Differential Incarceration by Race-Ethnicity and Mental Health Service Status in the Los Angeles County Jail System

ODR obtained sociodemographic and legal data from the LA County Sheriff’s Department (LASD) for all persons incarcerated in JMH housing on February 14, 2019, yielding a purposive, cross-sectional sample of 5,134 persons. JMH housing consists of high- and moderate-observation units and a certified psychiatric inpatient unit for any incarcerated person on an involuntary hold; residents are persons in the general jail population who are prescribed an array of psychotropic medications by jail psychiatrists (e.g., antidepressants, antipsychotics, mood stabilizers, and anxiolytics). Initial JMH placement is based on a brief mental health screening upon booking. If the initial screen is positive, or if concerning behaviors are observed, the defendant is seen by a clinician (e.g., registered nurse, psychiatric social worker, or clinical psychologist), who determines the defendant’s housing area and necessity for referral to psychiatry. These housing assignments are independent from security levels assigned by custody, and defendants move between JMH housing levels on the basis of reported or observed psychiatric symptoms during their detention.

Age, race-ethnicity, sex, and type of charge or charges on arrest (any felony charge or charges versus misdemeanor or other charge with no felony charges) from the JMH sample were compared with the overall LA County jail population (N=16,975) by using the most recent mean aggregate data reported on the LASD public Web site (6). In contrast with current U.S. Census and epidemiologic conventions, racial-ethnic categories used by LASD include black, Hispanic, white, Chinese, and other. Therefore, these are the racial-ethnic categories used in this study, given the lack of other racial-ethnic data from LASD. Observed proportions for categorical measures in the JMH sample were compared with the reported proportions in the overall jail population by using chi-square tests, with a p≤0.05 level of significance for dichotomous measures. Adjusted chi-square residuals, with Bonferroni correction for multiple comparisons, were examined for post hoc pairwise testing of individual measures within statistically significant nondichotomous variables. The study protocol was approved by the LA County Department of Public Health Institutional Review Board.

Table 1 summarizes sociodemographic and arrest-level data of the JMH sample and the overall jail population. Compared with the overall jail population, the JMH sample was significantly older (with a smaller proportion of persons ages 18–25 and a larger proportion ages ≥45) and featured a significantly higher percentage of black (41% versus 30%) and white (19% versus 15%) persons and a lower percentage of Hispanic persons (35% versus 52%). The proportion of women was slightly larger in the JMH sample, compared with the overall jail population (15% versus 13%), and the proportion charged with a felony or felonies was significantly smaller in the JMH sample (80% versus 91%).

Structural racism and differential criminal sanctioning in the carceral system are increasingly acknowledged in public health discourse (7). In addition, racial inequity in access to and quality of mental health treatment has been well described (8). However, the racially disparate patterns of individuals with mental illness in LA County jails, one of the nation’s largest mental health systems, have not been previously studied. These findings contribute to research on the racially distinct interplay between mental illness and incarceration.

According to the U.S. Census (2018), LA County’s population is 9% non-Hispanic black, 48.6% Latino, and 26.1% non-Hispanic white (9). In our study, disparities by race were evident between the overall jail population and the JMH sample, even when the analyses conservatively adjusted the significance level for multiple-hypothesis testing. Compared with the U.S. Census proportion, black persons were overrepresented in the LA County overall jail population (30%) and more disproportionately overrepresented in the JMH population (41%). Latino individuals were also overrepresented in the overall jail population (52%), whereas they were underrepresented in the JMH sample (35%). White persons were underrepresented in the overall jail population (15%), compared with the U.S. Census data; however, this underrepresentation was less prominent in the JMH sample (19%), which more closely approximated the proportion of white persons in LA County (26%). These racially disparate trends are not explained by differing prevalence in mental illness by race or ethnicity; prior studies have shown a higher prevalence of reported mental illness among white Americans, compared with black and Latino Americans (10). These data also contrast with literature suggesting minimal difference in the prevalence of mental illness by race among incarcerated individuals (10). The racial disparities present in our sample may reflect the potentiating effects of racism at various stages in the carceral and mental health systems.

Compared with white persons, black persons are at increased risk of criminalization (e.g., higher risk of arrest and harsher sanctioning) when committing the same offense (7). Moreover, black individuals are half as likely as white individuals to receive psychiatric treatment for diseases of similar severity (8) and white persons have greater access to and receive better quality of mental health care (8). Jails, in turn, have become de facto safety-net providers for underserved communities (11). Another contributor to the disparities between the general jail population and the JMH sample may be clinicians’ racial biases in diagnosis and treatment. Previous research demonstrated that black Americans have higher than expected rates of schizophrenia diagnoses and are prescribed higher doses of antipsychotics (12), an example of overpathologizing bias. However, affective disorders are underdiagnosed among black Americans, and they are less likely to be prescribed antidepressants (12, 13), an example of minimization bias. Because the JMH sample encompassed defendants receiving any psychotropic medication, with some diagnosed as having a schizophrenia spectrum disorder and some diagnosed as having an affective disorder, clinician bias alone is unlikely to explain the study’s findings.

