ilable at ScienceDirect

Social Science & Medicine 67 (2008) 1502–1511

Contents lists ava

Social Science & Medicine

journal homepage: www.elsevier .com/locate/socscimed

Do physicians care about patient choice?q

Ulrika Winblad*

Department of Public Health and Caring Sciences, Health Services Research, Uppsala Science Park, 75185 Uppsala, Sweden

a r t i c l e i n f o

Article history:Available online 9 September 2008

Keywords:Patient choiceImplementationMedical professional rolePatient rightsPatient empowermentSwedenReferralPhysician role

q The paper has benefited greatly from commentical Science Association (NOPSA) conference in RAugust 2005. Also, a great thank you to all the physthe survey and made the study possible.

* Tel.: þ46 18 611 35 88.E-mail address: ulrika.winblad@pubcare.uu.se

0277-9536/$ – see front matter � 2008 Elsevier Ltddoi:10.1016/j.socscimed.2008.07.016

a b s t r a c t

A new policy (patient choice) was introduced in Sweden in the early 1990s to give patientsthe right to choose their healthcare providers, however, evaluations show that fewpatients exercise this right. This paper analyses physicians’ roles in putting the patientchoice policy into effect. To examine attitudes, knowledge and behaviour among physi-cians, a questionnaire was sent to 960 physicians in one of the most populous counties inSweden. The results show that the physicians approve of the policy, yet only a minoritystate that they regularly help patients to choose healthcare providers by giving theminformation and letting them choose where they will be referred. Instead, referrals aremostly based on medical grounds; the patient’s wish to choose a specific provider isconsidered less important. In summary, we found that more than a decade after the policywas introduced, only a minority of physicians act according to the political intention. Thiscould be one explanation for why many patients still do not exercise their right to choosea hospital.

� 2008 Elsevier Ltd. All rights reserved.

Introduction

Patient choice was introduced in Sweden in the early1990s with the intention of giving patients the right tochoose healthcare providers freely across the country. Thiswas a new phenomenon in the Swedish context; patientshad previously been restricted to using the nearest medicalfacility. The most important political motive behind thepatient choice policy was to increase patient influence in thehealthcare sector. The Swedish healthcare system wascriticised during the 1980s for the discrepancy betweenservices supplied and those sought by patients. von Otterand Saltman (1990) describe how the rigidities of servicedelivery within the Swedish healthcare sector reflect deeplyrooted patterns of institutional authority and organisationalbehaviour in the Swedish welfare state in general and the

s at the Nordic Polit-eykjavik, Iceland in

icians who answered

. All rights reserved.

healthcare sector in particular. According to these authors,a patient had little opportunity to influence the caregivingprocess at this time. The introduction of patient choicewithin the healthcare sector was seen as a way to correct thisproblem and empower the patient. The same trend has beenseen in other Scandinavian countries. Both Denmark andNorway have introduced patient laws that include freechoice of hospital (Vrangbaek & Ostergren, 2006; Vrangbæk,Ostergren, Okkels Birk, & Winblad, 2007). From a Europeanperspective, Scandinavian patients’ new right to choosefreely must be considered far-reaching, and it was alsointroduced quite early. For instance, patients in Englandreceived the right to choose their provider of electivehealthcare as late as December 2005 (Lewis, 2005).

However, there are some indications that the effects ofthe Swedish policy have been limited. Studies have foundthat relatively few people are familiar with their right tochoose (Anell & Rosen, 1995; Health Care Barometer, 2004).Secondly, few patients have actually chosen other facilitiesfor care; most still seek care at the nearest hospital(National Board of Health and Welfare, 2005). Anotherindicator of the lack of goal attainment is the increasingnumber of complaints received by the county councils from

U. Winblad / Social Science & Medicine 67 (2008) 1502–1511 1503

patients who have tried but failed to exercise their right(National Board of Health and Welfare, 2003).

One possible explanation for the limited effects of thepolicy is the lack of policy implementation at the clinicallevel. In this article the physician’s role, especially, is seen asa major impediment to this implementation.

Even though the policy is meant to empower patients,physicians are highly involved when it comes to realisinga patient’s choice. The assistance of the physician is impor-tant in two respects when it comes to choosing a hospital.Firstly, a referral from the physician is usually needed whena patient wants to choose hospital care. Also, the past 20years have brought a technological revolution in the medicalfield and physicians are still the main source of patients’information about specific rules, waiting times and qualitydifferences between alternatives. Altogether, this showsthat patients depend on the benevolence of physicians.

Consequently, patient choice implies a new way oflooking at the patient from the perspective of the physician,and demands changes in the physician’s traditional way ofworking. Here, doctors need to involve and recognise theimportance of patients’ views instead of simply referringthem to a specific hospital based on the doctors’ ownmedical judgement or force of habit. Increased patientpower could therefore, from a doctor’s point of view, beconsidered something that threatens and interferes withthe traditional medical role.

Two questions are asked in this paper: Do physicians actin accordance with the policy intentions, informing andsupporting patients in their choice of hospital? If not, howcan this be explained?

In order to analyse physicians’ roles in putting the policyinto effect, a questionnaire was sent to 960 physicians.1

Before the theories, hypotheses and methods used in thestudy are outlined, the Swedish healthcare system and thepatient choice policy will be described.

The Swedish healthcare system and the choicerecommendation

Swedish healthcare is, in an international perspective,extremely decentralised (Saltman, Bankauskaite, & Vrang-baek, 2007) with a high degree of regional self-determi-nation. The healthcare is administered and financed byregional, political units called ‘‘county councils’’ that alsohave the right to levy taxes. Even if the central governmenthas established some basic principles for the whole country(most important is the Health and Medical Service Act), thenumber of laws has decreased in recent years. Instead,there is a strong tradition of what may be termed informalor consensus-based steering, so-called ‘‘soft law’’ (Morth,2004), by which the intentions of the Swedish parliamentand government are carried out in consensus throughdiscussion and negotiation with local government rather

1 The article is based on material from my doctoral dissertation (2003)which is titled ‘‘Från beslut till verklighet. Lakarnas roll vid imple-menteringen av valfrihetsreformer i halso- och sjukvården’’ (From deci-sion to practice. The role of the Physician in Implementing patient choicein Swedish Health Care).

than through compelling measures or binding legislation(Blomqvist, 2007).

