesswithin the community. This wouldresult in a negative relationship
between physician distribution andcommunity health levels. Reskinand Campbell found a moderate,
VARIABLES MAJOR FINDINGS
III.DECISION FACTORS (Cont'd) positive relationship between allcategories of physicians and the
C. Demand Related (Cont'd) three measures of communityhealth. Since the relationships
1. Population are positive, they conclude thatCharacteristics (Cont'd) the "attraction" of medical
doctors to morbidity is strongerg. Morbidity and than their "effect" on health
Mortality (Cont'd) levels.
h. Health Insurance In addition to knowledge ofCoverage medical facilities, the ability to
purchase care is necessary foraccess to the services of a physician. According to Sloan, theextent of insurance coverage (asindexed by the percentage of thepopulation with insurance coverage) influences the physiciansupply in a number of ways, mostparticularly in its strong effecton physicians' income (Sloan,1968). Data from other sourcesare incomplete but-suggest thatextensions of insurance coveragemay not be having a desiredredistributional effect (Long,1975).
Rimlinger and Steele founddemand for physicians' services tobe a function of income, age, sexand possession of health.insurance.Demand alone, however, was notfound to be sufficient explanationfor the existing disparities inphysician supply (1963). Parkerand Sorensen's anecdotal dataindicate that physicians are moresatisfied in a rural practice whenthey are involved in a prepaidhealth care system, and where thepatients' ability to pay is of nocc rn (1979).
187
2.3PvC)t..)
'VARIABLESMAJOR FINDINGS
4
kr-/
i. Per Capita HealthCare Expenditures
Visits,
(hospital, M.D.)
188
The Human Resources ResearchCenter study found public expendi-tures on physician services 'o becorrelated with numbers ofphysicians, with the highestcorrelation for nonprimaryphysicians in office-basedpractice (Yett, 1970). In a 1977study Fuchs and Kramer comparedthe expenditures rendered byphysicians iri private practice forthree periods of time (1948-56,1956-66, and 1966-68). They founda trend toward physicians settingup practice in mote affluent areas(1972).
According to Rimlinger andSteele, one of the factors thatdoes influence physician locationchoices is number of physicianvisits on a regional basis (1963).Education and income have beenshown to be associated with physi-cian use rates (for data on theireffect on physician location seeIIIC 1 e and f). One of the vari-able's included in Hadley's equa-tion as a proxy for the demand forphysician services was annualnumber of patient visits per physi-cian; he found that visits didaffect the location choice ofphysicians who had no priorcontact in the State, and to alesser degree who had their medicalschool and GME in the same State(Hadley, 1973). 'shirty -six per-cent of Crawford and McCormack'sphysicians saw over 40 patientsdaily and two-thirds cited over-work as the reason for leavingprimary practice (1971).
4
'VARIABLES MAJOR FINDINGS
2, Physician Incomein Area
189
Statistical studies whichrelate the stock of physicians ina State to average net physicianincome typically show a weak ornegative association between thetwo variables. This, they implythat economic motives are ethersecondary or negligible factors indetermining the distribution ofphysicians (Hambleton, 1971iHadley, 1973; Yett, 1973; Yett &Sloan, 1974). On the other \hand,a few studies which have employedsupply-and-demand models show asignificant positive associationbetween the number of physiciansand average physician net income(Sloan, 1968). These studies didnot segregate physicians byspecialty, nor specify gross incomeas a measure of net income oppor-tunties. On the other hand, it maybe that population characteristicsare more accurate proxies forphysician permanent net incomes inan area than are average currentnet incomes. That is, if thestructure of demand and supply forphysician's services is changingin an area, current average incomelevels may reveal little of futurenet income trends. (Yett, 1973;Long, 1975.)
