Examination of referral source and discharge outcomes among women in residential substance use disorder treatment

Abstract

Background:

Court-mandated substance use disorder (SUD) treatment, compared to nonmandated treatment, has been associated with increased retention and completion. However, due to limitations of previous studies, whether child protective services (CPS) and criminal justice (CJ) mandated treatment improve treatment completion and retention among women in residential treatment remains unclear.

Purpose:

This study investigated differences in treatment completion and progress based on three clinical discharge outcomes (i.e., completer, noncompleter with significant progress, and noncompleter without significant progress). We hypothesized that women mandated by (1) CJ will have a better treatment discharge outcome (i.e., treatment completer and noncompleter with satisfactory progress) compared to women who are CPS mandated; (2) CPS will have a better treatment discharge outcome (i.e., treatment completer and noncompleter with satisfactory progress) compared to nonmandated women.

Methods:

Study staff conducted multinomial logistic regression analyses on data for a diverse sample of 161 women mandated or nonmandated (CJ: N = 71, CPS: N = 66, nonmandated: N = 24) into residential SUD treatment to determine each group’s clinically defined treatment discharge outcomes while controlling for covariates.

Results:

Multinomial logistic regression analyses revealed that being mandated by the CJ system predicted being a treatment completer compared to those who were CPS mandated (RR = 9.88, p = .009). The study found no differences in discharge status of completer without satisfactory progress between those who were CPS mandated and those who were CJ mandated or nonmandated. For women mandated by the CPS system compared to nonmandated women, the risk of being a treatment completer relative to noncompleters with satisfactory progress was not significant (RR = 1.08, p = .897). Analyses showed that being mandated by the CJ system predicted an improved clinically defined discharge outcome of treatment completer compared to women who were nonmandated to treatment (RR = 10.74, p = .016). In several of the models, drug and alcohol craving was associated with increased odds of being a noncompleter of treatment without satisfactory progress

Conclusions:

This study demonstrates that improved treatment completion and discharge status cannot be assumed based solely on being mandated by the CJ or CPS systems. As evidenced by variability in treatment discharge outcomes within and among referral groups, the paper suggests directions for future research.

1. Introduction

Mandated residential substance use disorder (SUD) treatment is a common intervention used as an extrinsic motivator to support recovery among women involved in the criminal justice (CJ) and child protective services (CPS) systems who present with complex psychosocial needs. In such cases, mandated treatment is used as a motivator for entering and completing SUD treatment by CJ in lieu of prosecution or as a condition of parole, probation, or community reentry (National Institute of Justice, n.d.); and as a requirement stemming from child protective court orders, investigations, or conditions for retaining or regaining parental rights (Rittner & Dozier, 2000). While in both systems, mandated treatment can serve as an extrinsic motivator to enter and complete treatment (Klag, O’Callaghan, & Creed, 2005). Such motivation may be greater among CJ-mandated women with underage children due to the “double jeopardy” of consequences from both systems (Marlowe, Merikle, Kirby, Festinger, & McLellan, 2001).

Completion of a treatment episode is critical to recovery; however, completing treatment is a major challenge for women (Brorson, Ajo Arnevik, Rand-Hendriksen, & Duckert, 2013). Compared with residential patients who successfully meet their treatment goals and complete a treatment episode, noncompleters are generally challenged by continued substance use relapses (Brorson et al., 2013; Evans, Li, & Hser, 2009); criminal activity (Hser, Evans, Huang, & Anglin, 2004; Simpson, Joe, & Brown, 1997); and other economic, social, and health consequences (National Institute on Drug Abuse, 2004; ONDCP, 2004). Yet prior research on mandated treatment has provided limited information on the effects on women’s residential treatment completion due to: (a) predominantly male samples in studies of CJ-mandated treatment, along with a single focus on criminal recidivism and lack of investigation of any modality of SUD treatment completion (Mitchell, Wilson, Eggers, & MacKenzie, 2012; Wilson, Mitchell, & MacKenzie, 2006); (b) few studies of CPS-mandated treatment and its effects on any modality of treatment completion (Marlowe & Carey, 2012); and (c) no studies on SUD treatment completion (including residential) comparisons among women who CJ or CPS has mandated and those not mandated.

These are critical gaps in research, considering estimates indicating a high proportion of CJ-mandated women in SUD treatment and CPS-involved parents who have substance use problems and are referred to treatment. A national study of women in SUD treatment found that 34% of women were mandated by the CJ system (Longinaker & Terplan, 2014). In a review of data sources on CPS, Young, Boles, and Otero (2007) noted that 60%–80% of substantiated CPS cases involve parents with substance use problems. And a study of 10,909 CPS-involved parents in New Jersey found that among those who received an assessment, 59% were referred to SUD treatment (Traube, He, Zhu, Scalise, & Richardson, 2015).

To address some of these previous limitations, this study investigated differences in treatment completion and progress based on clinical discharge outcomes as determined by the study site clinical team protocol (i.e., completer, noncompleter with significant progress, and noncompleter without significant progress).¹ We evaluated these outcomes by three major referral sources: CJ mandated, CPS mandated, and nonmandated in a sample of women entering residential treatment.

1.1. Importance of specialized residential treatment for highly vulnerable women

With a significant lack of adequate access to specialized SUD treatment for women (Terplan, Longinaker, & Appel, 2015), research should highlight why such services, especially residential treatment, are critical for women mandated by the CJ and CPS systems. Residential treatment plays an important role in the continuum of care. In 2016, nearly 1.5 million individuals (65% of those receiving SUD treatment) received this SUD treatment modality (ASPE, 2017). Intensive treatment such as provided in the residential modality is highly relevant for women involved with the CJ and CPS systems, because women are the fastest growing incarcerated population (Swavola, Riley, & Subramanian, 2016), they compose the majority of those involved in the CPS system (Marlowe & Carey, 2012; Young et al., 2007), and many of them have a history of SUD and trauma history (Marlowe & Carey, 2012; Swavola et al., 2016; Young et al., 2007).

