Ethnographic Assessment of an Alternative to Incarceration for Women with Minor Children

Abstract

Allowing criminal justice-involved women to remain with their children in the community may decrease some of the negative intergenerational effects of incarceration. Little is known about potential program models to safely support community co-residence in this population. Ethnographic methods were used to explore the historical development of and life within a supportive housing alternative to incarceration (ATI) program for women with minor children and how they impact the health and social needs of resident families. Participants included 8 current and former adult tenant, 12 of the their resident children, 3 program staff, the program administrator, and 5 prosecutors who originally conceptualized it. Women also reported information about their 8 non-resident children. Analysis revealed three major themes: “The Cycle,” “This is My Home,” and “This Doesn’t Go With That.” While the program built on a core value of family preservation, results illustrate that keeping families together is only the beginning. Clinical and research implications for co-residence ATI programs are discussed in relation to the uniqueness of this context and population.

Jurisdictions across the US are exploring reform strategies to address the negative effects of incarceration and to manage unsustainable costs. This is occurring after almost four decades of explosive growth in incarceration rates (National Research Council, 2014). Traditionally, both carceral and community criminal justice (CJ) programs have been designed to manage individuals, mostly men, apart from their family systems (Arditti, 2012; Morash, 2010). Approximately 60% of incarcerated women report being the mothers of minor children (Glaze & Maruschak, 2010). While men continue to vastly outnumber women within all phases of the CJ system, incarcerated women are more likely to report being parents, having multiple children, and living with those children prior to their incarceration (Glaze & Maruschak, 2010; Mumola, 2000). CJ-involved mothers are also more likely to experience a range of health and social inequities that affect their own wellbeing and their abilities to care for their children (Arditti, Burton, & Neeves-Botelho, 2010). As CJ reform efforts increase, critical need exists for research on alternatives to incarceration (ATI) tailored to serve the unique needs of pregnant and parenting women. To expand the evidence in this area, this strategic ethnographic study explores the historical development of and life within a supportive housing (SH) ATI program for women with minor children and the health and social needs of resident families.

Multigenerational Effects of Maternal Incarceration

The number of women incarcerated in the US has grown dramatically over the past 35 years. Approximately 213,000 women were incarcerated in U.S. prisons and jails on any given day in 2013, representing a 657% increase since 1980 (Cahalan, 1986; Carson, 2014; Minton & Golinelli, 2014). Overall, U.S. imprisonment rates have begun to stabilize, but a holistic view hides increases in the female prison populations in 36 states, the percentage of women sentenced to more than a year in prison, and in the total female jail population (Carson, 2014; Minton & Golinelli, 2014). Increased incarceration rates in women have been linked to drug policy and law enforcement changes, not to increases in overall criminality or violent behavior during this time (Chesney-Lind, 2002). These changes have disproportionately affected women of color. The U.S. imprisonment rate for black women in 2013 (113 per 100,000) was over twice the rate for white women (51 per 100,000; Carson, 2014).

The dramatic growth in female incarceration rates has had wide-ranging, multigenerational effects. The number of children with mothers in prison more than doubled (131%) between 1991 and 2007 (Glaze & Maruschak, 2010). On any given day in the US, the mothers of approximately 147,400 children are in prison. By providing only point in time estimates within a dynamic system, federal statistics grossly underestimate the total number of children separated from their mothers due to imprisonment at some point over the course of their childhoods. Federal statistics for jailed mothers and their children are unavailable, leaving a large group of affected children completely uncounted. Additionally, an estimated three to 10% of women are pregnant upon admission into a correctional facility (Clarke, Hebert, Rosengard, Rose, DaSilva, & Stein, 2006; Maruschak, 2008). This is also likely an underestimate, given inconsistencies in pregnancy testing and reporting in jails (Clarke & Adashi, 2011). In 42 state prison systems and all but one jail, women who give birth while incarcerated are separated from their newborns at or before hospital discharge (Goshin & Byrne, 2009; Women’s Prison Association [WPA], 2009).