-
Several recent studies concludethat income does not seem to be animportant factor. Physicfan salaies are currently so high that the
lure of still higher income influ-ences few. (Bobula and Goodman,1978; Institute of Medicine, 1979)
In contrast to the hypothesisthat metropolitan physicians wouldemphasize income as one of the more
23r
VARIABLESMAJOR FINDINGS
III.DECISION FACTORS (Cont'd)
C. Demand Related (Cont'd)
2. Physician Incomein Area (Cont'd)
3. Price of Area Services
4.. Perceived ExcessDemand---
important factors in locationchoice, the nonmetropolitan orAppalachian physician sampled inone study reflected a greaterpreoccupation with income as a
principal motivating factor (Laveet al., 1975).
Thus, results are inconclusiveconcerning the relationship betweenincome and supply. It may be -hatphysician income is not an impor-tant influence in location choiceor that studies have not beendesigned in a way that would permita valid test of this relationship(Lave et al., 1975; Bobula andGoodman, 1979; Hadley, 1980).
The prevailing pricing prac-tices of physicians, which are toincrease feee in proportion to theincome of the area, contributesignificantly to the observeddifferences in physician-population ratios between urbanand rural areas. (Rimlinger andSteele, 1963)
According to_Long, it is notdemand that attracts physicians,but the "visible consequences ofexcess demand," such as the avail-ability of attractive opportunitiesfor medical practice. Indicators.include areas whose existingphysicians work relatively longhours, or where there has been arecent influx in population (1975).
190
Sloan looked at the percentageof male physicians who worked over49 weeks in 1959 as a substitutefor physicians' income andrationalized that new physicians
VARIABLES MAJOR FINDINGS
III.DECISION FACTORS (Cont'd)
C. Demand Related (Cont'd)
4. Perceived ExcessDemand (Cont'd)
5. Medicaid-Medicare-Research
will enjoy attractive startingincomes in States where there isexcess demand. He found that newphysicians have settled'in Stateswhere physicians incomes are high,or alternatively, where he propor-tion of physicians working over 49weeks a year is high (Slan,1968). Wiggins et al. surveyedgraduates of medical school from1915-55 and found perceived demandfor medical services as a generalfactor influencing the choice ofpractice location (1970).McFarland also identified excessdemand as a very importantvariable which is a potenti'al, butinfeasible variable for policymanipulation (1972).
Research grants to medicalschools constitute an importantFederal subsidy to doctors--thisfunding favors those areas withmore physicians (de Vise, 1972).Rushing found that Medicare andMedicaid nrovide incentives tolocate in more attractive commu-nities where there was previousunmet demand (1971). Yett statesthat these findings do not estab-lish that reimbursement opportuni-ties lead new physicians to enterhigh opportunity areas. Themajority of physicians in his studywere undoubtedly already in prac-tice before the beginning of reim-bursement programs, so that theassociations reported may indicateonly that total reimbursementincreases with the number ofphysicians eligible to receive it(Yett, 1973). Bobula and Goodmanpoint out that inability to payfor care is no longer a deterrent
191
2
VARIABLES MAJOR FINDINGS
III.DECISION FACTORS ( Cont'd)
C. Demand Related ( Cont'd)
5. Medicaid-MedicareResearch (Cont'd)
192
to obtaining a number of physi-cian's services. They believethat non-pecuniary factors aremore important in determiningpractice location (1979). JonGabel found that Medicareprevailing fees in counties withmore than 100 physicians per100,000 population are 21 percentgreater than prevailings incounties with fewer than 25physicians per 100,000 (1978).Hadley agrees that Medicareincentives do not support publicand private efforts to encouragephysicians to locate inunderserved areas (1978).
2 9t_.,
REFERENCES
Ball, D.S. and Wilson, J.W. C9mmunity health facilities and services:The manpower dimension. Am J Agric Econ 50:1208-1222, December, 1968.
Benham, L., Maurizi, A., and Reder, M.W. Migration, location andremuneration of medical personnel: Physicians and dentists. Rev.Econ Stat L:332-349, August, 1968
Bible, B.L. Physicians' views of medical practice in nonmetropolitancommunities. Public Health Rep 85:11-17, Jan 1970.
