PAYING DOCTORS: IMPACT OF A CHANGE IN … · puis cette époque, les facultés de médecine au...

LA REVUE CANADIENNE D'ÉVALUATION DE PROGRAMME 73The Canadian Journal of Program Evaluation Vol. 17 No. 1 Pages 73–96ISSN 0834-1516 Copyright © 2002 Canadian Evaluation Society

PAYING DOCTORS: IMPACT OF A CHANGEIN REMUNERATION METHOD AT ACANADIAN ACADEMIC HEALTH CENTRE

S.E.D. ShorttQueen’s UniversityKingston, Ontario

Almost a century ago, North American medical reformers ar-gued that clinician scholars could be relied on to do their bestacademic work only if they were freed from reliance on entre-preneurial activity. Since that time, Canadian medical schoolshave become increasingly reliant on medical fees to subsidizetheir academic mission. At present, clinical earnings no longerappear able to support medical education. This article describesthe early results from an innovation in physician payment atan Ontario academic health centre. Under the new system allfaculty abandoned fee-for-service billing and were funded froma global budget. This remuneration change had little impact onclinical service, education, or research. However, the system hasintroduced unprecedented stability and accountability to thefunding of medical education.

Il y a un siècle, en Amérique du Nord, les réformateurs du do-maine de la santé affirmaient que les cliniciens/professeurs demédecine libérés de leur dépendence sur des activités entre-preneuriales produiraient un meilleur travail académique. De-puis cette époque, les facultés de médecine au Canada sontdevenues de plus en plus dépendantes des frais médicaux pourfinancer leur mission d’enseignement. Présentement, les gainsde la pratique clinique ne semblent plus capable de soutenir lescoûts de l’éducation médicale. Cet article décrit les résultatspréliminaires de l’application d’une rémunération novatrice desmédecins d’un centre universitaire de santé de l’Ontario. Sousle nouveau régime, les professeurs de médecine sont rémuné-rés à partir d’un budget global plutôt qu’à l’acte. Ce change-ment de rémunération a eu peu d’impact sur le service à laclientèle, l’éducation et la recherche. Ce nouveau système a in-troduit une stabilité et une responsabilité financière face à l’édu-cation médicale qui demeurent sans précédent.

Abstract:

Résumé:

THE CANADIAN JOURNAL OF PROGRAM EVALUATION74

A recurrent theme in the health system restructuringthat has characterized Canada in the past decade is the need tocontrol costs by finding alternatives to the fee-for-service (FFS)method of paying physicians. Physician remuneration generally ac-counts for up to a quarter of provincial health care expenditures. Inprimary care, capitation payment for rostered patients is a popularmodel. Devising similar innovations for paying specialists has re-ceived far less attention. However, the Ontario government is cur-rently committed to creating alternative payment plans forspecialists at each of the province’s five academic health centres(AHCs). The studies described in this article explore the character-istics and initial consequences of the first of these plans, implementedin Kingston in 1994.

PROGRAM CONTEXT: PAYING ACADEMIC PHYSICIANS

Early in the 20th century, medical reformers began to argue thatclinician scholars could be relied on to do their best academic workonly if they were freed from reliance on entrepreneurial activity(Ludmerer, 1985; Rothstein, 1987). Despite a flurry of enthusiasmfor salaried payment, however, North American medical schools be-came increasingly dependent on faculty earnings. By 1991 faculty-generated clinical earnings accounted for 45% of American medicalschool budgets (Jolin, Jolly, Krakower, & Beran, 1992), and the As-sociation of American Medical Colleges estimated that by 1993, 28cents of every dollar of clinical faculty income was used to subsidizeacademic activities (Jones & Sanderson, 1996). Canadian schools,where clinicians are required to return to their faculties billings inexcess of an individually negotiated ceiling, are similarly reliant onfaculty earnings (Thorne, 1997).

Throughout North America this system of financing medical educa-tion is threatened. In the United States, the Balanced Budget Act of1997 will significantly reduce Medicare subsidies to medical educa-tion by 2002 (Iglehard, 1999). At the same time, confronted with thecompetitive managed-care market, teaching hospitals are no longerable to bill at rates that reflect the extra costs associated with theiracademic role, traditionally a premium of 30% (Fishman & Bentley,1997; Kuttner, 1999). In Canada, a constricted hospital sector dur-ing the 1990s has lead to fewer teaching beds and a reduction inpractice volume for hospital-based clinical faculty (Government ofOntario, 1997, 1998). As well, provinces imposed both individualand aggregate caps on physician billings (Dowdall & Ramchandar,

LA REVUE CANADIENNE D'ÉVALUATION DE PROGRAMME 75

1999). As most academic departments share in clinician earningsabove individually negotiated ceilings, capped billing had its great-est impact on that portion of an academic clinician’s earnings whichwould have gone to the medical school.

Program Objectives

It was in this context of diminished fiscal stability for medical edu-cation that the academic health centre in Kingston, Ontario, negoti-ated with the province an Alternative Funding Plan (AFP) in 1994and put in place a new governance structure, known as the South-eastern Ontario Academic Medical Organization (SEAMO), by whichto administer it. The AFP replaced FFS with annual envelope fund-ing initially based on previous aggregate clinical billings, but in-cluding several other sources of public funds as well. Thegovernment’s objective was to achieve budgetary predictability forthe provision of clinical care at the centre. The academic centre alsosought budgetary stability, but additionally believed the removal ofvolume-based financial incentives would result in changed patternsof physician behaviour. Marginally necessary care would be reduced,resulting in both enhanced quality of care and increased time forresearch and teaching (Queen’s Health Policy Research Unit, 1996).

