Selected Resource Insecurities and Abstinence Self-Efficacy Among Urban and Rural Incarcerated Women with Opioid Use Disorder

Abstract

Purpose.

To investigate the relationship between selected resource insecurities (i.e., food, housing, and transportation) and abstinence self-efficacy among women with criminal legal system (CLS)-involvement and opioid use disorder (OUD) and examine potential moderating effects of living in a rural or urban area.

Methods.

Data were collected as part of the NIDA-funded Kentucky Justice Community Opioid Innovation Network (JCOIN) trial. Women (N = 900) were randomly selected from nine Kentuckyjails, screened for OUD, enrolled, and interviewed. Measures included resource insecurity and a single-item measure of abstinence self-efficacy.

Findings.

Results from a multivariable regression analysis examining independent correlates of abstinence self-efficacy demonstrated significant associations with food insecurity (p = .008), controlling for sociodemographic covariates. Rurality moderated the relationship between transportation and abstinence self-efficacy (p = .001).

Conclusions.

Resource insecurity, particularly transportation insecurity in rural areas, may reduce abstinence self-efficacy. Addressing resource disadvantages is vital to improving treatment engagement and recovery among women.

The intersection between justice-involvement, substance use, and resource insecurities (or limited access to basic needs such as food and housing) is a complex and pressing issue, particularly for women.^1–3 Among many other consequences to health and well-being, constrained access to resources may have significant repercussions on women’s abstinence self-efficacy, or the perceived ability to remain abstinent from substance use.⁴ A number of women with criminal legal system (CLS) involvement are unable to meet their basic needs,^1,5 increasing strain on mental and physical health,^6,7 which in turn may have negative implications for women’s abstinence self-efficacy. Additionally, communities where women reside may further influence resource insecurity, especially for those living in rural as opposed to urban areas. Yet, literature is limited concerning the influence of resource insecurity on abstinence self-efficacy. This study will thus describe selected resource insecurities among women with CLS involvement in Kentucky with opioid use disorder (OUD) and explore the relationship between selected resource insecurities, abstinence self-efficacy, and rural-urban residence prior to incarceration (PTI).

Resource insecurity and health.

Resource insecurity has been shown to have significant implications for health and overall well-being. Research emphasizes the interconnectedness of broader impacts of economic insecurity (the inability or fear of being unable financially to afford basic needs) and its trickle-down effect on selected resource insecurities, such as housing and transportation.^8–12 This research highlights how the experience of having multiple resource insecurities (particularly food, transportation, and housing) may influence overall health and wellness.⁶

Other studies have examined the effect of food and housing insecurity on other health outcomes (cardiovascular health) and found a link with greater risk of mortality from cardiovascular disease (see recent review by Parekh et al.¹¹). Food insecurity, specifically, has been identified as a substantial health crisis affecting the United States.¹³ Prior studies consistently highlight the behavioral health issues and outcomes associated with food insecurity.^14–16 For instance, there is a positive correlation between food insecurity and substance use,^17–19 as well as a link to poor mental health.⁷ Additionally, the stress of poverty and economic insecurity increases vulnerability for substance use.²⁰ Transportation insecurity has also been underscored as a barrier to health care, negatively affecting care, medication use, and appointment attendance.²¹ Finally, housing insecurity has also been shown to increase risks of substance use disorders and overdose.²² Such studies highlight how selected resource insecurities may affect community health. This, however, will be the first study to examine resource insecurities and women’s self-efficacy for remaining abstinent after release from jail.

Justice involvement and women.

For individuals involved in the criminal justice system, resource insecurity may be further amplified. Challenges related to having a criminal record (e.g., difficulty finding steady employment) may contribute to a heightened risk of economic or financial insecurity^23–25 and thus, affect resource insecurity for such necessities as food, housing, and transportation insecurity. Additional literature has highlighted the role of justice system policies that may contribute to resource insecurity for individuals reintegrating into the community following incarceration. Having limited access to essential resources (e.g., transportation, housing, and food) compounds the health disparities already experienced by CLS-involved populations¹ and increases the risk of recidivism.^2,26

