Secondary Logo

Journal Logo

Winner of Resident Paper Award 2019

Race, History of Abuse, and Homelessness Are Associated With Forced Medication Administration During Psychiatric Inpatient Care

THOMAS, TINA E. MBBS, MRCPsych; LANE, SCOTT D. PhD; ELKHATIB, RANIA M. MD; HAMILTON, JANE E. PhD, MPH; PIGOTT, TERESA A. MD

Author Information
Journal of Psychiatric Practice: July 2020 - Volume 26 - Issue 4 - p 294-304
doi: 10.1097/PRA.0000000000000485
  • Free

Abstract

Health care disparities are defined as “differences in health care services received by 2 groups that are not due to differences in the underlying health care needs or preferences of members of the groups.”1,2 A plethora of research exists regarding disparities in mental health care, especially among the African American population. Studies have found that African Americans are overrepresented in inpatient facilities even when they have an adequate living situation and health insurance.3–5 African Americans have also been found to underutilize outpatient mental health services compared with non-Hispanic white patients6–8 and to be more likely to experience coercive measures, such as involuntary hospitalization3,9 and administration of emergency medication, as a result of violent incidents.10 Conversely, some studies have not found a statistically significant difference between racial/ethnic minority and white patient groups in the administration of forced medication (FM) during psychiatric hospitalization.11,12

A number of studies have reported that patients who receive FM are more likely to be diagnosed with schizophrenia, bipolar disorder, or other psychotic disorders.13–15 Other factors associated with the use of FM include increased inpatient length of stay (LOS),16 lower levels of depressive symptomatology, higher positive psychosis symptom scores on the Brief Psychiatric Rating Scale (BPRS), involuntary admission status, and poorer global functioning.13 In addition, research has suggested that African Americans are more likely to be diagnosed with schizophrenia and less likely to be given an affective disorder diagnosis compared with non-Hispanic white individuals.3,9,17

There is a paucity of data on the relationship between prior trauma and the use of coercive measures, despite research documenting the role of trauma in the development of psychiatric symptoms. A growing body of research has linked prior trauma to chronic depression,18 a poorer course of illness in bipolar disorder,19 posttraumatic stress disorder,20 and schizophrenia including more positive symptoms and greater evidence of paranoia.21–25 It would be reasonable to postulate that, given that positive psychotic symptoms and poorer functioning have been noted to increase the likelihood of coercive measures, a link may exist between prior trauma and the use of FM.

While living situation can be a measure of socioeconomic deprivation, data concerning the association of previous living situation and the use of coercive measures are conflicting. One study found cohabiting with a partner before hospital admission was correlated with a reduced risk of FM.26 In contrast, the living situation was not found to correlate with coercive measures in a large European study.13 Among psychiatric patients experiencing homelessness, court-ordered psychiatric treatment has been employed as an intervention to increase the utilization of mental health services.27 In addition, involuntary outpatient commitment has been implemented to reduce the risk of homelessness among psychiatric patients with severe functional impairments.28

The goal of this case-control study is to address a gap in the literature by examining the potential association of court-ordered FM with racial minority status, homelessness, and prior experience of trauma. The investigation of court-ordered FM represents an underreported area of research. The majority of earlier studies have either involved violent incidents with subsequent administration of medication on an emergent basis or they did not specify the context of the FM administration.14 Diagnoses were also examined to investigate their possible differential impact on FM administration in patients in different racial groups. In addition, participants’ LOS and hospital readmission rates during the sampled time period were included in the analysis to examine treatment outcomes related to FM.

METHODS

Sample

Data for this retrospective study were obtained from electronic medical records for the period January 1, 2010, to February 15, 2018, from Harris County Psychiatric Center, an academic psychiatric hospital affiliated with University of Texas Health Science Center and McGovern Medical School (UT-HCPC) which serves as a regional safety-net in Houston, Texas. This sample included data on 57,615 patient admissions, from which ∼6200 cases in which FM petitions were requested and approved were identified. Patients were not included in this study if they met the following exclusion criteria: under 18 years of age, presence of intellectual/developmental disability, dementia, or other neurological condition, or primary diagnosis of a nonpsychiatric medical condition or a substance-induced mood or psychotic disorder. After data on those patients were excluded, the final sample included records from 2569 patients in which the administration of FM was identified. A control group of patients without FM petitions or court approval was created using propensity score matching via the R MatchIt package.29 The sample was matched by age, sex, and admission date (to avoid secular trends in the use of FM). The final dataset, therefore, included 5138 cases (n=2569 in each group). As confirmed by t test, χ2 test, and histograms of propensity score frequency, matching on these variables was successful; group means and distributions were nearly identical and groups were not statistically different (Table 1).

