uo

ni

A

Patient Education and Counseling 79 (2010) 338343

Contents lists available at ScienceDirect

Patient Education

journa l homepage: www.e lsecDepartment of Biostatistics, Virginia Commonwealth University, Richmond, VA, USAdDepartment of Epidemiology and Community Health and Massey Cancer Center, Virginia Commonwealth University, Richmond, VA, USAeDepartments of Family Medicine and Epidemiology and Community Health, Virginia Commonwealth University, Richmond, VA, USA

1. Introduction

More Americans than ever before are overweight or obese,creating a daunting public health issue that cuts across age, gender,and ethnic domains in our society. Currently, 33% of adults areoverweight and 34% are obese; data that represent the signicantincreases in overweight prevalence rates from 1988 to 2004 [1,2].As such, this national epidemic requires immediate action andnovel solutions to slow the current trajectory. Reducing obesitywill require new perspectives and signicant changes in the way

the problem is addressed by physicians, researchers, parents,public health institutions, and community groups. Clinicians andpractices, as systems of care, face a growing need to offer patientshigh-quality resources on health behavior change and healthpromotion strategies; yet, medical practices may lack the time andresources to thoroughly assess the health behaviors of patients,their readiness to change, or the type of assistance they need [3,4].

Having been encouraged by clinicians to improve health habits,patients often seek additional resources regarding techniques,motivational strategies, and community programs. However,studies show that only 65% of obese patients receive the advicethey seek on weight loss [5]. Physician organizations and practicegroups are proactively developing and testing new approaches foraddressing overweight and obesity in patients who need morestructure, motivation, and resources [6].

The use of the electronic medical record (EMR) providesclinicians with technology to easily identify patients with

A R T I C L E I N F O

Article history:

Received 7 August 2009

Received in revised form 14 January 2010

Accepted 30 January 2010

Keywords:

Health behaviors

Weight loss

Clinician-delivered intervention

Electronic medical record

Community collaboration

The 5 As

A B S T R A C T

Objective: Obesity has become a public health epidemic in adults and children. Clinician practices need

new models to effectively address overweight in patients, yet, practices lack time and resources. We

tested a clinician-delivered intervention that utilized community resources for in-depth counseling for

unhealthy behaviors including overweight.

Methods: Eligible patients in nine primary care practices were identied using an electronic linkage

system (eLinkS) which also automated patient referrals to group (Weight Watchers), telephone

counseling (TC), or usual care. Pre/post-survey data were used to assess factors related to counseling

choices as well as changes in BMI (kg/m2) and weight-related behaviors using descriptive statistics,

unadjusted, and adjusted statistical analyses.

Results: Study sample (n = 146) was 70% female with a mean age of 57 years. More patients (57%)

selected WW, followed by usual care (27%) or TC (16%). Age, gender, clinician recommendation, and

counseling program characteristics were inuential in counseling selections. Weight Watchers

participants and those in TC, reported statistically signicant weight loss, WW participants also

reported signicant increases in fruit/vegetable intake; after 4 months compared with usual care.

Conclusions: This practice-based intervention utilizing community counseling referrals was associated

with positive health behavior change.

Practice Implications: Identifying inuential factors related to patient weight counseling choices may

help guide referrals to community programs.

2010 Elsevier Ireland Ltd. All rights reserved.

* Corresponding author at: Virginia Commonwealth University, Division of

Quality Health Care, Department of Internal Medicine, 1200 E. Broad St., PO Box

980307, Richmond, VA 23298-0306, USA. Tel.: +1 804 828 9891/241 8019;

fax: +1 804 828 4862.

E-mail address: dbwilson@vcu.edu (D.B. Wilson).

