Suicidal Ideation of Probationers. Gender Differences.

BACKGROUND: Gender is often related to different life stressors and mental health disorders, but a limited amount of research examines risks of suicidal ideation of probationers by gender. AIMS: The aim of this study was to examine gender differences in suicidal ideation of probationers.
METHOD: Using a national sample of 3,014 male and 1,306 female probationers with data from the National Survey on Drug Use and Health (2009-2011), multivariate regression analysis was conducted.
RESULTS: Male and female probationers display similar demographic characteristics although their life circumstances and experiences seem different. Female probationers in the study were more likely to experience financial, psychological, and residential stressors than male probationers were. Female probationers were also more likely to have received medical and/or psychiatric treatments. Female probationers were exposed to more suicidal ideation risks than male probationers were. Additionally, no protective factors to suicidal ideation were found for female probationers.
CONCLUSION: The findings suggest that a gender-specific approach to suicidal ideation of probationers may lessen the prevalence of suicidal ideation of this largely neglected population.

Probationers make up the largest segment of the criminal justice population, and there are approximately four million active probationers in the US in any given year (Maruschak & Bonczar, 2013). Unfortunately, almost 10% of probationers return to custody owing to probation violations or new arrests (Maruschak & Bonczar, 2013). Another 5% of probationers do not successfully complete their probation terms for other reasons. One of the most consistent predictors of technical violations during community supervision is mental health problems (Manchak, Skeem, Kennealy, & Eno Louden, 2014; Messina, Burdon, Hagopian, & Prendergast, 2004; Minor, Wells, & Sims, 2003). The presence of suicidal ideation (SI) reflects stressors or maladjustment to life (Fitzpatrick, Irwin, LaGory, & Ritchey, 2007), need for psychiatric care (Wasserman, McReynolds, Schwalbe, Keating, & Jones, 2010), or lack of hopeful outlook (Chin & Holden, 2013), and may interfere with successful completion of probation sentences. Therefore, it is important to examine the presence of SI among probationers to help them avoid revocation, which risks incarceration. Additionally, contact with the criminal justice system is related to suicide risk (Patterson, 2013; Webb et al., 2011). Studies show that offenders serving their time in the community have higher risks of death from suicides, accidents, and drug overdose (Sattar, 2003). Nevertheless, SI among probationers is an under-researched topic.

Despite a well-established gender difference in SI and suicidal behaviors (Canetto, 2009; Clark et al., 2013), until recently much of the research on suicide was carried out primarily on males or on both genders without examining gender differences (Gradus, Street, Suvak & Resick, 2013). This is understandable given that women make up only 7% of inmates (Carson & Sabol, 2012) and 10% of parolees (Maruschak & Bonczar, 2013). However, the sheer number of female probationers, about one million or 25% of four million probationers, justifies the need for a study to understand gender-specific predictors of SI among probationers. The current study addresses this gap by investigating gender differences of probationers in SI. Understanding gender-specific predictors of SI will help probation authorities identify at-risk population and aid in providing appropriate resources and intervention strategies to lower the prevalence of SI among probationers.

SI and Suicide Attempts

As mentioned, a higher risk of death from suicides, accidents, and drug overdose has been reported for offenders serving time in the community (Sattar, 2003). A seminal study on suicide of probationers shows that about a third of probationers have SI (Akehurst, Brown, & Wessley, 1995). A study in Sweden estimated over 20% of the criminal justice-involved population have attempted suicide (Hakansson, Bradvik, Schlyter, & Berglund, 2010).

Probationers with SI have high unemployment rates, major depression, marital separation, substance use, and history of suicide attempts (Akehurst et al., 1995). History of previous child abuse and poor educational adjustment are also identified as risk factors. While it is not identified as a risk factor, male probationers displayed high residential instability as a group, which presented difficulties for follow-up (Pritchard, Cox, & Dawson, 1997). In a recent study on suicide attempts among a community corrections population, residential instability was identified as a risk factor (McCullumsmith, Clark, Perkins, Fife, & Cropsey, 2013).