Physician Payment in the Literature

It is difficult to find compelling support in the existing literature forall the founders’ assumptions as to the likely effect of the AFP. Forexample, no literature explicitly links the payment method of fac-ulty members to educational processes or outcomes. The same maybe said for remuneration and research activity. The documentedpredictors of research productivity for individuals include age(Krumland, Will, & Gorry, 1979; Pearse, Peeples, Flora, & Freeman,1976), research training (Beaty, Babbott, Higgins, Joly, & Levey,1986; Levey et al., 1988; Pincus, Haviland, Dial, & Hendryx, 1995),and gender (Barnett et al., 1998; Carr, Friedman, Moskowitz, Kazis,& Weed, 1992), while for organizations the key factors recognized inthe literature are the presence of a well-established research cul-ture (Allison & Stewart, 1974; Bland & Schmitz, 1986; Crane, 1965;Long, 1978; Long & McGinnis, 1981; Manu, Landaw, Schwartz, &Williams, 1985), mentoring of junior researchers by senior colleagues(Cameron & Blackburn, 1981; el-Guebaly & Atkinson, 1996; Mills,Zyzanski, & Flocke, 1995), and assurance of sufficient protected re-


search time for faculty members (Harrington & Levine, 1986;Vardan, Smulyan, Mookherjee, & Mehrotra, 1990). However, a spe-cific relationship between payment method and research has notbeen described.

In contrast to the situation for research and education, importantaspects of the relationship between remuneration method and theprovision of clinical service have been studied, though predominatelyin American settings. There is ample evidence that FFS paymentencourages an increase in the volume of medical services comparedto other payment formats (Hemenway, Killen, Cashman, Parks, &Bicknell, 1990; Hickson, Altemeier, & Perrin, 1987; Hohlen et al.,1990; Nguyen & Derrick, 1997; Yip, 1998). Financial incentives havealso been shown to alter the content of practice. Surgeons paid byFFS, in comparison to salaried practitioners (Wilson & Longmike,1978) or participants in prepaid plans (Luke & Thomson, 1980), havebeen found to have different approaches to resource use and timingin the management of common surgical conditions. For example, theCaesarean rate under Medicare reimbursement is lower than undermore lucrative reimbursement from private insurance in the UnitedStates (Gruber, Kim, & Mayzlin, 1999; Keeler & Brodie, 1993). Simi-larly, a Medicare demonstration project revealed that bundling allcomponents of payment for bypass surgery into a negotiated globalsum resulted in significant cost savings due to changed surgeon prac-tice with respect to the use of nursing staff, laboratory tests, and in-tensive care beds (Cromwell, Dayhoff, & Thoumaian, 1997).

Some of the variations in patterns of care provision under differentremuneration systems may represent the elimination of least-nec-essary care. When physicians under prepaid plans were comparedto FFS surgeons, the prepaid groups performed fewer proceduresfor which there is generally a large discretionary component, suchas hysterectomies, tonsillectomies-adenoidectomies (LoGerfo, Efird,Diehr, & Richardson, 1979), or Caesarean sections (Price &Broomberg, 1990). A study of gynecologists who switched from FFSto a capitated contract found service volume decreased largely dueto a reduction in purely elective procedures (Ransom, McNeeley,Kruger, Doot, & Cotton, 1996). Similarly, when physicians paid bycapitation, who had an equity interest in utilization incentives paidto their group, were compared to those paid by traditional indem-nity insurance, it was found that capitated patients in the 25-to-44age group were admitted to hospital at a significantly lower rate forlow-acuity medical conditions (Josephson & Karcz, 1997).


In summary, there appears to be a consensus in the literature thatFFS practice, compared to salary or capitation methods of payment,encourages higher service volumes. Shifting an AHC to block fund-ing, therefore, might reasonably be expected to result in lower vol-umes of service provision. However, the literature is inconclusive asto whether the change in volume would represent the elimination ofexcess servicing or of essential medical care. Finally, existing stud-ies do not consider how a shift in clinical volumes would influenceresearch and educational activities.

EVALUATING THE AFP

Constraints

To what extent did the AFP realize its goals of enhanced care andincreased faculty time for educational and research activities? De-termining an answer to this complex question was hampered by anumber of significant constraints:

• Data: For many research questions either no pre-AFP datawere available, a problem that often persisted into the post-AFP period, or its retrieval was prohibitively costly and timeconsuming. This deficit is shared by most AHCs in Canadaand reflects an era in which management had relatively lit-tle external accountability for outcomes.

• Comparison site: No single AHC in Ontario or beyond hadthe characteristics to serve as an ideal comparison site. Dif-ferences in record coding, the presence of large non-AHCconsultant populations, and multiple hospital communitiesall created obstacles in selecting comparison sites.

• Provincial data: For some research findings, no provincialdata were readily available from which to create a context.For example, in Ontario there are no surgical waiting timedata against which changes documented at SEAMO couldbe compared.

• Confounders: The health care system has been in a state offlux over the lifetime of the AFP. Policies such as the cap-ping of physician billing or hospital restructuring signifi-cantly intruded on attempts to isolate a pure AFP effect.

• Time frame: Many significant effects of the AFP, particu-larly those focused on research and educational outcomes,were unlikely to become apparent during the lifespan of theAFP evaluation project.


Organizational Issues

The agreement with the Ministry of Health and Long-term Care(MOH) required, without stipulating further details, an evaluationof the AFP. Responsibility for the cost was debated during the firsttwo years of the program, a controversy that inhibited the organiza-tion of a coordinated evaluation effort. Initially it was deemed es-sential to involve participants in the evaluation process. A steeringcommittee was created, composed of interested faculty and mem-bers of the Faculty of Health Sciences administration. It createdfour working groups to focus on the clinical, research, educational,and organizational impacts of the AFP. Each group was composedof clinicians and a research associate seconded from the Queen’sHealth Policy Research Unit (QHP). An external scientific advisorycommittee was recruited to review findings.