For CJ- and CPS-involved women with SUD, mandated residential treatment may be a particularly important intervention, given this group’s vulnerability to dropout, in part as a result of comorbidities and a heightened level of SUD severity, social burdens, and health problems—a profile aligned with criteria for residential treatment set forth by the American Society for Addiction Medicine (Mee-Lee, Shulman, & Fishman, 2013). For instance, compared with men, women report significantly higher rates of physical and sexual abuse trauma (Du, Huang, Zhao, & Hser, 2013; Dunkerley, 2017), more severe SUDs (Grella, Hser, & Huang, 2006), and more complex medical comorbidities.

The most vulnerable women with SUD are served in sex-specific residential programs. For example, compared to women in mixed-sex programs, those in women-only residential programs are more likely to be homeless, on probation, and have more extensive histories of substance use (Dunkerley, 2017; Grella et al., 2006). Such programs address the unique needs of women (e.g., childcare and child services, trauma-informed care) and thereby, can facilitate treatment access and retention (Greenfield et al., 2007; Grella & Joshi, 1999; Grella, Polinsky, Hser, & Perry, 1999). In addition, women in SUD treatment with involvement in the CJ and CPS systems typically have more severe SUDs than their male counterparts (Greenfield et al., 2007; Grella et al., 2006).

1.2. Previous research on treatment completion among women mandated by CJ and CPS

Findings from the most recent meta-analytic review of outcomes from 154 independent drug court evaluations (Mitchell et al., 2012) indicate lower recidivism rates (i.e., overall and drug-related criminal behavior and drug use) among adult drug court participants compared to nonparticipants. However, in addition to caveats related to methodological limitations and highly variable findings, results are not generalizable to women because the vast majority (84%) of studies in Mitchell el al.’s meta-analysis had samples mostly comprising men.

Fewer studies have systematically investigated outcomes of mandated treatment in women. For example, reporting on characteristics and treatment outcomes among women in the Alcohol and Drug Services Study, encompassing a representative sample of 2,395 SUD treatment facilities, Brady and Ashley (2005) reported that the odds of SUD treatment completion were higher in women (primarily White sample) referred by the CJ system. Another study based on data from 461 participants in the Women, Co-Occurring Disorders, and Violence Study found that women who were court mandated stayed in residential treatment longer and had a 35% lower risk of dropout compared to those not court mandated (Amaro, Chernoff, Brown, Arévalo, & Gatz, 2007). A larger and more recent study by Longinaker and Terplan (2014) used the Treatment Episode Dataset Discharge (N = 582,671 treatment episodes) to explore the association between CJ-mandated treatment and treatment completion in a sample of predominantly White women (63% of nonmandated and 67% of mandated). They found those mandated to residential treatment had slightly higher odds (OR = 1.06, 95% CI = 1.02, 1.09) of completion or transferring to another treatment than those who entered voluntarily (Longinaker & Terplan, 2014). The only study on drug courts focused on assessing outcomes of mandated treatment among women was an evaluation of the Brooklynn Treatment Court (1996–1999). It found significantly lower substance use and criminal involvement at follow-up among female participants (n = 110) compared to otherwise-eligible women who did not qualify due to living outside zones targeted for service or not referred to the court (n = 26; Harrell, Roman, & Sack, 2001).

Fewer studies have examined the effects of CPS-mandated treatment on SUD program completion (Marlowe & Carey, 2012). This is a critical gap in the literature given the continuing expansion of family treatment courts (CPS-mandating agency) in the United States from two in 1994 to 495 in 2018 (Lemus & Ritcher, 2018). Marlowe and Carey (2012) conducted an in-depth review of the literature and concluded that emerging research suggests that family dependence court participants complete SUD treatment at rates 20 to 30 percentage points higher than that of other parents not mandated by family dependence court.

The lack of studies examining SUD treatment completion among women mandated by CJ and CPS compared to those not mandated, along with the noted limitations of previous studies on CJ- and CPS -mandated treatment, underscores the need for a greater understanding of how the treatment referral sources impact discharge outcomes among women in SUD treatment

1.3. Measures of treatment completion and progress

Research widely uses time in SUD treatment, measured as days or months, as a temporal measure of treatment retention (time from intake to date of discharge) for a predefined course of treatment (Choi, Adams, Morse, & MacMaster, 2015; Marcus et al., 2009). Although associated with measures of treatment outcomes such as lower rates of relapse (Arfken, Klein, di Menza, & Schuster, 2001; Choi et al., 2015), measures of time in treatment overlook individual variations in progress made during the treatment course. For example, the early stages of residential SUD treatment are vital, because prior work reveals a high percentage of treatment dropout occurs in the first 30 days (Loveland & Driscoll, 2014). Further, variability in treatment progress may occur for those who remain in treatment longer without achieving the same progress as those who remain in treatment for fewer days (Arnaudova, Jin, & Amaro, 2020). This variability may be in part due to alcohol and substance use severity, motivation to engage in treatment, or the differential treatment goals of patients and their mandating referral source (CJ, CPS, or nonmandated). Therefore, treatment progress is as equally important to examine as retention (number of days in treatment) and treatment completion (Arnaudova et al., 2020). Likewise, studies of CJ- and CPS-mandated treatment that have documented SUD treatment episodes as completed or not completed have failed to capture progress that those who did not complete treatment made. As is found in the current study, clinically defined discharge outcomes measured by three levels of progress (treatment completer, noncompleter with satisfactory progress, and noncompleter without satisfactory progress) may offer a direct approach for assessing progress made during the treatment course (Black & Amaro, 2019; Choi & Ryan, 2006).

1.4. Client characteristics associated with discharge status

Prior SUD treatment research has demonstrated that clients’ race/ethnicity and psychological characteristics are associated with noncompletion of SUD treatment. Although national completion rates in residential treatment are generally higher to the rates in outpatient treatment (Substance Abuse and Mental Health Services Administration, 2013), Hispanic and Black populations, in contrast to non-Hispanic White populations, are less likely to complete treatment, a disparity attributed to socioeconomic inequities (Saloner & Lê Cook, 2013). Regional studies have showed disparities exacerbated by higher rates of homelessness and mental health issues (Guerrero, Marsh, Duan, Oh, Perron, & Lee, 2013). In a large national study involving 318,924 cases, residential SUD programs reported a 65% completion rate compared to 52% for outpatient settings, significantly moderated by race or ethnicity (Stahler, Mennis, & DuCette, 2016).