TABLE 1
TABLE 1:
Means (±SD), Frequency Counts (%), and Statistical Outcomes (t Test or χ2, as appropriate) for Demographic and Clinical Variables Separated into Forced Medication (FM) and No-forced Medication (NFM) Groups

Data Analyses and Primary Hypotheses

Statistical analyses tested the following demographic and clinical hypotheses: (1) racial distribution would differ between groups, with a higher proportion of African American patients in the FM group than in the no-forced medication group (NFM) group; (2) living situation would differ between the groups, with more homelessness in the FM group than the NFM group; (3) history of abuse (physical and/or sexual) would differ between the groups, with more patients with a history of abuse in the FM group than the NFM group. The study also tested the following hypotheses related to treatment outcomes, on the basis of the observation that the need for FM is associated with greater acuity (both short term and long term); (4) LOS would differ between the groups, with longer LOS in the FM group than the NFM group; (5) readmissions over the sampled time period would differ between the groups, with more readmissions in the FM group than the NFM group.

Hypotheses 1, 2, and 3 were tested via χ2 frequency tests. For hypothesis 4, the outcome variable LOS was right-skewed and the residuals from initial linear models were not normally distributed, as estimated by skewness/normality tests and graphical inspection of q-q and k-density plots. Thus, LOS was log-transformed and modeled via linear regression with group (FM vs. NFM), living situation (homeless/not homeless), race, admission status (voluntary/involuntary), history of reported abuse, and substance use disorder (SUD) (present/absent) as predictors. While group status was the primary predictor (independent) variable of interest, the additional predictors were included in the model because they were significantly different between groups and therefore served as important covariates.

For hypothesis 5, the outcome variable number of hospital readmissions in the 8-year time sample was approximately Poisson distributed and thus readmission data were modeled using Poisson regression with group (FM vs. NFM), and race, living situation (homeless/not homeless), admission status (voluntary/involuntary), history of reported abuse, and SUD (present/absent) again serving as predictors. Both regression models for hypotheses 4 and 5 included robust-clustered SE estimates clustering by patient to be appropriately conservative against heteroscedasticity and leverage, and to account for within-patient autocorrelation across repeat admissions.

Exploratory Analyses

In exploratory analyses, the difference in the proportion of discharge diagnoses between FM and NFM patients was examined via χ2 frequency tests. The discharge diagnosis was categorized as bipolar disorder, schizophrenia, depression/anxiety disorder, or other. All Diagnostic and Statistical Manual of Mental Disorders (DSM)/International Classification of Diseases (ICD) diagnostic codes within that spectrum of disorders were included in creating the category.

RESULTS

General Sample Characteristics

The average age across both groups was 38.14 years (SD=13.06 y). The proportion of males and females across the 2 groups was 53% and 47%, respectively. The average LOS was 14.67 days (SD=9.07 d). Admission dates were distributed approximately equally per year across the time period of data sampling (January 1, 2010, to February 15, 2018), owing to equivalent annual hospital admission rates over that time period. The average percentage of involuntary admissions at the hospital was ∼60%.

Group Demographic and Clinical Variables

Using equal (1:1) weighting and nearest-neighbor matching (ie, the closest control match for each treated unit is chosen one at a time), the R MatchIt algorithm generated groups that were equivalently matched by age, sex, and admission date. Means for age were FM=38.24, NFM=38.05 (z=0.68, nonsignificant). Percentages for sex were FM=53% male, NFM=53% male (z=0.33, nonsignificant). Median admission dates were FM 09/2013 (range=01/2010 to 01/2018) and NFM=10/2013 (range=01/2010 to 01/2018), z=1.34 (nonsignificant). Table 1 presents means and/or frequency counts, as appropriate, for the FM and NFM groups on the following variables: age, sex, race, SUD (present/absent), admission status (involuntary/voluntary), reported history of physical or sexual abuse (present/absent), and living status (homeless/temporary or permanent housing). While by design (because of propensity matching), the groups did not differ in age, sex, or admission date, FM patients were less likely to have an SUD (P<0.001), and were more likely to be admitted involuntarily (P<0.001). Variables such as race, history of abuse, and living status were part of the study hypotheses and are described in greater detail below. Given that an exclusion criterion for the FM group was the presence of a substance-induced mood or psychotic disorder, it would be expected that the rate of SUDs would be lower in the FM group. This is explored further in the discussion.