CJ-involved women’s active parenting roles increase the risk of caregiving instability in these families. Caregiving mothers are immediately separated from their children upon incarceration. An estimated 61% of imprisoned women report living with their children immediately prior to their incarceration compared to 41% of men (Glaze & Maruschak, 2010). Within the group of parents who reported living with their children, mothers as opposed to fathers provided most of their children’s daily care. Relatedly, when fathers are incarcerated, almost 90% of children stay with the other parent, the mother, whereas only 37% of children stay with their fathers when their mother is incarcerated (Glaze & Maruschak, 2010). Relative to the family networks of incarcerated fathers, mothers’ networks are more likely to include other incarcerated members, and mothers more often report that their children’s other parent is concurrently incarcerated (Dallaire, 2007). This leaves women with fewer stable caregiving options. The rise in incarceration of women has been a primary contributor to the increase in nonrelative foster care caseloads (Swann & Sylvester, 2006). Maternal history of incarceration has also been associated with increased time in child welfare custody and decreased likelihood of reunification (Ehrensaft, Khashu, Ross, & Wamsley, 2003).

Health and social inequities experienced by incarcerated women affect their wellbeing and have the potential to compromise parenting. Incarcerated mothers are predominately single and have low levels of education (Johnson & Waldfogel, 2002). Relative to incarcerated fathers, incarcerated mothers are more likely to have been unemployed and homeless before their last arrest (Mumola, 2000). Mental illness, substance dependence, and a history of family violence are more prevalent in incarcerated mothers than incarcerated fathers or community-residing women (Glaze & Maruschak, 2008). Incarceration in mothers is often the end-point of a long period of instability, particularly for those who are substance dependent or mentally ill (Ehrensaft, Khashu, Ross, Wamsley, 2003; Brennan, Breitenbach, Dieterich, Salisbury, & Van Voorhis, 2012). CJ involvement may compound rather than improve health and social vulnerabilities, such that upon reentry women continue to struggle with mental health issues, substance use, family violence, and lack of resources, while also working to resume at least some level of care for their children (Arditti & Few, 2006). Within this context of cumulative disadvantage and family instability, substantial proportions of incarcerated mothers report commitment to their mothering identities and plans to reunify with their children (Enos, 2001; Ferraro & Moe, 2003; Richie, 2001; Stringer & Barnes, 2012). It is unclear from the current body of research in this area how many women do go on to resume care for their children after release from a correctional facility.

A growing body of research documents a range of negative outcomes in children whose mothers have a history of incarceration. In young children, insecure attachment representations (Poehlmann, 2005) and behavioral problems (Murray, Farrington, Sekol, & Olsen, 2009) are more likely in this group than samples of children with mothers not known to have a history of incarceration. In adolescence, maternal incarceration significantly predicts academic failure (Hagan & Foster, 2012), depression (Foster & Hagan, 2013; Lee, Fang, & Luo, 2013), and personal CJ involvement (Dallaire, 2007; Huebner & Gustafson, 2007). In fact, incarceration of either parent has been proposed as a primary means of intergenerational stress transmission in the US, especially in families of color (Turney, 2014).

And yet, newer research taking into account the other adversities experienced by children of incarcerated mothers provides a more nuanced view. This work calls into question the prior focus on the independent effects of the incarceration-related separation itself as opposed to the larger context of family cumulative disadvantage. Recent research findings suggest that maternal incarceration may have no behavioral effects on school-aged children whose mothers are at high risk for incarceration by virtue of demographic and behavioral indicators (Turney & Wildeman, 2015; Wildeman & Turney, 2014). This research does not address important, potentially confounding, variables in the areas of family process and the extent of prior maternal CJ contact (Arditti, 2015). Most critically, it focuses on one discreet, if important, outcome, and does not in the end demonstrate a net benefit of maternal incarceration. It does however provide an impetus to develop more nuanced clinical and policy responses for a population experiencing multiple, chronic, and intersecting vulnerabilities.

U.S. Responses to Maternal Incarceration

No national standards or best practices currently exist to guide a response to parental incarceration in general or maternal incarceration more specifically. For the most part, U.S. CJ policy is controlled by state and local government agencies. The large, de-centralized, and politicized nature of this system renders standardization difficult. Federal policy in this area only governs the custody of parents incarcerated in federal prisons. It does not address broader issues regarding maternal incarceration, or affect the much larger group of mothers in state prisons and local jails.