For women with CLS involvement, the intersection of gender and justice may complicate the relationship between resource insecurity and health and well-being. Women overall face gender-specific vulnerabilities, such as lower pay or lack of access to health care,²⁷ that may contribute to greater resource insecurity compared with men.^28,29 The health ramifications tied to resource insecurity are also more pronounced among women.^30,31 For example, Connor et al.³² concluded that gender-related health disparities experienced during the COVID-19 pandemic had a greater impact on women experiencing poverty and housing insecurity compared with men. Other work has highlighted the relationship between economic insecurity and an increased risk of substance use, notably stronger among women than among men.³³ Among women with CLS involvement, difficulty accessing resources and economic insecurity are linked with poorer mental health outcomes³⁴ and a recurrence of behaviors such as substance use.³⁵ Thus, women with CLS involvement seem to face an elevated risk of resource insecurity, potentially influencing their overall well-being.

Rurality and resource insecurity.

Selected resource insecurities among women may also be more pronounced if they live in a rural community compared with an urban community prior to incarceration. Rural communities often experience widespread poverty that, when coupled with the limited availability of community resources and health services,^36,37 may lead to poorer health and decreased life expectancy among residents.^38–40 Specifically, lack of transportation in rural communities increases the difficulty of accessing available resources and may cause isolation from sources of support.^41–43 Moreover, there is evidence that perceived stigma among individuals seeking public assistance serves as a significant barrier to resources in rural populations.^44–46 In rural areas where substance use is prevalent, the scarcity of resources may complicate access to substance use treatment.^3,46 Stigma and resource insecurity may place unique burdens on individuals living in rural areas related to their needs for assistance and healthcare, especially concerning substance use treatment. Thus, women living in rural settings prior to incarceration may experience greater resource insecurity compared with women living in urban areas, contributing to a lower likelihood of engaging in treatment services. Furthermore, the stress of resource insecurity may especially affect rural women’s perceptions of their ability to remain abstinent from substance use.

Self-efficacy and substance use.

Self-efficacy theory posits that an individual’s confidence in their abilities shapes their motivation and capacity to execute behavioral functions required to achieve a desired outcome.⁴⁶ Abstinence self-efficacy, in particular, pertains to an individual’s belief in their ability to refrain from substance use and has been identified as a crucial predictor of drug and alcohol use, as well as recovery more broadly.^4,47 Research suggests that abstinence self-efficacy can be a protective factor against return to use.^4,47,48 Majer et al.⁴⁹ found that low levels of abstinence self-efficacy at residential treatment entry predicted an increased likelihood of returning to use within two years following treatment. Other studies have also noted the significant connection between low abstinence self-efficacy and return to use within the first year of abstinence.^47,48 This suggests that lower abstinence self-efficacy is associated with higher rates of return to use, even years later, while higher self-efficacy may contribute to an increased likelihood of abstinence and recovery over time.

Indeed, women with high abstinence self-efficacy exhibit better treatment outcomes and are more likely to engage in supportive behaviors for treatment recovery, such as adopting coping mechanisms and seeking help from other support systems, such as mental health services.^47,50,51,52 Although research indicates that self-efficacy is closely connected with outcomes related to health and substance use, less is known about factors that may influence self-efficacy itself. It is possible that women experiencing resource insecurity may be less likely to feel confident in their abilities to remain abstinent while basic needs such as food, adequate housing, and reliable transportation remain unmet. However, to date, no research has examined associations between resource insecurities and abstinence self-efficacy among women, specifically among those with the particular vulnerabilities associated with criminal justice system involvement.

Current study.

While research has examined the effects of resource insecurity on general health and well-being, there is a gap in the literature regarding the relationship between selected resource insecurities and abstinence self-efficacy. This relationship may be especially important for women with CLS involvement—who often face significant barriers to accessing resources—and could vary by living in rural areas prior to incarceration, where the prevalence and impact of resource scarcity may differ from those in urban communities. This study aims to fill these gaps by 1) describing selected resource insecurities (i.e., food, housing, and transportation) among women with CLS involvement, 2) investigating the relationship between selected resource insecurities and abstinence self-efficacy in women with CLS involvement and opioid use disorder (OUD), and 3) examining how the relationship between selected resource insecurities and abstinence self-efficacy may vary among women from rural versus urban counties.