Psychological characteristics associated with higher noncompletion of SUD treatment include post-traumatic stress disorder (PTSD) symptomology (Tull, Gratz, Coffey, Weiss, & McDermott, 2013), stress (Andersson, Steinsbekk, Walderhaug, Otterholt, & Nordfjærn, 2018; Daughters, Richards, Gorka, & Sinha, 2009), low emotion regulation (Eddie et al., 2013), low distress tolerance (Tull et al., 2013), depression symptomology (Antony, Bieling, Cox, Enns, & Swinson, 1998), and drug and alcohol craving (Hopper et al., 2006; Law et al., 2016). Social network composition and support characteristics, such as the level and quality of social support and motivation to engage in treatment (Dobkin, Civita, Paraherakis, & Gill, 2002), have also been associated with treatment retention (Arnaudova, Jin, & Amaro, 2020). Emerging evidence also indicates that support from individuals in recovery, whether as formal peer coaches or not, plays a role in treatment retention and completion and other treatment outcomes such as relapse (Eddie, Hoffman, Vilsaint, Abry, & Bergman, Hoeppner, Weinstein, & Kelly, 2019).

1.5. Study aim and hypotheses

Guided by this prior empirical work, the primary aim of this study was to investigate differences in treatment completion and progress based on clinical discharge outcomes as determined by the study site clinical team protocol (i.e., completer, noncompleter with significant progress, and noncompleter without significant progress). We evaluated these outcomes by three major referral sources: CJ mandated, CPS mandated, and nonmandated in a sample of women entering residential treatment. ² Hypothesis 1: Women mandated by the CJ system will have a better treatment discharge outcome (i.e., treatment completer and noncompleter with satisfactory progress) compared with women who are CPS mandated. Hypothesis 2: Women mandated by the CPS system will have a better treatment discharge outcome (i.e., treatment completer and noncompleter with satisfactory progress) compared with nonmandated women. Hypothesis 3: Women mandated by the CJ system will have a better treatment discharge outcome (i.e., treatment completer and noncompleter with satisfactory progress) compared with women who are nonmandated.

2. Methods

2.1. Study design

This research study involved a secondary analysis of data from a randomized controlled trial. Because this study did not focus on intervention effects, we controlled for group assignment in the analysis. The parent study was a parallel-group trial (NCT02977988) conducted from 2016 to 2018 as a two-armed design comparing retention days in SUD residential treatment between women randomly assigned to one of two study conditions as adjuncts to their treatment as usual: (a) Moment-to-Moment in Women’s Recovery or (b) Neurobiology of Addiction, with the latter serving as a psychoeducational control condition (Amaro & Black, 2017). The study conducted baseline interviews prior to group randomization. All study participants received treatment services as normally delivered in an SUD residential therapeutic community setting without affecting the standard level of care typically provided to patients. The University of Southern California’s Institutional Review Board approved the parent study (Amaro & Black, 2017).

2.2. Study site

The study site is a publicly funded residential treatment facility for women diagnosed with SUDs in Southern California. It had the capacity to provide on-site housing and comprehensive services for up to 110 women and their children. Although women could remain in residential treatment for up to 12 months, the average stay was 5.5 months. The program includes biopsychosocial assessment, mental health and SUD diagnosis and treatment (including medication-assisted treatment), chemical dependence education and counseling, individual and group therapy, relapse prevention, random drug testing, specialized women’s groups, trauma education and support, family education and counseling, vocational training, educational support, case management, nutritional education and support, health and wellness activities, 12-step meetings, childcare, and children’s services. Referrals to the following services were made as needed: medical and dental, domestic violence, psychiatric care, GED classes, and Early Head Start. The residential treatment program is situated on a large 4-acre campus that includes residential quarters, space for group therapy and classes, a computer lab, a nursery, and a classroom for the pre-school program that the local school district provides.

2.3. Participants and procedures

All participants (N = 161) were adult women diagnosed with SUD, with children under age 18. Participants who the CJ system mandated entered treatment after incarceration or through a reentry court as a condition of probation or parole. Participants who the CPS system mandated entered treatment through a family treatment court as a means of retaining child custody. Upon admission to the treatment site, patients meet one-on-one with the site’s intake clinician coordinator (master’s-level licensed clinician) who conducts an assessment for substance use disorders, mental health disorders, and suicidality using the DSM-5 (American Psychiatric Association, 2013). Then, the clinician completes a psychosocial assessment using an in-house form to identify important aspects of patient history and patient needs to inform case management and treatment plans. The site psychiatrist and on-site clinician coordinator make the final diagnostic decision after discussing each patient’s diagnostic assessment. It is then that the final diagnoses are determined and recorded in the patient’s clinic record.

Approximately 3 weeks after residential treatment entry, the names of clients who assented to be contacted by the study team were provided to the research interviewer. The research interviewer made appointments with prospective participants, conducted the informed consent and HIPAA processes, and administered the baseline assessment interview. Trained study interviewers using computer-assisted interview protocols collected data from the participants (Amaro & Black, 2017). Research staff entered study data into Research Electronic Data Capture, an electronic data-capturing tool (Harris et al., 2009). In line with data-quality protocols, this study conducted double data entry for all interviews, and we selected 10% of data for manual confirmation of accuracy of data entry.