0738-3991/$ see front matter 2010 Elsevier Ireland Ltd. All rights reserved.doi:10.1016/j.pec.2010.01.025Patient weight counseling choices and ocare and community collaborative interv

Diane B. Wilson a,b,*, Robert E. Johnson b,c, Resa M. JSteven H. Woolf e, Sharon K. Flores b

aDepartment of Internal Medicine and Massey Cancer Center, Virginia Commonwealth UbDepartment of Family Medicine, Virginia Commonwealth University, Richmond, VA, UStcomes following a primaryention

nes d, Alex H. Krist b,

versity, Richmond, VA, USA

and Counseling

vier .com/ locate /pateducou

unhealthy behaviors, a counseling platform such as the 5 As(Assess, Advise, Agree, Assist, Arrange) [10], and the ability toautomate referrals to outside health behavior counseling services[7,8]. EMRs can be programmed to identify patients who smoke orare overweight and prompt clinicians to initiate counseling aboutimproving health behaviors. Such features can be valuable foridentifying groups of patients who would benet from moreintensive counseling through referral to community resources [9].More research is needed to assess and evaluate the features of low-intensity interventions that most practices could potentiallydeliver.

2. Methods

2.1. eLinkS intervention

This study was part of the Prescription for Health initiative,sponsored by the Robert Wood Johnson Foundation, to study newapproaches of addressing unhealthy lifestyle behaviors in primarycare practices [12]. Details of the study intervention are reportedelsewhere [11]. Briey, this study tested a clinician-delivered(clinicians included physicians, physician assistants, and nursepractitioners) intervention (eLinkS) that utilized an EMR-basedprompt and referral system to identify adult patients who wereoverweight, obese, or current smokers; prompt the clinician tocounsel the patient using the 5 As and, when a patient wasinterested, automate patient referrals to community resources for

individual weight loss counseling or smoking cessation, free, for 9months. This report focuses on patients referred for weight loss.Patients could select from: group classes offered through acommercial weight loss program (Weight WatchersTM [WW]);individual telephone weight loss counseling, provided by trainedcounselors at the University of Kentucky Wellness Center(Behavioral Health Improvement Program, BeHIP [13]); compu-ter-based counseling, provided by accessing an informative websitedeveloped by ACORN (Ambulatory Care Outcomes ResearchNetwork) for a previous project; or through an e-counselingservice that ACORN designed with BeHIP for this project [4,10].Usual care was also an option, which consisted of any alternativethe patient and clinician decided to pursue (e.g., counseling by theclinician or a decision not to address overweight).

The intervention was launched at the primary care practices, ina setting considered to be small town and semi-rural, on April 16,2006. Study recruitment was discontinued 5 weeks later (May 22,2006) when the number of patients referred for group weight losscounseling (WW) exceeded budgeted allocations. However,patients who were referred to either group or telephonecounseling within the 5 weeks were eligible to participate inweight loss counseling at no cost for up to 9 months.

2.2. Study sample

The sample for this analysis consisted of adult patientsparticipating in eLinkS, who were referred to and participated

pon

ion (

D.B. Wilson et al. / Patient Education and Counseling 79 (2010) 338343 339Fig. 1. Determination of sample for weight counseling study cohort. 60% of those res37% of the overweight cohort. Patients who selected the electronic counseling optding to the baseline survey (274) responded to the 4-month survey (164); this was

e-mail) were not included in the analysis due to low participation.

in, intensiveweight loss counseling or usual care (n = 440) andwhoreturned both baseline (2 weeks) and 4-month follow-up surveysthat had complete data, and reported a weight change that was notconsidered an outlier. Based on the distribution of our data, 18.2 kgwas the maximum weight change, either gained or lost, that wasnot considered an outlier.

Fig. 1 shows the total weight loss cohort and number ofparticipants that comprise the sample for analysis. Sixty-twopercent of participants (n = 274) returned baseline (2 weeks)surveys, and of these, 60% (n = 164) also returned 4-monthsurveys; nine were excluded for missing data and six for outlierdata, resulting in a nal analysis sample of n = 146. Due to the smallnumber of counselees selecting electronic computer counseling,these individuals were not included in the analysis.

2.3. Data collection and study variables

2.3.1. Survey instrument

All data for this study was collected using a self-report survey,sent to all patients referred forweight loss counseling or usual care.The overall health behavior survey instrument was a reliable andvalid instrument, developed by Glasgow et al. [14], which served asa set of common measures, and was required to be used by allPrescription for Health study sites funded by the Robert WoodJohnson Foundation. Items related to experiences in the ofceweredeveloped internally and were tested for readability. The survey

variables of interest included measures of factors that may havebeen inuential in patient counseling choices. These includedsurvey items measuring patients experiences during the ofcevisit including helpfulness of the clinician and receipt ofsufcient information; using a Likert scale [15]. A single surveyitem was used to specically assess factors affecting counselingchoice (e.g., convenience of the program, condentiality, type ofsupport, and doctors recommendation).