Methods

Participants and procedures.

This study uses data from the ongoing (started on December 2020) National Institutes on Drug Abuse (NIDA)-funded Kentucky Justice Community Opioid Innovation Network (JCOIN) cooperative project UG1DA050069⁵² Incarcerated women (18 and older) were randomly selected from nine study jails via daily census reports and privately screened by trained female research staff for eligibility (in-person or via videoconference; all procedures were contingent on COVID-19 protocols at each jail site). Eligibility criteria for participation included: 1) meeting criteria for opioid use disorder (OUD) according to the Diagnostic Statistical Manual (DSM-5) checklist (reporting the presence of two or more criteria; see DSM⁵³) or scoring four or higher on the NIDA-modified Alcohol Smoking Substance Involved Screening Test (NM-ASSIST) for street or prescription opioids;⁵⁴ 2) absence of active psychosis or cognitive impairment; 3) reported remaining period of incarceration between 7–60 days from the date of screening (for purposes of the study intervention); and 4) agreed to participate in study procedures.

Out of the 2,102 women screened for study eligibility, 1,202 were ineligible to participate: 220 had not used opioids in the 12 months PTI, one did not meet criteria for OUD, 105 were transferred to another facility, 780 did not have an anticipated release date within 7–60 days of screening, five were inaccessible due to facility restrictions, and two did not complete baseline interviews. Additionally, 89 women declined to participate in this study (reasons unspecified). Baseline interviews were completed with 900 eligible women directly following screening and informed consent. Interviews were conducted by female staff while women were incarcerated and took place in a confidential setting at each jail. During the baseline interview, participants were asked to respond to questions including demographics, resource insecurities (food, housing, transportation), substance use, and abstinence self-efficacy. Interviews ranged from one to two hours in length. Directly following completion, women were compensated $45 for their time. The study was approved by the University Institutional Review Board (IRB) and a certificate of confidentiality covered all study procedures.

Measures.

Sociodemographic characteristics.

Participants self-reported demographic information, including current age (recorded as a continuous measure), race/ethnicity (1 = non-Hispanic White, 0 = other race/ethnicity), sexual orientation (1 = heterosexual, 0 = non-heterosexual), marital status (1 = married/living as married, 0 = not married/living as married), employment status (1 = employed full or part-time prior to incarceration [PTI], 0 = unemployed PTI), and county of residence PTI. Counties were reclassified using Beale codes (2013 Rural-Urban Continuum Codes), wherein codes one to nine denote a county’s population size and its distance from a metropolitan area, as defined by the United States Department of Agriculture Economic Research Service. Subsequently, these codes were assigned dichotomous classifications, where codes 1 through 3 were designated as urban (= 0) and 4 through 9 were identified as rural (= 1), which aligns with other substance-use research.^55,56

Selected resource insecurities.

Selected resource insecurities prior to incarceration were measured in three areas: insecure access to food, housing, and transportation.

Food insecurity.

The USDA Food Security Scale²⁸ is a 10-item scale measuring reliable access to food that fulfills an individual’s requirements for maintaining a healthy lifestyle (e.g., “We couldn’t afford to eat balanced meals” or “Were you ever hungry, but didn’t eat, because there wasn’t enough money for food?”). The scale was administered to participants, referencing food access in a woman’s household in the 12 months PTI. Questions were rated on one of two three-point scales or on a dichotomous scale. For most questions, 0 = never true, 1 = sometimes true, and 2 = often true were used. Other questions that discussed timeframes (e.g., “How often did this happen?) regarding food insecurity were rated using a second scale (0 = in only one or two months, 1 = some months, but not every month, and 2 = almost every month). Finally, some close-ended questions (e.g., “Did you lose weight because there wasn’t enough money for food?”) used a dichotomous scale (0 = no, 1 = yes). Responses of ‘sometimes’, ‘often’, ‘some months, but not every month’, ‘almost every month’, or ‘yes’ are coded as affirmative to food insecurity and affirmative items are summed to calculate a raw score, ranging from 0–10. To characterize the sample, food security status was determined by cut points, consistent with USDA scoring guidelines: 0 = very low food insecurity, 1–2 = low food insecurity, 3–5 = marginal food insecurity, and 6–10 = high food insecurity. For regression and moderation analyses, food insecurity was analyzed as a continuous measure of 0 to 10, with higher scores indicating less secure access to food.