2.4. Inclusion and exclusion criteria

Inclusion criteria for the parent study were: client at the residential treatment study site, female, aged 18–65 years, diagnosed with an SUD, fluent in English, and agreed to participate in the study. Exclusion criteria were: inability to comprehend or sign the informed consent due to language reasons, cognitive impairment, untreated psychotic disorder, severe chronic mental health condition, past-30-day suicidality based on clinical assessment, older than 65 years, current prisoner, more than 6 months pregnant, enrolled in another study, and not willing to sign a HIPAA form or be audio recorded during interviews and parent study intervention sessions.³

2.5. Data sources and measures

2.5.1. Demographic, clinical, and referral data

The study research interviewer obtained baseline demographic information via in-person interviews; this demographic information included age, race and ethnicity, parental status, education, employment status, marital status, and housing status. Information abstracted from patient clinical records included DSM-5 (American Psychiatric Association, 2013) diagnoses for SUDs and other mental health disorders conducted by the site intake clinician and psychiatrist, and referral source (CJ mandated, CPS mandated, or nonmandated). The clinical team recorded discharge status in the patient clinical record as described in the Measures section 2.5.3. On a monthly basis, the on-site clinician study coordinator extracted this information from each patient clinical record and recorded it on a study form via a monthly batch from the beginning to the end of the study.

2.5.2. Self-report measures

Research interviewers assessed measures at baseline via in-person interviews with validated psychometric scales of drug and alcohol craving (Penn Alcohol Rating Scale; Black & Amaro, 2019; Flannery, Volpicelli, & Pettinati, 1999), perceived stress (Perceived Stress Scale; Cohen, Kamarck, & Mermelstein, 1983), PTSD symptomology (PTSD Symptom Scale; Foa, Cashman, Jaycox, & Perry, 1997), emotion regulation (Difficulties in Emotion Regulation Scale; Gratz & Roemer, 2004), distress tolerance (Distress Tolerance Scale; Simons & Gaher, 2005), depression symptomology (Depression, Anxiety and Stress Scale; Antony et al., 1998), and social support (Important People Drug and Alcohol Interview; Zywiak et al., 2009). Study staff also conductd additional assessments of items measuring drug and alcohol use severity, pending charges or sentencing, legal severity, importance of counseling or referral for legal problems (Addiction Severity Index-Lite [ASI-L]; Cacciola, Alterman, McLellan, Lin, & Lynch, 2007) and incarceration 8 months prior to treatment entry (Life Stressors Checklist Revised [LSC-R]; McHugo et al., 2005). This study quantified substance use in the 8 months prior to treatment entry using the Timeline Follow-Back measure (Robinson, Sobell, Sobell, & Leo, 2014) by calculating total days of use for alcohol to intoxication, specific drugs, and polysubstance use.

2.5.3. Discharge status outcomes

The study extracted data on clinically defined treatment discharge status outcomes from patient clinical records. The study based this determination on SUD treatment site process and protocol guidelines, and the clinical treatment team (i.e., certified SUD counselor, master’s-level clinician therapist, and team supervisor master’s-level therapist; the latter two registered with the Board of Behavioral Sciences) decided on a case-by-case basis for each participant if the client demonstrated the required behaviors and skills to complete residential treatment. The three treatment discharge status designations were: treatment completer, noncompleter with satisfactory progress, and noncompleter without satisfactory progress based on each patient’s treatment progress at discharge. A completer is a patient who completed the full range of treatment services outlined in the treatment plan, met all treatment objectives, and achieved satisfactory recovery stability (consistent negative drug screens; attended treatment groups outlined in treatment plan; improved coping skills; and adhered to both facility home visit rules and CPS, parole, or probation conditions). A noncompleter with satisfactory progress is a patient who left treatment before completing the treatment plan or achieving all treatment objectives, although clinicians determined them to have made satisfactory progress toward treatment goals while improving recovery stability (e.g., leaving treatment due to childcare responsibilities, regaining child custody from family members, or financial reasons due to being head of household). A noncompleter without satisfactory progress is a patient who left before completing the treatment plan and was clinically determined to have made little or no progress toward achieving treatment goals (e.g., substance use relapses, possession of drugs or paraphernalia in treatment facility, or left treatment having received minimal services; Black & Amaro, 2019).

2.6. Statistical analyses

For examination of sociodemographic characteristics between referral groups, we conducted comparative analyses using t-tests and analysis of variance for comparison of means and chi-square tests for comparison of percentages with categorical variables. To address hypotheses 1, 2, and 3, we employed a sequential multinomial regression model-building approach to determine if mandated (CJ or CPS) or nonmandated referral status predicted improved clinically defined discharge outcomes. Referral status was the study’s primary independent variable of interest, and the outcome variables were the three levels of clinically defined discharge. Study staff conducted univariate analyses to determine potential associations between being CJ mandated, CPS mandated, or nonmandated and the three levels of clinically defined discharge outcomes.

The study examined potential covariates and control variables to determine whether they contributed to a greater than 10% change in the effect estimate of the mandating variable. Based on prior research, the study evaluated potential covariates and control variables for possible inclusion in the final models: (a) sociodemographic characteristics, number of mental health diagnoses, SUD and AUD diagnoses, and intervention condition (from the parent study); (b) ASI-L items measuring drug and alcohol use severity, legal and counseling severity, and pending charges or sentencing and an LSC-R item measuring prior incarceration; and (c) hypothesized psychological and social support variables.

The control variables that study staff determined should be included in final models were race or ethnicity, alcohol severity, drug and alcohol craving, and incarceration during 8 months prior to treatment entry. For final multinomial logistic regression models, we present relative risk ratios (RR) and 95% confidence intervals (CI). Study staff conducted all statistical analyses using Stata version 15.1.

3. Results

3.1. Participant characteristics by mandated status

As Table 1 shows (total column), the sample comprises 161 women (all mothers), of whom 44.1% were mandated to SUD treatment by the CJ system, 41.0% were mandated by CPS, and 14.9% were nonmandated; they had a mean age of 31.17 years. Most participants were Hispanic (62.11%), followed by non-Hispanic White (19.25%) and non-Hispanic Black or African American (17.39%). Participants were characterized by significant vulnerability, as indicated by their demographic, psychosocial, and clinical characteristics. Average days of substance use in the 8 months prior to treatment entry was highest for methamphetamine, with the majority (94%) reporting use of more than one drug on one or more days. A total of 55.28% completed treatment, 14.29% were noncompleters with satisfactory progress, and 30.43% were noncompleters without satisfactory progress.