2.4. Statistical analysis

Statistical analyses were performed using SAS version 9.1.3.Descriptive statistics were used to determine frequencies andproportions for categorical variables (fruit/vegetable intake) andmeans for weight, BMI, and exercise at 2 weeks and 4 months;health behaviors were stratied by counseling type. Paired t-testswere performed to determine pre/post-differences in meanweight, BMI, and exercise MET-hours/week, within each counsel-ing type. MET-hours/week pre/post-differences were adjusted formissing or incomplete walking, moderate, or vigorous exercisedata in one of the two data collection points (2 weeks or 4months)by assuming no change; effectively imputing the response fromthe other (pre or post) data collection point. TheMcNemar test wasused to assess reported pre/post-change in categorical measures(fruit/vegetable intake) for each counseling type. Fishers exact testwas used to evaluate differences in clinician communication and

ne

81

tio

D.B. Wilson et al. / Patient Education and Counseling 79 (2010) 338343340instrument was then pilot-tested for face validity with patientswho did not participate in the study. We used a modied Dillmantechnique to enhance survey response rateswith a series of follow-up mailed reminders to study participants [15].

2.3.2. Key measures

The primary study outcome variables were weight (kg), BMI(kg/m2), fruit/vegetable intake (01, 24, 5 servings/day), andexercise. Exercise survey items included questions on days/weekand minutes/day of moderate, vigorous, walking, and totalexercise, which were measured in days/week, minutes per session,and converted to MET-hours/week (metabolic equivalents) foranalysis, in accordance with published standards [16]. Other

Table 1Demographic characteristics by counseling group and total sample.

Demographic characteristics Group counseling (N=83) Telepho

Gendera n (%)

Female 72 (87) 15 (63)

Male 11(13) 9 (38)

Ageb

Mean years (range) 53 (2382) 64 (40

Race

White 58 (70) 15 (63)

Black 24 (29) 8 (33)

Other 1 (1) 1 (4)

Income (annual)

Unknown 8 (10) 3 (13)

$75,000 10 (12) 3 (13)

Education

Unknown 2 (2) 2 (8)

Table 2Patient assessment of factors inuencing weight counseling choice by counseling group and total sample.

Communication with physician: Group counseling N(%) Telephone

counseling N(%)

Usual care N(%) Total sample N(%)

Was enough information provided?a 55 (81) 16 (73) 13 (50) 83 (72)

Was the doctor helpful in making the decision?b 69 (91) 17 (74) 18 (60) 103 (80)

Primary consideration in selecting counseling option

Convenience 32 (51) 12 (57) 10 (67) 53 (54)

Privacy/condentiality 12 (19) 7 (33) 3 (20) 21 (21)

Type support offered by programc 39 (62) 6 (29) 6 (40) 50 (51)

Doctors recommendation 31 (49) 11 (52) 6 (40) 47 (48)

Note: Fishers exact test was used to test for differences between three counseling options N refers to total number of yes responses. Percent (%) is number of yes responses

over total responses, omitting missing items.a p=0.014.b p

D.B. Wilson et al. / Patient Education and Counseling 79 (2010) 3383433424. Discussion and conclusions

4.1. Discussion

Research documents the need for testing new clinician-basedinterventions to help patients improve health behaviors [1721].In this study, patients referred to weight loss counseling over-whelmingly preferred group classes (WW) to telephone counselingand usual care. This preference was so substantial that interven-tion intake reached maximum capacity after only 5 weeks. Somefeatures of the counseling programs and characteristics of oursample may have contributed to this pattern. First, we found thatyounger females selected group classes over the other options, andwomen comprised 75% of our sample (Table 1). Also, type ofsupport offered was reported as the most signicant factorinuencing the counseling decision by patients selecting WW.Patients seeking in-person group support might have been morelikely to select WW. It is also possible that patients seekingprofessional weight loss counseling were familiar with thebranding of WW, given its national reputation. Older womenwere more interested in BeHIP telephone counseling. This groupwas 63% female with mean age 64 years; Blacks comprisedapproximately 33% of BeHIP participants. This program was basedat the University of Kentucky and may have been less familiar toVirginia patients than WW or usual care. Patients selecting BeHIPwere most likely to rate convenience and doctors recommen-dation as reasons for their choice. Usual care patients were morelikely to be male (62%).