Hypotheses 1, 2, and 3: Demographic and Clinical Measures

Table 1 provides frequency counts and statistical details. For hypothesis 1, which predicted that the FM group of patients would have a higher proportion of individuals of African American race relative to other races, the overall model was significant (Pearson χ2df=4=13.01, P<0.012). Comparison of adjusted residuals for individual cell counts revealed that African American patients were significantly more frequent in the FM group than in the NFM group (n=1284 vs. 1192, z=2.57, P<0.05). In contrast, non-Hispanic white patients were significantly less frequent in the FM group than in the NFM group (n=1062 vs. 1156, z=−2.65, P<0.05). The FM and NFM groups did not differ in expected versus obtained frequency for Hispanic, Asian, or Native American races. Asian and American Indian patients comprised <4% of the sample, thus the findings for these groups should be interpreted with caution. For hypothesis 2, predicting there would be a higher proportion of homelessness in patients in the FM group, the overall model was significant (Pearson χ2df=1=53.26, P<0.001). Patients in the FM group were more likely to be homeless than those in the NFM group (n=838 vs. 603, respectively). For hypothesis 3, that a higher percentage of patients in the FM group than in the NFM group would report a history of abuse, the overall model was significant (Pearson χ2df=1=27.21, P<0.001). Patients in the FM group were more likely to report a history of physical or sexual abuse than those in the NFM group (n=1858 vs. 1685).

Hypotheses 4 and 5: Outcome Measures

Hypothesis 4, which predicted that patients in the FM group would have an average longer LOS than those in the NFM group, was examined using linear regression, with log LOS as the dependent (outcome) variable and FM status, living situation (homeless/residence), admission status (voluntary/involuntary), SUD (present/absent), and history of abuse (present/absent) as predictors. The overall model was significant (Fdf=9,3011=101.40, P<0.0001, root mean square error=0.53, Akaike information criterion=1.58). Controlling for the other factors in the model, FM status was a significant predictor of longer LOS (t=19.74, P<0.001). LOS was also significantly associated with greater homelessness (P<0.03), more involuntary admissions (P<0.001), lower rates of SUD (P<0.001), and greater reported history of abuse (P<0.001). Table 2 presents details of the statistical results for hypothesis 4.

TABLE 2
TABLE 2:
Results for the Outcome Variable Length of Stay (LOS, Hypothesis 4), Expressed in Log Units and Modeled via Linear Regression

Hypothesis 5, which predicted that patients in the FM group would have more hospital readmissions over the 8-year study period than those in the NFM group, was examined using Poisson regression, with readmissions as the dependent (outcome) variable and the same predictors as hypothesis 4: FM status, living situation, admission status, SUD, and history of abuse. The overall model was significant (Wald χ2df=9=102.87, P<0.0001, maximum likelihood R2=0.25, Akaike information criterion=6.78). Controlling for the other factors in the model, FM status was a significant predictor of more readmissions (z=2.12, P<0.04). Homelessness was also significantly associated with more readmissions (P<0.001). When findings concerning race were examined, there were significantly more readmissions among African American patients (P<0.02), and significantly fewer readmissions among Hispanic patients (P<0.001) and Native American patients (P<0.03) when compared with non-Hispanic white patients as the reference group (note that there were only 3 Native Americans in the total sample so that the results are not likely to be credible). Table 3 presents details of the statistical results for hypothesis 5.

TABLE 3
TABLE 3:
Results for the Outcome Variable Number of Hospital Readmissions (Hypothesis 5), Modeled via Poisson Regression

For the exploratory analysis examining the proportion of FM to NFM across discharge diagnoses, the overall model was significant (Pearson χ2df=3=182.59, P<0.001). Comparison of adjusted residuals for individual cell counts revealed that patients with a diagnosis of schizophrenia were significantly more frequent in the FM group than in the NFM group (n=1613 vs. 1142, z=13.17, P<0.01). In contrast, the FM group had significantly fewer patients with diagnoses of bipolar disorder (n=310 vs. 427, z=−4.66, P<0.01) or with “other” diagnoses (n=312 vs. 553, z=−8.96, P<0.01) than the NFM group.

DISCUSSION

This study found that more African American patients received FM than non-Hispanic white patients after controlling for age, sex, and admission date, whereas there was no difference between Hispanic and Asian patients in terms of FM versus NFM groups. This finding is consistent with earlier research that reported racial disparities in the administration of FM. In a literature review, Jarrett et al14 reported ethnicity data regarding FM available in 3 of the 14 papers reviewed. In 2000, Gudjonsson et al10 published an analysis of 2180 incident forms for 165 patients in the United Kingdom that found an increased probability of racial/ethnic minority patients receiving medication after an incident. In particular, black patients were more likely than white patients to receive FM following assaultive behavior.10 In another study published in 2004, Gudjonsson et al30 reported that black patients had an increased likelihood of receiving emergency medication rather than physical restraint in response to violent incidents. However, after the data were adjusted for other variables, such as the reason for admission, history of substance use, and agitation, these differences disappeared. Finally, a 1995 survey of 28 patients by Lucksted and Coursey11 did not find a statistically significant difference between ethnic groups in the administration of FM.