Table 1.

Participant sociodemographic and pretreatment characteristics (N = 161).

Characteristic	CJ (n = 71) n (%)	CPS (n = 66) n (%)	Nonmandated (n = 24) n (%)	Total (N = 161) n (%)	p
Clinically defined discharge status					.027
Treatment completer	45 (63.38)	33 (50.00)	11 (45.83)	89 (55.28)
Noncompleter with satisfactory progress	3 (4.29)	15 (22.73)	5 (20.86)	23 (14.29)
Noncompleter without satisfactory progress	23 (33.39)	18 (27.27)	8 (33.33)	49 (30.43)
Demographics
Age, M (SD)	31.81 (7.47)	30.34 (6.49)	31.59 (5.70)	31.17 (6.83)	.431
Race and ethnicity					.580
Non-Hispanic White	10 (14.08)	14 (21.21)	7 (29.17)	31 (19.25)
Non-Hispanic Black	16 (22.54)	9 (13.24)	3 (12.50)	28 (17.39)
Hispanic	44 (61.97)	42 (63.64)	14 (58.33)	100 (62.11)
Other	1 (1.41)	1 (1.52)	0 (0.00)	2 (1.24)
Education					.406
Less than high school diploma	32(45.07)	33 (50.00)	15 (62.50)	80 (49.69)
Completed high school	20 (28.17)	22 (33.33)	5 (20.83)	47 (29.19)
Some education after high school	19 (26.76)	11 (16.67)	4 (16.67)	34 (21.12)
Employment					.191
Full-time	7 (9.86)	11 (16.67)	6 (25.00)	24 (14.91)
Part-time	8 (11.27)	7 (10.61)	5 (20.83)	20 (12.42)
Not working	56 (78.87)	48 (72.73)	13 (54.17)	117 (72.67)
Marital status					.441
Married or common law	3 (4.23)	7 (10.61)	1 (4.17)	11 (6.83)
Separated or divorced	11 (15.49)	10 (15.15)	6 (25.00)	27 (16.77)
Never married	57 (80.28)	49 (74.24)	17 (70.83)	123 (76.40)
Housing					.576
Stable	10 (14.09)	16 (24.24)	5 (20.83)	31 (19.25)
Unstable	61 (85.91)	50 (76.76)	19 (79.17)	130 (80.75)
Currently pregnant					.541
Yes	4 (5.6)	5 (7.6)	3 (12.5)	12 (7.50)
No	67 (94.4)	61 (92.4)	21 (87.5)	149 (92.5)
Number of children < 18, M (SD)	2.41 (1.39)	2.91 (1.62)	2.17 (1.40)	2.58 (1.51)	.052
Ever lost custody of 1 or more children					.029
Yes	36 (50.7)	25 (37.9)	5 (20.8)	66 (41.01)
No	35 (49.3)	41 (62.1)	19 (79.2)	95 (59.0)
Any children in own custody at treatment entry					.034
Yes	26 (36.7)	27 (40.9)	16 (66.7)	69 (32.9)
No	45 (63.4)	27 (40.9)	8 (33.3)	92 (57.1)
Any children living with someone else due to child protective court order					.03
Yes	36 (50.7)	42 (63.60)	6 (25.00)	84 (52.22)
No	35 (49.30)	24 (36.36)	18 (75.00)	77 (47.83)
Incarcerated anytime in 8 months prior to treatment entry					.000
Yes	66 (92.96)	15 (22.73)	6 (25.00)	87 (54.04)
No	5 (7.04)	51 (77.27)	18 (75.00)	74 (45.96)
Pending charges, trial, or sentencing					.000
Yes	38 (54.29)	8 (12.12)	4 (16.67)	50 (31.25)
No	32 (45.71)	58 (87.88)	20 (83.33)	110 (68.75)
Substance use, SUD diagnoses, and severity
Days used substances 8 months prior to treatment entry, M (SD)a
Alcohol to intoxication	16.37 (35.24)	22.36 (48.71)	33.71 (53.83)	21.41 (44.25)	.247
Cocaine or crack	14.31 (44.19)	1.85 (10.36)	10.79 (37.05)	8.67 (33.60)	.090
Methamphetamine	74.62 (71.37)	99.79 (90.27)	92.33 (80. 95)	87.69 (81.27)	.184
Any drug (among drug users)	97.35 (68.89)	122.47 (90.30)	123.0 (95.34)	111.47 (82.77)	.158
Polydrug (among users of >1 drugs per day)	35.32 (61.15)	44.94 (75.13)	49.33 (84.62)	41.35 (70.64)	.611
Substance use disorder diagnosis					.321
Alcohol use disorder only	5 (7.04)	4 (6.25)	4 (16.67)	13 (8.18)
Drug use disorder only	56 (78.87)	49 (76.87)	19 (79.17)	124 (77.99)
Both	10 (14.08)	11 (17.19)	1 (4.17)	22 (13.84)
Substance use severity (ASI), M (SD)
Drug use severity	.094 (.091)	.201 (.159)	.216 (.144)	.156 (.141)	< .001
Alcohol use severity	.040 (.090)	.177 (.240)	.220 (.253)	.123 (.204)	< .001
Mental health disorders and psychological characteristics
Mental health diagnoses (besides SUDs)					.079
None	29 (40.85)	18 (28.13)	6 (25.00)	53 (33.33)
One	27 (38.03)	38 (59.38)	11 (45.83)	76 (47.80)
Two or more	15 (21.13)	8 (12.50)	7 (29.17)	30 (18.87)
Mental health diagnosisb
Posttraumatic stress disorder	24 (33.8)	21 (31.8)	8 (33.3)	53 (32.4)	.974
Depressive disorder	15 (21.1)	11 (16.7)	6 (25.0)	32 (19.0)	.596
Other mental health disorder	31 (43.7)	31 (47.0)	15 (62.5)	77 (48.8)	.277
Psychological characteristics scales, M (SD)
Distress tolerance scale	3.09 (0.83)	2.68 (0.841)	2.72 (0.94)	2.90(0.86)	.053
Posttraumatic stress disorder symptomology	14.54 (12.35)	18.65 (11.58)	20.96 (15.79)	17.18 (12.77)	.048
Drug and alcohol craving	1.97 (1.58)	2.43 (1.69)	3.00 (1.96)	2.31 (1.71)	.029
Perceived stress	20.23 (7.75)	22.56 (6.91)	22.88 (7.04)	21.58 (7.37)	.116
Emotion regulation	80.09 (27.45)	90.64 (27.73)	103.25 (33.64)	87.86 (29.40)	.002
Depression symptomatology	4.06 (4.27)	5.36 (5.07)	7.58 (6.11)	5.12 (5.02)	.010
Social support, M (SD)
Recovering alcoholics and abstainers in network	.592 (.313)	.565 (.320)	.628 (.323)	.586 (.320)	.698
Recovering drug users and abstainers in network	.692 (.325)	.676 (.328)	.788 (.276)	.700 (.3120)	.331
Self-perceived seriousness of legal problemsc					< .001
Not serious	37 (52.11)	55 (83.33)	20 (83.33)	112 (69.57)
Serious	34 (47.89)	11 (16.67)	4 (16.67)	49 (30.43)
Self-perceived importance of counseling or referral for legal problems d					< .001
Not serious	31 (43.66)	53 (80.30)	19 (79.17)	103 (63.98)
Serious	40 (56.34)	13 (19.70)	5 (20.83)	58 (36.02)

