USE OF CONFIDENCE-INTERVALS AND SAMPLE-SIZE CALCULATIONS IN HEALTH ECONOMIC-STUDIES

OBJECTIVE: To discuss the calculation and application of confidence in tervals in pharmacoeconomic studies. DATA SYNTHESIS: The increasing fr equency with which pharmacoeconomic evaluations are made within clinic al trials makes it possible to obtain information on the outputs and c osts of an intervention in each patient of a sample under study. This allows the same statistical principles commonly used in clinical trial s to be applied to cost or cost-effectiveness data. The methodology de scribed in this article would allow expression of cost-effectiveness r atios in the form of confidence intervals. The calculation of the cost -effectiveness ratio by means of a confidence interval may have import ant practical consequences, both in decision-making on the choice of 1 intervention versus another and in calculating the size of the sample necessary to identify statistically significant differences, from bot h clinical and economic points of view. CONCLUSIONS: The complementary use of confidence intervals and sensitivity analysis makes it possibl e to measure uncertainty related and unrelated to variability in sampl e data, allowing the decision to adopt 1 technology or another to be b ased on the most objective information available. Although several eth ical and methodologic concerns remain to be addressed, this methodolog y may contribute to improving the more rational and efficient use of d rugs.