Stephen J.LurieMD, PhD, Senior EditorIndividualAuthorJody W.ZylkeMD, Contributing EditorIndividualAuthor
In Reply: In response to Dr Dat, the estimated distribution of time until HIV progression is presented in our article and includes the distribution estimable from the data. Median time to progression was not reached. Regarding Figure 5 A, one must be careful not to overinterpret individual points on such graphs.1 The lack of statistical significance means that the fluctuations are probably due to chance. The precision of the curves is considerably less at those times when only 10% of the patients (who constituted the subset) were sampled. The baseline characteristics of the subset of 250 patients are the same as the entire group, and their response was similar to the whole group, as we stated in our article. The lack of a consistent viral load effect of HIV-1 Immunogen is well documented in the article. We strongly disagree with Dat regarding the statistical method used to analyze the data. A slope metric, plotted for each individual, was the prespecified statistical method for this study and should not be replaced with other less precise methods, or with methods selected on the basis of the results that they yield.
Kahn JO, Lagakos S. Outcomes of a Trial of HIV-1 Immunogen in Patients With HIV Infection—Reply. JAMA. 2001;285(17):2191-2195. doi:10.1001/jama.285.17.2191