Explaining primary healthcare pharmacy expenditure using classification of medications for chronic conditions.

BACKGROUND The Valencian Autonomous Community (Spain) has implemented a scheme of purchasing services with the participation of public and private providers. Five districts are managed using public-private partnership. The financing model is capitation and inter-center invoice. The pharmaceutical benefits are not included in the per capita assignment. OBJECTIVES Modeling and explaining pharmacy expenditure using electronic prescriptions drug data. METHODS A database of electronic prescription corresponding to 625,246 patients between November 2008 and October 2009 was used to run four linear models that explain the pharmaceutical expenditures. We take as dependent variable the neperian log of total pharmacy annual cost per patient in the primary health setting. The independent variables used combined demographics with revised classification in 18 chronic conditions obtained from the anatomical therapeutic chemical classification index (ATC). RESULTS The retrospective model selected included: gender, pharmaceutical co-payment status and 8 dummy variables for the number of chronic conditions of each patient from 1 to 8 or more. The goodness-of-fit achieved is measured in R(2) of 57%. CONCLUSIONS These models must be considered in the current capitation system for pharmaceutical budgeting in a primary care setting established at regional level, as is the case in the Valencian Autonomous Community. The use of diagnostics and information regarding hospital encounters appears to be a complementary option for refining models of capitation of pharmaceutical and total health expenditure.