Objective: To measure regional variations in knee replacement (KR) utilization rates across the 49 counties of Ontario, Canada, since April 1, 1984, and to evaluate its relationship with the supply of health professionals.
Methods: Utilization data were acquired from the Canadian Institute for Health Information, and physician data were acquired from the Ontario Ministry of Health. There were 18,530 hospitalizations for KR in Ontario between April 1, 1984 and March 31, 1991. Utilization rates were defined as the number of KR procedures for all residents of each county (irrespective of where these procedures were performed) divided by the resident population. Direct methods were used to standardize utilization for age and sex. The extremal quotient, the weighted coefficient of variation, and the systematic coefficient of variation were used as measures of variation.
Results: Between fiscal years 1984 and 1990, the extremal quotient fell from 8.1 to 3.4, the weighted coefficient of variation fell from 0.430 to 0.315, and the systematic coefficient of variation fell from 0.215 to 0.142. Utilization was greater in southwestern Ontario (p < 0.001) and lower in Toronto and Durham (p < 0.001) than elsewhere. Utilization in remote areas was not different from that for the province as a whole (p < 0.001). The correlation coefficient between standardized KR utilization and the density of orthopedic surgeons was not significant (p = 0.26) at -0.16, but those for rheumatologists (r = -0.25, p = 0.08) and physical medicine specialists (r = -0.42, p = 0.002) were significant.
Conclusion: Regional variation in KR surgery has fallen since fiscal year 1984 from "very high" variability to "moderate" variability. KR utilization was shown to be negatively, but not significantly, related to orthopedic surgeon density. Therefore, proposals to modify surgeons' fees or to influence their practice locations may equalize the supply of surgeons, but are unlikely to affect the pattern of KR utilization.