VALIDATING LITERATURE-BASED MODELS WITH DIRECT CLINICAL-TRIAL RESULTS- THE COST-EFFECTIVENESS OF SECONDARY PROPHYLAXIS FOR PCP IN AIDS PATIENTS

Objective. To compare literature-based estimates of the cost-effective ness ratios of strategies for secondary prophylaxis of Pneumocystis ca rinii pneumonia (PCP) in AIDS patients with estimates obtained using d ata from a recent comparative clinical trial. Design. A decision-analy tic Markov model with data on drug efficacy and toxicity from both the medical literature and a national randomized clinical trial. Drug cos ts were from average wholesale prices. Discounted life expectancy, tot al direct medical costs, and cost-effectiveness were projected in doll ars per year of life saved (YLS). Setting. Hypothetical for the litera ture-based model, then the clinical trial results from the multicenter AIDS Clinical Trials Group (ACTG Protocol 021). Patient population. P atients with AIDS and a prior episode of PCP. Interventions. Strategie s included no prophylaxis, TMP-SMX (160/800 mg) daily, or aerosolized pentamidine (300 mg) monthly. Patients experiencing major toxic reacti ons to either medication would cross over to the other agent. Main res ults. In the literature-based model no prophylaxis was associated with a projected life expectancy of 1.430 years, and total direct cost of $42,080. TMP-SMX increased life expectancy to 2.051 years and cost to $42,300; for aerosolized pentamidine life expectancy was 2.066 years a nd cost $43,960. TMP-SMX had an incremental cost-effectiveness ratio o f $350 per YLS compared with no prophylaxis; the incremental ratio for aerosolized pentamidine was $2,950 per YLS when compared with no prop hylaxis, but rose to $110,880 per YLS compared with TMP-SMX. When data from ACTG clinical trial 021 were utilized in the model, the incremen tal cost-effectiveness ratio for TMP-SMX compared with no prophylaxis was $720 per YLS; aerosolized pentamidine was not cost-effective, and was ''dominated'' by TMP-SMX because it was associated with higher cos ts and shorter life expectancy. Conclusions. Literature-based cost-eff ectiveness models are useful in developing health policy before clinic al trials are completed. Clinical trial results, when available, can b e used to validate and revise these models. For secondary PCP prophyla xis in AIDS patients, TMP-SMX is substantially more cost-effective tha n aerosolized pentamidine.