UBC Faculty Research and Publications

The effects of British Columbia's physician payment initiatives : making sense of the dollars Evans, Robert G., 1942-; Barer, Morris Lionel, 1951-; Pascali, Marina V. Mar 31, 1995

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


52383-Evans_R_et_al_Physician_payment_initiative.pdf [ 14.53MB ]
JSON: 52383-1.0048515.json
JSON-LD: 52383-1.0048515-ld.json
RDF/XML (Pretty): 52383-1.0048515-rdf.xml
RDF/JSON: 52383-1.0048515-rdf.json
Turtle: 52383-1.0048515-turtle.txt
N-Triples: 52383-1.0048515-rdf-ntriples.txt
Original Record: 52383-1.0048515-source.json
Full Text

Full Text

THE EFFECTS OF BRITISH COLUMBIA'SPHYSICIAN PAYMENT INITIATIVES:MAKING SENSE OF THE DOLLARSRobert G. EvansMarina V. PascaliMorris L. BarerHPRU 95:40 March 1995The Effects of British Columbia'sPhysician Payment Initiatives:Making Sense of the DollarsRobert G. EvansMarina V. PascaliMorris L. BarerMarch, 1995Robert Evans is Professor, Department of Economics, and Faculty, Centre for Health Services andPolicy Research, atthe University of British Columbia. He is Fellow and Head of the Canadian Institute for Advanced Research Programin Population Health, and a Canadian National Health Scientist Marina Pascali is a Ph.D. Candidate in HealthServices Research, University of California at Berkeley, and a Canadian National Health Ph.D. Fellow. Morris Bareris Director, Centre for Health Services and Policy Research, and Professor, Department of Health Care andEpidemiology, at the University of British Columbia. He is an Associate of the Population Health Program, CanadianInstitute for Advanced Research.The Centrefor Health Services and PolicyResearch was established by the BoardofGovernors of theUniversity of BritishColumbia in December 1990. It was officially openedin July 1991. The Centre's primary objective is to co-ordinate. facilitate. andundertakemultidisciplinary research in the areasof healthpolicy. healthservices research, populationhealth, and healthhuman resources. It bringstogether researchers in a variety of disciplineswhoarecommitted to a multidisciplinary approach to research and to promoting wide dissem­inationanddiscussion of research results. in theseareas. The Centreaims to contribute tothe trnprovement of population healthby beingresponsive to the research needsof thoseresp­onsiblefor healthpolicy. To thisend. it provides a research resource for graduate students;develops and facilitates access to healthandhealthcaredatabases; sponsors seminars. workshops.conferences andpolicyconsultations; anddistributes Discussion papers. Research Reports andpublication reprints resulting from the research programs of Centrefaculty.The Centre's Health PolicyResearch UnitDiscussion Paper series provides a vehicle for thecirculation of preliminary (pre-publication) workof CentreFaculty and associates. It is intendedtopromote discussion and to elicitcomments and suggestions thatmightbe incorporated withinthework prior to publication. WhiletheCentre prints anddistributes these papers for thispurpose,theviews in thepapers are those of the author(s).A complete listof available Health Policy Research UnitDiscussion Papers andReprints. alongwith an address to which requests for copies should be sent. appears at thebackof eachpaper.The Changing Economic Climate: A Short Historical PerspectiveThe year 1981 appears, in retrospect, to have been something of a turning point in theevolution of the Canadian health care system. It was not obvious at the time -- the year didnot, like 1961 or 1971, mark the completion of a clearly defmed stage of public coverage or,like 1978, a major shift in funding policy.' In fact, nothing remarkable at all happenedwithin the health care system itself. But the beginning of the 1980s saw a major deteriorationin general economic performance in Canada. This in turn appears to have upset the balancewhich was established during the 1970s, between the expansion of the health care system andthat of the rest of the economy supporting itThis change is demonstrated very clearly in Figure 1, which plots the level of GDPper capita, in constant dollars, from 1960 to 1993.2 A simple logarithmic trend, representinga constant annual growth rate, is fitted to the values for the years from 1960 to 1980 and thenprojected through to 1993. The annual values fit this trend line remarkably well for the firsttwo decades, all lying within five percent of it But the major recession of the early 1980sdropped the Canadian economy sharply away from this twenty-year trend, and mostimportantly, the "recovery" of the mid-1980s did not bring it back. As shown in Figure 1, byabout 1983 real GDP per capita had resumed a rate of growth approximately equal to theprevious trend, but the income "lost" during the recession was not made up. Unemploymentand public sector deficits correspondingly remaining high by historic standards.[Figure 1 about here]1. Canada's federal-provincial public insurance systemprovides for federal financial contributions to provincesoperating programs that meet national standards. The provinceseach joined this system at different times, depending upon whentheir conforming programs were introduced. The hospitalinsurance system became national when the last province joined in1961, and medical insurance became complete in 1971. The year1978 saw a major change in the fiscal arrangements between thefederal governments and the provinces, as the federal governmentmoved from formula-based sharing of program costs to acombination of block grants and increased provincial access tothe revenues from income and other taxes.2. Data sources for all figures, and references in the text,are detailed in the Data Apppendix.2The next and even more severe recession of the late 1980slearly 1990s then droppedreal GOP per capita even farther below the 1960-1980 trend, with the current "recovery" sofar merely ending the decline. By the early 1990s, income and output in the Canadianeconomy were about 30% below where they would have been (in real terms), if the growthtrend of the 1960s and 1970s could have been maintained?The significance of this sustained decline for the health care system is shown in Figure2, which displays the familiar ratio of health care spending to GOP. But Figure 2 presentsonly the hospital and physicians' services components of health care. These are the servicescovered by the Canadian federal-provincial public insurance plans; moreover data on thesecomponents is available back to the late 1940s.4[Figure 2 about here]Figure 2 also includes a hypothetical line, from 1981 on, representing what the ratiowould have been, if the economy had continued to grow along the trend path shown in FigureI, while hospital and medical spending took on their actual values. If the general economyhad continued to grow at 1960-1980 rates, then Figure 2 shows that the share of hospital andmedical spending in GOP would have remained constant or drifted downwards slightly, as itdid in the 1970s. From this perspective it is clear that the "cost crisis" of the 1980s was aresult not of some change in trends within the medicare system itself, but of the deterioration3. The average growth rate during the period from 1960 to1980 was in fact somewhat higher than for the post-war period asa whole. 