Alternative Interpretation for the Early Detection of Psychosis Study

To the Editor: In the April issue, the Treatment and Intervention in Psychosis (TIPS) early-detection study reports 10-year results in a manner that overstates the impact of reducing the duration of untreated psychosis (1). The authors dismissed a 50% increase in hospitalization in the treatment group after 5 years as the result of regional policy differences. They did not describe the policy differences or analyze the effects of this impressive confound on the small difference in symptoms, instead claiming to have demonstrated “positive effects on clinical and functional status” (2, 3). They omit hospitalization results altogether at 10 years, despite this being by far the most impressive result at 5 years (1).

Perhaps because at 5 years the researchers reported a nonsignificant advantage in remission for the control group (2), at 10 years they introduce a new recovery metric, based largely on work function, which showed a significant advantage for the treatment group (1). Although they acknowledge a significant attrition bias by 10 years, they do not report that at 5 years there was no difference in work function, or suggest how reducing the duration of untreated psychosis at baseline would not improve work function at 5 years but double work function at 10 years.

The authors reported that the control group achieved independent living significantly more often at the 10-year mark, but dismiss this evidence of worse function in the treatment group, suggesting that independent living is not evidence of recovery because it is not included in the new metric. They do not analyze the possibility that failure to achieve independent living is evidence of poor function (1).