A Multisite Study of the Prevalence of Serious Mental Illness, PTSD, and Substance Use Disorders of Women in Jail

A total of 491 preconviction (51%) and postconviction (49%) women in jails in Colorado, Idaho, metropolitan Washington, D.C. (Maryland and Virginia), and South Carolina participated in structured diagnostic interviews. Interviews were conducted at nine jails across the four regions between June 2011 and March 2012. Facilities ranged in size, housing from 25 to 300 women. Forty-three percent (N=211) of the participants were housed in jails in rural or nonmetropolitan, low-population areas.

The women ranged in age from 17 to 62, with a mean±SD age of 35.00±10.65. Most (75%, N=369) had children under the age of 18. Before their incarceration, 33% (N=162) were employed full-time, whereas 46% (N=228) were unemployed. Approximately one-quarter reported attending some high school (26%, N=129), one-third had completed high school or obtained their GED (34%, N=165), and 36% (N=178) reported attending at least some college. Women endorsed the following ethnic identities: white or Caucasian (38%, N=184), African American or black (37%, N=183), Latina (15%, N=74), American Indian (4%, N=21), and other (6%, N=29). One-quarter (25%, N=119) were first-time offenders, and 16% (N=79) were charged with or convicted of a violent crime (including assault, battery, nonprostitutional sex offense, manslaughter, or homicide). One in five participants was incarcerated for two weeks or less (20%, N=98), and 49% (N=242) were incarcerated for fewer than five weeks at the time of the interview.

The Composite International Diagnostic Interview (CIDI) is a structured interview widely used by nonclinician interviewers (13). Interviewers are trained and supervised on how to proceed through the interview (how to follow skip patterns in questioning), but the interview is structured to eliminate the need for clinical judgments, thus allowing use across disciplines. The CIDI assesses lifetime and 12-month criteria. Several national prevalence studies have utilized the CIDI, allowing comparisons of prevalence rates between incarcerated and nonincarcerated populations.

The CIDI paper-and-pencil (PAPI version 7) modules for major depressive disorder, bipolar disorder, PTSD, and substance use disorders were used in this study. The CIDI screening items for serious mental illnesses (major depressive disorder, bipolar disorder, and schizophrenia spectrum disorders) were administered to all participants, and corresponding modules were administered to participants who screened positively. The PTSD and substance use disorders modules were administered to all participants. The CIDI has extensive screening items for psychotic spectrum disorders but does not have a module to assess all relevant criteria. Thus participants who screened positively on the CIDI psychotic items also were administered an adapted version of the psychotic disorders module of the Structured Clinical Interview for DSM disorders (SCID-I), Research Version, to assess schizophrenia, schizophreniform, or schizoaffective disorder (labeled in this study as schizophrenia spectrum disorders) (14). Adapted items included assessment of delusional thinking and observations of grossly inappropriate affect or behavior as well as items to assess duration of symptoms and overlap in mood and psychotic symptoms and to rule out psychosis attributable to other conditions.

Items from the Sheehan Disability Scale (15) are integrated into each CIDI module to assess impairment. Participants who endorsed problems in the past year within a particular module (depression, for example) were then asked to indicate the extent to which those problems, when at their worst in the past 12 months, interfered with their functioning on a scale of 0 to 10, where a 0 indicated no interference in functioning; 1–3, mild interference; 4–6, moderate interference; 7–9, severe interference; and 10, very severe interference. Participants indicated degree of functional impairment in their home management (including grocery shopping and cleaning), ability to work, ability to form and maintain close relationships, and social life. Average current impairment across areas is reported in this study.

Interviewers received extensive training and supervision, beginning with a DVD that demonstrated interview administration for standardization purposes across sites. Interviewers also had the opportunity to watch an experienced investigator conduct an interview and to be observed interviewing. Supervision occurred within and across sites through regular meetings. Institutional review board (IRB) approval was obtained from each academic institution and from jail or community IRBs as needed. Lists of inmate names were obtained from participating facilities and updated regularly throughout data collection. Offenders were randomly selected, called out, and invited to participate in a study of women’s pathways to jail, mental health, and life experiences. Inmates who were unavailable at the time their name was selected were invited on the subsequent interview date to participate in the study. Inmates who declined were removed from lists. Once informed consent to participate was obtained, we conducted interviews in private, enclosed rooms. Compensation for the interview varied by jail because of local regulations and included a snack, $10 deposited into the individual’s commissary account, or funds applied toward the purchase of materials for inmates’ use (such as for self-help books). Interviews lasted from one hour to six hours, or a mean±SD of 1.95±.91 hours.

