Criminal Justice Settings as Possible Sites for Early Detection of Psychotic Disorders and Reducing Treatment Delay

The sample included English-speaking patients who were receiving initial treatment for a nonaffective psychotic disorder. Exclusion criteria included known or suspected mental retardation, a Mini-Mental State Examination (13) score of <24, inability to give informed consent, hospitalization for psychosis more than three months before the index hospitalization, and a history of taking antipsychotic medications for more than three months. Most patients had received no treatment before the index hospitalization.

Data were collected as part of an ongoing study of the impact of premorbid cannabis use on the early course of psychotic disorders. Individuals (N=191) were recruited from hospitals in two U.S. cities between August 2008 and April 2012. In Atlanta, Georgia, participants were enrolled at an urban, public-sector hospital (N=137), a suburban crisis stabilization unit (N=33), and a state psychiatric hospital (N=11). In Washington, D.C., individuals were recruited from a private university-affiliated hospital (N=7) and an urban community hospital (N=3). After receiving a complete description of the study and before participation, patients gave written informed consent. The study was approved by all relevant institutional review boards.

The assessment was typically administered during the participant’s hospitalization, after initial stabilization. Demographic information and incarceration history were collected using a structured interview. Interviewers asked, “Have you ever been arrested?” and “Have you ever been incarcerated, or locked up?” Follow-up questions about the date, length, location, and reason for each incarceration were asked. Data were collected for up to six separate incarcerations. When available, one or two informants (family members or friends who had seen the participant regularly during the year before hospitalization) were asked to provide the same information (such informant data were available for 128 of 191 participants).

After data were collected from all available sources, the research team met to derive consensus-based best estimates for several variables, including the number of arrests and incarcerations. In the few cases in which there were irreconcilable discrepancies in the reports of the participant and informants or the chart information, data were regarded as missing. If a participant reported having been incarcerated multiple times and was able to provide information about specific incarcerations that occurred after psychotic symptoms emerged but could not recall the dates of earlier incarcerations, the occurrence of an incarceration during the duration of untreated psychosis was recorded but the incarceration variables for the prodrome and premorbid period were regarded as missing. Research assessors asked participants to report the charges for which they served time and recorded these verbatim. Charges were later classified as violent (for example, assault, robbery, arson, any sexual offense, illegal threats, or intimidation) (14) or nonviolent (for example, possession of illicit substances, driving violations, shoplifting, jaywalking, loitering, or child neglect).

Data on diagnoses and symptoms were collected with widely used, valid, and reliable instruments. The Positive and Negative Syndrome Scale (PANSS) was employed for current and past-month positive, negative, and general psychopathology symptoms (15). After an in-depth, semistructured interview, trained research assessors rated the severity of 30 symptoms on a scale from 1, absent, to 7, extreme. The PANSS has demonstrated criterion-related validity with concurrent and antecedent measures, predictive validity, and utility for dimensional and typological assessment (15).

The ages at onset of prodromal symptoms and of psychotic symptoms were determined in a standardized and rigorous method that used the Symptom Onset in Schizophrenia (SOS) inventory (16). Participants’ recollection of dates was enhanced through cross-referencing with memorable life events. Psychiatric diagnoses were determined with the Structured Clinical Interview for DSM-IV Axis I Disorders (SCID) (17), which was completed after the in-depth PANSS and SOS interviews and after information was collected from participants’ medical charts and informants.

Premorbid academic functioning was assessed with the Premorbid Adjustment Scale (18), rated by trained interviewers after a semistructured interview. Participants’ scores during childhood (≤11 years), early adolescence (12–15 years), and late adolescence (16–18 years) were based on their reported grades, social connections, and participation in extracurricular activities. Possible scores range from 0 to 6, with higher scores indicating poorer academic functioning. To conservatively ensure that truly premorbid, rather than prodromal, functioning was assessed, data for a time period were excluded if prodromal or psychotic symptoms emerged within one year of the period in question.

Independent-samples Student’s t tests were used to examine bivariate associations between incarceration during the duration of untreated psychosis, prodrome, or premorbid period and all linear variables of interest. Chi square tests were applied for categorical variables. On the basis of bivariate analyses, two predictors of poor insight (history of incarceration during the duration of untreated psychosis and severity of hallucinations) were entered into a stepwise multiple linear regression model. Analyses were conducted in SPSS, version 17.0, with p<.05 in two-tailed tests establishing statistical significance. Because the variables for duration of untreated psychosis and duration of untreated illness were highly right-skewed, statistical tests were performed with a log transformation of the variable. For interpretability, medians were reported rather than the log-transformed means.