An important link between the national governmentand the county councils is the Swedish Association of LocalAuthorities and Regions (formerly called the Federation ofSwedish County Councils, FCC), a co-ordinating politicalagency for all county councils; it initiates new policychanges and recommends them to the counties. The mostimportant policy concerning patient choice was a recom-mendation introduced in 1989 by the FCC (FCC, 1989). Therecommendation was not legally binding, but all countycouncils agreed to introduce it. From the beginning of 1991,patients had the right to choose care providers anywhere inthe country. The policy states that patients can choose notonly primary care centres but also which hospital or privateclinic to visit. However, if patients wish to choose careoutside their own county, a referral from the ‘‘home clinic’’is needed in most cases. The care is always paid for by thepatient’s home county council. The policy, which wasrevised and updated in 2000, is still in effect in 2008.

Other policies have also enhanced patient choice, forinstance certain changes in the Health and MedicalService Act in 1999 and a national waiting-time guaranteeintroduced in 2005. But here the focus is put on therecommendation from the FCC, since it can be consideredthe core of the different policies describing patient choice.It is also the only policy that focuses explicitly on the choiceof provider in specialist care (i.e. choice of hospital andprivate speciality clinic) which is the aspect of patientchoice that is studied here.

Physicians as implementers of the patientchoice policy

Implementation within the healthcare field is compli-cated due to the complex organisation consisting of anadministrative system, a professional system and a politicalsystem. These systems all have clear, but different, idealsand values on how healthcare is to be managed and whatthe result should be (Sahlin-Andersson & Ostergren, 1998).Still, doctors2 have historically held a unique position asregards to other professional groups as well as theadministrative and political actors (Freidson, 1989).

In this study we are interested in determining whetherdoctors have implemented the specific policy of patientchoice. That is, we ask if the physician actively supports andhelps the patient to choose a caregiver, for instance byinforming the patient of his or her right to choose or byexplaining different referral alternatives and waiting times.The Health and Medical Service Act states that healthcarepersonnel shall inform the patient about his or her healthstatus and methods for examinations, care and availabletreatments. This obligation to inform could of course alsoinclude information about different caregivers and waitinglists. It is also mentioned in the patient choice policy thathealthcare personnel are to act choice-supportively. Thereferral situation is of utter importance here – this is when

2 The terms doctor and physician are used equivalently in the text.

U. Winblad / Social Science & Medicine 67 (2008) 1502–15111504

the doctor needs to ask the patient to which hospital he orshe wants to be referred.

Sweden has no prior tradition of an explicit patientchoice in the referral situation. Instead, doctors have mostlymade referrals on the basis of medical concerns, knowledgeof the system and force of habit. According to the patientchoice policy, both public and private doctors are allowedto refer choice patients. Important to note is that the cost ofa referred patient never rests on the individual doctor;however, the ‘‘home’’ county of the patient is alwaysresponsible for paying for a patient when he or she seekscare in other counties. Most Swedish doctors are publiclyemployed by the county councils and receive a fixed salarywithout extra reimbursement for admitting choicepatients. Privately employed doctors, on the other hand, arereimbursed by the county councils for each treated patient,and therefore have a stronger financial incentive to admitchoice patients. Private care has increased in Sweden lately,and 25% of the ambulatory care is now privately run byenterprises commissioned by county councils.

However, it is quite common that the treating clinic athospitals receives revenue from out-of-county patients.Within the counties, on the other hand, reimbursement forchoice patients is not normally taken into consideration inthe budget (Winblad, 2007). To summarise, the Swedishhealthcare system creates few incentives, at least for theindividual public doctor, to either encourage or discouragechoice patients because of financial reasons.

If the result of the study shows that physicians do not takeaction to help patients choose a caregiver, the next task in thestudy is to explain this behaviour, i.e. to identify the reasonswhy physicians do not take actions to fulfil the politicalintention regarding the free choice of provider. Threedifferent hypotheses will be proposed below to explainphysicians’ behaviour. To achieve the policy goals, physiciansmust (1) be able to implement the changes; (2) understandthe intentions behind the changes and actual rules; and (3) bewilling to implement the changes. In the next sections, thesehypotheses will be discussed in the context of theories onimplementation and the medical profession.

Michael Lipsky and the theory ofstreet-level bureaucrats

The implementation literature has historically been splitinto two major schools, top-down and bottom-up. Amongthe bottom-uppers, it is Michael Lipsky who has mostexplicitly paid attention to officials working closest to thegeneral public, e.g. patients. He shows that these officials, inhis analysis called street-level bureaucrats (e.g. teachers,policemen, or doctors), are important political actors (Lip-sky, 1980). They shape policy through the large number ofdecisions they make in interacting with their clients andpossess a great amount of discretion since it is difficult fortheir superiors to control and monitor their activities.

Street-level bureaucrats perceive a huge gap betweentheir own resources and the demands of the public. Thisleads to a chronic feeling of insufficiency. In hopes ofdiminishing the gap between resources and demands, thegrass-roots bureaucrats develop certain coping strategies

such as complicating access to services, prioritizing easycases and dominating meetings with their clients.

Lipsky’s theory is important since it directs the search-light at professional groups in a way that the traditional top-down theorists fail to do. However, Lipsky has been criti-cised for ignoring the importance of political and structuralfactors affecting the actions of the street-level bureaucrats.‘‘Glaringly missing’’, writes Hasenfeld (1981, 156), ‘‘isa coherent explanation of the factors that shape and deter-mine the working conditions of lower-level staff to whichthey must adapt’’. The question asked is whether the servicepractice is more a function of the discretion of the street-level bureaucrats than of deliberate policies of the organ-isational and political elites. Translated to our case thepossibility for the doctors to implement the reforms, i.e.support the patients in choosing caregivers, is partlydecided on political and organisational levels other thanthose Lipsky suggests. The organisational structure createscertain prerequisites for the doctors’ abilities to help andsupport the patients when it comes to informing and sup-porting them in choosing a hospital.