1960-1980 is of particular significance for the healthcare system, however, in that the federal Hospital Insurance andDiagnostic Services Act of 1957 led to the establishment ofuniversal public hospital insurance coverage in all provinces by1961, and set the pattern for the coverage of physicians'services later that decade. The public system has thus alwaysfunctioned in the post-1960 economic environment.A logarithmic trend fitted from 1947 to 1980 and projectedto the present shows a similar, though less dramatic, patternafter 1980. Even using this longer run perspective, however, by1993 actual constant dollar GDP per capita had fallen more thantwenty percent below trend.4 Estimates of total expenditures on health care in Canadahave only been reported in public documents since 1960.3in the performance of the general economy, which resulted in large increases in the share ofnational income -- and of public budgets -- used to finance health care (Evans, 1993).The long-term data in Figure 2 also show clearly the break in the pattern of escalationin 1971. Variants of this figure, usually displaying total health care expenditures over GDP,have frequently been employed in discussions of the performance of the Canadian health caresystem, particularly in international comparisons (see, e.g. Culyer (1988); Ham, Robinson andBenzeval (1990». Figure 2 in conjunction with Figure 1, however, makes the point that themarked shift from a steadily escalating ratio during the decade prior to 1971 to an essentiallystable ratio during the decade afterwards takes place in the middle of a twenty year period ofsteady general economic growth. It is the numerator, not the denominator, of the ratio thatchanges its behaviour with the establishment of universal public insurance coverage.'Nevertheless, the denominators do come to matter later. The Canadian economy hasexperienced two severe recessions since 1981, linked by what one might call, relative to theexperience of previous decades, "growth without recovery". But the health care system waslargely protected against the impact of the first recession, and so continued to expand at muchthe same rate.There was much rhetoric about cost control, and pained responses from health care5. This observation has a bearing on controversies in theUnited States over the potentiai for healtQ care cost control insingle-source funding systems. In 1971, both Canada and theUnited States were spending about 7 1/2% of their GDP on healthcare; by 1993 the corresponding figures were 10.1% for Canada and14.2% for the United States. Opponents of universal publicinsurance in the United States have argued, however, that thisdivergence is the result, not of superior cost-controlperformance in Canada, but of more rapid general economic growth-- a difference in denominators, rather than numerators. Thedata here are not consistent with that argument.4providers, and certainly increased pressures were felt in particular sectors. But whatever painwas felt in the health sector during this decade paled in comparison with the labour painsbrought on by contractions elsewhere in the economy. The share of GOP devoted to healthcare rose sharply in the recession years, and then plateaued, but did not fall, when therecession ended. By the end of the 1980s total health care spending in Canada accounted forabout 10% of GOP, compared with about 7.5% at the beginning.Evidently the patterns of behaviour and the expectations of continuous expansionwhich providers of health care had developed during the previous two decades were carriedforward into the new, low- or no-growth environment effectively unchanged. Unless, andagainst all expectation, the rate of general economic growth in Canada accelerates back topre-1981 levels, the only choice appears to be between further escalation in the share ofnational income going to support the health care system, or a major change in the internalgrowth dynamics of that system.This elementary arithmetic, summarized by the British Columbia Royal Commissionon Health Care and Costs as "Business as usual is not sustainable" (British Columbia, 1991),is now generally understood among those involved in studying or making health care policyin Canada. It underlies many of the recent provincial initiatives in health care policy, amongwhose key objectives is the reduction of costs. But cost reductions translate in the service­intensive health care industry into job or income reductions. So one finds in response thequite predictable political manoeuvring intended either to keep the share of GOP going tohealth care on the increase, or to ensure that the burden of cuts (whether reduced incomes orreduced access to services) falls on someone else," But this broad understanding is a featureof the 1990s, not the 1980s, and in any case it is still not clear how far it has become part of.the public consciousness.6 These efforts underlie the renewed interest in variousforms of private delivery and private finance -- more money forproviders, and preferred access (preferably with public subsidy)for patients with money.5Roughly speaking, one might say that during the first half of the 1980s the generalnational response to what B.C.'s government labelled the "New Reality", was to hope that itwould go away. The recession was unusually severe, no doubt, but presumably, like allrecessions, it would end. There was no point in taking on the political dangers of majorreform of the health care system -- extremely popular the way it was, and fundamentallysound -- in response to what was, after all, a cyclical problem.By the mid-1980s, however, the celebration of recovery was being somewhat mutedby the realization that the usual problems of recession - unemployment, welfare and,especially, large public sector deficits, were persisting.' In the second half of the 1980s,provincial governments began to appoint Royal Commissions or equivalent independentbodies to review the functioning of their health care systems. Each, in its own way, came outwith some version of "Business as usual is not sustainable". Now, in the 1990s, muchtougher action is being taken by provincial governments, belatedly, to contain their healthcare systems,"The B.C. Response: Containing Physician Costs Any Way You CanThere