Note. Data are shown as n (%) unless otherwise stated as M (SD) in the variable column. CJ = criminal justice; CPS = child protective services.

^aSubstances used on average 10 days or more during previous 8 months.

^bOther MH diagnosis are grouped because individually they were each reported by < 10% of the sample.

^cItem from ASI asked “How serious do you feel your present legal problems are? Responses recoded as not serious (not at all an slightly) or serious (moderately, consistently, and extremely).

^dItem from ASI asked “How important to you now is counseling or referral for these legal problems?” Responses recoded as not serious (not at all and slightly) or serious (moderately, consistently, and extremely).

As Table 1 shows, bivariate analysis revealed significant differences between referral groups for clinically defined discharge outcome (p = .027), wherein those who were CJ mandated were more likely to complete treatment. Differences among referral groups were observed for drug (p < .001) and alcohol use severity (p < .001), PTSD symptomology (p =.048), drug and alcohol craving (p = .029), emotion regulation (p = .002), and depression symptomology (p < .01); all indicated greater severity in the nonmandated group. Bivariate analysis further revealed differences among referral groups in (a) having been incarcerated one or more days during the 8 months prior to treatment entry (p < .001); (b) having pending charges, trial, or sentencing (p < .001); (c) self-perceived legal severity (p < .001); (d) self-perceived importance of counseling or referral for legal problems (p < .001); and (e) having ever lost custody of one or more children (p < .029); all indicated worse severity in the CJ-mandated group. The study found no differences among mandated groups on days of use of various types of substances, SUD diagnoses, and number of or type of mental health diagnoses.

3.2. Findings for main hypotheses

Tables 2–4 present clinically defined discharge outcomes relative to mandated status. These separate regressions include outcome variables of clinically defined discharge outcomes (treatment completer or noncompleter without satisfactory progress), with the reference group being noncompleters with satisfactory progress. To address hypothesis 1, in Table 2 we present findings on discharge status comparing CJ-mandated and nonmandated women to those who were CPS mandated as the reference group. Multinomial logistic regression analyses revealed that being mandated by the CJ system predicted being a treatment completer compared to those who were CPS mandated (RR = 9.88, p = .009), while controlling for covariates. The study found no differences in discharge status of completer without satisfactory progress between those who were CPS mandated and those who were CJ mandated or nonmandated.

Table 2.

Multinomial regression predicting discharge status by CJ mandated and nonmandated compared to CPS mandated.^a

Clinically Defined Discharge Status	Treatment Completer RR (CI)	Noncompleter without Satisfactory Progress RR (CI)
Mandated statusb
Nonmandated	0.92 (0.26, 3.27)	1.18 (0.30, 4.74)
CJ mandated	9.88 (1.75, 55.70)**	5.15 (0.84, 31.57)
Race or ethnicityc
Non-Hispanic Black	0.44 (0.097, 1.95)	0.25 (0.04, 1.53)
Hispanic	1.18 (0.33, 4.26)	1.77 (0.44, 7.01)*
Alcohol severity	1.24 (0.10, 15.26)	0.48 (0.03, 8.20)
Drug and alcohol craving	1.17 (0.84, 1.64)	1.44 (1.01, 2.04)*
Incarceration prior to treatment entry	0.91 (0.24, 3.39)	2.30 (0.57, 9.28)

Note. RR = relative risk ratio; CI = confidence intervals. Control variables: race or ethnicity, alcohol severity, drug and alcohol craving, and incarceration prior to treatment.

^aReference group for model = noncompleter with satisfactory progress.

^bReference group is CPS mandated.

^cReference group is White.

^*p ≤ .05.

^**p ≤ .01.

Table 4.

Multinomial regression predicting discharge status by CPS mandated and nonmandated compared to CJ mandated.^a

Clinically Defined Discharge Status	Treatment Completer RR (CI)	Noncompleter without Satisfactory Progress RR (CI)
Mandated statusb
Nonmandated	0.09 (0.01, 0.65)*	0.23 (0.03, 1.77)
CPS mandated	0.10 (0.02, 0.57)**	0.19 (0.03, 1.19)
Race or ethnicityc
Non-Hispanic Black	0.44 (0.10, 1.95)	0.25 (0.04, 1.53)
Hispanic	1.78 (0.33, 4.26)	1.77 (0.44, 7.06)
Alcohol severity	1.24 (0.11, 15.26)	0.48 (0.03, 8.20)
Drug and alcohol craving	1.17 (0.84, 1.64)	1.44 (1.01, 2.05)*
Incarceration prior to treatment entry	0.91 (0.25, 3.39)	2.30 (0.57, 9.28)

Note. RR = relative risk ratio; CI = confidence intervals. Control variables: race or ethnicity, alcohol severity, drug and alcohol craving, and incarceration prior to treatment.

