Physician Fee Levels: Medicare Versus Canada W. Pete Welch, Ph.D., Steven J. Katz, M.D., Stephen Zuckerman, Ph.D. Adjusted for differences in purchasing fee levels and how they are adjusted over power and practice expenses, Canadian time that is likely to be the single most Im­ physician fees are, on average, 59 percent portant issue faced under the Medicare of Medicare fees. The general perception fee schedule. that Medicare fees are low is the result of Thus far, assessment of the appropri­ comparison with U.S. private fees, not to ateness of Medicare payment rates has the much lower Canadian fees. In the con­ been confined to comparisons with rates text of the current U.S. health care sys· paid by private insurers in the United tem, lowering Medicare fees to Canadian States (Physician Payment Review Com­ levels could jeopardize access to care by mission, 1992c). Because private rates are Medicare beneficiaries. However, if all higher, the comparison raises concerns payers used the same fee schedule, fees that the elderly's access to care could be that differed substantially from those cur· impaired if physicians prefer to treat pa­ rently used by private insurers might be tients covered by private insurance rather viable. than Medicare enrollees. The experience of the Medicaid program, whose rates are INTRODUCTION substantially lower than private insurance The new Medicare fee schedule, a ma­ rates, is an example of how low physician jor reform of physician payment in the fees can restrict access to care within the United States, has accomplished two U.S. health care system (Holahan, Wade, goals. First, by incorporating a resource· and Gates, 1992). Of course, given Medi­ based relative value scale (RBRVS), it par· care's larger share of physician revenues, !Iaiiy corrects historical inequities in pay­ physicians might respond differently to ment for procedure and cognitive serv· low Medicare fees than they do to low Ices. Second, by reducing these inequi­ Medicaid fees. ties, it focuses debates over Medicare Fees under private insurance are not fees on the overall generosity of pay­ the only reasonable source of compari­ ments, both now and in the future. Al­ son. Physician fees in other countries though the relative values may be revised may be just as appropriate a "bench­ and the extent of geographic equity is mark." Canada is an obvious source for sometimes questioned, it is the overall comparison for several reasons beyond its geographic proximity to the United States. First, physicians in both countries Support for this research was provided by the Health Care Fl· nancing Administration to The Urban Institute through Cooper­ are generally paid on a fee-for-service ba­ ative Agreement Number 99-C-98526 to the Brandeis University sis. Second, although the negotiations Health Polley Research Consortium. W. Pete Welch and Stephen Zuckerman are with the Urban Institute, Steven J. between the provincial medical associa­ Katz is with the University of Michigan. Any opinions ex­ tions and the governments are not di­ pressed are those of the authors and not The Urban Institute, University of Michigan, or their sponsors. rectly analogous to Medicare, they result Health Care Financing RevlewiSpring 1993/volume14,Number3 41 in fee schedules that are generally similar METHODS to Medicare's structure. Third, the Cana­ dian health care system has been sug­ To compare fees, we developed In­ gested by some as a solution to the prob­ dexes that are weighted averages of fees lems of the U.S. system. Finally, Canadian in Canada relative to those paid under fees are established in a single-payer con­ Medicare. Although our prlmary objective text and, as such, provide a useful con­ is aggregate comparisons of relative fees, trast to Medicare, which is forced to oper­ we also compare fees disaggregated by ate in a market with private insurers. In type of service. For this we need to define this sense, Canadian fees suggest what the unit of Medicare services, select ser­ U.S. fees might be under a system in vices, and ensure comparability of Cana­ which all payers adhered to the same rate dian and Medicare fees. Ensuring compa­ schedule. rability involves addressing differences In In spite of these reasons, few coding systems, payment rules, and eco­ Canadian-U.S. comparisons of physician nomic conditions. fees have been made. The present study Our basic unit of observation is the fills a void by comparing the physician Common Procedural Terminology (CPT) fees paid by Medicare to those paid by code. Although CPT codes are used by Canadian provinces. In addition to com­ most payers In the United States, each paring physician fees In the aggregate, Canadian province has Its own coding we compare fees by type of service. system. Because of the difficulty of iden­ One of the few contributions in this tifying Canadian provincial codes that are area Is by Fuchs and Hahn (1990), who equivalent to CPT codes, II was