Bobula, J.D. and Goodman, J.J. The Supply and Distribution of Physicians:A Look at the Issues. American Medical Association, Chicago, 1979.
Breisch, W.F. Impact of medical school characteristics on location ofphysician practice. J Med Ed 45:1068-1070, Dec 1970.
Breiger, W.R. Attitudes of Nigerian medical students toward ruralpractice, J Med Ed 54:427-29, May 1979.
Bynum, G., Sanchez, R., and Odegaard, W. Medical education: A causalagent in physician's maldistribution. J Med Ed 47:722, Nov 1972.
Cantwell, J.R. and Eisenberg, B.S. The Spatial Distribution ofPhysicians. A Literature Review. Center for Health ServicesResearch and Development, American Medical Association, Chicago,June, 1975.
Castleton, K.B. Are we building too many medical schools? JAMA216:198-192, June 1971.
Champion, D.J. and Olsen, D. Physician behavior in Southern Appalachia:Some recruitment factors. J. Health Soc Behav 12:245-252, Sep 1971.
Charles, E. Policies to increase the supply of physicians in rural care.Am Econ XVI:37-44, Fall, 1972.
Coker, R.E., et al. Patterns of influence: Medical school facultymembers and the values and specialty interests of medical students.J Med Educ XXXV:518-27, June 1960.,
193 9
Coleman, S. Physician distribution and rural ,iccest. :o medical services.Executive summary. Prepared for DHEW, DiT of Medicineunder Contract No. 231-750-0613 by the R,,nd Corp° ation, SantaMonica, CA. No. R-1887/1, April, 1976.
CONSAD Research Corporation. An Evaluation of the Effectiveness of LoanForgiveness as an Incentive for Health Practitioners to Locate inMedically Underserved areas. A report prepared for the Departmentof Health, Education and Welfare, Office of the Secretary, ContractNo. HEW-OS-73-68, Washington, D.C., 1973.
Cooper, J.K., et al. Rural or urban practice: Factors influencing thelocation decision of primary care physicians. Inquiry 12:18-25, Mar,1975.
Cooper, J.K., Heald, K. and Samuels, M. The decision for rural practice.J Med Ed 47:939-944, Dec 1972.
Affecting the supply of ruril physicians. Am. J. Public Health57:756-759, August, 1977.
Cordes, S. Factors influencing the location of rural generalpractitioners. A study in Washington State. West J of Med128:75-80, Jan 1978. \\,
Cotterill, Ph.D. and Eisenberg,, B.X. Improving access to medical care inunderserved areas: The role of group practice. Inquiry 16:141-53,
\Summer 1979. ,
Crane, S.C. and Reynolds, J. Physician Location Selection andDistribution. A bibliography of relevant articles, reports, and datasources. Health Manpower Policy Studies Group, School of PublicHealth, University of Michigan, Ann Arbor, November, 1974.
Crawford, R.L. and McCormack, R. Reasons physicians leave primarypractice. J Med Ed 46:263-268, Apr 1971.
Cullison, S., Reid, C., and Colwill, J.M. Medical school admissions,specialty selection, and distribution of physicians. JAMA235:502-55, Feb 2, 1976.
Dewey, D. Where the Doctors Have Gone (Research paper). IllinoisRegional Medical Program, Chicago Regional Hospital Study, 1973.
Dewey, D. and Petto, A. Age-Selective Migration Patterns of Physiciansin Chicago Metropolitan Area, 1950-70. 10 pp., no date.
DeVise, P. Physician migration from inland to coastal states: Antipodalexamples of Illinois and California. J Med Educ 48:141-149,Feb, 1973.
194
Ducker, D. The myth of professional isolation among physicians innonurban areas. J Med Educ 52:991-998, Dec 1977.
Elesh, D. and Schollaert, P. Race and urban medicine: Factors affectingthe distribution of physicians in Chicago. J Health and Soc Behav13:236-250, Sep 1972.