Housing insecurity.

Women were asked how many days (within 90 days PTI) they had “been homeless or had to stay with someone else to avoid being homeless” and “lived in a homeless shelter or emergency housing.” To describe the sample, results were dichotomized: women who reported one or more days living as homeless or precariously housed (e.g., sleeping on a friends’/family members’ couch) were considered to have experienced housing insecurity 90 days PTI (1 = living as homeless or precariously housed, 0 = not living as homeless or precariously housed). For regression and correlation purposes, however, answers to both questions were summed into a single variable (resulting in a 180-day range), allowing for a greater degree of variance.

Transportation insecurity.

To assess transportation insecurity, women were asked 1) whether they had regular access to a vehicle they could drive and 2) whether they lived in an area that had public transportation, both during the 12 months PTI. Answers to the two questions were combined into a single variable. To describe the sample, a dichotomous variable was created where an answer of “no” to both forms of transportation (e.g., no reliable access to public transportation or a car) indicated transportation insecurity (1 = no access to reliable vehicle or public transportation, 0 = access to reliable vehicle or public transportation). For regression and moderation analyses, results were analyzed using a continuous measure of 0–2 (2 = no access to reliable transportation, 1 = access to one form of reliable transportation, 0 = access to both forms of reliable transportation).

Abstinence self-efficacy.

Similar to previous research,⁵⁷ abstinence self-efficacy was calculated using a single-item measure. This item asked, “On a 10-point scale (1 = not at all confident, 10 = extremely confident) how confident are you that you will be completely abstinent in one year?” Results were analyzed on a continuum.

Analytic plan.

Some women (n = 23) from the baseline interview had incomplete data on one or more variables of interest (missing data: county of residence = 12, education = 8, self-efficacy = 1, housing insecurity = 1) and thus were removed from the study, resulting in a final analytic sample of N = 877. To address aim 1 of this study, descriptive statistics were performed to describe selected resource insecurities among incarcerated women. For aim 2, Pearson correlations and t-tests were calculated to examine the associations between selected resource insecurities (food insecurity, housing insecurity, and transportation insecurity) and abstinence self-efficacy. Additionally, a single linear regression was used to investigate food, housing, and transportation insecurities as independent correlates of abstinence self-efficacy while controlling for demographic variables that were identified prior to research as potentially relevant to abstinence self-efficacy (i.e., age, sexual orientation, marital status, employment, education). To assess multicollinearity, variance inflation factors (VIFs) were examined for all models, with none exceeding 1.1. To investigate aim 3, county of residence PTI (dichotomized as rural-urban) was used as a moderator of the relationship between selected resource insecurities variables and abstinence self-efficacy. Specifically, moderation was assessed using a series of three regression models, each with a unique interaction term (rurality*food insecurity, rurality*housing insecurity, and rurality*transportation insecurity), to evaluate the impact of rural-urban residence PTI on the relationship between selected resource insecurities and abstinence self-efficacy. Cook’s Distance (Cook’s D) was computed to identify influential cases and to avoid inflated associations. Influential cases, identified by Cook’s D values, were removed from the moderation models (Model_{transportation} = 45, Model_housing = 46, Model_food = 48), resulting in an improved R² value. Consequently, these influential cases were excluded from the final moderation analyses. Additionally, removal of these cases did not affect the significance or direction of these findings. The Breusch-Pagan/Cook-Weisburg test was used to evaluate the heteroskedasticity of residuals, and any heteroskedasticity was addressed in each model by employing robust standard errors. Data analyses were performed using IBM SPSS 28.0 (IBM, Armonk, NY) and STATA/SE 17 (© 2021 StataCorp LLC. 4905 Lakeway Drive, College Station, TX).

Results

Sample characteristics.

