Short-term inverse association between changes in government expenditure on retail pharmaceuticals and treatable mortality in Spain

Specific studies on the relationship between pharmaceutical expenditure and mortality are sparse, particularly those that consider mortality from causes most sensitive to medical treatments (treatable mortality). The purpose was to assess short-term associations between changes in government expenditure on retail pharmaceuticals (GERP) and other health goods/services (GEOH) and changes in premature and non-premature mortality by avoidability and cause-groups in Spain. A cross-sectional time series design was used with annual data from 17 Spanish regions during 2001–2019. Outcomes were age-sex-standardized rates for premature (< 75 years) and non-premature (≥ 75 years) mortality stratified by avoidability (treatable, preventable, and non-avoidable), by cause-groups (circulatory, neoplastic, respiratory, infectious, and other causes) and jointly by avoidability and cause-groups. Using weighted log-log first-difference regression adjusted for relevant covariates, we estimated mortality elasticities with respect to GERP and GEOH, interpreted as the percentage change in mortality rate associated with a 1% increase in the two-year mean change in expenditure. A sub-analysis (2001–2013) additionally adjusted for government social expenditure (GSE). For GERP, significant negative elasticities ( p < 0.05) were observed for premature treatable mortality (− 0.23%) and for premature mortality from respiratory (− 0.56%), infectious (− 0.49%), and circulatory causes (− 0.22%), and also for non-premature treatable mortality (− 0.20%) and for non-premature mortality from infectious (− 0.66%) and respiratory causes (− 0.50%). In jointly-stratified analyses by avoidability and cause-groups, the largest inverse associations were observed for premature treatable mortality from infectious (− 1.60%) and respiratory causes (− 0.88%). For GEOH, elasticities were generally statistically nonsignificant and weaker, except for non-premature mortality from certain causes (e.g. circulatory) and particularly with one-year lag GEOH. In the GSE-adjusted sub-analysis, negative elasticities for non-premature mortality with respect to GERP weakened considerably. Short-term changes in GERP were inversely associated with changes in premature mortality, particularly from treatable causes, especially respiratory and infectious ones. For non-premature mortality, associations attenuated after GSE adjustment.