Erdmann, J.B., et al. AAMC longitudinal study of medical schoolgraduates of 1960. National Center for Health Services Research.DHEW Pub. No. (PHS) 79-3235, Hyattsville, Md. Jan 1979.
Evashwick, C. The role of group practice in the distributionphysicians in nonmetropolitan areas. Med Care 14:808-823,
Fahs, I. and Peterson, 0. Towns without physicians and townsone: A study of four states in the Upper Midwest, 1965.Health 58:1200-1211, July 1968.
of
Oct 1976.
with onlyAm J Public
Fein, R. Factors influencing the location of North Carolina. GeneralPractitioners: A study in physician distribution Unpublished Ph.D.dissertation, Johns Hopkins University, Baltimore, 1956.
Fein, R. and Weber G. Financing Medical Education: An Analysis ofAlternative Policies and Mechanisms. McGraw-Hill, New York, 1971.
Feldstein, M.S. The rising price of physicians' services. Rev of EconStat 52:121-133, May 1970
Fuchs, V.R. and Kramer, M.J. Determinants of Expenditures for Physicians'Services in the U.S., 1948-68. DHEW Publication No. (HSM) 73-3013.Rockville, Md,-Dec 1972.
Gabel, J.R. Medicare-Medicaid Reimbursement and the Financing ofGraduate Medical Education. Paper prepared for GMENAC, Health CareFinancing Administration, June 12-13v1978.
Gabel, J.R. Medicare Reimbursement Policies and Graduate MedicalEducation. U.S. Tepartment of Health, Education, and Welfare.Health Care Financing Administration, Office of Research, Mar 21,1978 (Mimeo).
Goodman, L.J. and Mason, H.R., Physician Distribution and MedicalLicensure in the U.S., 1975, Part I, American Medical AssociationChicago, 1976.
Guzick, D.S. and Jahiel, R. Distribution of private practice offices ofphysicians with specified characteristics among urban neighborhoods.Med Care 14:469-88, June 1976.
195
236
Hadley, J. Models of Physicians' Specialty and Location Decisions.Technical Paper No. 6, DHEW, HRA, National Center for Health ServicesResearch, Rockville, Md., 1975.
Hadley, J. et al., Final Report Financing Medical Education: Issues andOptions, (Working Paper 5925-3) Urban Institute, Washington, D.C.,June 30, 1978.
Hadley, J., et al. Financing Medical Education: Issues and Options.Final report. (Working Paper 5925-3), June 30, 1978.
Hambleton, J.W. Determinants of Geographic Differences in the Supply ofPhysician Services. Doctoral dissertation, Univ. of Wisconsin,Madison, 1971
Harvey, E.B. The vanishing practitioner. J Med Ed 48:718-24, Aug 1973
Hassinger, E.W. Background and Community Orientation of Rural PhysiciansCompared with Metropolitan Physicians in Missouri. Univ. ofMissouri, College of Agriculture, Research Bulletin 822, Columnbia,Aug 1963.
Hassinger, E.W., et al., A Restudy of Physicians in Twenty Rural MissouriCounties. U.S. Dept. of Health, Education, and Welfare, HRA, Officeof Graduate Medical Education, DHEW Pub. No. (HRA) 79-634,Hyattsville, MD., 1979.
Haug, M.R., Lavin, B. and Breslau, N. Practice location preference atentry to medical school. J Med Educ (in press)
Heald, K.A., et al., Choice of Location of Practice of Medical SchoolGraduates: Analysis of Two Surveys. Rand Corp., Santa Monica, Cal.R-1477 Nov 1974.
Held, P.J. The mi ration of the 1955-65 :raduates of American MedicalSchools. Ford FoundationAdministration, Berkeley, Cal. 1973.
Program Research in University
Institute of Medicine. (See under Scheffler, R.M., et al.)
Institute of Medicine.. Physician Choice of Specialty and GeographicLocation: A Survey of the Literature, in Social Security Studies,Medicare-Medicaid reimbursement policies. pp. 289-325. NationalAcademy of Sciences, Washington, D.C., Mar 1976.
