New methods were published for relating treatments for HIV infection to favorable changes in AIDS incidence. These methods and associated data indicate that treatment has had a favorable effect on national AIDS incidence rates. Methods for correcting AIDS incidence series for reporting delays were published, and work is in press on rapid regression methods for """"""""backcalculation"""""""" of AIDS data. Sample size calculations were published for designing studies of complex aspects of the relative risk function. These calculations and other considerations were used to assess the feasibility of studies of the effects of domestic exposure to radon. A paper on cluster sampling of controls for case-control studies appeared, and work on weighted sampling schemes for case-control studies is in press. A paper is in press on determining the relative efficiencies of alternative procedures for sampling controls from a large cohort, and Poisson regression methods are under development for case-cohort data. Proceedings from a workshop on the effects of errors in exposure measurements on epidemiologic studies have been published, and related work was completed on the effects of assessments of dose-response in the presence of non-differential misclassification of exposures and on the value of additional error-prone exposure data to supplement """"""""gold standard"""""""" data. Papers are in press on inference for attributable risk based on logistic regression as well as on estimates of absolute risk. A paper appeared on the use of receiver-operating curves for the evaluation of diagnostic tests. The accuracy of surrogate and self respondents when reinterviewed after 5 years is assessed in an in press paper, which found both groups of respondents similar in their reliability. Final testing and documentation are underway for a system of computer programs for epidemiologic analyses on the IBM-PC, and a paper was published on the use of the PHGLM Procedure in SAS for specialized types of survival analyses.