The underrepresentation of Latinos in the JMH sample is notable and not attributable to varying prevalence of mental illness by race or ethnicity (10). Although both Latino and black Americans experience restricted access to mental health services and poor quality of care (8), their relative representation in the JMH sample was skewed in opposite directions. This finding may be explained by differential rates of homelessness in LA County by race and ethnicity. Vast literature evidences the bidirectional relationship between homelessness and incarceration for those with mental illness, with homeless individuals having significantly higher lifetime arrest rates than persons in the general population (14). In LA County, Latino individuals are at lower risk of homelessness, compared with black individuals (9), consistent with national patterns (10). This may be partially due to greater reliance on social networks for housing among Latinos; the lower rates of homelessness may thus be protective against incarceration. Lack of linguistically appropriate care may also lead to underutilization of psychiatric services within jails, as is true in the community (15).

Finally, despite LASD’s stated policy—that their jails “do not generally retain inmates on misdemeanor charges” (16)—these data suggest that individuals with mental illness charged with misdemeanors are disproportionately jailed. This differential criminalization of mental illness echoes Snowden’s (12) observations that “decisionmakers other than mental health professionals, including . . . police and courts, play an important role in assessing mental illness and in deciding whether troublesome behavior warrants treatment or punishment.”

Limitations of this study include the cross-sectional design, which limited causal inference, and the lack of individual-level data from the overall jail population, restricting the ability to control for confounding variables. Moreover, the overall jail population used for comparison purposes included the JMH sample; data for persons in the JMH sample could not be removed from the larger sample for the comparative analyses. However, the inclusion of persons in the JMH sample would be expected to diminish the magnitude of differences between the two samples, biasing study findings away from rejection of null hypotheses and leading to more conservative conclusions, rather than increasing the probability of type I error. In addition, the JMH sample was clinically heterogeneous, and detailed individual-level clinical data were not available. Finally, racial-ethnic categories used by LASD are limited, may not be congruent with self-identified race-ethnicity, and do not account for within-group ethnic heterogeneity. Nonetheless, these findings underscore ways in which structural racism differentially affects communities of color and reveal the disproportionate jailing of people with mental illness charged with misdemeanors in LA County.

The data reported here can be utilized by jurisdictions that intend to implement equitable diversion strategies. LA County’s Board of Supervisors, for example, recently formed the Alternatives to Incarceration (ATI) Work Group to help guide criminal justice reform. Key among ATI’s recommendations, and consistent with the implications of this study, is to use a racial equity lens when allocating county funds. In this vein, jail diversion can be a mechanism to reduce the racial health disparities exacerbated by incarceration. Because previous studies of jail diversion revealed that diverted defendants were disproportionately white (17), jurisdictions similar to LA County may consider focusing resources on diversion of black defendants.

ODR’s mission, in part, is to divert persons with serious mental disorders out of LA County jails. Pertinent to this study, early findings regarding ODR’s supportive housing program indicate that the majority of program participants were black; for all program participants, the study also demonstrated 6- and 12-month housing stability rates of 91% and 74%, respectively (18). Related to the disproportionate number of defendants in the JMH population charged with minor crimes, ODR operates a diversion program for those found incompetent to stand trial on misdemeanor charges. Since 2015, a total of 1,431 defendants have been released to this program. However, people charged with misdemeanors remain overrepresented in the JMH population, and diversion at earlier intercept points is critical. Expanding community services to provide robust and comprehensive care would likely reduce “quality-of-life” crimes.

Structural racism is deeply embedded in the legal and mental health systems and will not be solved solely by reducing the jail population. Community advocates, practitioners, and scholars have emphasized that overreliance on jail detention must be addressed while simultaneously diverting individuals at earlier intercept points. While working to bolster diversion across intercept points, LA County may consider investing resources in prioritized diversion of black individuals and of individuals with mental disorders who are charged with misdemeanors. Ultimately, all diversion efforts should aim to ensure health equity and reduce disparities. Future research should evaluate diversion programs to ensure equitable practices when considered in their local context. Longitudinal data could elucidate whether diversion across sequential intercept points has an impact on racial disparities in the jail population and the surrounding community.