Women who declined to participate (N=142 out of 633) did not differ from participants by age or offense type (violent versus nonviolent offense). Women who declined differed significantly by ethnicity. Individuals who identified themselves as African American (79% accepted) or Latina (85% accepted) participated at rates similar to whites (75% accepted) and at significantly higher rates than the small number of randomly selected American Indians (53% accepted). Also, individuals in the two jails where compensation was limited to a donation to a general fund declined at significantly higher rates (50%). A total of 15 women were excluded from the study: five were excluded because of threat of violence toward others, and five more were excluded because of acute distress at the time of the invitation to interview. Finally, an additional five were excluded because of IRB restrictions in one of the four regions prohibiting interviews with pre-sentence offenders charged with homicide, first-degree assault, or felony sex charges. Although this restriction was in place, it is important to note that overall participation by individuals who committed violent crimes was not lower in this region.

The prevalence of mental disorders in the full sample was high; 91% (N=446) met lifetime criteria, and 70% (N=343) met 12-month or current criteria for at least one disorder. Notably, 43% met lifetime and 32% met current criteria for a serious mental illness (Table 1), including major depressive disorder (28% lifetime, N=137; 22% current, N=109), bipolar disorder (15% lifetime, N=71; 8% current, N=41) and schizophrenia spectrum disorders (4% lifetime, N=21). Substance use disorders were the most commonly occurring lifetime (82%) and current disorders (53%). Lifetime and current PTSD rates also were high (53% and 29%, respectively).

There were no significant differences in serious mental illness, PTSD, or substance use disorders among participants in rural and urban locations. There also were few differences among individuals in different jails within the same regions (Table 1). In contrast, there were significant regional differences, with participants in the metropolitan D.C. area meeting criteria for serious mental illness significantly less frequently than those in Idaho and Colorado (χ²=17.13, df=3, p<.001). Idaho participants also met criteria for PTSD (χ²=24.82, df=3, p<.001) and a substance use disorder (χ²=16.57, df=3, p<.001) significantly more often than participants in the metropolitan D.C. area. Although it is not possible to determine with certainty the reason for these regional differences, our findings are similar to those from a recent national study. A 2011 Substance Abuse and Mental Health Services Administration report assessed serious mental illness rates by state and found 12-month rates of 5.8% in Idaho, 5.2% in Colorado, 4.1% in South Carolina, 3.9% in Maryland, and 3.6% in Virginia (15). In particular, that report notes that Idaho residents reported among the highest rates of serious mental illness in the country, whereas residents of Maryland and Virginia were among those reporting the lowest rates nationwide.

Participants attributed notable impairment in functioning in the past 12 months to mental health and substance use problems. For 152 participants who indicated problems associated with their serious mental illness in the past year, the average impairment was rated 6.21±2.81. However, 45% of these participants (N=68) rated their average impairment a 7 or higher, suggesting severe levels of impairment. Similarly, for 130 individuals who endorsed problems as a result of PTSD symptoms in the past year, the mean rating was 6.17±2.65, whereas 42% of these participants (N=55) reported an average of 7 or greater functional impairment. Finally, 30% (N=79) of the 263 individuals who reported difficulties associated with substance use in the past year indicated severe problems in functioning, whereas the average impairment in functioning was 4.94±3.37.

Many of the incarcerated women met criteria for multiple disorders. Lifetime rates of comorbidity were high: 29% (N=142) met criteria for lifetime serious mental illness and PTSD, 38% (N=187) met criteria for lifetime serious mental illness and a co-occurring substance use disorder, and about one in four (26%, N=127) met criteria for all three in their lifetime. In the past 12 months, 20% (N=98) of participants met criteria for a substance use disorder and serious mental illness, 14% (N=70) of participants met criteria for both PTSD and a serious mental illness, and 9% (N=45) met criteria for all three. Finally, 46% (N=225) of the sample met criteria for lifetime PTSD and a substance use disorder, whereas 18% (N=90) met criteria for both in the past 12 months.