Busy clinicians evinced limited time and capacity for launchingevaluation studies, and anticipated external funding for the evalua-tion did not materialize. After two and one half years, only five stud-ies had been initiated by the working groups to supplement severalpreliminary studies commissioned from QHP. In 1997 the workinggroups were collapsed into a faculty advisory group, and responsi-bility for completing the evaluation was given to QHP. Priority wasplaced on completing studies in progress, undertaking surveys tofollow up those previously done, and documenting key aspects ofthe clinical impact of the AFP not yet studied. By early 1999, withinthe constraints imposed by the practical and organizational issues,a broadly focused evaluation had been carried out. The followingsections of this article describe the way in which, deploying avail-able data and diverse techniques, researchers documented the im-pact of the AFP on the main components of the AHC’s tripartitemission — clinical care, education, and research — and on the or-ganization of the centre.

PROVISION OF CLINICAL SERVICE

Fee-for-service practice is generally acknowledged to encourage botha higher volume of practice and shorter waits than fixed remunera-tion schemes. Conversely, it is often argued that exemption fromvolume-based practice promotes a different mix and better qualityof clinical service. To determine whether the AFP altered clinicalservice provision within SEAMO, studies were done on the five broadareas discussed below.


Service Volume

Hospital discharges from SEAMO decreased 6.4% from 1992 to1996.Of that percentage, 2.5% occurred following the implementation ofthe AFP and almost 4% occurred prior to the AFP. Comparison withprovincial data establishes that changes within SEAMO were com-parable to changes elsewhere; however, the morbidity of hospital-ized patients, based on resource intensity weights as a very weakproxy for severity of illness, climbed steadily in SEAMO after 1994,in contrast to a more leisurely increase at other academic healthcentres in the province. This suggests AFP clinicians may haveadopted a higher acuity-of-illness threshold for admitting patientsto hospital (Parent, 1999a). In light of the substantial evidence ofthe high incidence of unnecessary hospitalizations in Canada (Work-ing Group on Health Services Utilization, 1994), this may representan improvement in appropriateness of care.

SEAMO clinicians provide the vast majority of specialist outpatientappointments to catchment area patients. Following a large drop inthe second quarter of 1992, a year after Ontario initiated caps onphysician billing, a relatively steady annual decline of –8.2% wasobserved in the total number of outpatients through 1996. The ratesof change in new appointments were neither uniform nor consistentacross the specialties and subspecialties within the hospitals. With-out comparable provincial data it is difficult to assess the degree towhich the AFP contributed to this decline, but it is clear that thetrend antedated the introduction of the new funding mechanism(Shaw, Jackson, & Farquhar, 1997).

Waiting for Service

Data on the amount of time patients spent waiting for their first elec-tive outpatient appointment with SEAMO specialists were gatheredfrom extant referring physician records for two years before and af-ter the implementation of the AFP. Although mean waiting timesincreased by a statistically significant average of 4.2 days for partici-pating clinicians, the changes varied widely across departments anddivisions. This suggested changes were most likely a reflection of lo-cal divisional or department circumstances, such as loss of physicianstaff, rather than an effect of the AFP. Moreover, the clinical impactof these relatively small changes for elective outpatients, that is,patients who are deemed by referring physicians to be non-urgent, islikely insignificant (Shaw, Jackson, & Farquhar, 1999).


Studies were also done of surgical waiting times at the two acutecare hospitals in SEAMO. Between 1994 and 1996 operating roomtime, surgical beds, and number of surgeons remained largely sta-ble. There was no effective change in mean waiting times in 3specialties, a statistically significant increase in 14 specialties, andin one instance a decrease. In aggregate these findings establishthat by 1996 patients waited an additional average 9.1 days for elec-tive surgery at one hospital and an additional 6.5 days at the other.It is difficult to determine whether the AFP contributed significantlyto the observed changes or whether the increased waits owed moreto restructuring in the hospital sector, as no provincial waiting timedata exist. None the less, it is evident that between 1992 and 1996there was an overall average 26% increase in the amount of timespent awaiting elective surgery (Shaw & Hall, 1998; Shaw & Shortt,1999).

Pattern of Clinical Care

Alterations in waiting times and patient volumes suggest that theremay have been change in the patterns of clinical care. To examinethe impact of the AFP on practice patterns, diagnosticgastrointestinal (GI) endoscopy was chosen as the focus for a pilotstudy. There was a decrease in the volume of upper GI endoscopies(–34%, p ≤ 0.01), which was similar to provincial trends, but thedecline in colonoscopies (–13%) was markedly less than the provin-cial decrease. Analysis of the changes findings suggested SEAMOendoscopists may have reduced volume by selectively eliminatingprocedures likely to provide diagnoses of debatable importance tothe patient’s management (Stanton & Parent, 1999).

To further explore the possibility that the AFP induced a change inpractice patterns and volumes, four sentinel surgical procedures,known to have a large discretionary component, were chosen forstudy in SEAMO and at the four other academic health centres inOntario that retained fee-for-service remuneration. Five years ofadministrative data, capturing all inpatient and outpatient recordsfrom 1992 through1996 at the five academic medical centres, wereanalyzed in this quasi-experimental, retrospective study using a lon-gitudinal, pre-post approach. Changes found at SEAMO after theimplementation of the AFP were similar to those found in the otherfour academic medical centres. No evidence was found that chang-ing the physician payment mechanism influenced the practice pat-terns of surgeons at SEAMO (Stanton & Shortt, 2000). Taken in


conjunction with the endoscopy study, these results suggest changesin practice patterns may be specific to individual specialties.

Population Health

Although there is significant debate as to the extent that the healthcare system has an impact on population health status, some mightargue that changes identified in service provision, patterns of clini-cal practice, and waiting times would be reflected in alterations tothe health of the catchment area population. Though deteriorationin health status would be difficult to ascribe simply to the initiationof the new funding formula, stable or improved health status wouldsuggest that service provision under the AFP had no detrimentalimpact. Population health status was measured by standardizedmortality ratios (SMRs), a measure sensitive to short-term influ-ences, for the whole population and for the populations under both65 and 75 years of age. A socio-economic risk index was developedto control for the confounding regional differences in socio-economicstatus. During the study period the health status of SEAMO’s popu-lation under 65 improved markedly, and a similar but non-signifi-cant improvement in health status was found when the SMRs werecalculated for those under 75 years (Stanton, 1999). Although therelationship between physician funding changes and populationhealth status is doubtless very indirect, these results document thatthere was no deterioration in the health of the catchment area popu-lation following implementation of the AFP.