Top-down theories about implementation

The top-down school, in contrast to Lipsky, has otheropinions about what is important in the implementationprocess. These scholars have focused their interest onfactors that can be manipulated from the top of the orga-nisation – hence the term ‘‘top-down’’ (Mazmanian &Sabatier, 1979; Pressman & Wildawsky, 1973). Solutions toimplementation problems have been defined in terms ofbetter control and co-ordination of the actors at loweradministrative levels. The policy-makers must be sure of theconnection between the efforts and the effects on the targetgroups. Top-down theorists also state that clear andunambiguous goals are a prerequisite for successful imple-mentation (Mazmanian & Sabatier, 1979). However, thiscriterion is difficult to fulfil. Vagueness and lack of clarity areoften part of the political process and are the result of a needto cover deviant meanings and opinions in politicalcompromises. One factor that is also said to be important fora successful implementation, according to the top-downschool, is the structuring of the implementation process(Sabatier, 1986). Firstly, only benevolent organisationsshould be involved in policy implementation. Localgovernments are often considered hindrances to nationalpolicy. The best solution is to design new, ‘‘immaculate’’organisations for the implementation of the program.Secondly, the numbers of actors involved should be mini-mised, thus decreasing the number of veto-points in thesystem. The healthcare system is a typical example ofa system that has become complicated by its many actorsand different political levels.

Hypothesis 1. (Physicians are not able to implement thepolicy) Thus far, we have shown that the top-downschool emphasizes that factors other than pressing workingconditions influence the behaviour of the street-levelbureaucrats. The prerequisites for our physicians to supportthe patients in their choices are, in accordance with thisdiscussion, determined on political levels other than that of

U. Winblad / Social Science & Medicine 67 (2008) 1502–1511 1505

the meeting between the doctor and patient. Instead,organisational factors are important for the doctor’s abilityto act in accordance with policy goals – i.e. support thepatients in the process of choosing. The county councils andthe departments are the organisational environment inwhich the physicians work. Does this environment preventphysicians from implementing the policy, i.e. to inform andhelp patients use their right to choose care providers? Thisis the reasoning that is the basis for Hypothesis 1:

Physicians are not able to help patients choose caregiverssince the organisational conditions for this behaviour arelacking in the county council.

Hypothesis 2. (Doctors do not understand the intentions of thepolicies) Hypothesis 2 draws mainly on insights from thebottom-up approach. This school has shown that street-level bureaucrats who are responsible for implementationmust interpret the policies and make certain trade-offsamong the policy goals. In the case studied, it seemsreasonable that the doctors’ understanding of the policyand knowledge of the rules plays an important role in theirability to implement it. Hypothesis 2 suggests:

Physicians do not understand the intentions of the policyand lack knowledge of the rule system regulating thepatient choice policy. This makes it difficult for physiciansto help patients choose a healthcare provider.

Sociological theories about the medical profession

One of Lipsky’s merits is that he has developed a theoryfor street-level bureaucrats. He shows systematic similari-ties when it comes to behaviour, independent of person-ality, nationality or policy field. However, Lipsky’s theoryhas been criticised for not explaining differences amongbureaucrats with similar workloads (Winter, 1990). Forinstance, it is possible that the policy field in itself influ-ences the behaviour of the street-level bureaucrat. Eachpolicy field with its varied professions creates specificprerequisites for the implementation of national policyinitiatives. For instance, the strictly biomedically orientededucation might leave its mark on doctors’ thoughts andactions. Lipsky’s theory does not take such professionalcultures into account, thereby lumping everyone together.

The healthcare field is characterized by a strongprofession, that of the physicians, which dominates thefield through its values and ideals. To understand imple-mentation problems in this area we must therefore knowmore about the medical profession per se. In his books, thewell known sociologist Elliot Freidson (1989, 1994) givesa picture of a strong, unthreatened medical profession. Heshows that there exists a knowledge gap between doctorand patient, which gives the doctor a strong position. Thisposition grows even stronger as medical technologyevolves. Even when the patient, for instance through theInternet, becomes more knowledgeable, the information ishighly technical and difficult for the patient to interpretwithout help from the physician.

Some sociologists have, however, argued that theknowledge gap between patients and physicians has

decreased over recent decades and that this has led toa decline in physicians’ authority over patients (Haug, 1988,1993; Haug & Lavin, 1983). According to the authors thereare several reasons for the decline, such as a rising educa-tional level among patients which gives them more confi-dence to address their own preferences or complaints todoctors, media reports on doctors’ medical failures, andpeople’s averseness to accept authority at large.

During the beginning of the 21st century Freidsonslightly revised his former hypothesis of a strong andunthreatened medical profession. In ‘‘Professionalism – thethird logic’’ he recognizes several changes in the status ofdoctors (Freidson, 2001). Among these can be mentionedthe growing support for non-medical practitioners andproducts to which access is not controlled by doctors. Thisdevelopment has, according to Freidson, complicated boththe health-related division of labour and the capacity of themedical profession to protect its own territory.

Moreover, he argues that the introduction of privateactors on the medical labour market has affected physi-cians’ autonomy. The bureaucratic management of privateinsurance companies and investor-owned health serviceproviders has come to influence the relation betweenphysicians and patients, meaning that the physician nolonger decides how much should be paid and for whichservices. A new form of cost and quality control has alsoenabled the bureaucratic administration to monitorphysicians’ effectiveness and compliance to national stan-dards in a way that was not conceivable earlier. Still, despiteintrusions on the territories of the medical core, Freidsonmaintains the idea of an influential profession, whichthrough different protective mechanisms, such as the legallicense to practice and the credential from medical school,manages to insulate itself from assaults.