Prior suicidal attempts and gender (being male) are the strongest predictors of completed suicide for the community corrections population, which includes parolees, probationers, and those released on work order (Clark et al., 2013). Additionally, lifetime exposure to physical injuries, depressive symptoms, and traumatic life experience before the age of 18 increase the risk of SI and suicide attempts (Gunter, Chibnall, Antoniak, Philibert, & Hollenbeck, 2011). Traditional suicide risk factors including unemployment, no access to health care, mental health issues, and substance dependence were all identified as risk factors for suicide attempts among a community corrections population (McCullumsmith et al., 2013). A recent study on SI of parolees showed that access to health care may lower the chances of SI among parolees (Yu, Sung, Mellow, & Shlosberg, 2014). Being older, White, and having a medical condition also increased the risk of suicide (Clark et al., 2013). Risk factors of SI and suicide attempts among community corrections population seem to be different by race and gender (McCullumsmith et al., 2013). Specifically, living alone is related to increased SI for African American male individuals but not for White male or female individuals (McCullumsmith et al., 2013).

For prisoners, being young, single, or having lower education or social or familial isolation increased SI and suicidal behaviors (Jenkins et al., 2005). A prior suicidal attempt (Fruehwald, Matschnig, Koenig, Bauer, & Frottier, 2004) and substance dependence were also related to increased risks for prisoner suicides (Jenkins et al., 2005; Way, Miraglia, Sawyer, Beer, & Eddy, 2005). A substantial number of prisoners who completed suicide had mental illness (Fruehwald et al., 2004; Mumola, 2005; Shaw, Baker, Hunt, Moloney, & Appleby, 2004). While a prospective study estimates about a third of jail inmates who completed suicide had a history of mental illness (Hayes, 2010), retrospective studies show that more than 70% of prisoners who had completed suicide had mental illness (Patterson & Hughes, 2008). The most commonly found mental disorders among prisoners of suicide were depression (Hayes, 2010) and neurotic disorder (Jenkins et al., 2005). In another study on criminal justice populations entering prison and jail, substance dependence, medical and mental health issues, and experience in sexual and physical abuse were identified as risk factors (Hakansson et al., 2010).

In the general population, both mood disorders and substance dependence are shown to be predictors of suicidal behaviors (Brown, Beck, Steer, & Grisham, 2000; Conner & Chiapella, 2004; Moscicki, 1997; Nock, Borges, Bromet, Alonso, et al., 2008). Poor physical health (Duberstein, Conwell, Conner, Eberly, & Caine, 2004) or living alone (Haw & Hawton, 2011) is associated with increased suicidal behaviors, whereas the opposite holds true for religious involvement (Dervic et al., 2004). Additionally, traumatic or stressful life events (Nock, Borges, Bromet, Cha, et al., 2008) and proactive and reactive aggression and impulsivity (Conner, Swogger, & Houston, 2009; Koller, Preuss, Bottlender, Wenzel, & Soyka, 2002) are also associated with increased suicidal behaviors.

Research Questions

Gender is correlated with the prevalence of certain mental disorders, including depression, anxiety, and somatic complaints (World Health Organization, 2000), the exposure to stressful life events (Kendler, Thornton, & Prescott, 2001), and the deployment of stress-coping responses and resources (Carlson & Grant, 2008; Taylor et al., 2000). Data and analysis in this study addressed a set of three basic questions: (1) What are the risk factors of SI shared by both male and female probationers? (2) What are the SI risks unique to male probationers? And (3) what are the SI risk factors unique to female probationers? Answers to these questions will inform the practice of probation supervision and community health.

Method

Data

Three years of National Survey on Drug Use and Health (NSDUH) data from 2009 to 2011 were analyzed to examine predictors of SI among probationers by gender. The Substance Abuse and Mental Health Services Administration (SAMHSA) conducts the NSDUH annually (SAMHSA, 2012) to obtain state and national estimates on subtance use. The sample size is about 70,000, utilizing a multistage area probability sampling of civilians aged 12 years and older. Young age groups (aged between 12 and 25 years) are oversampled to obtain estimates by racial subgroups (Gordek & Folsom, 2012). Since 2002, respondents have received US $30 as incentive for their participation. For the 2011 survey, the weighted response rate was 75%. The survey is conducted in person with a computer-assisted personal interview in the privacy of the respondent's own home. NSDUH aims to provide state and national estimates on the prevalence of legal and illegal drug and substance use. Therefore, each respondent was given person-level weight based on a multistaged sampling method calculated for each state (Morton, Martin, Shook-sa, Chromy, & Hirsch, 2012).