While an assortment of parenting-related supports are provided in individual correctional facilities across the US, the facilitation of ongoing parent-child communication comprises the main U.S. response to parental incarceration. Communication between incarcerated women and their children most often occurs through letter writing, phone calls, and in-person visits. Extreme variation exists across the country in policies and programming to facilitate parent-child contact (see Poehlmann, Dallaire, Loper, & Shear, 2010, for an excellent review). In visiting policy, for example, major differences are seen in the following areas important to maintaining the relationship between a mother and child: the frequency of scheduled visiting times, visiting hours (i.e. during or outside of school hours), how incarcerated people gain or lose visiting privileges, and the level of physical contact allowed between visitors and the incarcerated person. For example, in some jails children may only see their mothers through closed-circuit television transmission, whereas others provide the ability to touch, play games, or share food (Poehlmann, et al., 2010). With regards to programming, correctional facilities may offer child-friendly visiting spaces and some level of intervention to help families manage feelings of distress during and after visiting, whereas others provide little in the way of support.

Mother-Child Co-Residence Programs within the Criminal Justice System

A small number of state and local programs across the U.S. allow women who are incarcerated or facing incarceration to live with their children. Prison nurseries, also known as mother-baby units, represent the most well known option. These are segregated prison units in which incarcerated women live with their infant or toddler children. Eight state prisons have functioning nurseries, though the total number has expanded and contracted over the past 40 years (Goshin & Byrne, 2009). While rare in the US, prison nurseries are common in women’s prisons in other developed and developing countries (Pösö, Enroos, & Vierula, 2010; Robertson, 2008). Across U.S. programs, eligibility is generally limited to women who are pregnant upon entry, convicted of non-violent crimes, and have an expected release date within 12 to 18 months of the child’s due date (Goshin & Byrne, 2009). Women may also lose their eligibility due to a history of child welfare involvement (Correctional Association of New York, 2015).

Co-residence in a U.S. prison nursery is associated with generally positive outcomes for the small number of women and children allowed to access these programs. Researchers followed one hundred mother-child dyads during their stays in New York State’s two prison nurseries (one is now closed) and for up to five years after their return to the community (Byrne, Goshin, & Blanchard-Lewis, 2012). Infants and toddlers in this sample showed age-appropriate developmental trajectories (Byrne, 2010). Even though high rates of internalized insecure attachment representations were seen in participating mothers (Borelli, Goshin, Joestl, Clark, & Byrne, 2010), child attachment patterns matched those found in low-risk community samples (Byrne, Goshin, & Joestl, 2010). In the preschool period, children in this sample showed significantly lower scores for internalizing behavior problems compared to a non-equivalent comparison group of children who were separated from their mothers by incarceration (Goshin, Byrne, & Blanchard-Lewis, 2014). The New York program has served as a model for the rest of the country. Maternal recidivism rates after prison nursery release in New York, Nebraska, and Washington are consistently much lower than those seen for women released from the general prison population in those or other states (Carlson, 2009; Goshin, Byrne, & Henninger, 2014; Rowland & Watts, 2007).

Even with evidence of positive outcomes, ethical questions about prison nurseries remain. Children living in prison nurseries have been called “invisible” within larger institutions not designed for their care or wellbeing (Pösö, Enroos, & Vierula, 2010). No guidelines or best practices exist in this area, and monitoring efforts are spotty (Correctional Association of New York, 2015). Best interest of the child standards and correctional protocols do not always align (Byrne, Goshin, & Blanchard-Lewis, 2012). For example, children can be abruptly removed if correction officers observe relatively benign behaviors they determine to be poor parenting, such as co-sleeping or bottle propping. The small number of programs and child age restrictions mean that prison nurseries serve a small proportion of the total number of families in need. Finally, advocates for incarcerated women increasingly question the need to imprison the pregnant and parenting women who would be eligible for these programs, namely those convicted of non-violent crimes (Correctional Association of New York, 2015; Rebecca Project for Human Rights, 2010; WPA, 2009).

In response to these concerns, multiple small, scattered programs have been developed to allow women facing incarceration to instead live with their children in community settings. Mother-child ATIs on which papers have been published use a wide range of program models and institutional structures, with most integrated into residential drug treatment settings (Cassidy, et al., 2010; Lichtenwalter, Garase, & Barker, 2010; Siefert & Pimlott, 2001). Like prison nurseries, eligibility is also often limited to pregnant women with non-violent charges (Cassidy, et al., 2010; Siefert & Pimlott, 2001). Women may also be denied for taking psychotropic medications (Barkauskas, Low, & Pimlott, 2002) in this population of women with a very high burden of mental illness. Programs for women with children beyond infancy are rare, with only two peer-reviewed papers describing programs allowing school-aged children (Brennan, 2008; Wiewel, & Mosley, 2006) and one on a program that accepts children up to pre-adolescence (Lichtenwalter, Garase, & Barker, 2010). To my knowledge, no diversion programs besides the one described herein allow adolescent children to remain or reunite with their mothers.