^aReference group for model = noncompleter with satisfactory progress.

^bReference group is CJ mandated.

^cReference group is White.

^*p ≤ .05.

^**p ≤ .01.

To ddress hypothesis 2, in Table 3 we present findings on discharge status by CJ mandated and CPS mandated, with nonmandated as the reference group. Analysis revealed that being mandated by CPS did not predict an improved clinically defined discharge outcome compared with women who were nonmandated. For women mandated by the CPS system compared with nonmandated women, the relative risk of being a treatment completer relative to noncompleters with satisfactory progress was not significant (RR = 1.08, p = .897). Table 3 also shows findings for hypothesis 3, which the data supported. Analysis showed that being mandated by the CJ system predicted an improved clinically defined discharge outcome of treatment completer compared with women who were nonmandated to treatment (RR = 10.74, p = .016), while controlling for covariates. The study found no other differences in discharge status by mandated status. Due to concerns regarding large confidence intervals in some estimates presented in Tables 2 and 3, research staff conducted analyses using CJ-mandated treatment as the reference group and reran the multinomial logistic regression model. Table 4 presents findings on predictors of discharge status by CPS and nonmandated compared to CJ mandated. For both nonmandated (RR = 0.09, p = .016) and CPS-mandated (RR = 0.10, p = .009), compared to CJ-mandated women, the odds of being a treatment completer relative to being a noncompleter were lower, holding all other variables in the model constant. The study found no other differences in discharge status by mandated status. Consistent with results in Tables 2 and 3, results in Table 4 indicate that the covariate, higher drug and alcohol craving, was associated with increased risk of being a noncompleter without satisfactory progress (RR = 1.44, p = .047). In contrast to Tables 2 and 3, results in Table 4 show significantly narrower CIs, providing a more precise effect estimate and enabling more confidence in the true relative risk. Difference in CIs across models reflect the different sample size of the reference group for each model.

Table 3.

Multinomial regression predicting discharge status by CJ mandated and CPS mandated compared to nonmandated.^a

Clinically Defined Discharge	Treatment Completer RR (CI)	Noncompleter without Satisfactory Progress RR (CI)
Mandated statusb
CJ mandated	10.74 (1.55, 74.54)*	4.35 (0.57, 33.39)
CPS mandated	1.08 (0.31, 3.87)	0.84 (0.21, 3.38)
Race or ethnicityc
Non-Hispanic Black	0.44 (0.10, 1.95)	0.25 (.04, 1.53)
Hispanic	1.18 (0.33, 4.26)	1.77 (0.44, 7.06)
Alcohol severity	1.24 (0.10, 15.26)	0.48 (0.03, 8.20)
Drug and alcohol craving	1.74 (084, 1.64)	1.44 (1.01, 2.05)*
Incarceration prior to treatment entry	0.91 (0.25, 3.39)	2.30 (0.57, 9.28)

Note. RR = relative risk ratio; CI = confidence intervals. Control variables: race or ethnicity, alcohol severity, drug and alcohol craving, and incarceration prior to treatment.

^aReference group for model = noncompleter with satisfactory progress.

^bReference group is nonmandated.

^cReference group is White.

^*p ≤ .05.

^**p ≤ .01.

4. Discussion

4.1. Mandated treatment as a predictor of discharge status

This study compared the association between mandated referral groups for women’s residential SUD treatment with three clinically defined discharge outcomes. Findings showed that women who were CJ mandated to residential SUD treatment showed improved odds of completing treatment compared with those who were CPS mandated or nonmandated. Contrary to our hypothesis, we did not find that women who were CPS mandated into residential SUD treatment had an improved clinically defined treatment discharge outcome compared with nonmandated women.

Findings provide important insights into the role of three referral conditions with different impacts on treatment discharge outcomes (treatment completer, noncompleter with satisfactory progress, and noncompleter without satisfactory progress). These referral conditions may involve different extrinsic motivational factors that may contribute to treatment discharge outcomes. Consistent with empirically based studies (Carey, Pukstas, Waller, Mackin, & Finigan, 2008; Coviello et al., 2013; Longinaker & Terplan, 2014), the current study shows that women who were CJ mandated to SUD treatment had improved odds of completing treatment and expanded this finding to women in the residential modality of treatment.

This calls into question whether CPS-mandated residential SUD treatment, as an extrinsic motivator, induces treatment motivation and leads to improved retention and treatment discharge outcomes (Hachtel, Vogel, & Huber, 2019; Longinaker & Terplan, 2014; Wild, Yuan, Rush, & Urbanoski, 2016). Explaining the strength of extrinsic motivation stemming from being CJ or CPS mandated is difficult, given the distinct and differentiated social contexts of the respective referral sources. This study may exhibit the mixed effectiveness of using mandated referrals, which reflects the unmeasured variability in patient motivation. For example, women mandated by the CJ system may also have CPS cases running concurrently (Grella et al., 2006; Young et al., 2007), potentially increasing extrinsic motivation to engage in and complete residential SUD treatment. Alternatively, women who have had numerous SUD treatment and incarceration experiences, have already lost custody of their children, or have severe SUD conditions may not have the same extrinsic motivation as women who retain partial or full child custodial rights and do not have an extensive criminal record, for whom dismissal of current charges upon treatment completion would improve their life circumstances. Although mandated treatment, as an extrinsic motivator, is an accepted predictor of improved treatment retention and completion (Carey et al., 2008; Coviello et al., 2013; Longinaker & Terplan, 2014), the results of this study show that this may not apply when CJ- and CPS-mandated treatment is considered. As the variability in clinically defined discharge outcomes between referral groups demonstrated, future research should examine optimal and suboptimal motivational profiles associated with clinically defined discharge outcomes. More generally, future research should investigate patient motivation levels prior to and during treatment to develop a more robust understanding of the diverse extrinsic motivation-related patient characteristics that may promote or thwart mandated SUD treatment completion and outcomes. Further, CJ- and CPS-mandated treatment can occur at various stages of involvement with the CJ and CPS systems and can be associated with various levels of consequences. Future research should consider these factors and their relationships to patient motivation levels and the effectiveness of mandated treatment in promoting women’s treatment completion.