Study Sample

The questions related to mental health were asked only to those aged 18 years and older at the time of survey. Therefore, responders under the age of 18 were excluded from the analysis. The probation status was determined by the question: "Were you on probation at any time during the past 12 months?" Those who answered "yes" to this question were identified as probationers. The final sample consisted of 3,014 male and 1,306 female probationers, from 2009 to 2011.

Dependent and Predictor Variables

The outcome variable was measured using a question in the NSDUH. The respondents were asked: "At any time in the past 12 months … including today, did you seriously think about trying to kill yourself?" On the basis of the existing literature on SI, we selected 34 variables; five variables each from demographic, social, and economic characteristics. Additionally, we examined 13 variables related to health status and access to health care, and six variables in substance use and criminal behaviors. See Table 2 for a full list of variables included in the analysis.

Table 2

Probationers' suicidal ideation and characteristics by gender

Characteristics	% Male	% Female	OR	95% CI	p
Note. Unweighted data from the National Survey on Drug Use and Health, 2009–2011. DV = dependent variable. DUI = driving under the influence. ns = not significant.
DV: suicidal ideation	8.0	12.8	1.657	1.369–2.079	< .001
Demographics
Black	16.6	16.6	1.004	0.843–1.196	ns
Hispanic	17.8	17.8	1.006	0.849–1.192	ns
Aged 29 years and younger	78.3	77.3	0.940	0.804–1.098	ns
Married	13.2	14.2	1.092	0.905–1.317	ns
High school completion	66.2	69.2	1.150	1.000–1.323	.049
Social
Attended religious services in the past month	48.2	58.1	1.491	1.308–1.700	< .001
Religious belief is important in your life	59.8	68.1	1.439	1.254–1.651	< .001
Live alone (one-person household)	7.6	6.0	0.779	0.598–1.015	ns
Have children under 18 in household	41.5	55.1	1.729	1.517–1.971	< .001
Moved more than once in the past year	24.5	30.2	1.332	1.153–1.539	< .001
Economic
Employed	77.5	81.6	1.289	1.094–1.519	.002
Below poverty line	27.6	38.7	1.659	1.446–1.902	< .001
Personal income greater than US $19,999	25.6	14.2	0.478	0.401–0.570	< .001
Family income greater than US $19,999	66.3	54.0	0.596	0.522–0.680	< .001
Received government assistance in past year	33.7	50.0	1.967	1.723–2.244	< .001
Health and healthcare within the past 12 months
Poor health	1.7	3.0	1.753	1.152–2.669	.009
Missed work due to injury or illness in the past month	14.9	16.3	1.113	0.932–1.330	ns
Health-care coverage	56.7	70.8	1.847	1.607–2.213	< .001
Any time not covered by health insurance	10.9	14.8	1.415	1.169–1.713	< .001
Serious psychological distress	20.7	34.8	2.048	1.773–2.365	< .001
Major depressive episode	8.2	19.1	2.652	2.194–3.206	< .001
Been to emergency room	40.1	52.8	1.667	1.463–1.900	< .001
Stayed overnight as inpatient in hospital	8.2	16.8	2.279	1.876–2.770	< .001
Prescription medication for mood disorder	4.2	13.7	3.640	2.869–4.619	< .001
Received prescription medication for mental health	11.0	26.9	2.689	2.308–3.134	< .001
Received inpatient mental health treatment	2.7	4.2	1.572	1.110–2.226	.011
Received outpatient mental health treatment	8.0	19.0	2.709	2.238–3.280	< .001
Received alcohol or drug treatment	24.6	21.0	0.816	0.697–0.955	.011
Substance use and criminal behaviors in the past 12 months
Used alcohol	66.3	58.3	0.710	0.621–0.811	< .001
Used illicit drugs (marijuana, heroin, crack, cocaine, inhalants, hallucinogens, and psychotherapeutics)	36.8	31.9	0.802	0.698–0.920	.002
Attacked someone with intent to seriously hurt them	11.8	9.5	0.781	0.629–0.969	.024
Arrested and booked for any reason	48.2	41.7	0.768	0.674–0.876	< .001
Arrested and booked for violence	4.4	3.6	0.815	0.580–1.144	ns
Arrested and booked for DUI	12.4	9.3	0.725	0.585–0.899	0.003
Arrested and booked for possession/sale of drug	11.0	7.4	0.646	0.510–0.818	< .001