Freidson has been criticised for not fully taking intoconsideration those external macroeconomic circum-stances factors which threaten to undermine the profes-sional dominance. In the article ‘‘The end of the golden ageof doctoring’’ (McKinley & Marceau, 2002) the authorsargue that different social influences have even led to anerosion and decline of the medical profession. The mostimportant reason for this, according to the authors, lies inthe shifting agenda of the national and local state. Earlier,the state served as a guardian for the medical profession byits legitimatizing function for many professional activitiesand by ensuring the profession’s profit through differentpayment systems. Nowadays, the state is more reluctant tointervene in economic and social life, as well as in medicalcare. This means that the state is neither willing to be incharge of the medical healthcare sector, nor willing toprotect the medical professional core in the same way thatit used to do. Instead, the state has laid its’ efforts onadvancing the interests of the financial and profit-seekingproviders in the healthcare system with the aim of stimu-lating the private market place.

Another explanation to the erosion of the medicalprofession is believed to be the new information tech-nology that offers an opportunity to empower patients(McKinley & Marceau, 2002). Computerized informationsystems with records of various aspects of the clinicalencounter make it easier for the patient to compare

U. Winblad / Social Science & Medicine 67 (2008) 1502–15111506

different providers in terms of quality of care, patientsatisfaction and productivity. This might lead to a feeling ofpowerlessness among physicians and implies that they,intentionally or unintentionally, oppose policies thatempower patients.

Hypothesis 3. (Physicians are not willing to implement thepolicies) Hypothesis 3, which deals with the changingpower relations among physicians and patients, is con-structed on the basis of the reasoning above. The idea of freechoice of provider challenges the traditional medical role inwhich the physician is devoted to medical issues and is theperson who decides where the patient is to be referred.Instead, the patient choice policy means that when morethan one alternative is possible in the referral situation, thequestion of what is best for the patient no longer belongs tothe physician. This can be perceived as a loss of authorityand a limitation of the internal, clinical autonomy.

Furthermore, the policy can also threaten doctors’external autonomy, i.e. their working conditions. Informingabout patient choice can be very time-consuming, forinstance.

Finally, there is a chance that doctors are unwilling tohelp patients choose a certain care provider, fearing thatthe patients might make medically inferior decisions andthus risk their health and recovery. This can be formulatedin Hypothesis 3:

Physicians do not want to help patients choose since thechoice policy means that the autonomy of the physicians isthreatened and the patients risk making medically inferiordecisions.

Methods

To examine the hypotheses presented above, a ques-tionnaire was sent to 960 physicians in one of the mostpopulous counties in Sweden in June 2001. The physicians’addresses came from the register of the National Board ofHealth and Welfare.

Since choice of hospital is focused on in the study thereferral situation is of central importance. The moment thephysician signs a referral, the patient has the possibility toexpress desires about where he or she wants to be referred,meaning that the free choice of provider policy is put intoeffect. In Sweden, general practitioners (GPs) as well ashospital doctors, both private and public, are allowed torefer choice patients to hospitals within or outside thecounty. A common ‘‘choice situation’’ is that a GP refersa patient to a consultant at a hospital for an assessment. Ifthe consultant decides that a medical/surgical procedure isneeded, the patient has the right to choose where he or shewants to be operated on. The patient might choose anotherhospital because of shorter waiting lists, or to be closer torelatives in the rehabilitation phase. In order to includeboth hospital doctors and GPs in the study doctors withinfour broad specialities who are most likely to refer choicepatients were chosen: General Practitioners, Orthopaedics,Surgeons and Internists.

The questionnaire that was sent out consisted of 52forced-choice questions, of which 22 were used for the

purpose of this article, i.e. had relevance for the testing ofthe hypotheses. The remaining questions concerned otheraspects of patient choice not covered in this article. Thequestionnaire was tested on seven physicians who com-mented on the design and content of the questions. Itsvalidity was also tested through factor analyses, which ledto the exclusion of four questions. The Sphinx computerprogram was used for input of data, and SphinxSurvey andSAS were used for statistical analysis.

After three reminders, the response rate was 71%. Afterthe exclusion of answers from retired physicians, physi-cians who worked in other occupations, physiciansworking as researchers, etc. the effective response ratedropped to nearly 60% (59.8%). Analysis of the non-response shows that this group consists of a higherproportion of older physicians and women. The differences,however, are not significant. More remarkable is that thenon-response consisted of a significantly higher proportionof GPs. Only 48% of this group answered the survey.

In addition to the questionnaire, a case study was con-ducted in the same county council to investigate how theactual policy was implemented generally, and morespecifically to investigate if there existed organisationalhindrances to physicians referring patients in line with thedesires of the patients.

The county council selected for the study was theStockholm County Council. This case was chosen accordingto the ‘‘critical case method’’, which means that the casechosen is where the policy is ‘‘most likely’’ to be imple-mented (Eckstein, 1975). If it does not work here, thechances of success are even smaller in counties with lessfavourable conditions. Stockholm has many alternativecaregivers to which to refer patients (both private andpublic), a positive political climate for patient choice andhighly motivated patients. In the case study, policy docu-ments from the past 25 years concerning patient choiceissues were collected and analysed to determine how thepolicy had been implemented in the county council. Inaddition to this, 19 public officials in central positions in thecounty council were interviewed regarding organisationalhindrances to the physicians’ abilities to refer patients.

Do physicians act choice-supportively?

In the following part, the question ‘‘Do the physiciansact in accordance with the policy intentions by informingand supporting patients in their choices of hospitals’’ isanswered. This is done by investigating the physicians’attitudes regarding the free choice of provider and then byasking them how they act when it comes to informingpatients of their right to choose.

Do physicians approve of the policy?

Table 1 shows that the majority of the respondents hada positive attitude regarding patient choice.