Joroff, S. and Navarro, V. Medical manpoiwer. A multivariate analysisof the distribution of physicians in urban United States. Med Care9:428-38, Sep-Oct 1971.
196
2
Kane, R., Dean, M. and Soloman, M. An evaluation of rural health careresearch. Eval Q 3:139-189, May 1979.
Kaplan, R. and Leinhardt, S. Determinants of physician office location.Med Care 11:406-415, Sep-Oct 1973.
Krawelski, J.E. and Luke, R.D. Group Practice: Review andRecommendations for Planning and Research. Blue Cross and BlueShield Associates, Chicago, Ill. 1979.
Kritzer, H. and Zimet, C.M. A retrospective view of medical specialtychoice. J Med Educ 52:47-53, Jan 1967.
Lave, J.R., Lave, L.B. and Leinhardt, S. Medical manpower models: Need,demand and supply. Inquiry 12:97-125, June 1975.
Lavin, B.F., Haug, M. and Breslau, N. Other Influences in Physicians'Locat'on Choices (in press).
Lavin, J.4. Do those doctor-short towns really want help? Med Econ11911,10, Sep 11, 1972.
Lieberson, S. Ethnic groups and the practice of medicine. Am SociolRe, /23:542-549, Oct 1958.
Lloyd, S., Jr., Johnson, D. and Mann, M. Survey of graduates of .atraditionally black college of medicine, J Med Educ S3:640-650, 1978.
Long, E. The Geographic Distribution of Physicians in the U.S.: AnEvaluation of Policy Related Research. Prepared for the NationalScience Foundation, NSF-C814. Interstudy, Minneapolis, Minn Jan 1975.
Longnecker, D.P. Practice objectives and goals: A survey of familypractice residents, Fam Prac II:347-51, Oct 1975.
Madison, D.L. Recruiting physicians for rural practice. Health SeryRep 88:758-762, Oct 1973.
Mantovani, R.E., Gordon, T.L., and Johnson, P.G., Medical StudentIndebtedness and Career Plans, 1974-1975, DHEW Pub No (HRA) 77-21,Washington, D.C.: GRP, Sep 1976.
Marden, P. A demographic and ecological analysis of the distribution ofphysicians in metropolitan America, 1960. AJS 72:290-300, Nov 1966.
Marshall, C.L. Principal components analysis of the distribution ofphysicians, dentists, and osteopaths in a Midwestern State. Am J PubHealth 61:1556-1564, Aug 1971.
197
233
Martin, E.D., et al. Where graduates go. The University of KansasSchool of MaTEIne: A study of the profile of 959 graduates andfactors which influenced their geographic distribution. J Kans MedSoc 69:84-89, Mar 1968.
Mason, H.R. Effectiveness of student aid programs tied to a service'commitment. J Med Educ 46:575-583, July 1971.
May, L.A. The Spatial Distribution of Physicians: The Special Case ofthe City. B.A. honors thesis, Harvard University, Cambridge,Mar 1970.
McFarinnd, J. Toward an Explanation of the Geographical Location ofPhysicians in the United States. Center for Health Services Researchcnd Development, American Medical Association, Chicago, Dec 1972.
McGrath, E. and Zimet, C. N. Female and male medical students:Differences in specialty choice selection and personality.Educ 52:293-300, Apr 1977.
J Med
McMillan, A.W., et al., Assessing the balance of physician manpower in ametropolitan area. Public Health Rep. 85:1001-1011, Nov 1970.
Monmonier, M.S. Socioeconomic controls on spatial variations in thephysician population. Proc Penn Acad Sci 46:91-93, 1972.
Mountin, J.W., Pennell, E.H., and Brockett, G.S. Location and movement ofphysicians, 1923 and 1938: Changes in urban and rural totals forestablished physicians. Public Health Rep. 60:173-185, Feb 16, 1945.
Owens, A. Working at full capacity, A lot of your colleagues aren't.Med Econ 63-71, Apr 2, 1979.