Sociodemographic characteristics are shown for the full analytic sample (N = 877) in Table 1. The mean age of women was 37.2. Most women reported being non-Hispanic White (94.0%), with more than a third married or living as married PTI (36.0%) and nearly three-quarters identifying as heterosexual (72.7%). About two-thirds (73.8%) of participants completed high school or a GED and 23.9% of women reported being employed either part-time or full-time PTI. Of the 877 women, 54.5% of women reported ever being prescribed medications for opioid use disorder (MOUD), while only 14.5% reported being prescribed MOUD 90 days PTI. Finally, on average, women were incarcerated for 253 days during their current incarceration (SD = 308.5; range = 2 to 1950).

Table 1.

DESCRIPTIVE PROFILE OF INCARCERATED WOMEN (N = 877)

	M (SD)/%
Demographics
Age (years; range 18–62)	37.2 (8.7)
Race (non-Hispanic White)	94.0%
Heterosexual	72.7%
High school diploma/GED or higher	73.8%
Employed PTI	23.9%
Married/living as married	36.0%
Rural residence PTI	65.9%
Resource Insecurity
Housing	44.2%
Days homeless or had to stay with someone else to avoid being homeless PTI (range 0–90)	31.6 (40.2)
Days living in a homeless shelter/emergency housing PTI (range 0–90)	1.4 (10.3)
Food, overall (range 0–10)	5.0 (4.5)
Very low food insecurity	40.6%
Low food insecurity	1.5%
Marginal food insecurity	6.2%
High food insecurity	51.8%
Transportation (range 0–2)	1.05 (.7)
Transportation, insecure	20.9%
Abstinence self-efficacy (range 1–10)	7.4 (2.6)

Note:

PTI = prior to incarceration.

Additionally, 65.9% of women resided in a rural area PTI and 20.9% noted not having reliable access to a vehicle or public transportation in the year before jail. Nearly two-thirds (57.9%) of women reported food insecurity, with most women reporting either very low food insecurity (40.6%) or high food insecurity (51.8%) 12 months PTI. Finally, 44.2% of women noted experiencing housing insecurity for at least one day in the 90 days PTI, spending an average of 31.6 days (about one month) as homeless or staying with someone to avoid being homeless, and 1.4 days living in a homeless shelter or emergency housing.

Bivariate associations were examined using Pearson correlations and t-tests between variables of interest. Abstinence self-efficacy showed no significant relationship with living in a rural or urban area but was significantly associated with all selected resource insecurity variables. Specifically, lower self-efficacy ratings were associated with increased housing insecurity (r = −.068, p = .044), food insecurity (r = −.123, p < .001), and transportation insecurity (r = −.102, p = .003). Additionally, all selected resource insecurity variables showed significant intercorrelations (Table 2). Rurality was not significantly associated with housing insecurity and food insecurity, but showed a significant relationship with transportation insecurity, where women living in rural settings PTI were more likely to report greater transportation insecurity compared with those in urban settings (t(886) = −9.13, p < .001).

Table 2.

CORRELATIONS BETWEEN SELF-EFFICACY AND RESOURCE INSECURITY (N = 877)

Variable	1	2	3
1. Self-efficacy	—
2. Transportation security	−.102**	—	—
3. Housing insecurity	−.068*	.090**	—
4. Food insecurity	−.123***	.209***	.328***

Note:

^*p <. 05;

^**p <. 01;

^***p<. 001.

Independent correlates of abstinence self-efficacy.

A linear regression model (Table 3) was used to examine resource insecurities as independent correlates of abstinence self-efficacy while controlling for demographic differences. Women who were not married (or living as married) reported significantly lower levels of abstinence self-efficacy (B = −.369, p = .048) than their counterparts. Additionally, lower levels of abstinence self-efficacy were endorsed by women who reported food insecurity (B = −.175, p = .008). However, after controlling for sociodemographic covariates and other forms of resource insecurity, housing insecurity (p = .273) and transportation insecurity (p = .066) were not significantly related to abstinence self-efficacy.

Table 3.