Evidence suggests multiple positive outcomes related to co-residence in mother-child ATI programs for pregnant women and young children. Access to supportive services in the following critical areas appears higher than in correctional facilities: childbirth preparation (Siefert & Pimlott, 2001), parenting and child development classes (Siefert & Pimlott, 2001), and intensive intervention to enhance caregiving sensitivity and attachment (Cassidy, et al., 2010). In particular, delivery of the Circle of Security-Perinatal Protocol in one jail-based diversion program was associated with caregiving sensitivity comparable to mothers in a community comparison group, improved maternal depressive symptomatology, and rates of attachment security that closely match low-risk community samples (Cassidy, et al., 2010). Women and Infants at Risk (WIAR), a program that served as prison early release or diversion from incarceration for pregnant women, showed a higher likelihood of breastfeeding compared to eligible women who remained in prison (Barkauskas, Low & Pimlott, 2002). Women in this program were also no more likely to have positive urine toxicology screens, despite arguably higher access to drugs in the community. More women in this program were also arrest-free in the ten years following their babies’ births than women in the comparison group (Kubiak, Kasiborski, & Schmittel, 2010).

Research in this area also highlights the inherent complexities of designing effective programming for families with serious health and social needs and multi-system involvement. Women mandated to the WIAR program as an alternative to incarceration were less likely to successfully complete it than women placed there as an early release from prison option (Kubiak, Young, Siefert, Stewart, 2004). In House of Healing, a diversion program for women with children under 12, almost half (47%) of residents absconded or were discharged for rule violations, and 57% of all women in the program had post-release reconvictions (amount of time after program completion unclear). Child welfare involvement may also be promoted through increased surveillance of women’s parenting. In the WIAR program, 54% of target children were temporarily or permanently removed from their mother’s care by child welfare authorities in 10 years after birth compared to 33% of comparison children. Community co-residence programs also show a significant amount of instability over the years related to funding and policy changes, such that both the jail-based diversion and WIAR programs discussed above are now closed (Barkauskas, Low, & Pimlott, 2002; Cassidy, et al., 2010). Mother-child ATI programs have also not used comprehensive models, such as supportive housing (SH), that have shown great promise in populations with a similar set of intersecting vulnerabilities.

Supportive Housing for Criminal Justice-Involved People

Supportive housing (SH) is an umbrella term for a diverse set of programs that combine affordable housing with flexible, wrap-around services for people experiencing or at risk of homelessness (Corporation for Supportive Housing [CSH], 2013). These services are tailored to the needs of target clients and often include some mix of case management, clinical services, and connection to community resources. In addition to homelessness or risk of homelessness, eligibility may be limited to adults with mental illness, substance dependence, or other disabling or chronic health condition (CSH, 2013). People living in SH have their own private spaces and legal rights as tenants. In the SH model, stable housing is seen as a platform for improved health, increased employment and income, and improved social connection and support. In homeless adults with mental illness and substance use disorders, a substantial body of research demonstrates that SH increases housing stability and improves mental health, as evidenced by decreased emergency department use and psychiatric hospitalizations (see Rog, et al., 2014 for a review). Besides safe, quality housing itself, the most effective elements of SH remain unclear, possible due to the flexibility and diversity of programming in this model.

SH programs have also been tailored to the needs of CJ-involved populations. Early research on SH showed decreased incarceration days in people who frequently cycle between jails, shelters, and emergency departments (Culhane, Metraux, & Hadley, 2002). Subsequently, SH programs, such as Frequent Users Service Engagement (FUSE), were further refined to specifically address the intersectional issues of CJ involvement, unstable housing, mental illness, and drug use (Aidala, McAllister, Yomogida, & Shubert, 2014). A hallmark of these tailored programs is coordination between SH providers, the courts, and community correction agencies (CSH, 2011). Results from the evaluation of two FUSE programs showed decreased criminal recidivism (CSH, 2011; Aidala, et al., 2014). Programs have also been developed to serve people being released from prison. Compared to a matched control group, post-release SH was associated with decreased recidivism and increased use of appropriate mental health services in adults with a disability and a history of homelessness being released from prison in one Midwestern state (Fontaine, Gilchrist-Scott, Roman, Taxy, & Roman, 2012).