Statistical Analysis

Characteristics of probationers were compared using 34 variables by gender. Multivariate logistic regression analysis by gender was conducted using SI as the dependent variable in SPSS 21. Statistical significance was indicated by providing p value as well as 95% confidence interval (95% CI). Because 34 predictors would create a cumbersome model and dilute statistical power, many variables were dropped from the regression model. The process ensured that all four aspects (demographic, social, economic, and health-related variables) were represented in the model. All criminal justice-related variables were excluded from regression analysis; it was not possible to determine whether the stated criminal justice involvement was the cause of their probation sentence, or whether it happened after being sentenced to probation.

In the multivariate analysis, we dropped "attending religious services" but kept "religious belief" in the model. For economic characteristics, we included employment status. We included health-related variables shown to be related to increased SI: serious psychological distress, major depressive episode, and mental health treatment. The first variable was created by combining the K6 scales. The second variable was created using 22 questions to assess MDE symptoms based on the fourth edition of the Diagnostic and Statistical Manual of Mental Disorders (detailed information is provided in RTI International, 2010). Lastly, we incorporated residential mobility and engagement in violent assault in the past year in the model because they are known to have differential impacts between male and female subjects. On the one hand, women experience more stress from residential instability than men do and the stress resulting from residential mobility is more likely to result in depression and a lower sense of well-being in women (Magdol, 2002; Oishi, 2010). On the other hand, women who use violence in their intimate and social relationships are also more likely to display symptoms of depression and posttraumatic stress disorders (Graham, Bernards, Flynn, Tremblay, & Wells, 2012; Swan, Gambone, Fields, Sullivan, & Snow, 2005).

Results

Prevalence of SI and Gender Differentials

The first finding that surfaces from our data analysis is the magnitude of the problem of SI among probationers. The second finding relates to the gender differentials in the scope of this specific health crisis. Table 1 displays our national estimates of SI among probationers and the rest of the population aged 18 and older in the community based on weighted NSDUH data for the 2009–2011 period. Probationers under community supervision demonstrated prevalence rates that were alarmingly higher than those observed among the nonprobationer population. Over the 3-year period, the average annual prevalence rate of SI among probationers (9.7%) surpassed that among nonprobationers (3.6%) by a factor of 2.69.

Table 1

Prevalence of suicidal ideation of probationers and nonprobationers by gender

	Male	% Suicidal ideation	Female	% Suicidal ideation	Total	% Suicidal ideation
Note. Weighted data from the National Survey on Drug Use and Health, 2009–2011.
Probationers
2009	3,917,803	8.1	1,346,982	11.9	5,264,785	9.0
2010	3,759,662	7.9	1,696,678	16.5	5,456,340	10.6
2011	3,362,813	7.4	1,408,645	13.9	4,771,458	9.3
Average	3,680,093	7.8	1,484,102	14.3	5,164,194	9.7
Nonprobationers
2009	105,805,392	3.4	116,136,369	3.8	221,941,761	3.6
2010	107,209,704	3.6	116,606,533	3.7	223,816,237	3.7
2011	108,496,109	3.3	119,357,732	3.9	227,853,841	3.6
Average	107,170,402	3.5	117,366,878	3.8	224,537,280	3.6

It is impossible to ignore gender disparities in the risks of developing SI. Nonprobationer women consistently showed a higher annual prevalence rate than nonprobationer men in 2009, 2010, and 2011 by a factor of 1.12, 1.03, and 1.18 respectively (see Table 1). Nevertheless, the same ratios of female prevalence to male prevalence among probationers for the same period of time were 1.47, 2.09, and 1.88 respectively. A comparison of the average prevalence rate for the 3-year period reveals that women probationers were not only 1.83 times as likely as men probationers to have had SI in the past 12 months but also 3.76 times as likely as nonprobationer women in the community to have suffered SI, demonstrating that SI is clearly a correctional health crisis, in particular for women offenders.