In assessing pre/post-differences in health behaviors, bodyweight and BMI appeared to decrease in all three groups, however,only patients selecting group or telephone counseling exhibitedchanges that differed signicantly from baseline. Group counselingparticipants lost an average of (3.5 kg) and 1.3 kg/m2 (BMI) andalso reported signicantly increased daily fruit/vegetable intakefrom pre- to post-intervention. These results support other reportsof weight loss by patients enrolled in commercial weight lossprograms [2226].

Participants counseled by telephone also reported a signicantweight loss (2.0 kg); such results may hold promise. Telephonecounseling has shown similar success with individuals in

Table 4Summary of weight change (kg) by loss/gain status within counseling groups.

Group

counseling

Telephone Usual care

Patients who lost weight

Total N=97 63 15 19

Average weight change (kg) 5.3 4.1 3.8

Percent with 5% weight change 44% 27% 26%

Patients who gained weight

Total N=49 20 9 20

Average weight change (kg) 2.1 1.6 2.3

Percent with 5% weight change 5% 0% 20%improving health behaviors when compared to rates of successusing face-to-face counseling [27]. Advantages of telephonecounseling cited in studies include privacy and the fact that itcan be delivered by trained professionals, other than clinicians,including certied health coaches and registered dietitians [18].Further, in a 2007 systematic review by Eakin et al. [28] evidencefor telephone delivery of dietary and physical activity was strong,with 20 of 26 studies reporting positive changes in healthbehaviors. Privacy may have been an important feature for ourpatients who selected this modality. More research is needed tofurther evaluate other, newer delivery modalities such ascomputer-based programs to test their efcacy against moreconventional modalities for weight loss counseling.The pattern of overall pre/post-changes in the usual care groupwas of interest. This group was selected mainly by males (62%),who reported the lowest mean baseline BMI (31 kg/m2) and bodyweight (93.1 kg), compared to group counseling and telephonecounseling groups, and showed no signicant change in weight,exercise levels, or fruit/vegetable intake over the 4 months asshown in Table 3. This seems to support studies showing that,historically, women are more likely to seek weight loss counselingthan males. In addition, it may be that, given limited time andresources generally provided during a clinic visit, the clinicianmaynot have placed emphasis on delivering A4 (assist) and A5(arrange). Reasons could also be that not as many participants whochose this group wanted to prioritize weight loss during the studyperiod compared to those in the intensive counseling groups. Forall of these reasons, it is not surprising that patients in the usualcare group, when considering weight gainers and losers, did notreport a statistically signicant change in weight loss, or otherhealth behaviors, compared to those selecting intensive weightloss counseling groups. The level of success among participantsreferred to intensive weight counseling provides evidence that thenovel primary care and community program collaboration, alongwith its unique features, promoted successful weight loss andrelated health behaviors. For an example of just one feature,electronic prompts were programmed which reminded cliniciansthat a patient was participating in a specic community programso that they could reinforce patient progress or address barriers,when patients followed-up. Support and reinforcement can serveas important motivators for weight loss.

While this primary care and community collaborative inter-vention makes an important contribution as a novel model forclinicians to address overweight/obesity and to refer patients tocounseling, the study has some limitations. These include using anon-random/prepost-design and having at least two potentialsources of selection bias: services were free to patients, and moremotivated patients may have selected group counseling (WW)rather than usual care. Survey data were self-report and carry withthem inherent limitations related to response rates and validity.The response rate for each of our two surveys was respectable, butthe overall impact of the survey response rates and our studycriteria exclusions resulted in 37% of our original weight losscounseling cohort being represented in the analysis. However, wecompared age and gender demographics of our survey responderswith those of non-responders. We found the two groups to besimilar by gender (70% female) but survey responders tended to beslightly older than non-responders (57 years vs. 53 years). Also, asa validity check, we calculated a correlation coefcient of 0.69 incomparing weights and BMI reported by 76 patients on baselinesurveys and the earliest weights recorded by WW.