The study described here is unique in that it examined the use of FM that was court-ordered for treatment of mental illness rather than administered emergently. Clinical reasons to request court-ordered medication include refusal of medication by patients with psychotic or manic symptoms, a history of medication nonadherence postdischarge, and a history of criminal or aggressive behavior when not maintained on medication. Pharmacological treatment given in this context is not a result of acute violent incidents but rather the treatment is scheduled and administered intramuscularly if the patient refuses. The most common medications appropriate for this purpose are antipsychotic agents. Therefore, the finding that FM was associated with a diagnosis of schizophrenia-spectrum disorders in this study is understandable. This result is also consistent with earlier research suggesting that coercive treatment is more likely in patients with psychotic disorders.31

A possible explanation for the higher rates of coercive medication administration in African American populations could be barriers to accessing outpatient services with subsequent greater illness exacerbation and greater symptomatology on admission. Evidence suggests that minority groups are less likely to seek mental health services and more likely to present with greater symptom severity, in addition to receiving poorer quality mental health care.32

Another factor that could have influenced the finding of higher rates of coercive medication administration in African American than non-Hispanic white populations could be cultural bias resulting in an increased likelihood that a diagnosis of schizophrenia will be assigned in African American individuals3,9,17 and the recording of higher BPRS scores on measures that represent psychosis, such as hallucinations, suspiciousness, and hostility.3 These results have been reported despite consistent epidemiological data showing that schizophrenia is not more common in African American compared with white individuals after controlling for socioeconomic factors.33 This finding raises concerns about the misdiagnosis of and over-attribution of hallucinatory behaviors to African Americans.4 Furthermore, the lack of cultural competency can lead to predictive errors in violence risk assessment,34 which factors into a justification for medication enforcement. Nonwhite and male patients have been found to have higher rates of false-positive assessments of violence, overestimating risk.34 The implications are that African American patients are more likely than white patients to be assessed as aggressive and therefore requiring FM to manage risk.35

One aspect of cultural competence is an accurate understanding of cultural communication styles. African American individuals have been described as using a high-context communication style, indicating a greater reliance on external factors other than explicit speech to convey meaning.35 This can lead to misinterpretation of culturally appropriate communication as aggressive, resulting in forced or erroneous administration of medication. Indeed, it has been documented that African Americans are more likely to receive inappropriate antipsychotic treatment rather than antidepressant medication,17 and to be prescribed higher doses of antipsychotics than their white counterparts.36 The implications of the research presented here regarding the complexities of mental health care disparities experienced by African Americans are very concerning. Our findings highlight the relevance of culture to mainstream psychiatry, as is recognized in the fifth edition of the DSM,37 which introduced the Cultural Formulation Interview as an aid to gathering culturally relevant diagnostic information. However, further research on strategies for improving cultural competency and the effect of such efforts on reducing health care disparities is needed.

In the study presented here, homelessness and prior trauma from physical and/or sexual abuse were also more frequent in the FM group. Furthermore, FM was a significant predictor of homelessness, which in turn was significantly associated with hospital readmissions. Increased LOS was also significantly related to homelessness and a history of abuse. These results indicate that deprivation, both economic and relational, may be associated with treatment and outcomes. African American race, homelessness, and prior traumatization were all independent predictors of FM. Therefore, another unifying theme of the results is a possible lack of trust among disenfranchised patient groups as a predictor of FM. Such a lack of trust in patients could also affect providers’ perception of the need to use FM to provide treatment. Shared decision-making and patient-centered approaches that have been implemented to increase trust among this patient population could be of key importance in reducing the frequency of use of coercive treatment.38

As discussed earlier, trauma plays a role in the genesis and severity of psychiatric disorders.19,20 It is reasonable to suggest that the use of FM is necessary when symptoms are more severe, explaining the association of trauma with coercive treatment. However, when considering a trauma-informed approach to patients, it is important to recognize the detrimental effect coercive treatment could have by retraumatizing patients.39

Evidence also suggests that coercion is linked to patient perception of hospitalization-related trauma.40 Paksarian et al40 found that the traumatizing events that were most frequently reported in interviews of patients about experiences of previous psychiatric hospitalizations were involuntary hospitalization, use of restraints, and FM. They did not find any association between these self-reported experiences and race, living arrangements before admission, or rates of rehospitalization. However, being female and having an employment status of homemaker were correlated with patients’ perception of trauma secondary to coercion. In a study of 3093 European inpatients who were involuntarily admitted or felt coerced to hospital treatment despite a legally voluntary admission, Fiorillo et al13 found that female patients perceived more coercion in psychiatric hospital treatment. Self-reported FM administration was linked to less time engaged in treatment during the following decade, which could represent either disengagement or symptom improvement.40 The use of FM, in particular, as opposed to other coercive measures is related to poorer patient perception of treatment.16 While a multiplicity of factors affect treatment outcomes, FM in a previously traumatized population could potentially lead to poorer outcomes, as evidenced by increased LOS and readmissions.

