Use of the Markov model to estimate the waiting times in a modified WHO staging system for HIV infection.

Seroconverting men in the Vancouver Lymphadenopathy--Acquired Immune Deficiency Syndrome Study cohort were studied in this analysis to estimate the waiting times in the different stages of the modified World Health Organization (WHO) staging system for human immunodeficiency virus infection. A time-homogeneous staged Markov model was used, and it was assumed that infected individuals progress irreversibly from Stages I to IV and eventually to death. There were 130 individuals who seroconverted during the study (seroincident) and were included in this analysis. With use of the modified WHO staging system, the estimated mean waiting times for progression to Stage IV were 9.5, 6.5, and 2.0 years from stages I, II, and III, respectively. The estimated survival times were 11.2, 8.2, 3.7, and 1.7 years from Stages I to IV, respectively. Waiting times and survival were not significantly different when the lymphocyte count was used instead of the CD4 cell count in application of the staging system. The application of a staged Markov model to the modified WHO staging system suggests that this is a clinically sensible model. Furthermore, our results confirm that in the absence of CD4 cell counts, lymphocyte counts can be used as an alternative without substantial loss of information.