Characteristics of Probationers by Gender

Table 2 presents the characteristics of probationers by gender. Female probationers were more likely to have SI than male probationers (OR = 1.657, 95% CI = 1.369–2.079). While male and female probationers were similar in respect to their race, ethnicity, marital status, and age group, there were stark differences between them in social and economic characteristics. Female probationers were more likely to have graduated high school (OR = 1.150, 95% CI = 1.000–1.323), were more likely to have attended religious services (OR = 1.491, 90% CI = 1.308–1.700), and more likely to state that their religious belief was important in their life (OR = 1.439, 95% CI = 1.254–1.651) compared with male probationers. Additionally, female probationers were more likely to have children under the age of 18 in the household (OR = 1.729, 95% CI = 1.517–1.971). Nevertheless, they experienced more residential instability than men did; female probationers were more likely to have moved more than once in the past year alone (OR = 1.332, 95% CI = 1.153–1.539) compared with male probationers.

Although female probationers were more likely to be employed (OR = 1.289, 95% CI = 1.094–1.519) than their male counterparts were, female probationers were more likely than male probationers to fall below the poverty line (OR = 1.659, 95% CI = 1.446–1.902) and to have received government assistance – which includes Supplemental Security Income, food stamps (also known as Supplemental Nutrition Assistance Program), cash and non-cash assistance programs (OR = 1.967, 95% CI = 1.723–2.244). Furthermore, female probationers had lower personal and family income than male probationers; they were less likely to have personal annual income in excess of US $20,000 (OR = 0.479, 95% CI = 0.401–0.570), or family income in excess of US $20,000 (OR = 0.596, 95% CI = 0.522–0.680). In addition to having more financial stressors than male probationers, female probationers were more likely to have experienced serious psychological distress (OR = 2.048, 95% CI = 1.773–2.365) and a major depressive episode (OR = 2.652, 95% CI = 2.194–3.206). Female probationers were more likely to have received prescription medication for mood disorder (OR = 3.640, 95% CI = 2.869–4.619), and inpatient (OR = 1.572, 95% CI = 1.110–2.226) and outpatient (OR = 2.709, 95% CI = 2.238–3.280) mental health treatments. However, they were less likely to have received substance use (alcohol or drug) treatment (OR = 0.816, 95% CI = 0.697–0.955) than male probationers were.

There were significant differences between male and female probationers in health status and access to health care. Female probationers were more likely to have access to health care (OR = 1.847, 95% CI = 1.607–2.213), but they were also more likely to have had a lapse in the coverage (OR = 1.415, 95% CI = 1.169–1.713). Female probationers were more likely to report that their health was poor (OR = 1.753, 95% CI = 1.152–2.669). Furthermore, female probationers were also more likely to have stayed overnight in a hospital (OR = 2.279, 95% CI = 1.876–2.770), and to have been to an emergency room (OR = 1.667, 95% CI = 1.463–1.900) in the past 12 months.

Even among this criminal justice-involved population, female probationers were still closer to being "good girls," and male probationers were still closer to being "bad boys." Female probationers were less likely to have used alcohol (OR = 0.710, 95% CI = 0.621–0.811), or illicit drugs (OR = 0.802, 95% CI = 0.698–0.920) than male probationers. They were also less likely to have attacked someone with intent to seriously hurt them (OR = 0.781, 95% CI = 0.629–0.969), and less likely to have been arrested and booked in all crime types examined (OR = 0.768, 95% CI = 0.674–0.976).

Predictors of SI by Gender

Multivariate logistic regression analysis was conducted to examine predictors of SI of probationers. Table 3 presents the multivariate logistic regression results. For both male and female probationers, many sociodemographic characteristics were not associated with SI. One exception was race for female probationers: being Black doubled the odds of having SI (OR = 2.014, 95% CI = 1.247–3.254). However, this was not the case for male probationers. Having religious belief significantly reduced the odds of SI for male probationers (OR = 0.586, 95% CI = 0.433–0.791), but not for female probationers. Use of illicit drugs was related to increased SI for male probationers only (OR = 1.668, 95% CI = 1.236–2.252).