Variable outcomes concerning LOS have been reported in a small number of studies on the use of coercive medication.14 Greenberg et al41 did not find any significant difference between groups, whereas Greenberg and Attia42 found that patients who received FM due to medication refusal had significantly longer hospital stays. We are not aware of any studies that have reported on any correlation between hospital readmission rates and the use of coercive medication. However, it has been reported that African Americans are more likely to be hospitalized and to be admitted involuntarily.9 The cause for this is likely multifactorial, but one reason that has been suggested is a delay in seeking mental health treatment. However, the association with the use of coercive treatment, especially in previously traumatized patients, could be a mediating factor.

Homelessness is an example of social-selection downward mobility possibly experienced by those with serious mental illness.43 Schizophrenia has been reported to be overrepresented in homeless populations compared with nonhomeless populations.44 This finding reflects a number of issues, including economic barriers to care, lack of long-term facilities for the care of the chronically mentally ill, and the severity of illness. In our study, the association between FM and homelessness could reflect increased mental health needs, poorer insight, and barriers accessing community-based mental health services. There is a plethora of research linking homelessness among the seriously mentally ill population with high rates of utilization of emergency and inpatient services.45–49 In addition, studies have found that people with schizophrenia are 10 times more likely to become homeless compared with the general population.50,51 However, it has also been argued that the prevalence of mental illness is overestimated in individuals who are homeless and that instead, homelessness is due to lack of housing provision to the most disadvantaged.52 Nevertheless, the homeless population is especially vulnerable to imbalances in autonomy versus paternalism, which could promote forced treatment. Melamed et al53 described the tension between respecting a person’s wish to live a life free from external intervention and the clinician’s desire or obligation to help the needy, especially when capacity is in question due to mental illness. Interestingly, another study from the same institution as the authors of this paper (UT-HCPC) found that African Americans have a disproportionate inpatient representation despite adequate housing, pointing to treatment disparities beyond economic factors.3 Notably, earlier research found that living situation was not associated with coercive treatment.16

This study demonstrated that patients who received FM were less likely to have a comorbid substance use diagnosis. This is not consistent with previous research that found a relationship between substance abuse, acute intoxication, and the use of coercive measures.54,55 However, those studies appear to have examined the use of coercive measures in the context of emergent management rather than court-ordered medication. Pawlowski and Baranowski31 documented that the second most likely diagnosis of patients who experienced coercive measures was a mental disorder caused by the use of a psychoactive substance. This diagnosis was found in a significantly higher percentage of those who experienced coercive measures than in the general population of the hospital. One of the exclusion criteria for the FM group in our study was a primary diagnosis of a substance-induced disorder because these patients can be discharged shortly after detoxification once no underlying symptoms of serious mental illness are present. Furthermore, it would not be expected that courts would grant FM petitions for substance-induced mental illness, which could explain our findings. However, a history of substance use is typically associated with an increased risk of aggression,56 which could arguably increase the use of FM.

In terms of treatment outcomes, patients in the FM group fared worse reflected by increased LOS, which was also associated with involuntary admissions. This finding is consistent with previous research showing that, on average, FM increases LOS by ∼9 days compared with average LOS.16 In addition, this study found that higher readmission rates were associated with FM and with being African American, whereas Hispanic patients had fewer readmissions than non-Hispanic white patients. The LOS and readmission rates could represent the acuity and severity of illness in patients who require FM. McLaughlin et al16 found that higher BPRS scores and previous admissions were associated with coercive measures. Court-ordered FM due to medication refusal would also increase LOS due to the time involved in the legal process. As noted above, racial disparities in the diagnosis of psychotic disorders and in readmissions among African Americans may be related to racial biases regarding symptomatology.4,57 Research has shown that differences in the race between clinicians can impact both diagnostic assignment and conceptualization of psychotic symptom level resulting in significantly higher scores for total psychotic symptoms among African American patients.4,57 It is unclear why Hispanic patients had fewer readmissions. However, this finding is consistent with the results of earlier research showing that Hispanic and Asian patients tended to utilize inpatient and emergency services less frequently while having higher rates of outpatient service utilization, while African American patients used more emergency services and less outpatient care compared with white patients.6 This pattern could have influenced our findings concerning readmission rates.