Most of the physicians agreed that patients ought tochoose their own GP. Choice of hospital was not consideredas important as choice of GP, even if almost three-fourthsalso considered this important. An analysis of the back-ground variables shows no significant differences by

Table 1Physicians’ attitudes regarding patient choice in healthcare (%)

Choice of Importanta Neither Unimportantb N

GP 86.8 8.4 4.8 502Hospital 76.5 15.0 8.5 507Specialist 71.2 17.6 11.3 507

a The categories ‘‘Very important’’ and ‘‘Somewhat important’’ havebeen combined into ‘‘Important’’.

b The categories ‘‘Very unimportant’’ and ‘‘Somewhat unimportant’’have been combined into ‘‘Unimportant’’.

U. Winblad / Social Science & Medicine 67 (2008) 1502–1511 1507

gender or age. However, type of speciality made a greatdifference. Private specialists and physicians at the hospitalwere far more positive to the patients’ choices of hospitalthan the GPs were. This is remarkable since it is primarilythe GPs who refer patients to hospitals.

Other answers also indicate that physicians see thevalue of the free choice of provider. They reported thatpatient choice had led to more effective healthcare and thatthe quality of the care had increased through the intro-duction of competition in the system. A quotation illus-trates this attitude: ‘‘Patient choice increases thecompetition between the caregivers and stimulates a betterand more effective therapy’’ (Physician No. 235).

Are physicians active in helping patients choose caregivers?

Thus far it has been shown that most physicians are infavour of the policy that has introduced choice of provider.The next step was to investigate whether doctors also act inthis way, i.e. whether they inform patients of their right tochoose and of the alternatives to which they can bereferred. The formulation of the first question investigatingthe supportiveness of the physicians was: ‘‘Do you usuallyinform patients of their right to choose healthcareproviders?’’ As Table 2 shows, only 17% of the physiciansalways informed the patients of their right to choose.Instead, the majority (40%) only informed patients whoasked about this right.

Again, great differences existed, depending on where inthe health system the physician worked. Almost 40% of theprivate physicians stated that they always informedthe patients of their right to choose, whereas only 7% of thehospital physicians reported doing so.

The other question which investigated how ‘‘choice-supportive’’ the physicians were was: ‘‘Do you usuallyinform patients about the alternatives they can be referredto?’’ The same pattern is shown here; few physicians (15%)regularly informed about alternatives. An equal numberstated that they never informed patients about this. As in

Table 2Numbers of physicians informing patients of their rights to choosea caregiver (N¼ 496) (%)

Informs about patient choice %

Always 17.1When information is requested 40.7When waiting lists are long 11.1Occasionally 17.5Never 13.5

Total 100

the previous question, it was mainly when the informationwas requested by patients that the physicians informedabout different caregiver alternatives.

The results of the survey show that the physicians donot regularly help patients choose healthcare providers bygiving them information and allowing them to choosewhere they want to be referred. Instead, the referrals aremostly based on medical grounds; the desire of the patientis of less importance. The analysis also indicates thatprivate physicians are more willing to provide informationto patients about choice of provider than are, for instance,physicians working in hospitals. To sum up, the doctors donot in an active way appear to support patients to choosehealthcare provider.

Why are the physicians not supportive?

The next question in the empirical investigation waswhy physicians did not encourage patients, in the referralsituation, to choose their healthcare providers. Threehypotheses, described earlier, were proposed to explainthis behaviour. To achieve the policy goals, physicians must(a) be able to implement the changes; (b) understand theintentions behind the changes and actual rules; and (c) bewilling to implement the changes.

Are physicians able to implement the policy?

This hypothesis investigates whether there are anyorganisational hindrances to doctors referring choicepatients. This is investigated through a questionnaire todoctors and a case study in the Stockholm County Council.The latter showed clearly that the politicians had a positiveattitude toward the policy and had introduced moregenerous rules locally than what was stated in the nationalpolicy. The county also provided data, accessible by doctors,in a national database on waiting times for electivesecondary care, i.e. hospital care. One problem, though, isthat the county council had done very little to disseminateinformation about the rules. The case study also demon-strated that there existed many alternative providers, bothprivate and public, to which doctors could refer patients.

Whether the physicians felt restricted in referringpatients to other county councils was investigated in thequestion: ‘‘Has your clinic/department introduced anyrestrictions on referrals to other counties?’’, only 21% of thephysicians, mainly the GPs, answered that they feltrestricted to refer choice patients.

In addition to this, answers to the question ‘‘Have youbeen refrained from referring a patient to another clinic/department within the county because your district lacksa contract with that clinic/department?’’ support theconclusion that doctors did not feel practically restrictedfrom referring patients freely. Just one-fifth answered thatthey had been refrained. Altogether, there seemed to befew hindrances on the organisational level that preventedthe physicians from referring patients to the caregivers thepatients themselves desired. However, one area in whichthe physicians were dissatisfied concerned the informationavailable about the policy. Over 60% considered the infor-mation available to be ‘‘very bad’’. Many of the physicians

U. Winblad / Social Science & Medicine 67 (2008) 1502–15111508

were indifferent, and only a minority felt satisfied with theinformation. On the question ‘‘Have you been offered totake part in education/information concerning issues aboutpatient involvement and the rules about patient choice’’,only 10% answered that they had been offered this oppor-tunity. The doctors were also asked how much they knewabout the waiting times at other units within their ownspeciality. Half of the doctors were satisfied with theirknowledge, while one-fourth was indifferent. To summa-rise, the physicians themselves confirmed the faultydissemination of information but stated, in accordancewith the results of the case study, that they did not feelparticularly restricted in any way by organisational factorsfrom referring patients to whomever they wanted.

Do physicians understand the intentions of the policy?

The idea behind the ‘‘understand hypothesis’’ is that theless knowledge the physicians have about the motivesbehind the policy and the rules that follow it, the less likelythey are to inform the patients about their right tochoose. The result of the questionnaire seems to confirmthe hypothesis that physicians do not fully understandthe choice policy. When asked ‘‘What do you think is thepolitical intention of increasing patients’ opportunities tochoose provider of care’’ only half of the respondentsmentioned that the intention of the policy was to increasepatient empowerment. This is a low number, consideringthis was its main rationale.

