OBJECTIVE: To rank 39 surgical procedures in order of variation of inpatient surgical rates, according to a new index of variation and to test the hypothesis that there is greater variation for primarily discretionary operations than for primarily non-discretionary operations. DESIGN: A population-based retrospective cohort study. SETTING: Nine provinces (99.19% of Canada's population). PARTICIPANTS: All hospital inpatients who underwent any of 39 types of surgery and were separated from hospital between Apr. 1, 1988 and Mar. 31, 1990 (the most recent time period for which Canada-wide data were available at the subprovincial level analysed). MAIN OUTCOME MEASURE: Rankings of the 39 procedures according to the index of variation, calculated from inpatient surgery rates in 255 census divisions across Canada. RESULTS: The 13 procedures with the greatest variation were all primarily discretionary. Thirteen of the 14 procedures with the lowest variation were primarily non-discretionary. All but one of the procedures whose degree of discretion was deemed intermediate were in the middle third of the rankings. CONCLUSIONS: The greatest variation is found in primarily discretionary operations. Further investigation should be focused on identified geographical locations where rates for operations that are primarily discretionary are unusually high or low, and particularly on those operations for which there is disagreement regarding the indications for surgery.
OBJECTIVE: To rank 39 surgical procedures in order of variation of inpatient surgical rates, according to a new index of variation and to test the hypothesis that there is greater variation for primarily discretionary operations than for primarily non-discretionary operations. DESIGN: A population-based retrospective cohort study. SETTING: Nine provinces (99.19% of Canada's population). PARTICIPANTS: All hospital inpatients who underwent any of 39 types of surgery and were separated from hospital between Apr. 1, 1988 and Mar. 31, 1990 (the most recent time period for which Canada-wide data were available at the subprovincial level analysed). MAIN OUTCOME MEASURE: Rankings of the 39 procedures according to the index of variation, calculated from inpatient surgery rates in 255 census divisions across Canada. RESULTS: The 13 procedures with the greatest variation were all primarily discretionary. Thirteen of the 14 procedures with the lowest variation were primarily non-discretionary. All but one of the procedures whose degree of discretion was deemed intermediate were in the middle third of the rankings. CONCLUSIONS: The greatest variation is found in primarily discretionary operations. Further investigation should be focused on identified geographical locations where rates for operations that are primarily discretionary are unusually high or low, and particularly on those operations for which there is disagreement regarding the indications for surgery.