RESUMO
OBJECTIVES: To analyse the prescription patterns in rural, urban and semi-urban Primary Care teams (PCT) and to study the variables which are correlated with the pharmaceutical cost, with special reference to medicines of low therapeutic value (LTV). DESIGN: Descriptive, crossover study. SETTING: Avila Health Area, 1995. PATIENTS AND OTHER PARTICIPANTS: Prescription profiles of 199 Primary Care doctors from Avila's 21 Health Districts, using the number of holders of the Individual Health Card (IHC) to determine population size. MAIN RESULTS: There were notable differences in prescription patterns between urban, rural and semi-urban PCTs, with the "Drug cost per inhabitant per year" higher in rural than in urban PCTs (24,082 pesetas vs. 14,804), and the "Cost per prescription" less (1,417 vs. 1,577). In the simple regression analysis, there was a significant correlation between the variables "Number of LTV packages per inhabitant per year" and "Drug cost per inhabitant per year" (r = 0.68, 0.51 and 0.64 for urban, semi-urban and rural PCTs); between "Number of LTV packages per inhabitant per year" and "Number of prescriptions per inhabitant per year" (r = 0.79, 0.80 and 0.82); between "Drug cost per inhabitant per year" and "Number of prescriptions per inhabitant per year" (r = 0.81, 0.85 and 0.88); between "Number of IHCs per doctors" and "Drug cost per inhabitant per year" (r = -0.62, -0.57 and -0.31) and between "Number of IHCs per doctor" and "Number of prescriptions per inhabitant per year" (r = -0.51, -0.62 and -0.36). In the multiple regression analysis, if "Drug cost per inhabitant per year" was used as a dependent variable, the variable "Number of LTV packages per inhabitant per year" entered the model in urban and rural PCTs (Adjusted r2 = 0.55 and 0.60). If "Number of prescriptions per inhabitant per year" was used as dependent variable, the variable "Number of LTV packages per inhabitant per year" again explained, for all kinds of PCT, most of the variability (Adjusted r2 = 0.65, 0.65 and 0.76). CONCLUSIONS: The drug cost per inhabitant per year is less in urban than in rural Centres, while the cost per prescription is greater. The cost per inhabitant per year and the number of prescriptions per inhabitant per year have an inverse relationship to the population size. LTV medicines behave as pharmacy cost predictors, since a consistent relationship between their prescription, the cost per inhabitant per year and the number of prescriptions per inhabitant per year can be shown.