4.2. Conclusions

Patients choices for communityweight loss counseling serviceswere inuenced by age, gender, and messages delivered by theprimary care clinician in offering the options. Group counseling(WW) was the preferred option, patients who chose group classesand telephone counseling reported signicant weight reductionsover 4 months. More study is needed to assess the long-termimpact and sustainability of this novel primary care andcommunity collaborative intervention and other delivery mod-alities designed to assist individuals with weight loss and otherhealth behaviors.

4.3. Practice implications

Successful features of delivery models tested in this studyinclude the use of EMR prompts, consistent clinician/patient

communication related to weight loss, partnerships with com-munity programs, and identication of factors reported as beingimportant in patient choice of weight loss programs. Under-standing inuences on patient selection of weight loss programs,may have important implications for increasing referral andretention of patients in future weight loss counseling programs.

Acknowledgements

The authors thank the practices in RiversideMedical Group thatparticipated in this study: Bruton Avenue Family Medicine, EagleHarbor Primary Care, Elizabeth Lakes Family Practice, HiltonFamily Practice, Mathews Medical Center, Mercury West MedicalCenter, Patriot Primary Care, Riverside Family Medicine, andWilliamsburgMedical Arts Family Practice. We also thank the staffof Weight WatchersTM International, Riverside Wellness Center,

[7] Tang PC, Lansky D. The missing link: bridging the patientprovider healthinformation gap: electronic personal health records could transform thepatientprovider relationship in the twenty-rst century. Health Aff (Mill-wood) 2005;24:12905.

[8] Nemeth LS, Feifer C, Stuart GW, Ornstein SM. Implementing change in primarycare practices using electronic medical records: a conceptual framework.Implement Sci 2008;3(January):3.

[9] Krist AH, Woolf SH, Rothemich SF, Johnson RE, Wilson DB. It takes a partner-ship: the value of collaboration in developing and promoting a website forprimary care patients. Ann Fam Med (Supp) 2005;3:S4750.

[10] Whitlock EP, Orleans CT, Pender N, Altan J. Evaluating primary care behavioralcounseling interventions: an evidence-based approach. Am J Prev Med2002;22:26784.

[11] Krist AH, Woolf S, Frazier C, Johnson R, Wilson DB, Rothemich S, et al. Linkingclinicians and community counselors for health behavior change: the impacton the delivery of the 5 As in practice. Am J PrevMed 2008;35(Suppl.):S3508.

[12] Green LA, Cifuentes M, Glasgow RE, Stange KC. Redesigning primary carepractice to incorporate health behavior change. Am J PrevMed 2008;35:S3479.

[13] University of KentuckyHealth andWellness ProgramBeHIP: Overview: http://www. uky.edu/HR/wellness/behipoverview.html.

D.B. Wilson et al. / Patient Education and Counseling 79 (2010) 338343 343and the University of Kentucky Health &Wellness Behavior Healthimprovement Program (BeHIP) for their support of this study. Thestudy also received extensive support from the staff of thePrescription for Health National Program Ofce, the Prescriptionfor Health evaluation team at the University of Medicine andDentistry of New Jersey, the National Advisory Committee forPrescription for Health, and the project ofcers of the RobertWoodJohnson Foundation. Finally, we wish to thank Tina Cunningham,PhD candidate from the VCU Department of Biostatistics, for hervaluable contributions to the statistical analysis of the study data.

This work was funded by the Robert Wood Johnson Foundation(Grant # 053769) and the Agency for Healthcare Research andQuality under the Prescription for Health funding initiative.

Weight Watchers International Incorporated provided studysubjects with vouchers to enroll in local classes at no charge,committed in a written contract to review but not to control orinuence the content of publications.

References

[1] Khan LK, Sobush K, Keener D, Goodman K, Lowry A, Kakietek J, et al. Centers fordisease control. Recommended community strategies and measurements toprevent obesity in the United States. MMWR Recomm Rep 2009;58(July):126.

[2] Ogden C, Carroll MD, Curtin LR, McDowell JA, Tbak CJ, et al. J Am Med Assoc2006;295:154955.