were, however, some exceptions to this pattern of delayed response. Inparticular, the province of British Columbia moved quite forcefully during the 1980s tocontain the escalation of costs of physicians' services, with a broad range of different policies.Per capita spending on physicians' services in B.C. was by far the highest in the country in1981, 34% above the national average, and 25% above that in Ontario, the second-highestprovince. These expenditures corresponded to both the highest ratio of physicians per capita,7 There were other reasons for the latter, rooted inchanges in the tax system, but a real recovery to the historicgrowth path would have helped a great deal.8 Indeed, the "toughness" of today's policies is in part.brought on by the (in retrospect) inadequate responses earlier inthe 1980s .613% above the national average, and exceptionally high fees, more than 40% above average,"Payments to physicians were thus an obvious target for cost containment.In addition, the recession of the early 1980s was particularly severe and prolonged inBritish Columbia. GDP per capita in real terms dropped 6.9% in B.C. between 1981 and1982, compared with 4.4% in the country as a whole. Over the next three years, while theCanadian average increased 11.6%, B.C. real income rose only 1.8%. The province wentfrom its historic position as a relatively wealthy province -- average incomes 10% or moreabove the all-Canada level -- to below average. This deterioration in the general economicenvironment generated greater pressures on health care sector budgets than were felt in anumber of other provinces.The overall impact of the various policies introduced by the government of BritishColumbia to limit the escalation of payments to physicians is reflected in Figure 3. The on­coming recession was evident by the summer of 1982, and the provincial government hadbegun efforts to re-open previous agreements with the B.C.M.A. But the result of this'negotiation', a temporary "fee give-back" which took effect in the latter part of fiscal1982/83, was relatively small, and left the base fee schedule untouched (Barer et al., 1987).More substantial controls began to bite in 1983/84. As Figure 3 shows, per capitaoutlays by the British Columbia Medical Services Plan for medical services (fees and salaryor sessional payments, but excluding supplementary benefits paid to non-medicalpractitioners) were exactly the same in fiscal 1992/3 as they were nine years earlier (afteradjusting for increases in the general price level, as measured by the Vancouver city9. Fee schedule comparisons published by Health and WelfareCanada for July 1, 1985 show B.C. fees 32.29% above the nationalaverage. But average fees increased less rapidly in B.C. than inCanada as a whole between 1981 and 1985; applying this differenceto the 1985 differential yields a B.C. advantage of 41% in 1981.7Consumer Price Index)." By contrast the increase in real per capita outlays for Canada asa whole (including B.C.) was 23.9%.[Figure 3 about here]In essence, the government of British Columbia adopted a three-pronged approach tolimiting the growth of payments to physicians. Logically, one can factor per capita paymentsto physicians within a fee-for-service environment into three components:(1) payments per service, or fee levels,(2) services per physician, and(3) physicians per capita.Policies were developed to address each of these components -- to hold down fees, to limitphysicians' capacity to increase their volume of billings, and to reduce the rate of increase inthe number of physicians per capita in the province.The detailed description and history of these various policies, as they have made theirway through the negotiating, legislative, and judicial processes, is beyond the scope of thispaper (and can be found elsewhere)." A larger research project, currently underway, willboth document the introduction, modification, and removal of the various policies andcombinations, and attempt to disentangle their individual impacts. That effort draws upon thecomputerized payment tapes of the Ministry of Health, which record the type and date of eachreimbursable service provided, the amount reimbursed, and who provided and received theservice.10 The Vancouver Census Metropolitan Area includes almostone-half the province's population. Payments to physicians aretaken from the provincial Ministry of Health Annual Reports, asdescribed in the Data Appendix, with one exception. Aretroactive payment of $42 million was made in 1990/91, withrespect to billings in 1989/90. In Figure 3 and subsequentfigures, this sum has been added to the reported outlays for1989/90 and subtracted from those for 1990/91.11 See, for example, Barer (1988), Lomas et ale (1989), andLomas et ale (1992).8But these detailed billing records are, for obvious reasons, neither widely available noreasily processable -- a single year of data is now made up of nearly forty-five million records.In this paper, therefore, we present a preliminary analysis based upon data available inpublished records. The conclusions, based on observations drawn from information in thepublic domain, will in the larger project be checked and significantly extended using thecomputerized data set12We begin that process in this paper by comparing the public data on physician feeincreases in British Columbia with an alternative index developed as part of the preliminaryanalysis of the more detailed payment records. As explained below, we regard this new indexas a "gold standard", avoiding certain conceptual problems associated with the publishedindex. It turns out, however, that at the aggregate level and for the purposes of this paperthere is very little difference between the published data on fee increases, and the indexwhich we have developed from the actual payment data. (The latter, however, is essential fordisaggregated analysis, whether by type of service, date, or region.)Begin at the Beginning: Holding the Line on Physician FeesReturning to the three components of expenditures per capita listed above, the mostobvious place to look first for the cost impact of changes in policy is in the evolution of feelevels. If physicians' fees escalate faster than the general price level, then ceteris paribus onewould expect to fmd per capita outlays rising, and conversely. Thus, in Figure 4 we presentthe index of physicians' fees in British Columbia compiled by Health Canada (1981 = 100.0),divided by the all-items Consumer Price Index for Vancouver (1981 = 100.0), for the fiscalyears 1969nO to 1992/93.