The fact that Americans from racial and ethnic minority groups have had disparate experiences with mental health care, including disparities in access, has been reported as a major concern and a needed area for additional study (1). In a press release about his 2001 report entitled Mental Health: Culture, Race, and Ethnicity, U.S. Surgeon General David Satcher stated, "While mental disorders may touch all Americans either directly or indirectly, all do not have equal access to treatment and services. The failure to address these inequities is being played out in human and economic terms across the nation—on our streets, in homeless shelters, public health institutions, prisons, and jails" (2).
With the number of available civil psychiatric beds in the public health system shrinking, many persons in need of hospitalization ultimately find their way into beds that are operated by state forensic mental health systems. These systems are at a crossroads between the criminal justice system and public-sector psychiatric systems, both of which are susceptible to racial and ethnic influences that lead to the differential treatment of individuals. The central missions of state forensic mental health systems are the forensic evaluation and clinical treatment of pretrial defendants as well as the provision of information to courts. The evaluations play a crucial role in ensuring due process for criminal defendants with mental illness. Given the importance of these missions, it is critical to determine whether race and ethnicity play a role in decisions that affect how forensic evaluees are managed within forensic mental health systems. However, these factors have not been studied in depth. One aspect in particular warrants attention. Although it is current practice to provide civil mental health services in the least restrictive setting, the management of criminally charged forensic evaluees need not follow this dictum. Our study examined practices in one jurisdiction to assess the extent to which race and ethnicity affected disposition outcomes for pretrial defendants.
Economic and social factors, in combination with the debilitating nature of serious mental illness and clinicians' perceptions, may contribute to differences in dispositional outcomes for persons from ethnic and racial minority groups. In a study that examined racial and ethnic disparities in admissions to public and private psychiatric hospitals, nonwhites were found to be more likely than whites to be treated in public hospitals, rather than private hospitals (3). In a study from Washington State, persons with mental illness who were from locally predominant ethnic minority groups were sent to prison, rather than to a state psychiatric facility, more frequently than persons from non-minority groups (4). Latinos, African Americans, and non-Latino whites were also shown to differ in their mental health service use patterns (5). Specifically, poor Latinos had less access to mental health care than poor non-Latino whites. Similarly, African Americans who were not classified as poor appeared less likely than whites to receive mental health services.
Clinical decisions may also be influenced by racial and ethnic factors. For example, several studies have shown that blacks receive diagnoses of certain serious mental illnesses more frequently than whites (6,7,8,9,10). Furthermore, Iwamasa and colleagues (11) found that diagnoses of antisocial and paranoid personality disorder were more likely to be assigned to African Americans than to European Americans. A recent study found that Latinos were disproportionately given a diagnosis of major depression, compared with European Americans, even though Latinos tended to have greater numbers of self-reported psychotic symptoms (10). The study also found that African Americans with fewer self-reported psychotic symptoms were more likely to receive a diagnosis of schizophrenia. These findings raised further questions about disparities in clinicians' diagnoses across racial and ethnic groups. The findings may reflect culturally based differences in symptom presentations (7,8,12,13,14,15). In contrast, Neighbors and colleagues (16) suggested that the impact of race and ethnicity on psychiatric diagnosis could result from clinicians' different interpretations of symptoms exhibited by blacks and whites. This disparity may be especially true when clinicians are unfamiliar with the belief systems of cultural minority groups (17). These clinical factors may lead to higher rates of diagnosis of serious mental illness and to higher hospitalization rates among persons from ethnic minority groups, both of which have been found in psychiatric emergency services (18,19). Racial and ethnic differences among psychiatric inpatients may also affect length of stay (20). In clinical settings, it is therefore important to consider how racial and ethnic factors may influence diagnostic assessments and disposition recommendations.
As indicated above, the state forensic mental health system occupies a unique niche in the spectrum of psychiatric services. The role of race and ethnicity has not been extensively explored in this area, but opportunities abound for these factors to affect individuals who are processed through this system. Before a person is referred to a forensic mental health bed, they have already been through several decision points in the criminal justice system (21,22). Many factors are involved in these decisions, but previous studies suggest that race and ethnicity play some role. For example, a comparison of white and nonwhite defendants who were evaluated for competence to stand trial showed that the number of nonwhite defendants who have been found incompetent to stand trial increased disproportionately compared with whites since the beginning of the deinstitutionalization period (23). Furthermore, after the deinstitutionalization period, among defendants who were found incompetent to stand trial, nonwhites typically had a higher number of previous arrests and hospitalizations than whites.