RESOURCE INSECURITY AND RURAL-URBAN RESIDENCE AS INDEPENDENT CORRELATES OF ABSTINENCE SELF-EFFICACY (N = 877)

	B	SE	95% CI	p
Age	.002	.011	−.018, .023	.813
White	.040	.067	−.092, .172	.554
Married or cohabiting*	−.369	.186	−.734, −.004	.048
Heterosexual	.165	.189	−.206, .537	.383
Employed PTI	.152	.209	−.258, .563	.467
Education	−.208	.206	−.612, .196	.313
Rural (vs. urban)	−.169	.189	−.540, .202	.372
Food insecurity**	−.175	.065	−.303, −.046	.008
Transportation insecurity	−.256	.139	−.528, .017	.066
Housing insecurity	−.002	.002	−.006, .002	.348
R2	.030

Note:

^*p <. 05;

^**p <. 01.

Moderated associations between selected resource insecurities and abstinence self-efficacy.

To address the third study aim, rurality (i.e., living in a rural vs. urban area) PTI was examined as a moderator of the associations between selected resource insecurities and abstinence self-efficacy, controlling for age, race, sexual orientation, relationship status, education, and employment. The interaction between living in a rural area and transportation access in relation to self-efficacy was significant (B = −.544; p<.001; see Figure 1). Specifically, among women living in rural areas PTI, there was a negative relationship between abstinence self-efficacy and transportation insecurity, such that women who had secure access to some form of transportation reported higher levels of abstinence self-efficacy. Post-hoc stratified regression analyses demonstrated that the relationship between transportation and abstinence self-efficacy was significant for rural participants, but not for urban participants. No significant interactions were observed between rurality and housing insecurity (B = −.003, F[9, 831] = 4.35, p = .376) or food insecurity (B = −.157, F[9, 819] = 6.28, p = .129).

Figure 1.

Moderation of rural-urban residence on the relationship between abstinence self-efficacy and transportation insecurity (N = 832).

Note:

Overall model: R² = .054

F(9,822) = 4.47

p = .001

Discussion

Research regarding resource insecurity among populations with CLS involvement is limited, but existing studies highlight the negative impacts of resource insecurities, particularly homelessness and food insecurity, on continued CLS involvement and substance use.^18,60 However, research exploring the relationship between abstinence self-efficacy and resource insecurity is sparse, with only one study exploring the relationship between diverse housing situations and abstinence self-efficacy.⁶¹ The present study has addressed these gaps by examining the association between selected resource insecurities and abstinence self-efficacy in a sample of women with CLS involvement, which further contributes to the extant knowledge base by investigating rural-urban living as a moderating factor in this relationship.

The findings from this study indicate that women with CLS involvement experience high rates of food insecurity, housing insecurity, and transportation insecurity. Testa and colleagues⁶² noted that about 17.2% of participants with criminal justice backgrounds experienced moderate-to-severe food insecurity, but higher rates were reported by Dong et al.,² who found that 48.0% of adults on probation experienced food insecurity. Like Dong et al.’s findings, most women in the present study experienced marginal-to-high food insecurity (58.0%). This study also noted that nearly half of women experienced housing insecurity PTI (44.2%), with women on average living over a month as homeless and/or spending at least a few days with unstable housing (e.g., homeless shelter or emergency housing). This finding is consistent with the significant relationship observed among homelessness, incarceration, and recidivism in other research.^63,64 Finally, one in five women in this study reported experiencing transportation insecurity prior to incarceration, but there is a dearth of literature on transportation insecurity among women with CLS involvement. However, one state-level analysis of women in Kentuckyfound that about 53.0% of women in the welfare system reported experiencing transportation problems.⁵ While this is notably higher than the current study’s sample, Ramlow et al.⁵ does not exclusively focus on women with CLS involvement. Future research should examine transportation among women with CLS involvement in more depth.

The present study also examined differences in resource insecurity among women living in rural or urban communities prior to incarceration. At the bivariate level, a notable association emerged concerning transportation insecurity, with higher rates observed among individuals residing in rural settings who may encounter difficulties accessing transportation, especially public transit. However, housing or food insecurity did not differ by women living in rural or urban areas before incarceration. Future research should examine these resource insecurities in more nuanced detail, particularly to explore how rural and urban women may navigate or resolve potential resource insecurities in different ways (e.g., living in abandoned buildings or staying with acquaintances; visiting food banks or eating with friends or family; hitchhiking, walking, cycling).