Strengths and Limitations of the Study

One of the limitations of this study is the use of retrospective data, which obviously limits any causal inferences that can be drawn from the results. However, in datasets of this size obtained in high-acuity community-based hospital settings, experimental designs providing causal inference are impracticable. Psychometrically valid measures of severity of illness were not included in this study due to a lack of systematic assessment over the 8-year duration of the dataset and the multiple hospital units from which data were obtained. This aspect of a patient’s presentation could potentially influence the need for FM. As noted above, however, cultural factors also influence the perception of the severity of illness in African Americans. Also absent from the dataset was the race of the examining doctor. The likelihood that a schizophrenia diagnosis will be assigned when hallucinations are present is greater if the clinician is African American.4 Further studies regarding the effect of cultural competency training on measures of health care disparities appear warranted.

An additional limitation of this study was the restriction of the dataset to a single inpatient community-based psychiatric hospital. It is therefore not known how well the findings presented here will generalize to other regions of the United States, to other countries with different mental health care policies and practices, and to other psychiatric settings and patient populations. It would be of considerable interest to replicate this analysis using merged datasets from similar inpatient community hospitals in other large US metropolitan regions.

Further research to investigate the moderation or mediation each of the key predictors (race, trauma, and homelessness) has on each other would also be of interest. For example, is there an additive effect of being African American, homeless, and having a history of trauma? Furthermore, if mistrust of organizations plays a factor, can shared decision-making have an effect on the reduction of coercive treatment needs? Person‐centered care approaches, including shared decision-making, in mental health treatment models, have the potential to improve patient engagement and outcomes.38

The class and type of FM was not reported in this study, which could affect the reason for enforcement. However, we do know that the medications available in the intramuscular form in the UT-HCPC drug formulary are the antipsychotics chlorpromazine, olanzapine, haloperidol, and fluphenazine, and one benzodiazepine, lorazepam. Therefore, it can be concluded that the FM would have been one of these. However, it would be useful for future studies to investigate which medications are most frequently used for FM administration.

Furthermore, while the FM group represented patients for whom the court had ordered legal enforcement if a medication was refused, it is not known how many of those patients actually had to undergo the process of being physically forced to take medication. While this study represents important information about those who are ordered FM, further clarification on who received medication in a forced process (ie, who was given medication intramuscularly against their will) could affect the relationship with trauma among other variables. Nevertheless, a court-ordered medication that a patient no longer has a right to refuse exerts influence, even without physical implementation being necessary.

The strengths of this study include the large sample size and the time frame of data collection. Previous work in this area has typically analyzed smaller datasets over more restricted time periods. The outcomes of the statistical analyses are generally robust in support of the study hypotheses, and they support and extend the results of earlier research on the use of FM. While some attention has been paid to seclusion and restraint and the practice of trauma-informed care,58–60 the observation that past significant trauma was associated with being administered FM has received limited attention in earlier literature. This novel observation merits more extensive examination in future work.

CONCLUSIONS

This study contributes both new and supporting information to the large body of research regarding health care disparities, and to the less investigated area of court-ordered FM. African American race, homelessness, and a history of trauma were significantly associated with FM administration during inpatient hospitalization. These findings inform a number of issues related to clinical practice, including the role of cultural competency, the balance of autonomy and paternalistic approaches to care, and the need for shared decision-making and trauma-informed care to improve clinical outcomes. Our study highlights the imperative for clinicians who are involved with the administration of FM to be aware of the complexities involved and the possible effects on patient outcomes.