not feasi­ compared fee levels in the two countries. ble to work with all the CPT codes and all Beyond the Medicare focus, the methods the provinces. We therefore focused on employed in this study extend the work fees in the four largest provinces, which by Fuchs and Hahn in two respects. First, together have 83 percent of the Canadian we use fees from the four largest Cana­ population. In order of population size, dian provinces, whereas Fuchs and Hahn they are Ontario, Quebec, British Colum­ used fees from a single small province bia, and Alberta (Manitoba) adjusted to represent all of Selection of Codes and Construction of Canada Second, our U.S. fees are consis­ Index tently derived from a single source the Medicare fee schedule. Fuchs and Hahn We classified CPT codes by type of ser­ combined data from Health Insurance vice to help ensure that the selected Association of America and Blue Shield codes were representative of a wide plans in Iowa and California and treated range of services and to allow us to com­ this as representative of the country as a pare fees at a somewhat disaggregated whole. In light of the weakness of their level. In particular, we employed a type-of­ underlying fee data, and given the Intro­ service classification system deviSed by duction of the Medicare fee schedule, the Berenson and Holahan (1992). The sys­ Issue of relative fees across the two coun­ tem divides physician services into 6 tries needs to be revisited. (major) categories and 23 subcategories. 42 Health Care Financing Review/Spring 1993/volume 14, Number3 Evaluation and management services Within the procedure and Imaging sub­ comprise the first category. Its subcate· categories, we selected the eight CPT gorles are office visits, hospital visits, codes with the greatest share of Medi­ emergency room services, home and care expenditures for each type of ser­ nursing home visits, consultations, and vice. (These expenditure shares also specialist evaluation and management serve as weights in the index, as noted services. The last subcategory In this later.) A CPT code's Medicare expendi­ group Includes services provided by oph· ture was calculated as Its volume In 1989 thalmologlsts, psychiatrists, patholo· times its relative value units (RVUs) In the gists, and allergists. new Medicare fee schedule. A code's ex­ Procedures, the second category, in· penditure divided by the expenditure for eludes major procedures (subdivided Into all services yields the code's expenditure cardiovascular, orthopedic, and other), share. CPT codes with high volumes are ambulatory procedures (subdivided Into more likely to be represented than codes eye and other), and minor procedures with low volumes, and CPT codes with (subdivided into endoscopic, oncology high fees (and RVUs) are more likely to be procedures, dialysis services, and other). represented than codes with low fees. Imaging, the third category, Is divided Into One might be concerned that our selec­ standard Imaging (routine X-rays and nu· tion procedure Is biased toward selecting clear medicine), advanced imaging (com· CPT codes with high Medicare fees (and puterized tomography (CT] scans and RVUs). To test this, we replaced Medicare magnetic resonance imaging [MRID, so· fees with Ontario fees in calculating ex­ nographlc imaging, and Imaging proce· penditure shares. Viliually the same CPT dures (largely cardiac catheterization). codes would be selected under the two The remaining three categories are tests methods. (subdivided into laboratory and other), an· Within the other subcategories (i.e., esthesiology services, and other services. evaluation and management), we se­ Of the Berenson and Holahan group· lected CPT codes by group: the 10 codes lngs, three categories and three subcate· for office visits, the 6 codes for hospital gories were dropped for this analysis. An· visits, the 5 codes for emergency room estheslology services were dropped visits, the 3 codes for subsequent nursing because In some fee schedules payment facility care visits, the 3 codes for rest varies by the duration of the procedure, home visits for established patients, the 4 making pricing difficult. Laboratory tests, codes for eye exams, the 2 codes for psy­ other tests, and "other services" were chotherapy, and the 5 codes for office dropped because of the difficulty of lden· consultations. Both initial and followup tllying equivalent Canadian codes. Oncol­ visit codes were included (Table 1). ogy services were dropped because they The CPT codes incorporated in this in­ are often included In Canadian global dex represent 63 percent of the RVUs in hospital budgets and cannot, therefore, these 17 subcategories. Within each of be separately priced. Dialysis services the three categories, this expenditure were dropped because of problematic share Is 83 percent for evaluation and Medicare data on volume. This left 17 sub­ management services, 43 percent for pro­ categolies in 3 categories.