Table 3

Multivariate logistic regression analysis on probationers' suicidal ideation by gender

	Male (N = 3,014)			Female (N = 1,306)
Predictors	OR	95% CI	p	OR	95% CI	p
Note. Unweighted data from the National Survey on Drug Use and Health, 2009 – 2011. ns = not significant.
Black	0.934	0.613–1.422	ns	2.014	1.247–3.254	.004
Hispanic	0.650	0.407–1.036	ns	0.772	0.411–1.451	ns
Aged 29 years and younger	1.135	0.768–1.677	ns	0.853	0.542–1.345	ns
Married	1.135	0.698–1.845	ns	1.439	0.830–2.496	ns
High school completion	0.924	0.672–1.270	ns	0.949	0.632–1.424	ns
Employed	0.935	0.658–1.328	ns	0.822	0.525–1.287	ns
Religious belief is important in my life	0.586	0.433–0.791	<.001	0.958	0.644–1.425	ns
Used illicit drug in the past month	1.668	1.236–2.252	.001	1.343	0.914–1.973	ns
In the past year:
Moved more than once	0.981	0.704–1.366	ns	1.896	1.290–2.787	.001
Serious psychological distress	5.222	3.759–7.257	<.001	4.969	3.130–7.888	< .001
Major depressive episode	5.125	3.578–7.341	<.001	3.080	2.031–4.672	< .001
Received inpatient mental health treatment	2.605	1.470–4.616	.001	3.801	1.951–7.407	< .001
Attacked someone with intent to seriously hurt them	1.181	0.793–1.759	ns	1.691	1.009–7.407	.046
Constant	.035		<.001	.030		< .001
Cox–Snell R2	.123			.175
Nagelkerke R2	.289			.328

For both groups, having serious psychological distress increased the odds of SI: for male probationers, the odd ratio was 5.222 (95% CI = 3.759–7.257) and for female probationers, it was 4.969 (95% CI = 3.130–7.888). Experiencing a major depressive episode (for males, OR = 5.125, 95% CI = 3.578–7.341; for females, OR = 3.080, 95% CI = 2.031–4.672) and receiving inpatient mental health treatment (for males, OR = 2.605, 95% CI = 1.470–4.616; for females, OR = 3.801, 95% CI = 1.951–7.407) were predictors of SI for both groups.

The two variables hypothesized to behave differently based on gender displayed a different relationship in male and female probationers. Residential instability, measured by moving more than once within the past year, increased the odds of SI for female probationers (OR = 1.896, 95% CI = 1.290–2.787) whereas it was not significant for male probationers. Attacking someone with intent to seriously hurt them was related to increased SI for female probationers (OR = 1.691, 95% CI = 1.009–7.407) but not for male probationers. Overall, male probationers had four risk factors and one protective factor for SI. Female probationers had six risk factors but no protective factors.

Discussion

The current study investigated gender-specific predictors of SI of probationers, a largely neglected research topic. Analysis showed that female probationers were almost twice as likely to have SI than male probationers were. To understand the gender differences, we analyzed various characteristics of probationers. The analysis showed that while demographic characteristics of female and male probationers are similar, their life circumstances and experiences are significantly different. Even religious involvement, a common protective factor against suicidal behaviors (Dervic et al., 2004; Stack & Kposowa, 2011), was not related to reduced SI for female probationers in our study. This is surprising since female probationers were more likely to attend religious services and express higher religiosity compared with male probationers. While we do not have any data to confirm it, this may be a reflection of female probationers relying on religion owing to lack of alternatives or social capital. It is also possible that female probationers simply experience more life crises than male probationers do. Despite a higher percentage of female probationers having achieved a minimum of high school education, they were more likely to have lower income than their male counterparts. Financial strain on female probationers is even more apparent: Female probationers had decreased odds of .4 of having a family income greater than $19,999 compared with male probationers. A possible explanation may be that female probationers have dependents in the household (i.e., children under 18 or those not employed) whereas male probationers have partners in the household. However, this cannot be verified using the current data.

Female probationers also displayed residential instability: Almost a third of them moved more than once within the past 12 months. Female probationers were more likely to have experienced serious psychological distress or a major depressive episode. They were more likely to have received medication or treatment for mental and physical health conditions, whereas male probationers were more likely to have also received substance use treatment. This difference seems to reflect the higher level of substance use among male probationers than female probationers, and higher mental stressors among female probationers. Additionally, while more female probationers had access to health-care coverage than male probationers, a higher number of female probationers also reported lapses in health-care coverage, which signals instability in their life. Higher percentages of female probationers reported having poor health, utilizing emergency room services, or staying overnight in a hospital. Given that women are more likely to have failed suicide attempts than men (Canetto & Sakinofsky, 1998; Schrijvers, Bollen, & Sabbe, 2012), it is possible that their utilization of emergency medical services or overnight stay in hospital was due to attempted suicide. However, there are no relevant data to confirm it.