In forensic settings, clinicians often play a significant role in advising courts about defendants, which influences key decisions about defendants' receipt of mental health services and legal outcomes. Thus any racial or ethnic bias among clinicians could further affect the proportion of persons from ethnic minority groups who are found in such settings.
Our study examined the relationship between race and ethnicity and dispositions in forensic mental health evaluations that were conducted in Massachusetts court clinics. In our study, race and ethnicity were identified by clinicians. Because research suggests that defendants who are from racial and ethnic minority groups are more likely to be incarcerated than hospitalized and to receive diagnoses of serious mental illness and possibly antisocial personality disorder, we hypothesized that defendants from racial and ethnic minority groups would be more likely than whites to be referred for inpatient evaluation and that defendants from racial and ethnic minority groups would be more likely than whites to have the evaluation occur in a strict-security inpatient setting rather than in the non-correctional mental health system. This measurement is especially significant in light of research showing that persons with mental illness who are from an ethnic minority group may be more likely to be referred to the criminal justice system than the mental health system.
All district and superior courts in the Commonwealth of Massachusetts are staffed by court clinicians, including psychologists, psychiatrists, and social workers. Whenever a defendant's mental state is at issue, parties to the legal proceeding may raise questions about the defendant's competence to stand trial and his or her criminal responsibility. These questions relate to the defendant's current mental state and the defendant's mental state at the time of the alleged offense. Once these questions have been raised, defendants are referred to a court clinic for a screening evaluation that is conducted by designated forensic psychologists and psychiatrists, who have been qualified by the Massachusetts Department of Mental Health to perform these evaluations (24,25). Dispositional recommendations are offered on the basis of clinical variables, the nature of the charges, and the defendant's history.
The screening evaluations produce four possible dispositional outcomes: no further evaluation; further evaluation on an outpatient basis; further inpatient evaluation in a hospital under the auspices of the Massachusetts Department of Mental Health; or, for males only, further evaluation in a strict-security setting.
In Massachusetts, male defendants who are referred for inpatient evaluation within a strict-security setting are sent to Bridgewater State Hospital, a 300-bed facility operated by the Department of Correction. Strict security is not defined by statute, nor is it clearly articulated in case law. Generally, the severity of a defendant's offense and assessment of current risk of violence and escape are considered when a recommendation is made to the court about the need for strict security. Although the ultimate decision rests with the judge, the courts most often concur with the court clinicians' dispositional recommendations. Both clinicians and judges have substantial latitude in the decision process about strict security.
Data were obtained from a computerized record system that was maintained by the Massachusetts Department of Mental Health. The records were based on encounter forms that were completed by court clinicians after each examination of a defendant who was referred for a court-based screening evaluation of competence to stand trial or for evaluation of criminal responsibility. Included in these data are the defendants' age, gender, criminal charges, and diagnosis as well as either race or ethnicity or both race and ethnicity. The terms "race" and "ethnicity" are not specifically defined on the encounter forms, but data on this variable are assigned by court clinicians on the basis of a broad list of possible racial and ethnic categories, with a column marked "other" in the event that the defendant does not appear to fit into any identified category. Data on disposition decisions are also recorded, including whether the defendant was sent for an inpatient evaluation in a facility of the Massachusetts Department of Mental Health or in a strict-security facility (Bridgewater State Hospital). The data used in our study cover referrals from July 1, 1994, through June 30, 2001.
Data used in our study were obtained in a deidentified, archived format. The central office research review committee of the Massachusetts Department of Mental Health and the committee on the protection of human subjects in research of the University of Massachusetts Medical School both reviewed the nature of the study and deemed it exempt from review.
Two sets of logistic regression models were developed to assess the relationship between the defendants' race and ethnicity and the likelihood that they would be referred for inpatient evaluation and, if they were referred, the likelihood that they would be evaluated within a strict-security facility. To test whether race and ethnicity influenced a defendant's disposition, we needed to be able to adjust for two other factors that could affect this decision—the severity of the defendant's offense and his or her diagnosis.