Findings from the linear regression analysis indicated that food insecurity was significantly and negatively connected with abstinence self-efficacy, even after controlling for demographic covariates and other forms of selected resource insecurities. Such findings suggest that food insecurity may pose special barriers to addiction recovery for women with CLS involvement and OUD. Women who experience food insecurity may experience malnutrition,⁶⁵ health concerns,⁶⁶ and vulnerability to mental health issues and cognitive decline.⁶⁷ Food insecurity may also create time constraints, requiring women to allocate more energy towards acquiring food (e.g., visiting food banks or searching through dumpsters). Moreover, women with substance use disorders who experience food insecurity may be more likely to experience an increase in related stressors, thus possibly affecting both abstinence self-efficacy and subsequently, a possible return to use.

While food insecurity was significantly associated with abstinence self-efficacy within the linear regression, housing insecurity and transportation insecurity were not. This lack of association could be related to a myriad of factors. First, because housing insecurity and transportation insecurity were significant at the bivariate level and not in the regression model, it is conceivable that the relationship between housing insecurity and abstinence self-efficacy is better explained by other factors within the model. Additionally, how housing insecurity was defined may differ from how women self-reported their subjective experience.

Of the covariates included in the linear regression model, only marital status was significant; however, to our knowledge, there is little to no literature examining marital status and abstinence self-efficacy. Some research does suggest a positive significant relationship between abstinence social support and abstinence self-efficacy,⁶⁸ yet do not indicate the type of relationship (e.g., married, friend) of these social supports. Nonetheless, this study suggests that this relationship may be important to consider, as women who were married or reported living as married noted higher levels of abstinence self-efficacy. Such a relationship may aid in providing social support^69,70 and increasing levels of personal fulfillment and life satisfaction.⁶⁹ Alternatively, this association may be related to social comparisons or expectations, such that unmarried women may feel a sense of loss or unfulfillment within societies and cultures where marriage is valued.^71,72 Thus, women who are married or living as married and who perceive themselves as having healthy social and communal supports, personal fulfillment, and life satisfaction may be more likely to experience greater levels of abstinence self-efficacy. Additionally, it is possible that women who are married or living as married may also experience greater economic support from their partners and thus, be less likely to experience resource insecurity.

Ongoing research should examine the relationship between married or cohabiting women and abstinence self-efficacy, as well as relational factors within married or cohabiting relationships that may augment or reduce abstinence self-efficacy. Additionally, research should consider exploring the extent of support that women receive from their partners while in recovery and to examine the potential relationship between partner support and abstinence self-efficacy. Such investigations might enhance understanding of how relational factors within marriages or cohabiting relationships influence abstinence self-efficacy, as well as possible interventions aimed at positively boosting abstinence self-efficacy among married women, and potentially extending to non-married or cohabiting women.

To investigate this study’s final aim, a series of moderated linear regressions were examined to assess living in a rural or urban area PTI as a moderator between selected resource insecurities and abstinence self-efficacy. Results suggest that rurality moderates the relationship between transportation insecurity and abstinence self-efficacy, but not for housing and food insecurities. Some research suggests that transportation is a significant barrier for women living in rural areas, such that they may be unable to meet criminal justice requirements (e.g., probation) or access substance use treatment.⁷³ Thus, it seems that transportation insecurity has potentially greater negative consequences on rural women’s abstinence self-efficacy compared with women living in urban areas. The significance of transportation insecurity in relation to abstinence self-efficacy may well be related to the geographical infrastructure and economic factors of rural areas, in that rural areas often lack well-developed public transportation infrastructure and systems, as well as often have terrains (e.g., mountains, forests) and large geographical distances that may pose challenges for women wanting to achieve abstinence from substance use. Additionally, rural women experiencing transportation insecurity may be less likely than women with transportation to experience higher levels of abstinence self-efficacy due to limited opportunities to attend treatment (e.g., long distances, lack of sidewalks/walkways), which may be less of a concern for women in urban areas, who may have treatment providers and recovery support services available within walking distance.³ Therefore, transportation insecurity can undermine abstinence self-efficacy because it directly affects rural women’s abilities to engage in treatment services and recovery support systems, potentially leading to a decline in women’s confidence in their ability to sustain abstinence. Thus, it is imperative for treatment centers and health care services and providers to provide a means of access for clients (e.g., bus, community transport initiatives, telehealth), as well as to advertise and provide information regarding available transportation services, particularly in rural areas. Non-emergency medical transportation is an underused, mandatory Medicaid benefit that can provide women with access to treatment services (e.g., therapy sessions, prescription pickups, medical care).⁷⁴ Expanding this service in rural areas does not require additional state or federal authority, making it a plausible avenue for reducing barriers to treatment or health care.⁷⁴ Additionally, clinicians and providers should collaborate with clients to identify alternative interventions (e.g., telehealth) aiming to boost abstinence self-efficacy and address challenges related to transportation insecurity.