More interesting is perhaps what actual knowledge thephysicians had about the actual rules concerning patientchoices. Several questions in the survey tested the physi-cians’ knowledge of the content of the policy, i.e. the rules.Table 3 shows that the knowledge of the rules within thecounty council is quite good; 7 of 10 physicians knew, forinstance, that patients were allowed to choose a publichospital within the county council. The physicians’knowledge of the rules was less accurate concerning othercounty councils, however. Only 7%, for instance, knew thatpatients have the right to choose care providers anywherein the country.

In sum, even if the respondents were quite knowl-edgeable about the rules within the county council, theirlevel of understanding of the motives behind the policy wasstill quite low.

Are physicians willing to implement the policy?

The ‘‘willing’’ hypothesis assumes that the physicians donot approve of the new policy and are therefore unwillingto help patients choose care providers. As shown in the

Table 3Physicians’ knowledge about patients’ formal rights to choose caregivers(%)a

The patient is allowed to choose Yes No Do not know N

Public care within the county council 72.1 15 12.8 506Private care within the county council 63.9 15.3 20.7 502Care in the whole country 6.9 60.3 32.8 476

a The correct answer is ‘yes’ to all questions.

beginning of the article, this unwillingness can for instancebe connected to a fear that the patient might take medicallyinferior decisions.

First, it was investigated how the physicians regardedthat patients might choose inferior caregivers. One questionasked, ‘‘Can a patient’s desire to be referred to a specificcaregiver be justified even if you consider the medical careinferior at that clinic?’’ Nearly 70% admitted they wouldrefer a patient anyway, 20% denied that they would do so,and the rest were unsure. This question generated a hugenumber of comments (125 comments). Most respondentsplaced conditions on their answers, saying they acceptedthe patient’s medically inferior choice, but only if they hadbeen informed of the risks or if the quality of the alternativeswas good enough. One group of physicians declared thatpatients have the right to make their own choices: ‘‘It is thepatient who decides about his or her own life’’ (PhysicianNo. 27). Another question showed clearly that physiciansprefer deciding themselves where to refer the patients;almost 60% said that the physician should have the decisiveinfluence on where the patient is referred. Altogether, theresults give quite an unclear picture. But apparently, whenmore specific questions about the referral situation areasked, physicians prefer to decide themselves where to refertheir patients. The traditional role of the physician as theperson who makes the medical decisions seems to linger on.

A second dimension of the hypothesis studied whetherthe doctors felt that the policy had changed their internal,clinical autonomy in a negative way. One question,formulated as: ‘‘Has the possibility of carrying out a plan-ned treatment been changed due to the patients’ oppor-tunities to choose’’, showed that only 15% agreed that theirclinical work had been influenced by the policy. Onecomplication mentioned was that patients, instead of fin-ishing a treatment they had begun, chose to seek a secondopinion. Other questions also proved that few doctorsadmitted that the policy had changed their inner rela-tionships with their patients.

A third dimension of the hypothesis investigatedwhether physicians felt that their external autonomy wasnegatively influenced by the policy, i.e. if helping patientschoose a caregiver required too much time and resources.Table 4 shows that opinions on whether it takes too muchtime to inform about the policy were divided. One-third ofthe physicians stated that helping the patients choose takestoo much time, one-third were indifferent and one-thirddid not agree at all that the policy was time-consuming.There were huge attitude differences depending on wherein the system the physicians worked. The GPs were, as alsoshown in other questions, significantly more negative.Another question showed that hardly any of the physiciansregarded the policies as too costly for their clinics (5.9%).Few of the respondents also admitted that their patients(from their catchment areas) ‘‘too often’’ chose hospitals orclinics elsewhere.

To summarise, the third hypothesis stated that thephysicians do not approve of the new policy and aretherefore unwilling to help patients choose a care provider.The physician’s answers did not offer any confirmation forthis assumption. Rather, the results suggest that physicianstoday do not feel particularly threatened by the policy

Table 4Physicians’ opinions about the policies’ influence on external autonomy

The statement is: (%)

N Gooda Neither goodnor bad

Badb

It takes too much time to informabout the policy

459 29.7 31.8 38.5

The policy is too costly for my clinic 472 5.9 28.6 65.4Patients in our geographical catchments

areas too often choose other clinicsor hospital than ours

481 4.1 17.9 77.9

a The alternatives ‘Perfectly right’ and ‘Quite right’ are combined into‘Good’.

b The alternatives ‘Not at all correct’ and ‘Not correct’ are combined into‘Bad’.

”able”

”understand”

”willing”

”Choice-supportive”

0.14**

0.14**

0.29***

0.08

0.12**

0.28***

0.09*

0.13**

0.26***

0.04

0.12**

0.26***

r r whenotherindicesare heldconstant

r whencon-foundersare heldconstant

r whenotherindices andconfoundersare heldconstant

Fig. 1. The coefficient of correlation (r) between ‘able’, ‘understand’, ‘willing’and whether physicians are ‘‘choice-supportive’’ (n¼ 504, *p< 0.05;**p< 0.01; ***p< 0.001).

U. Winblad / Social Science & Medicine 67 (2008) 1502–1511 1509

investigated here. They are willing to let patients choosemedically inferior options, if the patients are fullyinformed. Also, it does not seem that the policies take toomuch time away from work or that the choice negativelyinfluences the doctors’ inner clinical work. One exception isthe GPs, who are, as a group, more willing to admit that thepolicies have negatively changed their working conditions.

Correlation among the hypotheses

In the last part of the empirical investigation, the focusis shifted to answering the question of whether there isa correlation between the separate hypotheses ‘‘able’’,‘‘understand’’ and ‘‘willing’’, and whether the physiciansassist patient choice through information about their rightsin general and about which alternatives patients canchoose (‘‘choice-supportive’’). In order to investigate thecorrelation between the hypotheses, four indices wereconstructed from the question results in the survey. Eachindex was the mean value of answers to a group of ques-tions, measured in an ordinal scale from 0 (negativeregarding patient choice) to 1 (positive regarding patientchoice). The selection of questions to be included in eachindex was validated by a factor analysis. The partial corre-lations between indices, i.e. when the effects of otherindices and the confounders (physician’s age, specialitywithin which the physician worked, and type of employerof the physician) were held constant, were computed usingthe Spearman rank test (Fig. 1).