Quality of Care

Although patterns of care may have shown little change, the impactof the AFP on quality of care was unknown. A before-and-after studyused the College of Family Physicians of Canada Practice Assess-ment Program (PASS), a consensus-derived tool for measuring proc-ess indicators, to assess the quality of care provided by familyphysicians in the office setting. The subjects were patients of nineparticipating physicians and three non-participant controls, whowere seen for an indicator condition both before and after the fundingchange. Eleven dimensions of care were assessed for eight indicatorconditions. There was no significant change in the quality-of-carescores of the participating physicians when the before period wascompared to the after, but non-participating physicians improvedon one dimension of care and on all dimensions taken together(Godwin & Shortt, 2000).


Views of Referring Physicians and Consultants

In the provision of health services, perceptions of change may be asimportant as more objective measures. To explore this area, all re-gional referring family physicians (n = 323) and the 208 academicconsultants at SEAMO were surveyed two and one-half years afterinitiation of the AFP. Four of ten referring physicians indicated theirreferrals to AFP consultants had decreased, and over a quarter saidtheir referrals to local non-AFP consultants or other secondary carecentres had increased. Almost one-quarter of regional doctors feltthe volume of work in their own practices had increased, and three-quarters stated they provided more care of complex patients and co-ordinated more community care. Eighty-six per cent of referringphysicians and 58% of consultants felt outpatient waiting times hadincreased. Though 39% of consultants stated their time spent onpatient care had increased, there was substantial agreement amongreferring physicians that consultants discharged patients from fol-low-up earlier and admitted patients to hospital less often. The ma-jority of referring physicians felt the funding change had a negativeeffect on local health care, and 40% of consultants agreed. However,the majority of physicians in both groups ranked the AFP last inimportance out of five trends in the health care system, includingbed closures and funding cuts (Godwin, Shortt, McIntosh, & Bolton,1999).

Based on the above findings, it is clear that although participantperceptions can serve as a useful guide to program areas requiringfurther research, they may not accurately reflect reality. For exam-ple, other studies previously referred to failed to document signifi-cantly increased outpatient waiting time and found preliminaryevidence of an improvement, rather than a decrease, in the qualityof care. In general, the directly attributable impact of the AFP onclinical activity was surprisingly scant.

PROVISION OF MEDICAL EDUCATION

Education is a second element in the academic medical centre’s mis-sion that, according to the advocates of the AFP, would be influ-enced by a change in the way in which clinical faculty are funded. Astudy comparing perceptions of the respective learning environmentby undergraduate medical students in Kingston and in Ottawa,where the FFS system continued, failed to detect significant differ-ences (Cosby, O’Connor, & Myers, 1998). Surveys of graduating stu-


dents in Kingston did, however, reveal increasingly positive viewsof the learning experience from 1994 to 1997.

Results on Part 1 of the Medical Council of Canada (MCC) qualify-ing exam, written upon completion of undergraduate training, dete-riorated in the three-year period from 1995 to 1997. However, it islikely that these changes primarily reflect alterations to admissionscriteria, which placed less emphasis on past academic achievement,and to adjustment to a new curriculum, which is know to cause aperiod of performance impairment while students adjust. Resultsfrom Part 2 of the MCC exam, written a year after graduation, aswell as from the certification examination of the College of FamilyPhysicians of Canada and from the specialty examinations of theRoyal College of Physicians and Surgeons of Canada, failed to showany significant change following the introduction of the AFP.

The health science centre also accomplishes its educational missionby providing continuing education programs for practising physi-cians. Between 1993–94 and 1997–98 a steadily increasing numberof programs were offered. This may reflect an enhanced willingnessof faculty to participate in such voluntary educational offerings, inthe absence of the previous financial disincentives to pursue suchactivity (Shortt, Shaw, & Keresztes, 1999).

Documenting an educational impact is confounded by a number offactors, including reductions in university funding, cutbacks in hos-pital teaching beds and outpatient clinics, and alterations in admis-sions policy and curriculum content. As well, no useful classroomlevel survey data are available to permit an assessment of changesin the quality of teaching. Finally, the four years under review maybe too short a period to permit identifiable change to emerge. How-ever, based on the available evidence, the introduction of the AFPappears to have had little impact on educational activities atSEAMO.

PURSUIT OF RESEARCH

The final component of the academic health centre’s mission is thepursuit of research. Though the AFP was introduced at a time whenpublic funding for medical research was declining in Canada(Bergeron, 1998), it is clear that it fostered an increase in what maybe considered research effort. Less compelling proof of research pro-ductivity than peer-reviewed publications or receipt of grants, ef-


forts such as non-peer-reviewed publications and presentations in-creased after 1993 and may represent a precursor stage to more for-mal research outputs. As well, episodes of collaborative research andthe number of grant submissions rose very substantially after 1994.None of these changes were related to the recruitment of new re-searchers. Collectively, these developments suggest increased fac-ulty research effort.

A more conventional measure of research activity is the number anddollar value of research grants actually received. Both of these meas-ures increased after 1994 at SEAMO, though the trend, at lesservelocity, antedated the introduction of the AFP. Similarly, thenumber of peer-reviewed publications by clinical faculty increasedtwo years after the AFP was initiated. This two-year lag is compat-ible with the amount of time required to write and publish researchthat was likely conducted before or immediately after AFP intro-duction. When overlapping five-year averages are used to report thenumber of published papers, the increase is seen to be part of a con-sistent long-term trend shared by other Ontario academic healthscience centres. Finally, the quality of these publications, based onthe widely accepted standard of citations received, is commensu-rate with the performance of peer institutions in Ontario (Shortt &Parent, 1999).