As shown in Table 1, this sample of 191 first-episode patients consisted primarily of single or never married (88%), African-American (85%), young-adult (mean age of 24.2) males (73%). A substantial portion (42%) did not graduate from high school, even though all were at least 18 years old. In the month before hospitalization, participants lived with family (66%), a friend or romantic partner (9%), or alone (8%), or they were homeless or in a structured living arrangement (17%). SCID diagnoses included schizophrenia (59%), psychotic disorder not otherwise specified (15%), schizoaffective disorder (12%), schizophreniform disorder (11%), delusional disorder (2%), and brief psychotic disorder (1%). The mean age at onset of psychosis was 21.2 years, and the median duration of untreated psychosis was 45 weeks.

Most participants (70%) reported a history of arrest, and a majority (59%) had been incarcerated at some point before the index hospitalization. As shown in Table 2, incarcerations occurred during the premorbid period for 22%, during the prodrome for 13%, and during the duration of untreated psychosis for 37%. Among those who had been incarcerated during the duration of untreated psychosis, the mean number of incarcerations during this period was 2.0, the median number of days detained was 30.5, and three-fourths (76% of the 62 participants with available data on this variable) were detained for a nonviolent, often petty, crime. In a subgroup of participants incarcerated during the duration of untreated psychosis for whom help-seeking behaviors were recorded (N=48), 33% reported having any health care or help-seeking contacts in the period after prodromal or psychotic symptoms emerged but before their incarceration.

Comparisons of individuals incarcerated during the premorbid period, prodrome, or duration of untreated psychosis with individuals who were not incarcerated are presented in Table 3. The median treatment delay for the group incarcerated during their duration of untreated psychosis was more than two years longer than the delay for participants with no incarceration during this period (130.5 versus 12.0 weeks, p<.001). The proportion of males in the group with incarcerations during this period was larger and the severity of positive symptoms at initial hospitalization was greater than in the group with no incarceration during this period. A post hoc analysis of specific positive symptoms at the time of initial hospitalization indicated greater hallucination severity in the group with incarcerations during their treatment delay (PANSS score of 4.6 versus 4.0, p=.010); however, the severity of delusions and disorganization did not differ between the groups.

Participants who had been incarcerated after the onset of prodromal or frank psychotic symptoms had better illness insight than those who had not been incarcerated during this period. To clarify the associations between insight at the time of initial hospitalization, hallucination severity, and incarceration history, a multiple linear regression model was conducted with a history of incarceration during the duration of untreated psychosis (yes or no) and PANSS hallucinations score entered as predictors of participants’ insight score. After stepwise elimination, hallucinations remained a statistically significant predictor of insight (R²=.079, df=172, p<.001)—although accounting for little of the variance in insight—and incarceration history was no longer statistically significant. Individuals who were incarcerated during the prodrome did not have a longer duration of untreated psychosis than those who were not incarcerated during the prodrome; however, they did have a longer duration of untreated illness (inclusive of the prodrome) and better insight at the time of initial hospitalization. Poorer premorbid academic adjustment was associated with incarceration during the premorbid period and during the duration of untreated psychosis (Table 3).

The past prevalence of incarceration in this sample was very high. More than one-third of participants (37%) were incarcerated during the duration of untreated psychosis, spending an average of one month in a criminal justice setting (most often a city or county jail or a juvenile justice setting). Those who had been incarcerated during this period had a much longer median duration of untreated psychosis (approximately two years, compared with three months for those with no incarceration during this period). Compared with those with no incarcerations, those with incarcerations during the duration of untreated psychosis were more likely to be male, had more severe positive symptoms, and had better insight. At their first incarceration, two-thirds had previously had no help-seeking contacts after the onset of prodromal or frank psychotic symptoms.

Several factors might account for the high incarceration rate and for its association with a longer treatment delay. Of note, the sample comprised mostly young African-American males (63%). It is well known that this demographic group is disproportionately arrested and incarcerated—for reasons that are complex and beyond the scope of this report (19). It is possible that involvement with the criminal justice system delays treatment or that individuals with untreated psychosis are more likely to offend. It is also possible that symptom-related, cognitive, socioenvironmental, or other factors explain both treatment delay and risk of incarceration. It is noteworthy that few individuals in the sample were incarcerated for a violent crime, despite evidence suggesting that a substantial portion (35%) of individuals with first-episode psychosis commit some sort of violence (often minor) before their first contact with mental health providers (20). It is possible that individuals who exhibited more severe violence were excluded from our study because they were incarcerated for longer periods or were identified and treated for psychosis while still in jail or prison.