There was an insignificant correlation (p¼ 0.04)between the organisational factors (‘‘able’’) and the degreeto which physicians help patients choose care providers(‘‘choice-supportive’’). This means that a physician whofeels restricted by organisational factors is not less willingto help patients choose care providers, compared toa physician who does not feel restricted.

As for the ‘‘understands’’ hypothesis, the analysisreveals a (moderately) significant correlation between thephysicians’ understanding of the policies and their actualbehaviour when it comes to informing patients of theirright to choose (p¼ 0.12**). Physicians who are moreknowledgeable about the actual rules and motives aremore liable to assist patients in choosing a care provider atthe hospital level.

The statistical analysis also shows a significant correla-tion between whether the physicians perceived their

working conditions to have changed negatively andwhether they were willing to help patients choose careproviders (p¼ 0.26***). The physicians most threatened bythe policy were much less willing to inform patients aboutpatient choice and vice versa. However, it should beremembered that, as demonstrated earlier, few physiciansoverall felt threatened by the policy.

Discussion

The aim of this paper has been to analyse physicians’roles in implementing a policy in Sweden on patient choice.A questionnaire completed by a large number of physiciansshowed that even if the physicians had a positive attituderegarding the policy few had taken action to put it intoeffect. Few physicians have changed their behaviour toconform to the policy directives, and patients are onlyhelped to choose care providers if they request this them-selves. To summarise, more than a decade after the patientchoice policy was introduced, the results illustrate that onlya minority of physicians act in accordance with the politicalintentions. This could be one explanation for why so manypatients are still unaware of their right to choose theirhospital.

Once this was established, the next question was whyphysicians did not support patients’ right to choose. Forinstance, did organisational hindrances in the system makeit difficult for the physicians to refer choice patients? Thecase study in the Stockholm County Council showed thatfrom a formal perspective there were no direct limitationsin the organisation that would prevent the physicians fromdoing so. The questionnaire confirmed that physicians didnot feel restricted in referring choice patients. Secondly, itshowed that knowledge of the rules and motives con-cerning the policy was poor among physicians in general.The statistical analysis confirmed that physicians who weremore knowledgeable about the policy were more liable toassist patients in choosing a healthcare provider at thehospital level. Thirdly, when the physicians were askedwhether they felt threatened by the empowered patients,the results showed that only a minority experienceda conflict between the policies and their daily clinicaldecision-making.

One idea behind the patient choice policy was thatdoctors were to be more service-oriented due to increased

U. Winblad / Social Science & Medicine 67 (2008) 1502–15111510

contestability. In reality, the fear that patients would ‘‘exit’’did not seem especially threatening to our doctors. Thismay be because Stockholm is a place where the demand isvast, and there is always a new patient seeking care. At thesame time, some physicians commented that losingpatients always involves a loss of prestige.

Worth considering is that even if the study showed thatfew physicians felt threatened by the policy, the correlationanalysis showed that when physicians feel threatened bythe policy they were less willing to help patients choose.However, there are many signs that the patients oftomorrow will be more knowledgeable about their condi-tions, as well as more demanding (Coulter & Jenkinson,2005; Neuberger, 2000). This might lead to a situation inwhich physicians experience the studied policy as morethreatening to their autonomy than they do today. Themore negative attitudes regarding the choice policy amongthe general practitioners already show this. One interpre-tation is that this reflects their negative experiences fromthe GP reform introduced in Sweden at the beginning of the1990s, which increased their work load.

The study also shows how difficult it is to change clinicalpractice with political initiatives. The success of a policy ishighly dependent on what happens when the policybecomes reality and is incorporated into daily routines.Although the policy was introduced more than 15 yearsago, the implementers – in this case, the physicians – havedone little to change their behaviour in accordance withthe policy directives. Physicians still do not involve theirpatients in decisions about referrals. Instead, strictlymedical assumptions seem to be the main basis for referral.Even if political support has been strong, it does not seemthat this enthusiasm has reached all the way down to theactual physicians who are to put into practice the initiativesregarding patient choice.

The failure of the implementation also raises questionsabout the ability of the Swedish national state and thecounty councils to manage the healthcare sector throughpolitical initiatives. Some studies have shown that doctorsgenerally respond reluctantly to management initiativesfrom above and resist changes that affect their autonomy(Blomqvist, 2007; Immergut, 1992). In addition to thisgeneral trend, the specific character of the Swedishhealthcare system with its unclear division of power, weakpolitical control and loose steering makes it even harder forthe politicians to govern the doctors. The system is highlydecentralised with strong regional units (county councils),which makes national management of the system difficult.National legislation is rare, and the relationship betweenthe national state and the counties is often built onagreements and negotiations, so-called ‘‘soft law’’ (Morth,2004). An example of this mode of governance is thestudied patient choice policy with no legally binding force.This in combination with no national follow-up of thepolicy, and weak structuring of the implementation processon the side of the county councils, created an environmentthat left great scope for the doctors to implement thereform in the direction they wanted. Supposedly, thepoliticians considered the patient choice policy mainly tobe a concern for patients, rather than something thatinvolved the healthcare personnel. As shown in this paper,

this was quite a naıve assumption – patients are in facthighly dependent on their physicians when it comes toexercising their right to choose a healthcare provider.