From this brief overview, it is clear that research effort increasedafter the AFP was introduced, and that previous grant and publica-tion trends continued. The new funding mechanism may have pro-vided a facilitating environment for research, but for several reasonsit is difficult to ascribe a more direct causal role to the AFP. First,the natural history of research is such that any effect of a new pro-gram should not be sought until a time elapse of at least two to fouryears. Second, the literature previously referred to identifies cer-tain organizational characteristics as necessary for the augmenta-tion of research activity. Without simultaneous alterations in thesefactors, it is unlikely that a change in funding mechanism alonecould induce a change in research activity.

ORGANIZATIONAL ASPECTS OF THE HEALTH SCIENCE CENTRE

The pursuit of an academic health centre’s mission occurs within anorganizational context and culture that may, in some important re-spects, be determined by the manner in which participating clini-cians are remunerated. The FFS format enjoys the power of tradition


and is viewed by physicians as conferring a high degree of clinicalautonomy. To succeed, an alternative funding approach must beperceived as providing at least offsetting advantages. Such percep-tions will derive not only from the new plan’s intrinsic merits, butalso the manner in which it is implemented and managed.

To gain an appreciation of this aspect of the AFP, in 1996 and 1998clinical department heads were interviewed and revealed a surpris-ing consensus towards the new funding plan (Anderson & Cosby,1997; Parent, 1999b). The critical issue identified by these clinicianswas how to meet ongoing clinical demand while at the same timeallowing enhanced time for research and educational activities. Theydid not believe that marginally necessary care, the component ofFFS practice that the new funding format sought to eliminate, com-prised a significant portion of faculty activity. Indeed, some clinicaldepartment heads believed that the emphasis on research placedclinical activities and quality of care in jeopardy. It was said thatclinical workloads were not necessarily reflected in the manner inwhich funding was allocated to departments or to individual clini-cians, and a fixed annual budget inhibited flexibility in matters suchas recruitment. Although they conceded the important advantagesof funding stability, equity and flexibility were seen as diminishedin comparison to traditional FFS funding.

In order to determine whether the views of department heads wereshared by their department members, clinical faculty were surveyedat two- and four-year intervals following implementation of the AFP(Cosby & Middleton, 1999; Keresztes, 1999). Clinical faculty indi-cated that they wanted to spend more time on research in 1996, butless time on it in 1998; conversely, they wanted to spend less timeon clinical practice in 1996, but wanted to spend more in 1998. Al-though faculty perceived a lack of incentives to respond to increaseddemand for clinical service, they also sensed disapproval by theirdepartments of any attempt to replace clinical work with academicactivity. These mixed messages may have created a degree of roleinertia, as there were no significant differences between periods inthe proportion of professional time that respondents reported spend-ing on research, education, clinical practice, and administrative du-ties. Yet in both the 1996 and 1998 surveys, more than one third ofthe respondents reported that the AFP had changed their researchand education activities, suggesting less a quantitative than an as-yet-undefined qualitative impact.


The views derived from these interviews and surveys may help toexplain why the new incentives associated with the AFP, such asresearch kudos or recognition for teaching excellence, have failed todisplace traditional concerns with service demands and financialrewards for both medical administrators and faculty members. Thisis an issue of managing organizational change, a crucial task in theintroduction of any new program, and a subject for ongoing research.

REMAINING QUESTIONS AND ISSUES

Studies to date have necessarily left unexplored relevant questionsand have identified a number of important evaluation issues. Thefollowing section suggests areas that should be addressed duringthe second iteration of the AFP, which commenced in early 2001:

• As Canadian medicare enters the era of accountability, itwill be increasingly necessary to collect relevant local andsystem-wide data. It is difficult to understand the signifi-cance of, or interpret, certain clinical service trends withinSEAMO without access to comparable provincial data. Of-ten, requests for such data cannot be gratified in a timelymanner or the data, as in the case of waiting times, simplydo not exist. Equally important, local data should be col-lected not just on clinical matters but also education andresearch activities, and should be made readily accessibleto evaluators.

• This article has only superficially addressed the criticalquestion of whether the AFP induced quality-enhancingchanges in practice patterns. It is essential to recognize boththe importance of and the significant obstacles to doing thistype of research. If the AFP cannot be shown to enhancethe provision of care, one of its key goals is called into ques-tion. Yet clinicians are very sensitive to and often resistantto any external attempt to assess the quality of their care.A resolution to this potential conflict will demand an or-ganization-wide commitment from SEAMO.

• Another critical question related to clinical care is the na-ture of the relationship between service provision and popu-lation health status. There is a large and often contentiousliterature that examines this complex connection, usuallyat a national or provincial level. The challenge for futureevaluations will be to design and validate indicators capa-ble not only of accurately detecting small changes in regional


health status, but also of linking such outcomes to the pro-vision of specific services.

• To fully appreciate the impact of alternative payment onscholarly activities will demand the use of individual-leveldata. In the case of educational activities, the impact is bestfollowed longitudinally at the level of the individual instruc-tor through well-designed course evaluations. At present,such data are only sporadically available for some faculty,and do not necessarily address the relevant pedagogicalquestions. Similarly, for research activities, current data-bases do not capture subtle yet critical processes such asmentoring of new faculty by senior researchers, enhancedinterdepartmental collaborations, or the production of non-peer-reviewed research efforts. Only by collecting the finedetails of individual faculty member’s academic activity overtime will the long-term effects of the AFP on scholarshipbecome apparent.

