The true incidence of treatment-resistant schizophrenia (TRS) remains uncertain, largely because earlier studies relied on heterogeneous definitions and retrospective assessments. The Treatment Response and Resistance in Psychosis (TRRIP) consensus established standardised diagnostic criteria, but these have rarely been applied prospectively.

To estimate the incidence of TRS prospectively by using operationalised TRRIP criteria within a large community-based trial, and to examine whether baseline demographic or clinical variables predict its emergence.

We applied TRRIP criteria to 1334 participants with chronic schizophrenia enrolled in the Clinical Antipsychotic Trials of Intervention Effectiveness (CATIE) study. TRS was defined as persistent symptoms and functional impairment following two adequate antipsychotic trials, each meeting TRRIP thresholds for duration, dose and adherence. Incidence was estimated with bootstrap resampling, and logistic regression examined associations between baseline variables and TRS onset.

Of 1334 participants, 782 (58.6%) met TRRIP absolute symptom thresholds at baseline; 77 (9.8%) fulfilled full criteria for TRS during follow-up. The incidence rate was 5.68 per 100 person-years (95% CI 4.51–7.02) overall and 9.20 per 100 person-years (95% CI 7.43–11.39) among above-threshold participants. Higher baseline Positive and Negative Syndrome Scale positive and negative subscale scores and greater global illness severity were associated with TRS, although predictive performance was poor.

This first prospective application of TRRIP criteria to a randomised trial data-set provides robust estimates of TRS incidence. Routine clinical and demographic variables have limited predictive value, supporting the need for longitudinal, biologically informed studies that use TRRIP-aligned frameworks.