REFERENCES

1. LêCook B, McGuire TG, Zaslavsky AM. Measuring racial/ethnic disparities in health care: methods and practical issues. Health Serv Res. 2012;47:1232–1254.
2. Nelson A. Unequal treatment: confronting racial and ethnic disparities in health care. J Natl Med Assoc. 2002;94:666–668.
3. Hamilton JE, Heads AM, Cho RY, et al. Racial disparities during admission to an academic psychiatric hospital in a large urban area. Compr Psychiatry. 2015;63:113–122.
4. Trierweiler SJ, Neighbors HW, Thompson EE, et al. Clinician attributions associated with the diagnosis of schizophrenia in African American and non-African American patients. J Consult Clin Psychol. 2000;68:171–175.
5. Snowden LR, Hastings JF, Alvidrez J. Overrepresentation of Black Americans in psychiatric inpatient care. Psychiatr Serv. 2009;60:779–785.
6. Hu TW, Snowden LR, Jerrell JM, et al. Ethnic populations in public mental health: services choice and level of use. Am J Public Health. 1991;81:1429–1434.
7. Snowden LR, Yamada AM. Cultural differences in access to care. Annu Rev Clin Psychol. 2005;1:143–166.
8. Thornicroft G. Stigma and discrimination limit access to mental health care. Epidemiol Psichiatr Soc. 2008;17:14–19.
9. Lawson WB, Hepler N, Holladay J, et al. Race as a factor in inpatient and outpatient admissions and diagnosis. Hosp Community Psychiatry. 1994;45:72–74.
10. Gudjonsson GH, Rabe-Hesketh S, Wilson C. Violent incidents on a medium secure unit: the target of assault and the management of incidents. J Forens Psychiatry Psychol. 2000;11:105–118.
11. Lucksted A, Coursey RD. Consumer perceptions of pressure and force in psychiatric treatments. Psychiatr Serv. 1995;46:146–152.
12. Bennewith O, Amos T, Lewis G, et al. Ethnicity and coercion among involuntarily detained psychiatric in-patients. Br J Psychiatry. 2010;196:75–76.
13. Fiorillo A, Giacco D, De Rosa C, et al. Patient characteristics and symptoms associated with perceived coercion during hospital treatment. Acta Psychiatr Scand. 2012;125:460–467.
14. Jarrett M, Bowers L, Simpson A. Coerced medication in psychiatric inpatient care: literature review. J Adv Nurs. 2008;64:538–548.
15. Raboch J, Kališová L, Nawka A, et al. Use of coercive measures during involuntary hospitalization: findings from ten European countries. Psychiatr Serv. 2010;61:1012–1017.
16. McLaughlin P, Giacco D, Priebe S. Use of coercive measures during involuntary psychiatric admission and treatment outcomes: data from a prospective study across 10 European countries. PLoS One. 2016;11:e0168720.
17. Barnes A. Race and hospital diagnoses of schizophrenia and mood disorders. Soc Work. 2008;53:77–83.
18. Wiersma JE, Hovens JGFM, Van Oppen P, et al. The importance of childhood trauma and childhood life events for chronicity of depression in adults. J Clin Psychiatry. 2009;70:983–989.
19. Leverich GS, McElroy SL, Suppes T, et al. Early physical and sexual abuse associated with an adverse course of bipolar illness. Biol Psychiatry. 2002;51:288–297.
20. Jonas S, Bebbington P, McManus S, et al. Sexual abuse and psychiatric disorder in England: results from the 2007 Adult Psychiatric Morbidity Survey. Psychol Med. 2011;41:709–719.
21. Heins M, Simons C, Lataster T, et al. Childhood trauma and psychosis: a case-control and case-sibling comparison across different levels of genetic liability, psychopathology, and type of trauma. Am J Psychiatry. 2011;168:1286–1294.
22. Holowka DW, King S, Saheb D, et al. Childhood abuse and dissociative symptoms in adult schizophrenia. Schizophr Res. 2003;60:87–90.
23. Rubino IA, Nanni RC, Pozzi DM, et al. Early adverse experiences in schizophrenia and unipolar depression. J Nerv Ment Dis. 2009;197:65–68.
24. Steel C, Marzillier S, Fearon P, et al. Childhood abuse and schizotypal personality. Soc Psychiatry Psychiatr Epidemiol. 2009;44:917–923.
25. Üçok A, Bikmaz S. The effects of childhood trauma in patients with first-episode schizophrenia. Acta Psychiatr Scand. 2007;116:371–377.
26. Girela E, López A, Ortega L, et al. Variables associated with the use of coercive measures on psychiatric patients in Spanish penitentiary centers. Biomed Res Int. 2014;2014:928740. Doi: 10.1155/2014/928740.
27. McQuistion HL, Finnerty M, Hirschowitz J, et al. Challenges for psychiatry in serving homeless people with psychiatric disorders. Psychiatr Serv. 2003;54:669–676.
28. Compton SN, Swanson JW, Wagner HR, et al. Involuntary outpatient commitment and homelessness in persons with severe mental illness. Ment Health Serv Res. 2003;5:27–38.
29. Ho DE, Kosuke I, King G, et al. MatchIt: nonparametric preprocessing for parametric causal inference. J Stat Softw. 2011;42:8.
30. Gudjonsson GH, Rabe-Hesketh S, Szmukler G. Management of psychiatric in-patient violence: patient ethnicity and use of medication, restraint and seclusion. Br J Psychiatry. 2004;184:258–262.
31. Pawlowski T, Baranowski P. How patients’ characteristics influence the use of coercive measures. Indian J Psychiatry. 2017;59:429–434.