Paiva, R. and Haley, J. Intellectual personality and environmentalfactors in career specialty preference. J Med Educ 46:281-289, Apr1971.
Parker, R.C., Rix, LA.. and Tuxill, T.G. Social, economic and demographicfactors affecting physician population in upstate New York. NY StateJ Med 69:706-712, 1969.
Parker, R. C. and Sorensen, A. A. The tides of rural physicians. Theebb and flow, or why physicians move out of and into smallcommunities. Med Care 16:152-166, Feb. 1978.
Physician attitudes toward rural practice: Answers and questionsthat were not asked. Forum on Med. 411 -416, June, 1979.
Parker, R. C. and Tuxill, T.G. The attitudes of physicians toward small-community practice. J Med Educ 42:327-344, Apr 1967.
198
Perlstadt, H. Internship placements and f ulty influence. J Med Educ47:862-868, Nov 1972.
Petersdorf, R.G. Issues in primary care: The academic perspective.J Med Educ. Part 2, 50:5-13, Dec 1975.1
Peterson, G.R. A Comparison of Selected Profess]. al and SocialCharacteristics of Urban and Rural Physicians 1 Iowa. Health CareResearch Series, No. 8, University of Iowa, Iowa City., 1969.
Peterson, 0.L., et al., Appraisal of medical students bilities asrelated to training and careers after graduation,N ngl J MedCCLXIX: 1174-1182. Nov 23, 1963. J.
Presser, C. Factors affecting the geographic distribution ofphysicians. J Leg Med 12-18, Jan 1975.
Reskin, B. and Campbell, F. Physician distribution across metropolitanareas. Am J Sociol 79:981-999, Jan 1974.
Riley, G.J., et al. A study of family medicine in upstate New YorkJAMA. 208:2307-14, 1969.
Rimlinger, C.V. and Steele, H.B. An economic interpretation of thespatial distribution of physicians in the U.S. South Econ J 30:1-12,July 1963.
Robertson, L.S. On the intraurban ecology of primary care physicians.Soc Sci Med 14:227-238, Aug 1970.
Roemer, M.I. Health resources and services in the Watts area of LosAngeles, Calif Health 23:123-143, Feb-Mar 1966.
Roleston, R.I. Michigan Medical Manpower Study: An Investigation ofDecision Making Factors Among Medical Students, Interns and Residentsand an Evaluation of Proposed Loan Programs for Retaining MedicalManpower, Blue Shield of Michigan, 1973.
Royce, P.C. Can rural health 'education centers influence physiciandistribution? JAMA 220:847-49, May 8, 1972.
Rushing, W.A. Community, physicians and inequality. Lexington Books,Lexington, Mass., 1975.
Schaupp, F.W. A Study of the Factors Influencing Outmigration of MDGraduates of West Virginia University. Graduate Research Monographin Business and EcOnomics, Vol. 1, No. 3. Morgantown, Bureau ofBusiness Research, West Virginia University, Morgantown, 1969.
199
Scheffler, R.M., et al. Physicians and New Health Practitioners: Issuesfor the 1980's. Final Report of contract, DHEW No. 282-78-0163-EJM.DREW Pub. No. (HRA) 80-621. National Academy of Sciences, Instituteof Medicine, Washington, D.C. May 1979.
Scheffler, R.M. An Empirical Investigation into the GeographicDistribution of Physicians and Specialists. "Doctoral dissertation,New York University, New York, 1971.
Shannon, G.W. and Dever, G.E.A. Health Care Delivery: SpatialPerspectives. McGraw-Hill Book Co., New York, 1974.
Sloan, F.A. Economic Models of Physician Supply. Doctoral dissertation,Harvard University, Cambridge Aug 1968.
Lifetime earnings and physicians' choice of specialty. Ind LaborRels Rev 24:47-56, Oct 1970.
Sloan, F.A. and Yett, P. Income and Other Factors Affecting PhysicianLocation Decisions: an Econometric Analysis. Final Report onsubcontract with New York University, contract HEW 105 69-94.Nov 10, 1970.