• Any large-scale program will demand complex organiza-tional change in order to succeed. How alternative paymentalters the profession incentives of physicians, how it mayinfluence recruitment patterns, and in what manner newforms of accountability are balanced with traditional no-tions of professional autonomy are examples of the types ofquestions that must be asked about the SEAMO experience.Answers to these important questions are unlikely to besupplied by quantitative methodology, an approach famil-iar to and trusted by physicians. Instead, the questions canonly be answered by relying on the sophisticated qualita-tive methods developed in the social sciences.

The five points discussed above are by no means the only areas re-quiring future attention. Rather, they are cited here as examples ofthemes the present evaluation has identified as demanding furtherelucidation if the impact of alternative funding on the activities ofthe academic health science centre is to be fully understood.

CONCLUSION

The AFP is a form of fixed-payment practice that is intended to en-hance the quality and productivity of a career in academic medicineby liberating it from the necessity of fee generation. Surprisingly, ithas had relatively little impact on the Ontario academic health cen-tre at which it was introduced. Surgical waiting times increased and


patient volumes decreased, though in neither case is it possible toestablish the degree to which the new plan was responsible. Thereis some evidence that marginally necessary care was reduced andquality of care remained stable, and reasonable evidence that theplan exerted no detrimental effect on the health of the regional popu-lation. It had, beyond increased continuing education programs, noimpact on undergraduate or postgraduate education at the centre,and a slight positive effect on research activities. Despite the plan’slargely innocuous results, regional physicians had a negative viewof its effects, and participating faculty and administrators were frus-trated by issues of financial equity and clinical workload.

It is possible that the AFP’s relative lack of impact, particularly inthe case of education or research, is a function of insufficient timeelapse since its initiation. As well, the lack of discernable impactmay reflect the constraints imposed on evaluation by a lack of ap-propriate data by which to detect subtle changes in the AFP envi-ronment. In this respect, however, the evaluation was not dissimilarto much health services research. This multidisciplinary field ofscholarly investigation copes by using a variety of available datasets, each often imperfect, and various methodologies to analyze thedata. Trends consistently found using multiple sources and meth-ods acquire a validity that transcends the limitations otherwise as-sociated with inadequate data. In the case of the AFP, no informationsource revealed an exception to the general trend of limited impact.

It might be argued that the new funding method was not signifi-cantly different from the preceding ceiling system to result in markedchanges. Under that system, individual clinicians annually returnedto their departments fees in excess of a negotiated amount. This, ithas been argued, provided a disincentive to maximize FFS billingin a manner similar to the AFP. However, there are important dif-ferences between the ceiling system and the fixed remuneration ofthe AFP. First, under the ceiling system maximizing billing in or-der to contribute large amounts to one’s department was seen as away of gaining influence within the department. Second, large“overages” could be used by clinicians to argue that their individualceilings had been set at an inappropriately low level. Finally, therewas no limit to the amount of billing revenue in excess of the ceilinglimit that could be retained by clinicians if spent on expenses ofpractice as accepted by tax authorities. All of these factors, despitethe ceiling arrangement, provided incentives to generate high FFSbillings; none of them were present after the introduction of the AFP.


In a broader sense, however, the inability of a series of evaluationstudies to document definitive impact does not suggest the AFP hasbeen without very significant effect. On the contrary, in two impor-tant ways it has been a marked success as an alternative to FFS pay-ment of academic clinicians. First, it has introduced predictability, acharacteristic of immense importance to government in a publiclyfunded health care system, into the increasingly volatile world ofAHC financing. Government is aware in advance of the funds re-quired to pay clinicians at the centre for a five-year period, knowledgeit could not have had under the previous FFS system. At the sametime, the academic health centre no longer relies on variable clinicalearnings to subsidize a portion of academic activities, but rather re-ceives annual block funding. Second, in marked contrast to the FFSsector, the AFP assumes an accountability on the part of the AHC tothe provincial funder. Previously, the medical faculty was a loose af-filiation of individual entrepreneurs, each of whom submitted billingsto the provincial medical insurance plan, but who could not be heldto fiscal account as a collectivity for service provision. In the finalanalysis, the AFP is a still-evolving but very significant innovationin academic health centre funding, and a harbinger of further changeto come in the fiscal environment of Canadian academic medicine(Haslam & Walker, 1993; Hollenberg, 1996; Stoddard & Barer, 1992).The initiation of predictability and accountability in the funding ofmedical education must be regarded as its major achievement to date.

ACKNOWLEDGEMENTS

The work of researchers at Queen’s Health Policy Research Unitand of the Queen’s clinicians upon whose studies this article reliesis gratefully acknowledged. Support for various components of theevaluation was provided by the Southeastern Ontario AcademicMedical Organization, core funding to Queen’s Health Policy Re-search Unit from the Ontario Ministry of Health and Long-term Care,and a research grant from the Canadian Health Services ResearchFoundation.

REFERENCES

A complete listing of all evaluation studies is available at <http://chspr.queensu.ca>. Any of the unpublished studies referred to, ex-cluding those currently under publication review, may be obtainedfrom the Centre for Health Services and Policy Research (formerlyQueens Health Policy) at Queen’s University by contacting<[email protected]>.


Allison, P.D., & Stewart, J.A. (1974). Productivity differences among sci-entists: Evidence for accumulative advantage. American Sociologi-cal Review, 39, 596–606.

Anderson, M.J., & Cosby, J.L. (1997). Changing management to adapt to anew compensation culture: Implementing an alternative funding planat Queen’s Faculty of Medicine. Kingston: Queen’s Health PolicyResearch Unit.

Barnett, R.C., Carr, P., Boisnier, A.D., Friedman, R.H., Moskowitz, M.A.,& Szalacha, L. (1998). Relationships of gender and career motiva-tion to medical faculty members’ production of academic publica-tions. Academic Medicine, 73(2),180–186.

Beaty, H.N., Babbott, D., Higgins, E.J., Joly, P., & Levey, G.S. (1986). Re-search activities of faculty in academic departments of medicine.Annals of Internal Medicine, 104, 90–97.