[Figure 4 about here]12. In particular the published data on payments tophysicians are based on the date of payment rather than ofservice; reported year to year movements are thus affected bychanges in payment lags. One can compile expenditure seriesbased on date of service, from which to calculate more accuratemeasures of average physician workloads and patient utilization,only from the individual payment records. These more detailedrecords will be used in the larger project.9What is perhaps most interesting about Figure 4 is how little change there has been inthis ratio, over the period of twenty-four years. There has been a slow downward trend;Figure 4 also includes a linear trend fitted over the twenty-four year period (R = .99436 ­.00172T where R =the ratio, and T (time) runs from 1 to 24.) In sixteen of the twenty-fouryears, the actual value of the ratio lies within 2 1/2% of the trend value. Over the wholeperiod this linear downtrend amounts to a decline in inflation-adjusted fees of about 4%, from99.3% to 95.3% of the (arbitrarily set) value of 1.0 in 1981/82. The actual decline from1969nO to 1992/93 is much larger, from 103.5% to 90.4%, but as indicated in Figure 4, theinitial year is well above the longer-period pattern, and the fmal year is well below.Three sub-periods stand out as exceptions to this generalization about stability. First,there is a downward adjustment in the level of fees in real, or purchasing-power terms, in theyears immediately after the introduction of universal public coverage in 1968. It is unclear,however, to what extent this was an adjustment after a de facto increase in real fee levelsresulting from changed billing practices and improved collections ratios (Barer and Evans,1983).Second, the period from 1980/81 to 1986/87 was characterized in the first three yearsby a sharp increase in fee levels, possibly as a result of over-estimates of the general rate ofinflation. The relatively rapid "squeezing out" of inflationary expectations from NorthAmerican economies, through severe recession, caught many people, including thoseresponsible for negotiating physician fee schedules, by surprise. Over the subsequent threeyears, however, the pre-1980/81 ratio of fees to the CPI was restored. Third, in the harshereconomic climate of the 1990s the ratio has declined again, falling sharply below the lineartrend in 1992/93.Taking Figures 3 and 4 together, one is struck by the relatively limited role of fees.Over the period as a whole, there was an increase of 105.7% in the "real" volume of services-- that is, total payments adjusted for changes in fees -- per capita. But fees fell by 12.6%relative to the general price level, so that inflation-adjusted expenditure per capita rose by10only 79.7%. The use of physicians' services was thus rising at an average rate of 3.2% peryear, while fees were falling by 0.6%.13A somewhat different story emerges, however, if we compare the two sub-periodsbefore and after 1983/84. As observed above, all of the increase in expenditures per capita(adjusted for general inflation) takes place in the earlier period, with an average annualincrease of 4.2% per year. But the "spike" that peaked in 1983/84 brought fees almostexactly back to their 1969nO level in real terms (actually 2.3% higher) so that taking thisearlier period as a whole, virtually the entire increase in expenditures is a consequence ofincreases in use rates.From 1983/84 on, however, fees increasingly lag behind the general inflation ratewhile use rates rise much less rapidly. Taken together, the two effects cancel each other outFees fall by 14.6%, or about 1.7% per year, while per capita use rises 17.6%; their product isa net increase of 0.4% in per capita, inflation-adjusted, expenditures over nine years.Accordingly one might say that the change in the rate of expenditure escalation after1983/84, from 4.2% per year to zero, is accounted for by less rapid growth in both use ratesand fees. Per capita use had been rising at 4.1% per year; this fell to 1.8%. Fees hadincreased by 0.2% over the period prior to 1983/84; afterwards they fell by 1.7% per year.The reduction in annual growth rates is roughly similar for both components, although thereduction in the rate of growth of use is somewhat larger (2.2% compared with 1.9%).On the other hand. a large proportion of the decline in fees after 1983/84 occurs onthe downside of the "six-year hill" at the beginning of the 1980s. That period might be13. Rates of growth must be combined by multiplication -­they cannot simply be added and subtracted -- and their averagesare calculated geometrically, not arithmetically. For smallrates, rounding-off will give the (false) impression of similarresults from either addition or multiplication. All calculationsherein are multiplicative.11thought of as an aberration, a one-time bargaining "error" caused by a large and unexpectedchange in general inflation rates, and subsequently recouped. After this rise and fall, theratio of physicians' fees to the CPI during the rest of the decade looks much the same as itdid in the 1970s. From this perspective, and taking the 1980s as a whole, the decline in realfees looks substantially less significant than the reduction in rates ofgrowth of utilization, as contributors to the flattening of expenditure growth.That is not to say that changing fee levels are no part of the story, much less that thewhole process of fee negotiation which absorbs so much of the effort and attention of theprovincial Ministry of Health and the B.C. Medical Association is without effect Even ifthat process appears, over much of the period, to have ended in fee increases matching (orrunning just behind) the general inflation rate, there is no reason to assume that that resultwould have emerged in the absence of negotiation. Moreover it is clear that when feeincreases and general inflation rates diverge, there~ observable effects on trends in percapita costs.Fees Are Only Half the Story: Trends in the Use of CareBut even during the clearest example of such divergence, the six-year period from1980/81 to 1986/87, changes in "real" fees account for only a part of the changes thatoccurred in overall costs. In the three years of run-up, to 1983/84, costs per capita increasedby 22.3% relative to the CPI, while fees rose only 12.3%. Over the next three years, fees fellin real terms by 8.8%, while real costs per capita fell only 1.4%. During this most dramaticexample of large swings in the real value of physicians' fees, important changes were alsotaking place in the volume and/or mix of services received per capita.These changes are illustrated in Figure 5, which overlays on FIgure 3 a correspondingseries of total expenditures per capita divided by the index of physicians' fees. This latterseries thus represents the real volume of services received per capita - the numbers ofdifferent services weighted by their values. The two series show roughly similar patterns, butthe volume of services rises more