Abstract
BACKGROUND: Despite similarities in demography and health care delivery, Canadian provinces exhibit significant variation in population adjusted-rates of blood product usage. A study was conducted to identify factors which underlie this variation.
METHODS: Potential drivers of blood product consumption were identified through literature review, and then mapped to parameters for which provincial data were available from analytic reports published by the Canadian Institute for Health Information. Blood product issues per province for the year 2001 were obtained from annual reports published by Canadian Blood Services and HémaQuébec. All data were controlled for provincial population using data from Statistics Canada. Linear regressions for each parameter were calculated separately for erythrocyte, plasma and platelet issues. Variables that were statistically significant by univariate regression analysis were subjected to multiple regression modeling.
RESULTS: Provincial variation in erythrocyte demand was statistically correlated with the following by univariate regression analysis: population density, coronary artery bypass graft (CABG) procedures and peptic ulcer diagnoses. By multivariate analysis, only CABG procedures continued to have a statistically significant correlation. For plasma issues, univariate analysis revealed average physician age, average specialist age, and CABG procedures as statistically significant drivers. However, by multivariate analysis only CABG procedures retained statistical significance. For platelet issues, average specialist age, cardiovascular procedures, CABG procedures, and number of full-time equivalent general surgeons were statistically significant drivers of provincial variation. However, none of these variables retained statistical significance by multiple regression modeling.
CONCLUSIONS: The population-adjusted rate of coronary artery bypass procedure graft surgery was a statistically significant driver of provincial variation in erythrocyte and plasma issues in the year 2001, a finding which may be of use in demand forecasting.