Smith, S. An analysis of the locationyear 1950. Doctoral dissertation, State University of Iowa, Ames,1961.
of physicians in the U.S. for the
Stefanu, C. et al. Hospitals and medical schools as factors in theselection of location of practice. J Med Educ 54:379-383, May 1979.
Steinwald, B. and Steinwald, C. The effect of preceptorship and ruraltraining programs on physicians' practice location decisions. MedCare 13:219-229, Mar 1975.
Taylor, M., Dickman, W., and Kane, R. Medical students' attitudes towardrural practice. J Med Educ 48:885-895, Oct 1973.
U.S. Department of Health, Edti ation, and Welfare, Health ResourcesAdministration, Bureau of ealth Manpower. Factors InfluencingPractice Location of Profs sional Health Manpower: A Review of theLiterature. DREW Pub. No. (HRA) 75-3, July, 1974. Available NTIS,Springfield, Va., No. PB-236 950/AS.
U.S. General Accounting Office. Report to the Congress: Progress andProblems in Improving the Availability, of Primary Care Providers inUnderserved Areas, Washington,'D.C. Aug 22, 1978.
Weber, G. An essay on the Distribution of Physicians AmongstSpecialties. Prepared for the U.S. Department of Health, Education,and Welfare, Washington, D.C., 1972.
200
2
'..(r,iskotten, G.G., et al., Trends in medical practice: An analysis of thedistribution and characteristics of medical college) graduates,1915-1955 J Med Educ 35:1071-1121, Dec 1960,
Wendling, W., Werner, J., and Budde, N. R onal Variations in theDeterminants of Physician's Location C ice. Presented at SouthernEconomic Association Meetings, Nov 9 978.
Werner, J., Wendling, W., and Budde, N. hysician Location and SpecialtyChoice: A Simultaneous Logit Approac Paper presented at the 42ndmeeting of the Midwest Economic Association, 1978.
Williams, R.C. and Uzzell, W.E. Attracting physicians to smallercommunities. Hospitals 34:49-51, July 1960.
Wilson, S.R. An Analytical Study of Physicians' Career DecisionsRegarding Geographic Location. Digest of the Final Report. AmericanInstitutes for Research, DREW Contract No. HRA 231-77-0088.Department of Health, Education and Welfare, Health ResourcesAdministration, Hyattsville, Md. 1979.
Wooldridge, J., Differences in Pricing Policy, Modes of Practice andLocation Between White Male and Minority Physicians. 'Paper presentedat the Annual Meeting of the Western Economic Association, SanFrancisco, Cal June 24-27, 1976.
Yett, D.E. and Sloan, F.A. Migration patterns of recent medical schoolgraduates. Inquiry 11:125-142, June, 1974.
Yett, D.E. The Specialty and Location Choices. The Human ResourcesResearch Center, University of Southern. California, Los Angeles, Cal1973.
201
240
Additional Recommended References
Anderson, J.G. and Bartkus, D.E. Physician location and distribution: Asocial systems approach. Socio -Econ Plan Sci 10:213-31, 1976.
Anderson, J.G. Construc4ng causal models: Problems of uniitls ofanalysis, aggregation,-and specifications. Health Sery RiaJ 13:50-600,Spring 1978.
Anderson, J.G. and Marshall, H.H. The structural approach to physiciandistribution: A critical evaluation. Health Serv. Res 9:195-207, Fall1974.
Aring, C.D. The distribution of physicians. JAMA 219:606-07, Jan 31,1972.
Bass, M. and Copeman, W.J. An Ontario solution to medically under-serviced areas: Evaluation of an ongoing program. Can Med Assn J113:403-06, Sep 6, '1975.
Bridgers, W. Rural physician distribution in Alabama. An historicalperspective--1923 and 1976. Policy considerations for the future. Ala.J. Med. Sci.