rapidly than expenditures when fees are falling in realterms, and conversely.12[Figure 5 about here]Any interpretation of this Figure is critically dependent on the reliability of the'utilization per capita' series generated by deflating expenditures using the publicly availablefee index. This index is calculated by linking together the estimates of average fee increasesthat are generated as part of the periodic (usually annual) negotiation between the B.C.Medical Association and the B.C. Ministry of Health. During negotiations a calculation ismade of what the previous period service patterns and volumes would have cost, under thenew schedule, and this is then compared with the actual outlays, to yield the percentageincrease. It is thus a quantity-weighted index, but with constantly changing weights.As part of the larger project noted above, we have developed a constant-fee quantityindex for the period 1979/80 to 1990/91. Every service provided during that period has beenvalued at the applicable fee as of April, 1988, and then aggregated to yield the total value ofservices provided in each month of each of those fiscal years. This aggregation thenrepresents the total value of all services provided during each month, if valued at constantApril 1, 1988 fee levels. Dividing actual monthly expenditures by this constant fee-levelexpenditure series yields a current-quantity-weighted Paasche index which provides a moreaccurate record of fee movements over the period and which, more importantly, provides ameans of 'deflating' actual expenditures to yield information on changes over time in(weighted) service use."14 This would have been a straight-forward computationalexercise, if the individual service items in the fee schedule hadremained constant over time. But they did not. New services, orvariants on existing services, are introduced; older ones arephased out. Moreover, as part of the negotiation process,services may be "bundled II or "unbundled II from one period to thenext. Thus there is a substantial amount of effort required togather information and make judgements as to what the fee for aparticular service, or bundle of services, would have been, inApril, 1988, if they had been in the fee schedule at that timewhen they were not . Typically (and this project was noexception) such problem items account for a relatively smallproportion of all billings, and a relatively large amount ofresearch effort.13This constructed deflator can then be compared with the fee index that we have usedfor the whole 1969nO to 1992193 period, as calculated from the periodic published revisionsto the fee schedule. It turns out, fortunately, that these two indices differ remarkably littlefrom each other. In the twelve years for which our Paasche price index is available, it alwayslies within 1% of the longer series we have used. This implies that there have not been largechanges over time in the quantity weights, such as to bias the published index. The latterthus appears to be quite satisfactory for aggregate-level analysis of annual data.How Many Physicians. How Many Bills From Each?Returning to per capita utilization, this can be factored into the product of the numberof physicians per capita, and the volume of utilization, or workload, per physician. Both ofthese components rose substantially during the period under study. The number of physiciansper person in B.C., as reported by Health Canada's measure of "Active Civilian Physicians,excluding Interns and Residents", rose 47.7% between 1969nO and 1992193, while averageMSP payments per physician, divided by the index of fees (henceforth fee-adjustedpayments), rose 39.3%.But the rate of increase of both components slowed markedly after 1983/84. In theearlier period, the supply of doctors per capita increased slightly faster, at an annual rate of2.2% compared with 1.8% in fee-adjusted payments per doctor, yielding an overall increase of4.1% in fee-adjusted payments per capita. But after 1983/84 those increases slowed to 0.9%per year for both components, combining to yield a 1.8% annual increase in utilization percapita. Figures 6 and 7 show the pattern of growth of physician supply and volume ofutilization.[Figures 6 and 7 about here]These observations suggest that reductions in the rate of increase in both componentscontributed significantly to the much slower growth of payments to physicians after 1983/84.Figures 6 and 7 also show, however, that the reported data imply a much less stable patternof change for utilization per physician than for physicians per capita. The number ofphysicians per person in B.C. increases quite steadily over almost the whole period, haltingonly after 1988 and rising again in 1992. Reported utilization per physician, on the other14hand, fluctuates markedly from one year to the next.ISB.C. Ministry of Health annual reports also include data on the total number ofdiscrete services reimbursed during each fiscal year, unweighted by fees. The trend in thiscount of "apples and oranges" will deviate from that of the {fee-adjusted} utilization above, ifthe mix of services provided shifts toward either higher or lower average fee per service.These service count data are only available from 1971/72 on; they do in fact show shifts overtime in the mix of services reimbursed. But the changes are not large, relative to the overallchanges in costs per capita.Figure 8 shows the ratio of total payments to physicians (not including salaried andsessional payments in this case), divided by the fee index, and by the reported number ofservices provided. The result is the average fee level (in constant fees) per service provided.Again there are a couple of years in which sharp and unsustained changes cast doubt on theprecision of the connection between annual activity levels and annual payments. But there isalso quite a clear If-shaped pattern, with a decline in the average value per service providedfrom 1971/72 until the early 1980s. The decline then stops, and the later trend is upwards,such that the values in the first and last years of the Figure are virtually identical.