Bergeron, M.G. (1998). “Scientists Inc.” — Illusion or disillusion? ClinicalInvestigative Medicine, 21(1), 39–44.

Bland, C.J., & Schmitz, C.C. (1986). Characteristics of the successful re-searcher and implications for faculty development. Journal of Medi-cal Education, 61, 22–31.

Cameron, S.W., & Blackburn, R.T. (1981). Sponsorship and academic ca-reer success. Journal of Higher Education, 52, 369–377.

Carr, P., Friedman, R.H., Moskowitz, M.A., Kazis, L.E., & Weed, H.G.(1992). Research, academic rank, and compensation of women andmen faculty in academic general internal medicine. Journal of Gen-eral Internal Medicine, 7(4), 418–423.

Cosby, J.L., & Middleton, M.J. (1999). The perceptions of the alternativefunding plan. Kingston: Queen’s Health Policy Research Unit.

Cosby, J.L., O’Connor, M., & Myers, K. (1998). Clinical faculty paymentsystems and learning environments at the Queen’s University andUniversity of Ottawa medical schools. Kingston: Queen’s HealthPolicy Research Unit.

Crane, D. (1965). Scientists at major and minor universities: A study ofproductivity and recognition. American Sociological Review, 30, 699–714.


Cromwell, J., Dayhoff, D.A., & Thoumaian, A.H. (1997). Cost savings andphysician responses to global bundled payments for medicare heartbypass surgery. Health Care Financing Review, 19(1), 41–57.

Dowdall, C., & Ramchandar, L. (1999). The application of physician bill-ing thresholds in Ontario and across Canada. Ontario Medical Re-view, 66(8), 17–21.

el-Guebaly, N., & Atkinson, M. (1996). Research training and productivityamong faculty: The Canadian Association of Professors of Psychia-try and the Canadian Psychiatric Association Survey. CanadianJournal of Psychiatry, 41(3), 144–149.

Fishman, L.E., & Bentley, J.D. (1997). The evolution of support for safety-net hospitals. Health Affairs (Millwood),16(4), 30–63.

Godwin, M., & Shortt, S. (2000). The effect of payment method on quality ofpatient care in an academic family practice setting. Kingston: Queen’sHealth Policy Research Unit.

Godwin, M., Shortt, S.E.D., McIntosh, L., & Bolton, C. (1999). Perceptionsby referring and consultant physicians of the effects of an alterna-tive funding plan at an Academic Health Science Centre. CanadianMedical Association Journal, 160, 1710–1714.

Government of Ontario. Management Board Secretariat. (1997). Expendi-ture estimate, 1997–1998. Vol. 1. Toronto: Queen’s Printer.

Government of Ontario. Management Board Secretariat. (1998). Expendi-ture estimate, 1998–1999. Vol. 1. Toronto: Queen’s Printer.

Gruber, J., Kim, J., & Mayzlin, D. (1999). Physician fees and procedureintensity: The case of cesarean delivery. Journal of Health Econom-ics, 18, 473–490.

Harrington, M.S., & Levine, D.U. (1986). Relationship between facultycharacteristics and research productivity. Journal of Dental Educa-tion, 50(9), 518–525.

Haslam, R.H.A., & Walker, N.E. (1993). Alternative funding plans: Is therea place in academic medicine? Canadian Medical Association Jour-nal, 148(7), 219–222.


Hemenway, D., Killen, A., Cashman, S.B., Parks, C.L., & Bicknell, W.J.(1990). Physician’s responses to financial incentives — Evidence froma for-profit ambulatory care center. New England Journal of Medi-cine, 322, 1059–1063.

Hickson, G.B., Altemeier, W.A., & Perrin, J.M. (1987). Physician reimburse-ment by salary or fee-for-service: Effect on physician practicebehavior. Pediatrics, 80, 344–350.

Hohlen, M.M., Manheim, L.M., Fleming, G.V., Davidson, S.M., Yudkowsky,B.K., Werner, S.M., & Wheatley, G.M. (1990). Access to office-basedphysicians under capitation reimbursement and medicaid case man-agement — Findings from the Children’s medicaid program. Medi-cal Care, 28, 59–68.

Hollenberg, C.H. (1996). The effect of health care reform on academic medi-cine in Canada. Canadian Medical Association Journal, 154(10),1483–1489.

Iglehard, J.K. (1999). Support for academic medical centers: Revisiting the1997 balanced budget act. New England Journal of Medicine, 341,299–303.

Jolin, L.D., Jolly, P., Krakower, J.Y., & Beran, R. (1992). US medical schoolfinances. Journal of the American Medical Association, 268, 1149–1155.

Jones, R.F., & Sanderson, S.C. (1996). Clinical revenues used to supportthe academic mission of medical schools, 1992–93. Academic Medi-cine, 71(3), 300–307.

Josephson, G.W., & Karcz, A. (1997). The impact of physician economicincentives on admission rates of patients with ambulatory sensitiveconditions: An analysis comparing two managed care structures andindemnity insurance. American Journal of Managed Care, 3(1), 49–56.

Keeler, E.B., & Brodie, M. (1993). Economic incentives in the choice be-tween vaginal delivery and cesarean section. Milbank Quarterly,71(3), 365–404.

Keresztes, C. (1999). Faculty perceptions of the alternative funding plan.Kingston: Queen’s Health Policy Research Unit.


Krumland, R.B., Will, E.E., & Gorry, G.A. (1979). Scientific publicationsof a medical school faculty. Journal of Medical Education, 54, 876–883.

Kuttner, R. (1999). Managed care and medical education. New EnglandJournal of Medicine, 341, 1092–1095.

Levey, G.S., Sherman, C.R., Gentile, N.O., Hough, L.J., Dial, T.H., & Jolly,P. (1988). Postdoctoral research training of full-time faculty in aca-demic departments of medicine. Annals of Internal Medicine, 109,414–418.