[3] Capriano RM, Flocke SA, Frank SH, Strange KD. Tools, teamwork, and tenacity:an examination of family practice ofce system inuences on preventiveservices delivery. Prev Med 2003;36:13140.

[4] Woolf SH, Krist AH, Johnson RE, Wilson DB, Rothemich SF, Norman GJ, et al. Apractice-sponsored web site to help patients pursue health behaviors: anACORN study. Ann Fam Med 2006;4:14852.

[5] Thande NK, Hurstak EE, Sciacca RD, Giardina EG. Management of obesity: achallenge for medical training and practice. Obesity 2009;17:10713.

[6] Leverence R, Williams R, Sussman A, Crabtree B, RIOS Net Clinicians. Obesitycounseling and guidelines in primary care. Am J Prev Med 2006;32:3349.[14] Glasgow RE, Ory MG, Klesges LM, Cifuentes M, Fernald DH, Green LA. Practicaland relevant self-report measures of patient health behaviors for primary careresearch. Ann Fam Med 2005;3:7381.

[15] Dillman D. Mail and Internet surveys: the total design method, 2nd ed.,Hoboken, NJ: Wiley and Sons, Inc.; 1999.

[16] Guidelines for data processing and analysis of the International PhysicalActivity Questionnaire (IPAQ). The International Physical Activity Question-naire Research Committee; 2005.

[17] Lyznicki J, Young D, Rigg J, Davis R, Council on Scientic Affairs, AmericanMedical Association. Obesity and management in primary care. Am Fam Phys2001;63:213945.

[18] GoldsteinM,Whitlock E, DePue J. Multiple behavioral risk factor interventionsin primary care. Am J Prev Med 2004;27:619.

[19] Sussman A,Williams R, Leverence R, Gloyd P, Crabtree B. The art and complex-ity of primary care clinicians preventive counseling decisions: obesity as acase study. Ann Fam Med 2006;4:32733.

[20] McAlpine D, Wilson A. Trends in obesity-related counseling in primary care:19952004. Med Care 2007;45:3229.

[21] Alexander S, Osthye T, Pollak K, Gradison M, Bastion L, Brower R. Physicianbeliefs about discussing obesity: results from focus groups. Am J Health Prom2007;21:498500.

[22] Dansinger ML, Gleason JA, Grifth JL, Selker HP, Schaefer EJ. Comparison of theAtkins, Ornish, Weight Watchers, and Zone diets for weight loss and heartdisease risk reduction: a randomized trial. J Am Med Assoc 2005;293:4353.

[23] Heshka S, Anderson J, Atkinson R, Greenway FL, Hill JO, Phinney SD, et al.Weight loss with self-help compared with a structured commercial program. JAm Med Assoc 2003;289:17928.

[24] Tsai AG, Wadden TA. Systematic review: an evaluation of major commercialweight loss programs in the United States. Ann Intern Med 2005;142(Janu-ary):5666.

[25] TrubyH, Baic S, deLooy A, Fox KR, LivingstoneMB, Logan CM, et al. Randomisedcontrolled trial of four commercial weight loss programmes in the UK: initialndings from the BBC diet trials. Br Med J 2006;332:130914.

[26] Rock C, Pakis B, Flatt S, Quintana E. Randomized trial of a multifacetedcommercial weight loss program. Obesity (Silver Spring) 2007;15:93949.

[27] Digenio AG, Manusco JP, Gerber RA, Dvorak RV. Comparison of methods fordelivering a lifestyle modication program for obese patients: a randomizedtrial. Ann Intern Med 2009;150:25562.

[28] Eakin E, Lawler S, Vandelanotte C, Owen N. Telephone interventions forphysical activity and dietary behavior change. Am J Prev Med 2007;32:41934.

Patient weight counseling choices and outcomes following a primary care and community collaborative interventionIntroductionMethodseLinkS interventionStudy sampleData collection and study variablesSurvey instrumentKey measures

Statistical analysis

ResultsDemographic characteristics and selection of weight counseling programsFactors considered in choice of counseling optionChange in BMI, body weight, and health behaviors

Discussion and conclusionsDiscussionConclusionsPractice implications

AcknowledgementsReferences