Incarcerated Women’s Experiences and Beliefs About Psychotropic Medication: An Empirical Study

Mental illness and psychotropic medication use among incarcerated persons represent significant concerns for psychiatric service providers. Although reported rates of mental illness in this population differ across studies, research has consistently shown that people with mental illnesses are incarcerated at disproportionate rates and that women are especially affected (1). In the United States, recent estimates from the Bureau of Justice Statistics indicate that 73% of female inmates in state correctional facilities experience some form of mental health–related difficulty, ranging from serious mental illnesses and personality disorders to sleep disorders and adjustment disorders (2). Partly because of a scarcity of resources, psychotropic medication has been the major vehicle for addressing these issues (3–5). For example, among those in state correctional facilities with a diagnosed mental health condition, 68.6% are prescribed psychotropic medications (6), with several studies reporting even higher rates of use among incarcerated women (7,8).

A few studies have explored issues related to psychotropic medication use among incarcerated populations, examining such topics as perceived quality of mental health care (9), medication management (10), undesirable side effects (11) and, most frequently, adherence to medication (3,12,13). However, there is a marked absence both of studies that focus on incarcerated women and studies that examine the lived experiences of women who use psychotropic medications. For example, discussion is lacking of their perceptions of medication’s effects and impact, their perceptions of personal agency, and their related experiences of stigma. In addition, although several studies have identified locus of control as an important variable for predicting mental health and recidivism outcomes among offenders (14–16) and another study has suggested that locus of control may influence treatment adherence among nonincarcerated psychiatric patients (17), no studies have examined locus of control in relation to psychotropic medication use among incarcerated individuals. The purpose of this project was to more fully understand these complex experiences with the hope of giving greater voice to incarcerated women and generating a richer knowledge base that might promote more effective and compassionate responses to their mental health–related needs and challenges.

Methods

Participants

During a single day in September 2015, all 535 women at a medium-security correctional institution in a Mid-Atlantic state were invited to participate in the study. Security staff escorted the women to a large gymnasium in five separate groups over the course of about two hours. Informed consent forms and questionnaires were offered upon entry. Once the women were seated, we described the study in detail, reviewed the consent form, and responded to any questions. Prospective participants were informed that we were looking for volunteers who had taken or were currently taking psychotropic medication in the prison setting. The women chose whether to complete the consent form or the survey. After about 15 minutes, they were thanked and asked to place their blank or completed informed consent forms and questionnaires anonymously in separate cardboard boxes as they exited. The effort produced 274 usable questionnaires. The study was approved by the full board of Virginia Commonwealth University’s Institutional Review Board as well as the research review board of the Department of Corrections in the state where the research was conducted.

Measure

We developed an original questionnaire for the study, crafting items around relevant topics from several sources. First, literature on medication use with incarcerated populations was consulted (3,12), as well as literature on the effects of psychotropic medication among persons with serious mental illness (18,19). Some questionnaire items were drawn from existing measures of attitudes toward medication (20–23). The questionnaire for this study included 22 items assessing beliefs about and experiences with psychotropic medication (Cronbach’s α=.762).

In addition, during development of the questionnaire, we conducted a focus group with mental health clinicians and administrators from the Department of Corrections to solicit their feedback on the research questions and instrumentation. Participants in the focus group indicated that they believed locus of control might be a fruitful avenue of investigation. Specifically, their experiences as service providers suggested that incarcerated women, who often have histories of victimization, might see themselves as having less personal agency in life, which might affect their beliefs about and experiences with medication. On the basis of this observation and discussions of locus of control in the literature, we added six items from Rotter’s Internal-External Control Scale (24), which were assessed to be most relevant for incarcerated populations. Three items were intended to measure external locus of control, and three were intended to measure internal locus of control. These items used the same 10-point Likert scale as the other items, rather than the forced-choice format used in the original instrument.

The complete questionnaire consisted of 35 items: six on demographic information, six measuring locus of control, and 22 measuring experiences with psychotropic medication. A final open-ended item asked participants to respond to the question: “What is the most important thing you would want people to know about your experiences with mental health medication?”

Data Analysis

Using SPSS Statistics 23, we began with various univariate analyses to assess frequencies of demographic and clinical variables, including current psychotropic medication use and self-reported psychiatric diagnosis. Next, to discern the underlying factor structure of the data, we conducted an exploratory principal-components factor analysis with varimax rotation of the 22 questionnaire items pertaining to experiences with and perceptions of psychotropic medication. Factors with an Eigenvalue greater than one were retained, as suggested by Costello and Osborne (25). Composite scores were calculated for the six factors that emerged. To incorporate locus of control into the statistical models, the average score for internal control items was subtracted from the average score for external control items, yielding a single score for which higher values indicated more external control. We conducted linear regression analyses to determine whether demographic variables, incarceration history, locus of control, current psychotropic medication use, and number of mental health diagnoses predicted experiences with and perceptions of psychotropic medication use. An analysis of covariance (ANCOVA) examined the role of polypharmacy. Responses to the open-ended item were subjected to a content analysis involving quantitative word counts, as well as an inductive process of organizing the narrative data into themes.

