Surrogate markers, particularly measures based on HIV-1 RNA copy number, currently are standard primary endpoints in AIDS clinical trials. However, because total eradication of the virus resulting in a cure for AIDS is not yet known to be possible, long-term clinical outcomes such as survival and functional status remain of primary importance to patients. Several recent papers have highlighted the inadequacy of existing statistical methods for evaluating whether clinical trials based on surrogate endpoints will lead to the same conclusions as would have been reached had clinical endpoints been used. The overall objectives of this proposal are: 1. To develop and implement Bayesian models that will improve on existing statistical methodology for evaluating surrogate markers as candidate endpoints in AIDS clinical trials. 2. To fit such models to data from clinical trials in pediatric AIDS in which both clinical and surrogate endpoints were observed: a. to compare several virologic, immunologic, and combination measures as surrogates for clinical progression. b. To explore the usefulness of incorporating other types of clinical trial data (e.g. early low-grade toxicities) into composite markers to capture treatment effects not mediated by virologic and immunologic measures. 3. To develop strategies, recommendations, and possibly software, to assist other investigators in using Markov chain Monte Carlo methods for fitting the models developed.
| Cowles, Mary Kathryn (2002) Bayesian estimation of the proportion of treatment effect captured by a surrogate marker. Stat Med 21:811-34 |