LoGerfo, J.P., Efird, R.A., Diehr, P.K., & Richardson, W.C. (1979). Ratesof surgical care in prepaid group practices and the independent set-ting: What are the reasons for the differences? Medical Care, 17, 1–7.

Long, J.S. (1978). Productivity and academic position in the scientific ca-reer. American Sociological Review, 43, 889–908.

Long, J.S., & McGinnis, R. (1981). Organizational context and scientificproductivity. American Sociological Review, 46, 422–442.

Ludmerer, K.M. (1985). Learning to heal: The development of Americanmedical education. New York: Basic Books.

Luke, R.D., & Thomson, M.A. (1980). Utilization of within-hospital serv-ices: A study of the effects of two forms of group practice. MedicalCare, 18(2), 219–227.

Manu, P., Landaw, S.A., Williams, W.J., & Schwartz, S.E. (1985). Analy-sis of publication output of internal medicine faculty members. Jour-nal of Medical Education, 60, 860–864.

Mills, O.F., Zyzanski, S.J., & Flocke, S. (1995). Factors associated withresearch productivity in family practice residencies. Family Medi-cine, 27(3), 188–193.

Nguyen, N.X., & Derrick, F.W. (1997). Physician behavioral response to aMedicare price reduction. HSR: Health Services Research, 32, 283–298.

Parent, K. (1999a). Acute care utilization trends. Kingston: Queen’s HealthPolicy Research Unit.


Parent, K. (1999b). Physicians’ perceptions of organizational change withinan alternative funding environment. Kingston: Queen’s Health PolicyResearch Unit.

Pearse, W.H., Peeples, E.H., Flora, R.E., & Freeman, R. (1976). Medicalschool faculty productivity. Medical Education, 51, 201–202.

Pincus, H.A., Haviland, M.G., Dial, T.H., & Hendryx, M.S. (1995). The re-lationship of postdoctoral research training to current research ac-tivities of faculty in academic departments of psychiatry. AmericanJournal of Psychiatry, 152(4), 596–601.

Price, M.R., & Broomberg, J. (1990). The impact of the fee for service reim-bursement system on the utilisation of health services, part 3: Acomparison of caesarean section rates in white nulliparous womenin the private and public sectors. South African Medical Journal,78(4), 136–138.

Queen’s Health Policy Research Unit. (1996). Alternative funding planevaluation: Documents model. Kingston: Author.

Ransom, S.B., McNeeley, S.G., Kruger, M.L., Doot, G., & Cotton, D.B.(1996). The effect of capitated and fee for service remuneration onphysician decision making in gynecology. Obstetrics and Gynecology,87, 707–710.

Rothstein, W.G. (1987). American medical schools and the practice of medi-cine. New York: Oxford University Press.

Shaw, R.A., & Hall, S.F. (1998). Waiting for surgery at the Hotel Dieu Hos-pital, 1994 through 1996. Kingston: Queen’s Health Policy ResearchUnit.

Shaw, R., Jackson, A., & Farquhar, D. (1997). The waiting time study: Theimpact of the Alternative Funding Plan on waiting times for firstappointments with specialists at SEAMO outpatient clinics. King-ston: Queen’s Health Policy Research Unit.

Shaw, R., Jackson, A., & Farquhar, D. (1999). Waiting-times for first ap-pointments with clinical specialists at a Canadian academic healthsciences centre before and after the implementation of an alterna-tive funding plan. Annals of the Royal College of Physicians andSurgeons, 32(6), 331–336.


Shaw, R.A., & Shortt, S.E.D. (1999). Waiting for surgery at the KingstonGeneral Hospital: A pre-AFP / post-AFP examination of waiting timesfor elective surgery. Kingston: Queen’s Health Policy Research Unit.

Shortt, S.E.D., & Parent, K. (1999). The alternative funding plan and re-search. Kingston: Queen’s Health Policy Research Unit.

Shortt, S.E.D., Shaw, R., & Keresztes, C. (1999). The alternative fundingplan and medical education. Kingston: Queen’s Health Policy Re-search Unit.

Stanton, S. (1999). Population health and health care provision: An over-view of trends in SEOHSC. Kingston: Queen’s Health Policy ResearchUnit.

Stanton, S., & Parent, K. (1999). The impact of change in physician remu-neration on clinical practice: A feasibility study focusing on upperGI endoscopy. Research in Healthcare Financial Management, 5(1),41–56.

Stanton, S., & Shortt, S. (2000). The influence of payment method on pat-terns of physician practice: Experience at a Canadian academic healthscience centre. Kingston: Queen’s Health Policy Research Unit.

Stoddard, G.L., & Barer, M.L. (1992). Toward integrated medical resourcepolicies for Canada: 5. The roles and funding of academic medicalcentres. Canadian Medical Association Journal, 146, 1919–1924.

Thorne, S. (1997). Medical schools seeking new ways to cope with fundingcutbacks. Canadian Medical Association Journal, 156, 1611–1613.

Vardan, S., Smulyan, H., Mookherjee, S., & Mehrotra, K.G. (1990). Fac-tors encouraging research productivity in a division of general in-ternal medicine. Academic Medicine, 65, 772–774.

Wilson, S.E., & Longmike, W.P. (1978). Does method of surgeon paymentaffect surgical care ? Journal of Surgical Research, 24, 457–468.

Working Group on Health Services Utilization. (1994). When less is better:Using Canada’s hospitals efficiently. Paper presented at the Con-ference of Federal/Provincial/Territorial Deputy Ministers of Health,Ottawa.


Yip, W.C. (1998). Physician response to Medicare fee reductions: Changesin the volume of coronary artery bypass (CABG) surgeries in theMedicare and private sectors. Journal of Health Economics, 17, 675–699.

PAYING DOCTORS: IMPACT OF A CHANGE IN … · puis cette époque, les facultés de médecine au...

Documents

Transcript of PAYING DOCTORS: IMPACT OF A CHANGE IN … · puis cette époque, les facultés de médecine au...