Factor structure and item response of psychosis symptoms among Kenyan adults.

Background: The aim of this study was to evaluate the construct validity of the psychosis module of the Mini International Neuropsychiatric Interview version 7.0.2 (MINI-7).

Method: We utilized data collected from 2738 participants with a primary psychotic or bipolar disorder. Participants were drawn from two Kenyan sites of a large multi-center neuropsychiatric genetic study. The factor structure of the MINI-7 psychosis items were explored using confirmatory factor analyses (CFA) and Item Response Theory approach, for the full sample and by gender.

Results: The CFA revealed that a 1-factor model provided adequate fit for the MINI-7 psychosis items for the full sample (x = 397.92, df = 35, p < .0001; RMSEA = 0.06; CFI = 0.92; TLI = 0.90) as well as for the female (x = 185.16.92, df = 35, p < .0001; RMSEA = 0.06; CFI = 0.93; TLI = 0.91) and male groups (x = 242.09, df = 35, p < .0001; RMSEA = 0.06; CFI = 0.92; TLI = 0.89). Item thresholds for the full sample, and female and male groups were highest for 'odd beliefs' (-1.42, -1.33, and -1.51 respectively) and lowest for 'visual hallucinations' (-0.03, -0.04, and -0.01 respectively).

Limitations: Our study used a hospital-based population, which may have excluded patients with milder psychotic symptoms. Findings may therefore not be generalizable to the community setting.

Conclusions: Our findings indicate good construct validity of the MINI-7 psychosis module, and provides support for use of the tool in diagnosing psychotic disorders in clinical settings in Kenya.