[Figure 8 about here]15 It seems very doubtful that annual physician workloads,averaged over the whole province, actually fluctuate to theextent suggested by Figure 7. These variations reflect the factthat paYments within each fiscal year, and reported in the publicrecord, do not always correspond closely to the volume ofservices actually provided by physicians during that year. Forexample, if a target rate of increase for utilization in a givenyear has been established in annual negotiation, and is exceeded,some portion of the excess is held back from paYments during thefollowing fiscal year. In the larger project described above wewill be able to identify from the computerized billing recordsthe actual dates on which particular services were provided. Inthis paper we are simply taking the published record as given,with the exception of the retroactive paYment of $42 million in1990/91 (Note #10 above).15The irregularities in the series make it difficult to identify an exact turning point. Butthe up-trend since 1983/84 is quite clear. From then until the end of the period, the mix ofservices provided shifted towards more expensive services, at an average annual rate of 0.9%per year. During the previous twelve years, the average value per service fell at a rate of0.7% per year.It would appear then that during the earlier period the escalation of average paymentsper physician (at constant fees) was associated with a decline in the average fee per service,implying an even more rapid increase in the number of discrete services provided perphysician. Apparently the less remunerative services were more rapidly increased.After 1983/84 this process was reversed, and the mix of services shifted back towardshigher-fee services. In this latter period, however, there was very little change in the averagenumber of services per physician. The shift to higher-valued services appears to haveaccounted for most of the (much smaller) increase in fee-adjusted payments per physicianduring this period.By 1992/93, the average fee per service had risen virtually back to where it startedwith the first reported data in 1971/72. Thus over the period since that date, the unweightedcount of discrete services provides a measure of volume increase that is very little differentfrom the dollar value of payments in constant fees. But the U-shaped pattern in Figure 8does suggest that changes in the mix of services provided were of some importance within thetwo sub-periods. Their effect was to moderate somewhat the effect of the change in the trendin service volume per physician -- from rapid increase in the first period to virtually flat inthe second -- on the trend in the average dollar value of payments (in constant fees) perphysician. The pattern of increase of average services per physician, from 1971n2 on, isdisplayed in Figure 9.[Figure 9 about here]Pulling All the Levers at Once...Taken together, these data indicate that less rapid increases in all three components16average fee levels (relative to inflation), numbers of physicians per capita, and averagepayments per physician (in constant fees) -- contributed to the sharply reduced rate ofescalation of inflation-adjusted physician payments per capita after 1983/84.Limits on fee increases were obviously a necessary part of cost containment, althoughit is notable that for the most part, and with the exception of the seemingly aberrant period inthe early 1980s, fee levels in real terms were not very different in the 1980s from what theyhad been for most of the 1970s. The expenditure turning point at 1983/84, however, isclearly associated with the beginning of a period of declining real fees.Explicit efforts to limit the growth in numbers of physicians per capita began inBritish Columbia as early as 1983, although the long-established growth trend does not flattenuntil near the end of the decade. The national series flattens at the same point, even thoughno other province had yet tried to limit the numbers of practising physicians. But the numberof physicians per capita grew much less rapidly in B.C. than in the rest of Canada after1983 -- 8.7% from 1983 to 1992 compared with 15.3% in the country as a whole ­suggesting that, whatever the fate of the explicit restriction policies, Canadian physiciansviewed B.C. as a relatively less attractive practice environment.But Are All the Levers Connected?The third component of the changed trend in total outlays is the much less rapidincrease in the volume of payments per physician, adjusted for fee change. Arguably, this toois a consequence of explicitpolicies, introduced at various times during the 1980s as part ofthe fee negotiation process, to discourage physicians from increasing their activity, or at leastbillings. An alternative hypothesis, that the steady increases in the supply of physicians percapita had simply "saturated" British Columbia with physicians, such that average physicianworkload could not continue to increase, seems implausible for a couple of reasons.First, "saturation" would imply falling workloads as supply increased, and there is nosign of a downtrend in the reported average payment data. In the 1970s, physician supplyand payments per physician rose rapidly together; in the 1980s physician supply continued to17rise for much of the decade. But payments per physician, while moving irregularly in thelater period, also continued to trend slowly upward.Secondly, that irregular movement of average payments is itself suggestive ofresponses to a series of "on again, off again" policy initiatives. If "saturation" wereoccurring, one might expect to see a slow tailing off of the rate of increase in payments perphysician, followed perhaps by a decline. But that is clearly not what is shown in Figure 7.But the presumption that physician behaviour changed in response to global policiesdesigned to discourage increased billings, raises another interesting question. A number ofthese policies imposed various forms of ex post "give-backs" -- temporary reductions fromnegotiated fees -- on the profession as a whole, when utilization increased more rapidly thansome pre-determined rate. Yet each individual physician continued to benefit directly fromthe whole of any increase in his or her own billings, with the subsequent give-back costsbeing borne by the profession as a whole. The amount of any subsequent fee discount thatcould be attributed to a physician's own billing patterns, is too small to be of interestThus a purely self-interested physician, as typically postulated in economic theories,would not respond to any of these policies. For that matter, a more "professionally"motivated physician, concerned about the well-being of patients as well as of self, would stilldisregard the negative effects of his or her increased activity levels on the incomes ofcolleagues. If global policies have influenced individual behaviour, that implies some formof implicit or explicit coordination of physician behaviour.The absence of any obvious explicit mechanisms for coordination then raises deeperquestions, particularly for economic analysis, as to how physician behaviour should berepresented and understood. The idea that a small group of firms might recognize theirmutual interdependence, and accordingly modify their behaviour from that implied by purelymyopic self-interest, is old news in oligopoly theory. But an "industry" comprised of severalthousand physicians, all actually or potentially