A second look at pharmaceutical spending as determinants of health outcomes in Canada.

Per capita spending on pharmaceutical products has increased substantially in recent decades in Canada. Recent Canadian research by Crémieux et al. concludes that there is a strong statistical relationship between pharmaceutical spending and health outcomes (Health Econ. 2005a; 14: 117, Health Econ. 2005b; 14(2): 107-116). This paper takes a second look at pharmaceutical spending as determinants of health outcomes in Canada. In doing so, it examines the robustness of the findings of Crémieux et al. by considering the appropriateness of the data used and statistical approach utilized. Particular attention is given to the potential for non-stationarity and spurious regression, issues related to unit heterogeneity and the choice of estimators. In contrast with earlier findings, on the whole, no discernable relationship between spending on private or public pharmaceutical products and infant mortality or life expectancy at 65 is observed.