Criminal charges were grouped a priori by two of the authors (DP and IP) into three levels of increasing severity that were incorporated into the model as dummy variables. Because no official categorization of the severity of the criminal charge exists, the authors used a process that was similar to that used in clinical evaluations to estimate the level of severity and consider risk. Level 1 included serious crimes of violence. Level 2 included crimes that were not considered overtly seriously violent but that placed others at some risk of injury—for example, breaking and entering at night; nonserious crimes involving weapons; sexual crimes against children, elders, or disabled persons; and indecent assault and battery. Level 3 crimes included offenses that were related to property, finances, and vagrancy. Level 1 was used as the comparison category in the regression models.
Diagnostic categories that were used in the model included schizophrenia and psychoses, bipolar disorder, depression, personality disorder, and substance use disorder; schizophrenia and psychoses were used as the comparison category. Race and ethnicity variables were black (which included African and African American categorizations), Hispanic, and white. White was used as the comparison group. Age was included in all models as a continuous variable, and gender was included in the first model.
For general background purposes, model 1 was developed to include court clinic-referred defendants of both genders. This model assessed the effect of race and ethnicity on disposition decisions in the general defendant population. However, because Bridgewater State Hospital accepts only male patients, two additional hospitalization models (models 2 and 3) were constructed that included only men. In model 2 the dependent variable measured whether or not the defendant was referred for an inpatient evaluation. In model 3 the dependent variable measured whether the defendant was sent for an inpatient evaluation in a setting of strict security.
As t1 shows, of 12,289 cases, 3,892 defendants (31.7 percent) were referred for inpatient evaluation. Additional analyses showed that 32.4 percent of men (N=3,224) were referred by the court clinic for inpatient evaluation. Of these men, black defendants had the highest referral rate (666 defendants, or 36.8 percent of the 1,808 black male defendants evaluated) and Hispanics the lowest (397 defendants, or 26.7 percent of the 1,437 Hispanic male defendants evaluated). Whites had a referral rate of 32.2 percent (1,978 defendants of the 6,151 white male defendants evaluated). Thirty-eight percent of men who were referred for inpatient evaluation (N=1,225) were assigned to a strict-security facility for their evaluation. Black men (246 defendants, or 36.9 percent of the 666 black men sent for inpatient evaluation) and Hispanic men (150 defendants, or 37.8 percent of the 397 Hispanic men sent for inpatient evaluation) were more likely than white men (510 defendants, or 25.8 percent of the 1,978 white men sent for inpatient evaluation) to be referred to a strict-security facility; the rates for black and Hispanic men were roughly equivalent.
Odds ratios (ORs)—which were adjusted for age, diagnosis, and severity of offense—and their 95 percent confidence intervals (CIs)—which were derived from the logistic regression analysis for the race and ethnicity variables in the three models—are shown in t2. In model 1, which included both genders, black defendants were 1.26 times as likely to be hospitalized as white defendants after adjustment for other factors in the model. Hispanics had slightly lower odds of hospitalization than whites, but the difference was not statistically significant.
In model 2, black and Hispanic men had adjusted odds of hospitalization that differed significantly from those for whites, although these effects were in opposite directions. The adjusted OR for blacks indicated that they were 1.25 times as likely to be referred for inpatient evaluation as whites. However, the adjusted OR for Hispanic men indicated that they were only .82 times as likely to experience such referrals.
Model 3 shows the adjusted ORs for being referred to a strict-security facility for inpatient evaluation. As in model 2, the effects of race and ethnicity were significant, but model 3 shows a different pattern. In model 3, the OR for black men shows them to be 1.87 times as likely as white men to be evaluated in a strict-security facility. However, unlike the results found in model 2, model 3 indicated that Hispanic men were 1.37 times as likely as whites to be referred to a strict-security facility.
In our study we analyzed data from one jurisdiction, Massachusetts, to examine potential racial and ethnic disparities in the dispositions of pretrial defendants for inpatient forensic mental health evaluations. Our findings showed statistically significant racial and ethnic effects in the referral patterns of defendants after a court clinic screening. Specifically, black defendants were more likely than white defendants to be referred for inpatient evaluations; no significant difference was found between Hispanic defendants and white defendants of either gender, although Hispanic men were less likely than whites to be referred for inpatient evaluations. In addition, compared with white male defendants, a significantly greater percentage of both Hispanic men and black men were sent for inpatient evaluation in a strict-security facility. Our data showed no significant racial and ethnic differences in defendants' diagnoses or in the severity of criminal charges that could account for the difference in dispositions, and any such effect was controlled for statistically in our analyses.