self-employed in "owner-managed" rums,18hardly looks like a textbook oligopoly.On the other hand, physicians are not in "atomistic competiton" either. Quite apartfrom the fact that they are all reimbursed through a common fee schedule, negotiated fromtime to time between their association and the provincial government, physicians also work indifferent regions, specialties, and sub-specialties. A general practitioner in Fort Nelson, anobstetrician in Prince George, and a cardiac surgeon in Victoria are not in competition witheach other in even the most tenuous of senses -- except insofar as their reimbursements allcome from the same public (and in recent years partially capped) budget. The referencegroup of colleagues/competitors for each of them is very much smaller -- perhaps smallenough for mutual interdependence to be recognized.But such questions are impossible to address at the aggregate level. Implicitly treatingthe physicians of British Columbia as if they were one big single physician, or equivalentlyone "representative" physician whose behaviour scales up to the aggregate, is unsatisfactory,both theoretically and empirically." If a major factor "explaining" the dramatic change inthe rate of escalation of physician payments is a change in the (trends in) billing activity ofphysicians, these changes must then be documented, and if possible understood, at a muchmore disaggregated level.It is that task that is being taken up in the larger project, which will involve not onlyanalyses of the more detailed service data using a more fme-tuned fee index, but alsointerviews with physicians 'at the coal-face', who have had to decide how, or whether, to16. The fallacy of this assumption shows up clearly in thefindings of Hurley et ale (1994) in Ontario. They found thatphysicians were in aggregate able to compensate for a steadilyfalling number of people per physician by increasing the averagevolume of services per patient, through either or both ofincreased services per visit, or increased visits per patient.But the pattern of adaptation varied widely from specialty tospecialty, so that the aggregate result was achieved in verydifferent ways.19react to the rich policy mix that has been thrown their way over the past decade.Figure 1Canada Real GDP per Capita$1986, 1960 - 199330~--------------------,~!:) _ _ -.._.._-_.- _._ __ _..-.._..__ _.._ _ _ _.._ - _............ . -.....-..~ ~o -.-..-..- -..--.-..-..-..- -..-..- -..-..- -..- - -............. ...- -. .-- -..~~croen:::Jo-'== 1!:) -..-..- - -- - ..-.-.- - - - .c10 -........ .. -.- - - -.- - -- -..-.-..- -..-..- -..- -..- _.-.-_ _..-61 63 6!:) 67 69 71 73 7!:) 77 79 81 83 8!:) 87 89 91 93yearI • GDP + Time Trend (60-80) IFigure 2Canada Hosp. and M.D. Exp. over GOP1947 -19930.055.,--------------------------;;:11=10.05 - -.- - --.--.- .o.o~ - --.- .o.o~ - ~..0.0~5 . - .O.O~ . -..- - - .O.O~5 ~.. -- -..- - - .O.O~---FO.015~--r-r..,.....,...-r-r--~r-T""""I--r--r-"""T"""T'"~-r--r--r--T""".,.-,--r-T""""I---,--,..--w--r-...,....,..........-.--~r-T""""I--r--r--,---,-~~7 49 51 5~ 55 57 59 61 6~ 65 67 69 71 7~ 75 7779 81 8~ 85 87 89 91 93yearI • Actual + Trend GOP ProjectedFigure 3B.C. MSP Payments per CapitaMedical Fees plus Alternative Payments260-r----------------------.240220,...~ 200,...1:-m 180c::816014012069170 72173 75(76 78(79 81/82 84/85 87/88 90/9192193Fiscal YearFigure 4B.C. Physicians' Fees in Constant $Fee Index over Vancouver CPI1.2-r----------------------,1.15o 1.1•,...~ 1.05~co 1m,...B 0.95enII: 0.90.850.869170 75/76 78/79 81/82 84/85 87/88Fiscal YearI • Actual Ratio + Linear Trend IFigure 5B.C. MSP Payments per CapitaMedical Fees plus Altemative Payments280-r-------------------------,260(17 240C\I~ 220CD0),.... 200"E-m 180c:oo 16014012069170 72/73 75176 78179 81/82 84/85 87/88Fiscal YearI • De1. by CPI + Def. by Fee Index IFigure 6B.C. Physicians per (000) CapitaExcluding Residents and Interns2~-----------------....., 78179 81/82 84/85 87/88 90/91 92/93Fiscal YearFigure 7B.C. MSP Payments per Physicianat Constant 1981 Fees140...,.----------------------,135130125.-..~ 120c:m115:::::Jo~ 1101051009578T19 81/82 84/85 87/88 90/9192193Fiscal YearFigure 8Average Payment per Physician ServiceAt 1981/82 Fees20-,---------------------,19.51878(79 81/82 84185 87188 9019192193Fiscal YearFigure 9Average Services per Physician(Not Weighted by Fees)6800-r--------------------.,66006400620060005800560054005200500078/79 81/82 84/85 87/88Fiscal Year20Data AppendixPopulation data are from Statistics Canada, quarterly estimates, as reported inCANSIM series 1.1 (Canada) and 1.1.10 (British Columbia). They include the revisions forCensus undercount released in September, 1993, back to 1971. Unrevised data prior to 1971have been linked by the authors; the percentage revision for the 1971 value was simplyapplied to each of the earlier years. The Canada data are July 1 (mid-calendar year); the B.C.data are October 1 (mid-fiscal year). Gross Domestic Product at 1986 prices is taken fromReference Table 4 of the 1994 issue of the Department of Finance Economic and FiscalReference Tables. National expenditures on hospitals and physicians' services from 1975 to1993 are reported in the 1994 issue of National Health Expenditures in Canada, compiled byHealth Canada. Earlier values of these series were assembled from earlier publications andreports from Health and Welfare Canada, as described in the data appendix to Barer andEvans (1986).Payments to physicians from the Medical Services Commission of British Columbiafor fiscal years 1983/84 to 1992/93 are reported in Appendix Table 2, page 83, of the AnnualReport of the British Columbia Ministry of Health for 1992-1993. Total services reimbursedare reported in Appendix Table 4, pages 86/87. Earlier years' data are compiled fromprevious issues of the Ministry Annual Report. As noted in the text (Note #10 above), aretroactive payment of $42 million has been transferred from the 1990/91 figure to the1989/90 figure; otherwise the data are as reported by the Ministry.Numbers of physicians in British Columbia (as of December 31) for the years 1981 to1991 are reported in Health Personnel in Canada. 