While clinically defined discharge status did not differ by mandated status, bivariate findings did show that in this sample, women who were not mandated to SUD treatment appeared to have greater substance use severity and mental health symptomatology (e.g., post-traumatic stress symptoms, drug and alcohol craving, depression symptoms) than those mandated by the CJ system. These findings are consistent with those of previous studies (Kelly, Finney, & Moss, 2005; Marshall & Hser, 2002). Such differences may reflect variation in stages of the recovery process and/or level of pretreatment engagement with services, which is likely greater among those mandated to treatment. Treatment providers should consider variations in clinical profile by mandated status in treatment planning to improve treatment outcomes.

4.2. Role of other factors in discharge status

In line with prior work, the current study found drug and alcohol craving to be associated with increased odds of being a noncompleter of treatment without satisfactory progress due to its well-known association with substance use relapse and early treatment dropout (Hopper et al., 2006; Law et al., 2016). Prior work has demonstrated that craving mediates the effects of stress in predicting relapse during alcohol treatment (Law et al., 2016).

The role of stress in the activation of craving of alcohol and drugs has been well documented (Cavicchioli, Vassena, Movalli, & Maffei, 2020; Sinha, 2001; Sinha & Li, 2007; Wemm, Larkin, Hermes, Tennen, & Sinha, 2019). Increased craving due to exposure to heightened levels of stress is associated with relapse (Law et al., 2016; McCarthy, Curtin, Piper, & Baker, 2010; Wemm et al., 2019). Women in the early stages of residential treatment are particularly susceptible to stress and craving, facilitating substance use relapse and treatment dropout (Daughters et al., 2009). Future research should examine the complex relationship of stress, early treatment dropout, and drug and alcohol craving, which may differentially mediate these associations among the three referral groups (CJ, CPS, and nonmandated).

4.3. Implications for future research

Our study’s findings of clinically defined discharge outcomes highlight the scientific and clinical value of measuring treatment discharge with a direct approach to assessing progress made during treatment (Black & Amaro, 2019). Beyond length of stay (days or months in treatment) as a retention measurement used in prior research (Marcus et al., 2009), this study extended that measure by investigating three levels of progress made at treatment discharge, which provides important clinical insight. For instance, long-term SUD recovery is often a cyclical treatment process involving several treatment episodes (residential or outpatient; Volkow, Poznyak, Saxena, & Gerra, 2017). As Volkow et al. (2017) noted, “drug addiction is a complex multifactorial health disorder characterized by [a] chronic and relapsing nature” (p.213), which requires a chronic care model approach in which SUDs and treatment are addressed just as with other chronic health conditions. Measures often used to assess mandated treatment outcomes (e.g., days in treatment, treatment completion, and substance use) do not reflect the accepted understanding of SUDs as a chronic condition. Research on mandated treatment would benefit from more nuanced measures of progress resulting from discrete treatment episodes (e.g., improvement in addiction severity, mental health symptoms, craving, and recovery-related skills; Cacciola et al., 2012; Goodman, McKay, & DePhilippis, 2013).

4.5. Limitations and strengths

As with any scientific study, there are limitations to the current paper. Generalizability may be limited due to the single-site treatment program and modality of women’s residential SUD treatment, and the sample size may have affected the reliability of estimates and limited the ability to detect an effect with certain psychological variables. Although qualified study site clinicians determined well-defined criteria for discharge outcomes, we could not verify the internal consistency of its use. In addition, the current study did not distinguish between the CJ process stage (e.g., presentencing, probation, parole) or program-level factors associated with early treatment dropout (i.e., insurance limitations for nonmandated women and family financial obligations; Guerrero, Cepeda, Duan, & Kim, 2012; Guerrero, Khachikian, Kim, Kong, & Vega, 2013). A strength of this study is its rigorous approach to measurement of psychological and diagnostic variables, having drawn data from a randomized controlled trial, in addition to its diverse representativeness of women in residential treatment and the comparison of CJ- and CPS-mandated and nonmandated women. Thus, findings are likely generalizable to women served in most publicly funded SUD residential treatment programs with similar demographic characteristics.

5. Conclusion

There are several implications for women’s mandated residential SUD treatment. The current study demonstrates that women who were CJ mandated had a higher likelihood of completing treatment while also having an increased risk of discharging from treatment without satisfactory progress compared with CPS-mandated or nonmandated women. These findings highlight the importance of examining clinically defined discharge outcomes, in contrast to days in treatment as a measure of retention, which enables evaluation of associations both within and between referral groups with discharge outcomes. The current study further suggests that measuring clinically defined discharge outcomes provides a direct approach to assessing progress made during the treatment course, which may inform subsequent mandating agencies and SUD treatment programs of patients’ prior treatment episode progress. Further, future research should identify potential risk factors within and between referral groups (i.e., race or ethnicity, SUD or AUD severity, drug and alcohol craving, and prior incarceration history) that necessitate targeted approaches for mandated and nonmandated subgroups that are susceptible to noncompletion of treatment. Clinics, for example, may incorporate treatments focused on managing or reducing drug and alcohol craving and stress. However, research should examine these directions before programs can broadly implement them. This study is an important starting point in understanding the complex psychosocial structures related to mandated and nonmandated women’s residential substance use treatment and their associations with clinically defined discharge outcomes.

Highlights.

• Criminal justice mandated women are more likely to complete residential treatment compared to women not mandated.

• Criminal justice mandated women are more likely to complete treatment compared to child protective services mandated.

• Drug and alcohol craving is associated with the noncompletion of treatment without clinical progress.

• Improved treatment completion and treatment progress cannot be assumed based on mandated status.