Our data must be interpreted with caution. It is worth noting that disposition decisions were influenced by the defendant's history of criminal charges, which was available to the legal professionals and clinicians involved. That history, which is an important component in determining a defendant's risk, was not part of our database. Therefore, we were unable to ascertain what influence past criminal charges might have had on dispositions and whether this history differed across racial and ethnic groups.
Another caveat relates to data collection. Screening evaluations and initial diagnoses were based on a limited clinical examination and limited information, and there may be a tendency for clinicians to overdiagnose or underdiagnose defendants simply because of the nature of the screening. Moreover, race and ethnicity, which have recently been recognized even in the U.S. Census as increasingly difficult to categorize, were assigned by clinicians—who may themselves be in error—during the screening. This factor reflects a critical issue in all studies that are related to race and ethnicity (26). It is thus unclear whether clinical assignments to a particular racial and ethnic category correspond to how defendants view themselves.
As noted, the path for defendants to forensic mental health beds involves numerous complicated decisions. Our data described only one point in the process, and one cannot interpret our study's findings on the basis of only this point. It is clear from numerous reports that racial and ethnic disparities exist in criminal justice decision making that have nothing to do with mental health factors. No clear data are available from Massachusetts that examine the percentage of persons from ethnic minority groups who are arrested relative to the percentage of persons from minority groups who are hospitalized. Valid data are also lacking about the race of arrestees in general. However, the Bureau of Justice Statistics noted that the lifetime chance of going to prison is higher for blacks (18.6 percent) and for Hispanics (10 percent) than for whites (3.4 percent) (27). Furthermore, a bulletin of the Bureau of Justice Statistics indicated that in 2001 black men in their twenties and thirties were incarcerated at higher rates than both Hispanic men and white men (28). Our data examined decisions that were made only after a person had been arraigned on criminal charges. Data collected at this level cannot be used to ascertain how many persons were diverted from the criminal justice system to the mental health system, either before or after their arrest, or what the racial and ethnic makeup of such a group might be. Such information would be useful in shedding further light on decisions in the processes in which racial and ethnic factors may play a role. Thus, although we found an association between race and ethnicity and dispositional outcome, we did not find any evidence of a causal connection.
Importantly, a request for recommending where a defendant should be evaluated may be the first point in the process in which a clinician's input is sought. As such, this input may be a minor factor that is interjected late in a defendant's processing. The disparate rates of referral for inpatient evaluation of defendants who were black, Hispanic, or white should be interpreted in light of the possibility that these results are a function not of clinician variables but rather of a difference in the threshold—that is, more or less severe symptoms—that is required for certain minorities to be referred by legal professionals for a mental health evaluation (29). Although our data are only from Massachusetts, this interpretation is consistent with data obtained in Alabama. In that study, Cooper and Zapf (30) found that black defendants were more likely than white defendants to be referred for competency evaluations. They also noted that the racial disparities found in their study were more likely attributable to systemic variables than to clinician bias.
We do not have data about the potential role of other factors, such as socioeconomic status, level of education, living conditions, and availability and willingness to obtain resources that are essential to understanding the nuances of these types of findings (4,26). What does seem to be clear, however, is that racial minority groups are overrepresented throughout the criminal justice system and that this disparity likely extends to the forensic mental health system. Our study highlights yet another venue in which racial and ethnic disparities require further research attention. Our findings point to the need for increased awareness among clinicians about who ends up in the forensic mental health system. Informing forensic mental health professionals about these disparities as well as assisting them to achieve cultural competence represent but the initial steps toward understanding these complex issues.
The authors are affiliated with the law and psychiatry program in the department of psychiatry at the University of Massachusetts Medical School, 55 Lake Avenue North, Worcester, Massachusetts 01655 (e-mail, firstname.lastname@example.org).
Characteristics of defendants who were referred to a screening evaluation by the court to determine their competence to stand trial and their criminal responsibility (N=12,289)
Results of regression analyses to examine associations between race and ethnicityand disposition referral among defendants who were referred to a screening evaluationby the court to determine their competence to stand trial or their criminalresponsibilitya
a The analysis adjusted for diagnosis, age, and severity of criminal charge. Comparisons were with whites.