Results

Sample Characteristics

As shown in Table 1, most respondents were Caucasian (80%) and between ages 25 and 44 (70%). The average time incarcerated during adulthood was 6.24 years, with a range from six months to 30 years. The most frequently self-reported psychiatric diagnoses were anxiety disorder (76%) and depressive disorder (72%), but a large percentage of women also reported trauma-based disorders (58%), substance use disorders (49%), and bipolar disorder (47%). Sleep disorders (31%) and personality disorders (20%) were also reported. Over three-quarters (77%) reported that they were currently taking psychotropic medication (Table 2), which accounted for 40% of the institutional census on the day of data collection. Antidepressants were the most common medication (93%), followed by mood stabilizers (33%). More than half (51%) reported taking more than one type of psychotropic medication. No statistically significant differences in race-ethnicity, age, or incarceration history were found between participants currently taking medication and those who took it only in the past.

Univariate Analyses

We calculated mean scores and standard deviations for responses to the questionnaire items pertaining to experiences with psychotropic medication use (Table 3). Perhaps most interesting was the overwhelming endorsement of items related to positive effects of psychotropic medication, such as “For me, the good things about medication outweigh the bad” (mean=7.48 on a scale of 1 to 10, with higher scores indicating stronger agreement) and of items related to personal agency in decision making regarding medication use, such as “I take medications of my own free choice” (mean=9.17)—the highest mean score among all items. On the other hand, items pertaining to seemingly negative aspects of medication use were not endorsed. For example, participants generally disagreed with the statements “My medication is a way the prison controls me” (mean=2.18) and “I feel ashamed that I take medication” (mean=2.28). Mean scores on the three items intended to measure internal locus of control were higher than those on the three items intended to measure external locus of control, indicating that respondents in general had greater internal locus of control.

Exploratory Factor Analysis

The exploratory factor analysis of the 22 questionnaire items pertaining to experiences with psychotropic medication yielded a six-factor solution that explained 62% of the variance, suggesting that the questionnaire captured six dimensions of the women’s experiences with psychotropic medication use. One item (“I know the names and dosages of all my medications”) was not assigned to a factor because of its low communality (<.3) and factor loading score. Based on a factor loading threshold of .40, we assigned items to six factors, which were named thematically as six dimensions of psychotropic medication–related experience: perception of therapeutic effects, perception of the impact of medication on life, perception of medication as a means of control, perception of personal agency regarding medication use, experience of stigma related to medication use, and feelings of biological vulnerability.

Linear Regression

Six linear regression analyses were performed to examine predictors of women’s scores on subscales representing of each of the six dimensions of experience with psychotropic medication (Table 4). For each analysis, the independent variables were age group, race-ethnicity, number of incarcerations, years incarcerated, external locus of control, current versus past medication use, and number of psychiatric diagnoses. External locus of control significantly predicted a greater likelihood of experiencing stigma related to medication use (B=.35, t=3, df=266, p=.003), an increased perception of medication as a means of control (B=.33, t=3.65, df=266, p<.001), a decreased perception of personal agency regarding medication use (B=–.16, t=−2.09, df=266, p=.038), and increased perception of biological vulnerability (B=.35, t=.08, df=266, p=.001). Not surprisingly, current medication use significantly predicted both a more positive view of therapeutic effects (B=10.33, t=5.28, df=266, p<.001) and a perception of a high impact of medication on life (B=8.2, t=5.88, df=266, p<.001). Current medication use also significantly predicted a stronger sense of personal agency related to medication use (B=1.37, t=2.44, df=266, p=.015). Number of self-reported psychiatric diagnoses significantly predicted both an increased perception of medication’s impact on life (B=1.03, t=3.26, df=266, p=.001) and an increased perception of biological vulnerability (B=1.22, t=6.82, df=266, p<.001). Number of years incarcerated was also a statistically significant predictor of the experience of stigma (B=.22, t=2.89, df=266, p=.004) as well as a perception of more positive therapeutic effects (B=.35, t=2.05, df=266, p=.042). In addition to current medication use, age group was a highly significant predictor of perception of high impact of medication on life (B=2.52, t=3.61, df=267, p<.001). Race-ethnicity and number of incarcerations were not statistically significant predictors in any model.

ANCOVA

To explore the potential role of polypharmacy in women’s experiences with medication use, we conducted an ANCOVA to examine differences in mean scores on dependent variables for which current medication use was a statistically significant predictor between those taking one type of medication and those taking more than one type. The number of medications currently being taken had a significant effect on perceived impact of medication (F=13.35, df=3 and 258, p<.001); the more types of medication being taken, the more likely the woman was to perceive medication as having a high impact on her life.