Califano, J.A., Jr. Remarks of the Secretary, USDHEW, before theNational Academy of Sciences. Washington, D.C. Oct. 26, 1978.
Eisenberg, B.S. and Cantwell, J.R. Policies to influence the spatialdistribution of physicians. Med Care 14:455-68, June 1976.
Egelston, E.M. and Camp, L. Physician distribution: Solving thespecialty-location dilemmas. Hospitals 50:101-06, Dec le, 1976.
Fenderson, D.A. Health Manpower development and rurah services. JAMA225:1627-31, Sep 24, 1973.
1kFox, J.G. and Richards, J. Physicians dominance and location of foreignand U.S. trained physicians. J Health Soc Behav 18:366-75, Dec 1977.
Garrard, J. and Verby, J.E. Comparisons of medical student experiencesin rural and university settings. J Med Ed 52: 02-10, Oct 1977.
Ginzburg, E. Paradoxes and trends: An economist loo s at health care.N Eng J Med 275:85-87, July 14, 1966.
202
2
\Ginzburg0 E. Physician shortage reconsidered. N Eng J Med 275:85-7,July .14, 1966.
Goldberg, J.H. Which states need which specialists: Med Econ pp.129-34, Dec. 13, 1976.
Hassinger, E.W. and Hobbs, D.J. Distribution of health services inMissouri, Research Bulletin 917, Univ. of Missouri, College ofAgriculture, Columbia March 1967.
Hassinger, E.W., Changes in number and location of health practitionersin a 20-county rural area of Missouri. Rural Health 90:313-18,July-Aug. 1975.
Heald, K.A. and Cooper, J.K. An Annotated Bibliography on Rural MedicalCare. The Rand Corporation, Santa Monica, CA. 1972
Heil, R.A. Geographic maldistribution: A case study and analysis forpodiatric medicTne. J of Podiatr Med Educ 6:6-10, Summer 1979.
Holden, W. and Levit, E. Migration of physicians from one specialty toanother. JAMA 239:208-09, Jan 16, 1978..
Inglehart, J.K. Health report/maldistribution of physicians is focalpoint of manpower debate. Natl J Rep 6:837-460 June 8, 1974.
Kane, W.J. Rural health care, medical issues. JAW, 240:2647-90,Dec. 8, 1978
Knox, P.L. The accessibility of primary care to urban patients: Ageographical analysis. J Royal Coll Gen Prac 29:160-168, Mar. 1979
Lankford, P.M. Physician location factors and public policy. Eco Geog50:244-255, July 1975.
Lynch, M. The physician shortage: The economics mirror. Am Acad PolSoc Sci 399:82-88, Jan 1972.
Mattson, D.,
physicians.
Mawardi, B.H.1965.
et al., Evaluation of a program designed to produce ruralJ Med Educ 48:323-331, Apr 1973.
Career study of physicians. J Med Educ 40:658-666, July
a ell, R.I Health care: The growing dilemma. McKinsey and Company,c., New 'ork, June 1975.
t.icMillan0 A. et al., Assessing the balance of physician manpower inmetropolitan area. Pub Health Rep 85:1001-1011, Nov. 1970.
203.
214
a
![Welcome [corporateapps.gsa.gov] · Web viewWelcome to Financial Data Reporting System Business Objects (FDRS BO) Training. FDRS Business Objects (Business Objects for short) is a](https://img.dokumen.tips/doc/110x75/6132dab2dfd10f4dd73ab707/welcome-web-view-welcome-to-financial-data-reporting-system-business-objects.jpg)
![[014] ass 014 [1881]](https://img.dokumen.tips/doc/110x75/5695d38d1a28ab9b029e5607/014-ass-014-1881.jpg)
![[ Nov 70] lap. FDRS Price M7-50.25 1.3C-$1.0 DESCRIPTORS](https://img.dokumen.tips/doc/110x75/6277f6a3c76a9c1ad0408798/-nov-70-lap-fdrs-price-m7-5025-13c-10-descriptors-.jpg)