1991 issued by Health Canada. Data for1992 were provided by Ministry staff; earlier years' data were compiled from earlier annualissues. The index of physicians' fees for British Columbia is assembled from thecompilations of changes in physician payment schedules issued from time to time by HealthCanada. The most recent is for July, 1994 and reports fiscal-year indices from 1981/82 to1993/94. These tabulate only the changes in average fee levels in each province.Comparisons of levels of fees in the different provinces are also issued from time to time; theone referred to in the text is as of July 1, 1985.The Consumer Price Index for Vancouver from 1971 to 1993 is reported in Reference21Table 49 of the 1994 issue of the Department of Finance Economic and Fiscal ReferenceTables. Values for 1969 and 1970 were estimated by assuming that the Vancouver index hadchanged in the same proportion as the national index in those years. Calendar-year data wereconverted to fiscal years by adding three quarters of one year's value to one quarter of thenext.22ReferencesBarer, M.L. and RG. Evans (1983), "Prices, Proxies and Productivity: An Historical Analysisof Hospital and Medical Care in Canada," in Diewert and Montmarquette, eds., Price LevelMeasurement: Proceedings from a conference sponsored by Statistics Canada, pp. 705-777,Ottawa: Minister of Supply and Services Canada.Barer, M.L., and RG. Evans (1986), "Riding North on a South-Bound Horse? Expenditures,Prices, Utilization, and Incomes in the Canadian Health Care System" in R.G. Evans and G.L.Stoddart (eds.) Medicare at maturitv: Achievements. lessons & challenges Calgary: Universityof Calgary Press for the Banff Centre.Barer, M.L., R Labelle, S. Morris, RG. Evans and G.L. Stoddart (1987) "The Impact onMedical Services Utilization of British Columbia's 1982/83 Physician Fee "Giveback":Preliminary Results" Canadian Journal of Public Health 78:1, 37-42.Barer, M.L. (1988) "Regulating Physician Supply: The Evolution of British Columbia's Bill41" Journal of Health Politics. Policy and Law, Vol.l3, no. 1, (Spring) pp. 1-25.British Columbia (1991) Closer to Home: The Report of the British Columbia RoyalCommission on Health Care and Costs, Victoria: Crown Publishers.Canada, Health and Welfare Canada, Health Information Division (1986) July 1, 1985Payment Schedule Comparison (provincial Supplement) Ottawa: Health Canada (January).Canada, Health Canada, Health Information Division (1993) Health Personnel in Canada.1991 Ottawa: Minister of Supply and Services Canada.Canada, Health Canada, Health Information Division (1994) National Health Expenditures inCanada. 1975-1993 Ottawa: Minister of Supply and Services Canada (June).Canada, Health Canada, Health Information Division (1994) Changes in Pavment Schedulesto Provincial Medical Insurance Plans: Canada. 1986/87 - 1993/94 Ottawa: Health Canada(July).Canada, Department of Fmance (1994) Economic and Fiscal Reference Tables. September1994 Ottawa: Department of Finance.Culyer, A.J. (1988), Health Care Expenditures in Canada: Myth and Realitv: Past and Future,Canadian Tax Paper No. 82, Toronto: Canadian Tax FoundationEvans, R.G. (1993) "Health Care Reform: 'The Issue from Hell'" Policy Options (TheInstitute for Research on Public Policy) Vol. 14, no. 6 (Special Issue, July-August 1993) pp.34-41.23Ham, C; R. Robinson, and M. Benzeval (1990), Health Check: Health care refonns in aninternational context, London: King's Fund Institute.Hurley, J., C. Woodward and J. Brown (1994) "Changing Patterns of Physician Services Usein Ontario, Canada and Their Relation to Physician, Practice and Market-Area Characteristics"Hamilton, Ont.: McMaster University Centre for Health Economics and Policy AnalysisWorking Paper #94-17.Lomas, J., C. Fooks, T. Rice, and R.J. Labelle (1989) "Paying Physicians in Canada: MindingOur Ps and Qs" Health Affairs Vol. 8, no.l Spring pp. 80-102.Lomas, J., C. Charles, and J. Greb (1992), The Price of Peace. The Structure and Process ofPhysician Fee Negotiations in Canada. Working Paper No. 92-17, Hamilton: McMasterUniversity Centre for Health Economics and Policy Analysis.HPRU 93:11RHPRU 93:110HPRU93:12DHPRU 93:130HPRU 93:140HPRU93:150HPRU 93:160HPRU93:170HPRU94:10HPRU94:1RHPRU94:2RHPRU94:30HPRU94:40Hertzman, C. "EnvilOnT1l8lJt &Health: InC6ntra/ sndEastern Europe" Q)st $12.00"Environment &Health: InCentral andEastsm Europe". August 1994. (C. Hertzman) cost $12.00Why Not User Charges? The Real Issues. December 1993. (G.Lstoddart, M. L Barer, R. G. Evans, V.Bhatia)cost $8.00Who Are theZOmbie Masters, andWhat Do They Want? December 1993. ( R.G. Evans, M. L Barer, G.Lstoddart, V. Bhatia) cost $5.00User Charges, Snares andDelusions: Another Look attheUterature. December 1993. (G. L stoddart, M.LBarer, R.G. Evans,) cost $8.00The Remarkable Tenacity ofUser Charges: AConcise History ofParticlpaUon, PosiUons, andRationales ofCanadian Interest Groups IntheDebate over "Direct Patient Partlclpatlon·ln Health Care Ananclng. December1993 (M.LBarer, V.Bhatia, G. L stoddart, R.G. Evans,) cost $8.00IrsNot the Money, IrsthePrinciple: Why User Charges for Some ServIces, and Not Others. December. ( R.G.Evans, M. L Barer, G.LStoddart, V. Bhatia) December 1993 Cost $8.00Charging Peter toPay Paul: Accounting for theAnanclal Effects ofuser Charges. December 1993 ( R.G. Evans,M. L Barer, G.LStoddart) Cost $12.00IIp-toelng Toward Health Care Reform IntheUnited states. January 1994. (M. L Barer, E. Morrison, I.Morrison) Cost $5.00Barer, M.L, MorrI$on, E.,Morrison, I.(1994) ·Canadlan Physicians may hear footsteps ofchange asUS tiptoestoward health care reform·, CanadIan M8dIcaI Association Journal, Vol 150 (6), P~987.This IsarsvIssd sndshortened V91sIon of94:1DEvans, Robert G. ·Lsss IsMore: Contrasting Styles InHealth GarB' Canada andtheUnited states, Differencesthat Count 21 -41. No chargeAvalanche orGlacier. HeaI1h Care andtheDemographic Rhetoric. March 1994. (M.LBarer, R.G. Evans, C.Hertzman) Cost $8.00DiagnosUc andSCreening Programs for Fetal Anomalies andGenetic DIsease: How shouJd we evaluate them?April 1994. (P. BaIrd) Cost $8.008HPRU94:5DHPRU94:6RHPRU94:7DHPRU94:8DHPRU94:9RHPRU 94:10RHPRU95:1DHPRU95:2DHPRU95:3DItAln' Neoessarlly So: The Cost Implications ofHealth care Reform IntheUnited Stales. AprU 1994. (M.LBarer,R.G. Evans, Matthew Holt, J.I.MooIson) Cost $8.00Blencwelg,A.Y. and Pagllccla N. (1994), "Utilization Patterns ofCohorts ofElderty Clients:ASlructural EquaUonModel", HeaJlh S8rvIces Research 29(2):225-245.Health, Hierarchyand Homlnlds: Biological Correlates ofthe SockrEconomlc Gradient In Health, August 1994(R.G. Evans) Cost $8.00An Empirical Evaluation ofComputerized Databasss forEmerg91lG)' Csre. October 1994. (George E. Pugh,Joseph K.Tan) Cost $8.00Evans, R.G. (Fall 1994)"HeaJlh and WeaJlh", Daedalus, Journal ofAmerIcan Academy ofArts andSCiences123 (4):21....2Hogg, R.G.,Stralhdee, SA, Craib, KJ.P.,O'Shaughnessy, M.V.andMontaner, J.S.G., SChechter, M.T. "Lowersocioeconomic status and shorter survival following HIV InfecUon", The Lancst Vol 334, P11~1124.So Near, And Yet SoFar: January 1995 (Morris L Barer) Cost $5.00Marketing the Market RegUlating RegUlators: Who GaIns? Who Loses? What Hopes? What Scope? January1995. (Robert G.Evans) $8.00Research InHuman Genetics: Promise, Pitfalls and Policy Challenges. February 1995 (Patricia BaIrd, M.D.).Cost$5.00.9


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items