Thematic Analysis of the Open-Ended Question

A total of 197 participants (72%) responded in some way to the single open-ended questionnaire item, “What is the most important thing you would want people to know about your experience with mental health medication?” A content analysis of the responses yielded three dominant themes and three less prominent but still important ideas. The dominant themes can be encapsulated with the sentence “It helps me [do this and not do that] under certain “right” conditions, and only through a time-consuming trial-and-error process.” Of those who responded to the question, 42% (N=83) used some form of the word “help” in their response; the word “right” appeared 59 times in the responses. Participants specifically wanted others to know that psychotropic medication helped them achieve positive change, such as staying calm or “sane” and staying stable or focused, and helped them cope, get rest, “face what comes,” control their temper, and “be a good person.” Medication likewise helped these women avoid negative symptoms or outcomes, such as depression, mood swings, suicidal ideation, apathy, being scared, or just “being a worse person.”

A caveat to the strong endorsement of the positive effects of psychotropic medication was the strong cautionary note that medication is helpful only under the “right” circumstances—that is, if the person needs it, and it is the right medication, in the right combination, at the right dosage, taken in the right manner, and for the right reasons. Respondents repeatedly indicated that they wanted people to realize that ensuring this set of circumstances is a lengthy and imperfect process of trial and error. Many women also offered direct advice to others (don’t be ashamed of taking medication, and work with your doctor) and expressed concerns about the structural issues in prisons and in society that influence their mental health care (limited choices of medication types behind bars, prison schedule dictating when medications are distributed, and stigma).

Discussion

Findings indicate that these incarcerated women were experiencing a range of often co-occurring mental, emotional, and behavioral disorders, which echoes the literature attesting to the widespread occurrence of mental health issues in correctional settings (1). Findings from the statistical and content analyses affirmed that most of the women associated psychotropic medication use with positive effects and high impact, even though they experienced unwanted physical side effects, especially for those currently taking medication. There was a surprisingly high level of personal agency (decision making and choice) reported, which is in contrast to studies suggesting that incarcerated women and men perceive medication as a means of carceral control (10,26). High endorsement of the item “I know the names and dosages of all my medications” (mean=8.37) further supports the picture that emerged of these women’s active involvement in their mental health care.

The linear regression analyses identified locus of control as a significant predictor of aspects of medication experiences and beliefs related to decision making, power, and interpersonal interactions, suggesting that locus of control represents a potentially important concept for understanding incarcerated women’s experiences with mental health treatment. Future research should investigate how locus of control interacts with specific mental health treatment approaches and institutional policies to predict treatment outcomes.

Regarding the findings from the open-ended question, it is important to note that the question asks specifically what these women wanted other people to know about their medication use. In this respect, the fact that the women wanted people to realize that their medication helped them a great deal and in many ways may indicate they do not think that others—including psychiatric service providers—know that or believe that medication is important to them. Indeed, data from mental health staff who participated in the prestudy focus group suggested that these clinicians are concerned about the medicalization of distress and overreliance on psychotropic medication by these women. Further research should explore whether these women have a sense of urgency to be understood and accepted for decisions they make about psychotropic medication use. Do the women perceive medications in particular to be more important because of the limited access to other forms of mental health treatment? Because perceived stigma was predicted by years of incarceration, more research may also be needed to tease out whether stigma is related to mental illness, medication, or criminal justice involvement in general.

The study was an exploratory study using a nonrandomly selected group of women offenders, and generalizability may be limited. Idiosyncratic institutional factors may have shaped participant experiences with psychotropic medication. In addition, data were self-reported, and we were not able to check data against existing records. Our measure included several items drawn from existing instruments that use different response formats, and thus most items were adjusted from their original format and may not have retained their reliability and validity. It is important to note that the low endorsement of the questionnaire item about bartering of medication (mean=1.09) may suggest social desirability bias because this finding is inconsistent with literature stating that bartering is a common practice in prisons (3,10). Because the mean and standard deviation for this item were lower than for the other items, it may indicate that social desirability bias did not extend to other items. Finally, we cannot rule out a nonresponse bias. We speculate that women not currently taking medication were less likely to participate because the research may have seemed less relevant to their current experiences.

On the other hand, this is the first study of incarcerated women’s experiences that focuses on beliefs and perceptions about psychotropic medication rather than adherence. The correctional facility provided unprecedented access to these women, a substantial number of whom participated. The large sample also created opportunities for robust statistical analysis as well as a unique content analysis of a large database of narrative responses.

Conclusions

This study's aim was to enrich the current knowledge base regarding incarcerated women's experiences and beliefs about psychotropic medication. The high rate of psychotropic medication use in this group of women—as well as the interesting findings pertaining to locus of control, perceptions of therapeutic effects, impact, personal agency, and stigma—support the need for fuller exploration of the topic. Specifically, future research should examine the individual and institutional characteristics that influence incarcerated women's perceptions about medication and other forms of mental health treatment. The hope is that a more nuanced and comprehensive understanding of women's experiences might enable psychiatric service providers, program developers, and advocates to design and deliver more tailored, effective, and compassionate mental health care in forensic settings.