Limitations.

The results of this study should be considered in light of some limitations. First, data were collected via self-report. While existing literature indicates that self-report data are reliable and valid among CLS-involved and female populations,^75,76 it is important to consider issues of recall bias and inhibited self-disclosure due to the jail setting and sensitive nature and the complexities of certain questions, despite confidential administration of interviews. Additionally, women were not asked about current/any withdrawal symptoms, nor was stability in incarceration addressed, which may be a limitation to this study.

Second, the generalizability of findings from this study may be limited in that women were recruited from jails and had to meet criteria for opioid use disorder; thus, findings may not be generalizable to prison or other populations with CLS involvement (e.g., those on community supervision), or to other substance-using populations. Women in this study were also primarily White and heterosexual. Literature commonly supports the differential marginalization of ethnic, racial, and sexual minorities in resource insecurity;^64,78–80 thus, it is important to consider additional influences on resource insecurity related to personal characteristics and identities, particularly those that may be culturally and socially marginalized. Third, selected resource insecurity measures varied temporally, where transportation and food insecurity were measured 12 months PTI and housing insecurity was measured 90 days PTI.

Fourth, because this study is ongoing, outcomes related to selected resource insecurity and abstinence self-efficacy post-incarceration are not yet available. However, the findings from this study will be informative for future analyses of follow-up data and should be used to inform future research. Furthermore, it is crucial to recognize the spectrum of recovery pathways (e.g., harm reduction, abstinence [with or without MOUD]). Additionally, participants’ abstinence self-efficacy was assessed using a Likert scale, with no subsequent follow-up questions probing response rationale. Future research should aim to fill this gap by 1) understanding the rationale behind high and low abstinence self-efficacy ratings; and 2) studying the relationship between abstinence self-efficacy and preference for treatment modalities and recovery approaches, including harm reduction strategies

Fifth, this study focused on only three areas of resource insecurity (transportation, housing, and food insecurity). It may be important for future studies to examine other areas of insecurity that may affect abstinence self-efficacy, such as access to childcare or medical and insurance insecurity.^80,81 Lastly, while food insecurity was measured using a validated scale from the USDA, housing and transportation insecurity measurements did not originate from a validated scale. Future studies should consider using scales, such as the transportation security index⁸² or the Housing Instability Scale,⁸³ both scales which are relatively new and were not available for use during the design of this study.

Conclusion.

Criminal legal system-involved women and women with OUD face heightened risks of resource insecurity. Women residing in rural areas face especially elevated levels of transportation insecurity, potentially compromising their abstinence self-efficacy and impeding their access to substance use treatment. Furthermore, selected resource insecurities may influence the functioning of women, and this study expands the understanding of their pervasive effects on women, including abstinence self-efficacy. This study highlights the widespread disempowerment experienced by women with CLS involvement who use opioids, which can profoundly affect confidence in personal autonomy, including to achieve and maintain abstinence. Examining these factors during women’s incarceration is crucial to community re-entry, as experiencing limited resources prior to incarceration may contribute to an elevated risk of return to use post-release and/or recidivism. Future research should expand studies on selected resource insecurities among rural and urban women with CLS involvement, with the objective to identify strategies to alleviate related barriers (e.g., transportation).