Etiological sources of SI could be varied and multiple among probationers. Even when SI evolves from thoughts and preoccupation to intend or attempt suicide, the cognitive-behavioral evolution of self-harm can be interpreted as a means of relieving stress, anxiety, depersonalization, or gaining control of one's environment (Farmer, Felthous, & Holzer, 1996; Serin, Motiuk, & Wichmann, 2002). SI among offenders under community supervision itself could be associated with one of the many clinical diagnoses including depression, schizophrenia, panic disorders, substance use disorders, posttraumatic stress disorder, and psychopathy (Bainbridge, 2012; Miranda et al., 2008). Female offenders are known for showing disproportionately higher levels of childhood abuse and intimate partner violence (Lake, 1995; Walsh, DiLillo, & Scalora, 2011), thus they may be at a heightened risk of developing SI (Thompson et al., 2012).

The current study showed that serious psychological distress, major depressive episodes, and receiving inpatient treatment were risk factors of SI. The medical and psychiatric implications of these observations render the incorporation of mental health care a requirement (Mann et al., 2005) in the community supervision of probationers, who have historically faced the enduring challenge of obtaining government-sponsored or employment-based health insurance. The Affordable Care Act (ACA) may potentially help probationers at risk of SI and suicide attempt to receive the kind of care and services they need. ACA allows states to expand Medicaid coverage to all uninsured residents younger than 65 years of age with incomes below 133% of the federal poverty level (Bainbridge, 2012). Furthermore, preventive care, early intervention, and treatment of mental health problems and substance use disorders will be considered essential health benefits under the ACA. Correctional authorities in general, and probation agencies in particular, will have to ride the waves of the expansion of Medicaid, investments to be made in health information technology, establishment of health insurance exchanges, and minimum essential coverage to serve the correctional populations.

Each year probation officers and supervisors are in direct contact with millions of criminal offenders at a heightened risk of suicide across the country. Therefore, it is critical for probation officers and supervisors to become effective gatekeepers. The current study demonstrates a need for a gender-specific approach to lessen SI of probationers – in particular, female probationers. Specifically, ensuring residential stability, offering anger management courses, or continued access to medical care to female probationers may lessen some stressors in their life. The probation officers need to be trained to recognize probationers with risk factors for SI and suicidal behaviors, in the practice of interventions appropriate for non-mental health professionals. The probation agencies need to develop a partnership with qualified mental health professionals for assessment and crisis intervention services. Training curricula should increase the awareness of risk factors, address issues related to making appropriate referrals to qualified mental health professionals, and provide a practical model for taking action to prevent suicide. Additionally, a case management system of at-risk probationers ensuring sufficiently increased exposure to various resources may lead to decreased SI and successful completion of probation (Guydish, 2011).

There are some limitations in the study. The data used in the current study are from a survey relying on self-reporting. Therefore, the data may suffer from underreporting owing to social stigma or social desirability (Heerwegh, 2009). Additionally, it is possible that being placed on probation may make one depressed or stressed, and it may be a main cause of SI. It is also possible that women may perceive being placed on probation as more stressful than men. Nevertheless, the cross-sectional nature of the data does not allow us to test the causal relationship. Furthermore, studies investigating the impacts of suicide attempts on probationers and predictors of progression of SI to suicide attempt will help explain how SI and behaviors may be reduced.

Conclusion

Probation gives individuals a chance to avoid a prison term, and it is important to help probationers adhere to various conditions imposed upon them so that they can desist from criminal involvement. The threats of homelessness, unemployment, social isolation, reoffending, incarceration, and criminal justice surveillance are among the significant life stressors pressing these vulnerable lives. It is likely that probationers with SI may have a more difficult time to successfully complete their sentences. Therefore, knowing predictors of SI will aid probation officers and agencies to identify the at-risk population for SI and suicidal behaviors. The current study revealed significant gender differentials in SI among probationers; for instance, residential instability was a predictor of SI for female probationers, but not for male probationers. Therefore, it may be necessary to expend more resources on housing services for female probationers to decrease SI. Additionally, violent behavior of female probationers should alert probation officers to other stressors in their life, inviting intervention to decrease SI and increase chances of successful completion.