Frequently Asked Questions

In this study, 41 of 377 adults (11%) with first-episode psychosis reported criminal legal involvement in the month before baseline assessment. That included 12 people who had been arrested, 3 who had been incarcerated, and 31 who were on probation or parole.

Over the 2-year follow-up period, 52 of 378 participants (14%) had new criminal legal involvement after entering treatment. These events included 31 participants who reported arrest and 41 who reported incarceration; the study defined follow-up criminal legal involvement as any arrest or incarceration reported during follow-up or clinician-reported dropout due to incarceration.

Yes. Participants with criminal legal involvement at baseline were about 3 times more likely to have criminal legal involvement during follow-up than those without baseline involvement. Specifically, 13 of 39 participants (33%) with baseline criminal legal involvement had later involvement, compared with 39 of 335 participants (12%) without baseline involvement (c7b251 = 13.73, p < .001).

Yes. Participants with criminal legal involvement at baseline were much more likely to drop out due to incarceration than those without baseline involvement: 7 of 39 (18%) versus 9 of 374 (2%) (c7b251 = 13.63, p < .001). Among the 16 participants reported to have dropped out because of incarceration, 14 (88%) did so within the first 6 months of treatment.

Before treatment entry, recent criminal legal involvement was associated with lower educational attainment, longer duration of untreated psychosis, lifetime alcohol use disorder, lifetime noncannabis drug use disorder, and higher PANSS excitement scores in univariate analyses. In the multivariable model, lifetime alcohol use disorder and lifetime other drug use disorder remained the significant unique predictors of baseline criminal legal involvement.

The study also found that participants with longer duration of untreated psychosis were more likely to have criminal legal involvement than those with shorter duration of untreated psychosis: 16% versus 6%.

After adjusting for baseline criminal legal involvement, univariate predictors of later criminal legal involvement included schizophrenia diagnosis, lifetime history of noncannabis drug use disorder, alcohol use in the last 30 days at baseline, cannabis use in the last 30 days at baseline, and higher baseline PANSS positive and excitement symptom scores. In the multivariable baseline model, cannabis use in the month before baseline was the only predictor that remained significant.

In post hoc longitudinal analyses using averaged follow-up clinical measures, higher average positive symptoms, higher average excitement symptoms, greater alcohol use, and greater cannabis use were each associated with criminal legal involvement during follow-up. When those longitudinal variables were modeled together, only average excitement symptoms remained significantly associated with follow-up criminal legal involvement (B = .20, SE = .09, p = .02, OR = 1.22, 95% CI = 1.031.44).

No. Criminal legal involvement rates were very similar in the two treatment groups both at baseline and over follow-up. During follow-up, rates were 13.88% in NAVIGATE and 13.61% in Community Care, and after adjusting for treatment site and baseline criminal legal involvement, treatment group was not significantly associated with follow-up criminal legal involvement (B = .02, SE = .38, p = .96, OR = 1.02, 95% CI = 0.551.87).

Criminal legal involvement was measured with the SURF-Q at baseline and every 3 months. Participants reported any arrest, incarceration, and probation or parole in the past 30 days, although the questionnaire was only administered every 3 months.

For follow-up analyses, the study defined new criminal legal involvement as any reported arrest or incarceration on the SURF-Q or clinician-reported dropout due to incarceration. Probation during follow-up was excluded from the definition of new criminal legal involvement because it could have reflected an offense already present at baseline.

The study has several important limitations. It did not collect a comprehensive history of criminal legal involvement before study entry, did not record actual criminal or aggressive behavior, charges, or convictions, and assessed criminal legal involvement only as past-month events sampled every 3 months, which the authors noted likely underestimates true criminal legal involvement in the sample.

The study also did not collect data on the types of charges, such as violent or substance-related offenses, and lacked detailed information on criminal legal involvement before the onset of psychosis.