To the Editor.
—Our recent findings on the relationship between depression and change in CD4 cell count1 and Dr Lyketsos and colleagues'2 failure to replicate our findings have added to the controversy concerning psychological distress and disease progression. Space limitations allow us to address only a few of the issues raised by the differences in articles and by the Editorial by Drs Perry and Fishman.3Perry and Fishman asserted that our significant findings may be a type I error attributable to our smaller sample size. The basis for this assertion is unclear. Smaller sample size diminishes power, thus making it more difficult to find a statistically significant effect (a type II error). Differences in results are therefore likely to be attributable to differences in procedures or in sample characteristics. The procedures of the two studies, though not identical, are similar. The study populations do differ on mean