These data are from the Terry Beirn Community Programs for Clinical 
Research on AIDS, which is a network of community-based (i.e. not
research hospital based) physicians involved in AIDS-related research.
Jim Neaton here in Biostatistics is the head of the Statistical and 
Data Management Center for the CPCRA.

The CPCRA 010 Clinical Trial was a randomized, prospective, double-blind study
comparing fluconazole vs. placebo for primary and secondary prophylaxis
of mucosal candidiasis in HIV-infected women. The repeated response
measurements are binary indicators for a positive vaginal culture for the candidiasis 
fungus. There are many baseline variables available, including age, anti-
retroviral use, diagnosis of vaginal or of oral candidasis at baseline, 
intravenous drug use, progression of disease at baseline, race (white/non-white), 
treatment (fluconazole vs placebo), and clinical unit.

The primary results paper, which tracked time until occurrence of
the first disease event (survival analysis, not longitudinal analysis), is 
Annals of Internal Medicine 1997;126;689-696.

NOTE: The file fungal.dat is an ascii file of the data where missing values are
denoted by a period.  The file fungal.s.dat is an ascii file where missing
values are denoted by NA.