As motivation to reduce treatment delay for psychosis in the United States increases, we must consider strategies for reaching affected individuals through existing institutions. Of 11 interventions designed to reduce duration of untreated psychosis that were recently reviewed, only two were successful: the TIPS campaign in Norway and the EPIP Project in Singapore (21). Both were multifocused campaigns, targeting the general public and specific groups to improve attitudes toward and knowledge of psychosis (21). Although neither focused on criminal justice settings, the EPIP Project found that during its campaign, patients were less likely to be referred to mental health services by law enforcement, perhaps as a result of increased self-referral. Pathways to care traversed by first-episode patients and their families should be a focus of future research, as well as interventions to decrease negative experiences (such as contact with police), because these initial experiences can shape individuals’ longer-term view of and engagement in services (21). Our data suggest that an outreach campaign to reduce treatment delay should include criminal justice settings. Patients in our sample were largely unemployed, not in school, and not receiving primary care services. Prisons and jails may be among the few institutions that come into contact with some individuals who have undiagnosed psychosis, especially those from low-income and socially disadvantaged communities.

The deleterious effects of incarcerating young people with serious mental illnesses and the moral, legal, and economic arguments for referring them to medical treatment are well established (22). However, creating an early intervention service for psychosis in jail and prison settings would present logistical challenges and ethical issues. Although prisons are coordinated at the state and federal levels, county and city jails, where a vast majority of the individuals who had been incarcerated in this sample had been detained, are operated and regulated at a local level in the United States, and each is governed differently. Current policies regarding involuntary treatment for mental illnesses require evidence of dangerousness to warrant involuntary hospitalization. Yet these laws appear to extend the duration of untreated psychosis (23), which is marked by economic, social, and educational decline, as well as by increased risk of harming others (24,25). Criminal offenses alone do not warrant compulsory treatment, and psychiatric symptoms do not automatically negate responsibility for illegal actions. However, it may be beneficial to the individuals in question and society at large if offenders are screened for mental illnesses and referred or diverted to services.

Prebooking jail diversion may follow the crisis intervention team model, in which officers are trained to recognize signs of mental illnesses and divert individuals to services rather than jail when appropriate (26), or the assisted outpatient treatment model, which involves court-ordered treatment for individuals who have a history of nonadherence, as a condition of their remaining in the community (27). In addition, although early intervention for individuals with undetected and untreated psychosis is warranted, this does not detract from the overwhelming need to improve services for individuals with chronic serious mental illnesses in criminal justice settings.

Several inherent methodological limitations must be acknowledged. First, although the research team has expertise in interview-based, retrospective measurement of key constructs preceding treatment seeking, data on arrests and incarcerations were obtained through self-report, which is subject to recall error. However, arrests and incarcerations are likely to have been underreported rather than overreported, and it is unlikely that a systematic bias influenced hypothesis tests in one direction over the other. Although state-level administrative criminal justice data could have been used, only arrests that warranted fingerprinting and that occurred within the state would have been reported, and details of incarcerations would not have been reliably available. Second, given our sample’s sociodemographic characteristics (for example, predominantly African American, low-income, and socially disadvantaged) and clinical characteristics (hospitalized mainly in public-sector settings), generalizability is limited. Nonetheless, internal validity is likely high, and our findings in this understudied group contribute to new and important discussions about the role of criminal justice settings in early detection and early intervention efforts. Third, given the dearth of research on prior incarcerations among young people with emerging psychotic disorders, our hypotheses were grounded largely in our own experience and suppositions rather than on previous findings. However, this somewhat exploratory analysis sets the stage for further work.

Early detection and intervention for psychotic disorders has become a focus of research, clinical programs, and policy in recent years. Early intervention is now thought by many to represent an ethical imperative (to reduce the psychosocial impairment that comes with untreated psychosis), and it may also hold promise for improving outcomes (given some evidence that early intervention has long-lasting beneficial effects). We found a high prevalence of prior incarcerations among patients experiencing a first episode of psychosis (11), these incarcerations were associated with longer treatment delays (12), and more than a third of the individuals in our sample were incarcerated during their duration of untreated psychosis. These findings suggest that discussions of early intervention must extend to criminal justice settings. Young people with an emerging psychotic disorder, who are in a “critical period” (28) in terms of early initiation of evidence-based pharmacological and psychosocial treatments, are hidden not just in private homes, classrooms, and primary care settings but also in our jails and prisons. They have a right to receive care, and we have an obligation to develop effective means for early identification and engagement in treatment in this as in other settings.

Research reported in this article was supported by the National Institute of Mental Health of the National Institutes of Health under award R01MH081011. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

The authors report no competing interests.