32. Office of the Surgeon General (US); Center for Mental Health Services (US); National Institute of Mental Health (US). Mental Health: Culture, Race, and Ethnicity: a Supplement to Mental Health: a Report of the Surgeon General. Rockville, MD: Substance Abuse and Mental Health Services Administration (US); 2001. Available at: www.ncbi.nlm.nih.gov/books/NBK44243. Accessed May 24, 2020.
33. Earl T, Williams D, Anglade S. An update on the mental health of black Americans: puzzling dilemmas and needed research. J Black Psychol. 2011;37:485–498.
34. McNiel DE, Binder RL. Correlates of accuracy in the assessment of psychiatric inpatients’ risk of violence. Am J Psychiatry. 1995;152:901–906.
35. Campinha-Bacote J. Cultural considerations in forensic psychiatry: the issue of forced medication. Int J Law Psychiatry. 2017;50:1–8.
36. Walkup JT, McAlpine DD, Olfson M, et al. Patients with schizophrenia at risk for excessive antipsychotic dosing. J Clin Psychiatry. 2000;61:344–348.
37. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5). Arlington, VA: American Psychiatric Association Publishing; 2013.
38. Dixon LB, Holoshitz Y, Nossel I. Treatment engagement of individuals experiencing mental illness: review and update. World Psychiatry. 2016;15:13–20.
39. Reeves E. A synthesis of the literature on trauma-informed care. Issues Ment Health Nurs. 2015;36:698–709.
40. Paksarian D, Mojtabai R, Kotov R, et al. Perceptions of hospitalization-related trauma and treatment participation among individuals with psychotic disorders. Psychiatr Serv. 2014;65:266–269.
41. Greenberg WM, Moore-Duncan L, Herron R. Patients’ attitudes toward having been forcibly medicated. Bull Am Acad Psychiatry Law. 1996;24:513–524.
42. Greenberg WM, Attia S. Nonemergent forcible medication in an acute hospital. Bull Am Acad Psychiatry Law. 1993;21:465–473.
43. Perry MJ. The relationship between social class and mental disorder. J Prim Prev. 1996;17:17–30.
44. Folsom D, Jeste DV. Schizophrenia in homeless persons: a systematic review of the literature. Acta Psychiatr Scand. 2002;105:404–413.
45. Dhossche DM, Ghani SO. A study on recidivism in the psychiatric emergency room. Ann Clin Psychiatry. 1998;10:59–67.
46. Hamilton JE, Passos IC, De Azevedo Cardoso T, et al. Predictors of psychiatric readmission among patients with bipolar disorder at an academic safety-net hospital. Aust N Z J Psychiatry. 2016;50:584–593.
47. Han KT, Kim SJ, Jang SI, et al. The outcomes of psychiatric inpatients by proportion of experienced psychiatrists and nurse staffing in hospital: new findings on improving the quality of mental health care in South Korea. Psychiatry Res. 2015;229:880–886.
48. Pasic J, Russo J, Roy-Byrne P. High utilizers of psychiatric emergency services. Psychiatr Serv. 2005;56:678–684.
49. Young AS, Chinman MJ, Cradock-O’Leary JA, et al. Characteristics of individuals with severe mental illness who use emergency services. Community Ment Health J. 2005;41:159–168.
50. Lin WC, Bharel M, Zhang J, et al. Frequent emergency department visits and hospitalizations among homeless people with Medicaid: implications for Medicaid expansion. Am J Public Health. 2015;105:716–722.
51. Susser ES, Lin SP, Conover SA. Risk factors for homelessness among patients admitted to a state mental hospital. Am J Psychiatry. 1991;148:1659–1664.
52. Hamid WA, Wykes T, Stansfeld S. The homeless mentally ill: myths and realities. Int J Soc Psychiatry. 1993;39:237–254.
53. Melamed Y, Fromer D, Kemelman Z, et al. Working with mentally ill homeless persons: should we respect their quest for anonymity? J Med Ethics. 2000;26:175–178.
54. Andersen K, Nielsen B. Coercion in psychiatry: the importance of extramural factors. Nord J Psychiatry. 2016;70:606–610.
55. Keski-Valkama A, Sailas E, Eronen M, et al. Who are the restrained and secluded patients? A 15-year nationwide study. Soc Psychiatry Psychiatr Epidemiol. 2010;45:1087–1093.
56. Dack C, Ross J, Papadopoulos C, et al. A review and meta-analysis of the patient factors associated with psychiatric in-patient aggression. Acta Psychiatr Scand. 2013;127:255–268.
57. Arnold LM, Keck PE, Collins J, et al. Ethnicity and first-rank symptoms in patients with psychosis. Schizophr Res. 2004;67:207–212.
58. Barton SA, Johnson MR, Price LV. Achieving restraint-free on an inpatient behavioral health unit. J Psychosoc Nurs Ment Health Serv. 2009;47:34–40.
59. Hammer JH, Springer J, Beck NC, et al. The relationship between seclusion and restraint use and childhood abuse among psychiatric inpatients. J Interpers Violence. 2011;26:567–579.
60. Muskett C. Trauma-informed care in inpatient mental health settings: a review of the literature. Int J Ment Health Nurs. 2014;23:51–59.
Keywords:

health disparities; forced medication; African American; homelessness; trauma; cultural competence; shared decision-making; trauma-informed care

Copyright © 2020 Wolters Kluwer Health, Inc. All rights reserved.