Lastly, the study also raises the difficult question of whatpatient choice truly implies. Does the policy place morevalue on the doctor’s politeness and behaviour than on thequality of care? There will always be moments when thedoctor feels that he or she knows better than the patientwhich hospital can give the best treatment. Our results showthat doctors solved this quite troublesome situation byletting the patient choose medically inferior options, underthe condition that the patients were fully informed.However, one ought to remember that decisions withinhealthcare are not only of a medical nature but deal with thewhole life situation and well being of a patient. The medicalprofession should acknowledge that a patient may considershorter waiting times in another county more importantthan high quality of care. In the long run the whole health-care has much to gain from involved and engaged patients –something that especially doctors can help in creating.

References

Anell, A., & Rosen, P. (1995). Valfrihet i vården. En empirisk undersokning avpreferenser hos allmanhet och vårdpersonal. [Patient choice withinhealth care. An empirical investigation of preferences among citizens andhealth care personnel], Vol. 4. Lund: Institutet for halso- och sjukvår-dekonomi, IHE.

Blomqvist, P. (Ed.). (2007). Vem styr vården? Organisation och politiskstyrning inom svensk sjukvård. [Who is in charge of Swedish health care.Organisation and political management within Swedish health care].Stockholm: SNS forlag.

Coulter, A., & Jenkinson, C. (2005). European patients views on theresponsiveness of health systems and healthcare providers. TheEuropean Journal of Public Health, 15(4), 355–360.

Eckstein, H. (1975). Case study and theory in political science. In F. I.Greenstein, & N. W. Paulsby (Eds.), Strategies of inquiry. Handbook ofpolitical science, Vol. 7. Reading, MA: Addison-Wesley PublishingCompany.

FCC (Landstingsforbundet) Forbundsstyrelsebeslut A 89:41 [Board deci-sion A 89:41].

Freidson, E. (1989). Medical work in America, essays on health care. NewHaven/London: Yale University Press.

Freidson, E. (1994). Professionalism reborn. Theory, prophecy and policy.Oxford: Blackwell Publishers.

Freidson, E. (2001). Professionalism – The third logic. Cambridge: PolityPress.

Hasenfeld, Y. (1981). Book review of street-level bureaucracy: dilemmasof the individual in public services. Social Service Review, 55(1),155–156.

Haug, M. R. (1988). A re-examination of the hypothesis of physiciandeprofessionalization. The Milibank Quarterly, 66(2), 48–56.

Haug, M. R. (1993). The role of patient education in doctor–patient rela-tionships. In J. M. Clair, & R. M. Allman (Eds.), Sociomedical perspec-tives on patient care. Lexington: The University Press of Kentucky.

Haug, M. R., & Lavin, B. (1983). Consumerism in medicine: Challengingphysician authority. Beverly Hills: Sage Publication.

Health Care Barometer [Vårdbarometern]. (2004). www.vardbarometern.nu/main.asp.

Immergut, E. M. (1992). Health politics, interests and institutions in WesternEurope. Cambridge: Press Syndicate of the University of Cambridge.

Lewis, R. (2005). More patient choice in England’s national health service.International Journal of Health Services, 35(3), 479–483.

Lipsky, M. (1980). Street-level bureaucracy: Dilemmas of the individual inpublic services. New York: Russel Sage Foundation.

McKinley, J. B., & Marceau, L. D. (2002). The end of the golden age ofdoctoring. International Journal of Health Services, 32(2), 379–416.

Mazmanian, D. A., & Sabatier, P. A. (1979). The conditions of effectiveimplementation: a guide to accomplishing policy objectives. PolicyAnalysis, 5(4), 481–504.

Morth, U. (2004). Soft law in governance and regulation. An interdisci-plinary analysis. Cheltenham, UK: E. Elgar Publishing.

U. Winblad / Social Science & Medicine 67 (2008) 1502–1511 1511

National Board of Health and Welfare (Socialstyrelsen). (2003). Patient-namndernas rapportering till Socialstyrelsen. [The patient boards’ reportto the National Board of Health and Welfare].

National Board of Health and Welfare (Socialstyrelsen). (2005). Var vårdaspatienten? Analys av patientstrommar i Sverige. Studie omfattandepatientregistret 1998–2002. [Where is the patient treated? Analysis ofpatient mobility in Sweden. The study includes the patient registry1998–2002].

Neuberger, J. (2000). The educated patient: new challenges for themedical profession. Journal of Internal Medicine, 247, 6–10.

von Otter, C., & Saltman, R. (1990). Valfrihet som styrmedel. [Patient choiceas means of control]. Stockholm: Arbetslivscentrum.

Pressman, J. L., & Wildawsky, A. (1973). Implementation. How greatexpectations in Washington are dashed in Oakland; or why its amazingthat federal programs work at all. Berkley: University of CaliforniaPress.

Sabatier, P. A. (1986). Top-down and bottom-up approaches to imple-mentation research: a critical analysis and suggested synthesis.Journal of Public Policy, 6(I), 21–48.

Sahlin-Andersson, K., & Ostergren, K. (1998). Att hantera skilda varldar:lakares chefskap i motet mellan profession, politik och administration.[Dealing with different worlds: the meeting between doctors, politics andadministration]. Stockholm: Landstingsforbundet.

Saltman, R. B., Bankauskaite, V., & Vrangbaek, K. (2007). Decentralizationin health care: Strategies and outcomes. Open University Press.

Vrangbaek, K., & Ostergren, K. (2006). Patient empowerment and theintroduction of hospital choice in Denmark and Norway. HealthEconomics, Policy and Law, 1, 371–394.

Vrangbæk, K., Ostergren, K., Okkels Birk, H., & Winblad, U. (2007). Patientreactions to hospital choice in Norway, Denmark and Sweden. HealthEconomics, Policy and Law, 2, 125–152.

Winblad, U. (2007). Valfriheten: En misslyckad sjukvårdsreform?.[Patient choice: a failed healthcare reform] In P. Blomqvist (Ed.),Vem styr vården? Organisation och politisk styrning inom svensksjukvård (pp. 132–155) Stockholm: SNS forlag.

Winter, S. (1990). Integrating implementation research. In D. J. Palumbo,& D. J. Calista (Eds.), Implementation and the policy process. Opening upthe black box. New York: Greenwood Press.