Currently, there does not appear to be any overlap between the literatures on CJ-focused and family SH programs. The research on CJ-focused SH programs best reflects the experience of single men, who comprise the majority of this population. Any discussion of the gender-specific needs of women or of tenant’s children, either residing with them or not, is absent. The research on family SH programs has focused on the mental health needs of resident children (Gewirtz, 2007; Gewirtz, DeGarmo, Plowman, August, & Realmuto, 2009; Gewirtz, Hart-Shegos, & Medhanie, 2008), with no mention of parental CJ-involvement. This leaves gaps in our knowledge of potential SH models to address the unique needs of CJ-involved primary-caregiving mothers and their children.

Family Supportive Housing as an Alternative to Incarceration

A single-site family SH program was developed in one of the largest jurisdictions in the country specifically as an ATI for women with minor children. Families have their own individual apartments, with leases and tenancy rights. They can enter and exit without approval during curfew hours. Women can work, or attend school and social events. Partners, other family members, and friends can visit but cannot reside. Because of the SH structure families can remain housed there after the completion of the mother’s term of community CJ supervision.

Women are initially identified by the local district attorney’s (DA) office or the criminal courts. Eligibility criteria include the following: female defendant who has been charged with a felony; willingness of the defendant to take a plea bargain offered by prosecutors; judicial approval for a suspended sentence; and custody or legal ability to reunify with a minor child. Women can bring up to three minor children. With violent charges, the crime cannot have resulted in serious injury, and the victim must approve of an alternative sentence. Funding at the time of this study required women to also meet the following criteria: 1) homelessness defined as living in a space not appropriate for human habitation, in transitional housing, or in an emergency shelter; imminent loss of primary nighttime residence; fleeing or attempting to flee domestic violence or other dangerous conditions; and 2) a physical or mental disability. The Housing and Urban Development (HUD) definition of disability was used. It is broad and “may include, but is not limited to, conditions such as visual or hearing impairment, mobility impairment, HIV infection, mental retardation, drug addiction (except current illegal use of or addiction to drugs), or mental illness” (HUD, n.d.). History of mental illness and drug dependence were the most commonly reported disabilities. The program received no CJ funding, and had no onsite CJ staff or programming.

The length of community CJ supervision is usually between 12 to 18 months. Each woman has a unique agreement with the court regarding the conditions that must be met prior to the completion of her term of community supervision. Mandates vary by women’s charges and needs as determined by the courts. They can include requirements to obtain mental health, drug or alcohol treatment, and to seek education or employment training. Women regularly report to a non-governmental, third party agency that monitors compliance with court mandates in people sentenced to community supervision in lieu of incarceration in this jurisdiction. This liaison agency serves as the formal link between the courts and mandated treatment providers. Charges are dropped or reduced to misdemeanors when women complete the terms of their agreements.

Case management and brief counseling are provided at the program site. A case manager conducts intake assessments, develops individualized service plans, provides limited individual and group counseling, and connects tenants to outside services. She also assists women in meeting their court mandates. This includes the following: 1) monitoring and supporting women’s attendance at mandated services; 2) ensuring that women present at scheduled meetings with the court liaison; and 3) submitting written progress reports to the court liaison. A live-in housing manager is responsible for ensuring the overall maintenance and general safety of the building and monitoring visitors. A part-time licensed clinical social worker is also available for brief counseling with resident women and to assist the full time staff members in managing behavioral disruptions, interpersonal issues between tenants, and rule violations. All other services, including health services and any that are mandated by the court, are off-site.

This study sought to describe from multiple stakeholder perspectives the program’s historical development, life within the program, and their connection with the health and social needs of tenant families. It responds to a national push for research on ATI models that might better address the complex, intersecting vulnerabilities of CJ-involved people than prison or jail settings (National Research Council, 2014), and to the call for research specific to women with children (Arditti, 2015). By telling the story of this program from its inception, the study also aims to provide actionable information to clinicians, policy makers, and advocates interested in partnering to develop programs for CJ-involved women and their children.

Research Design

Participants

All women and children who lived in the SH ATI program described above during its first two years of existence, all program creators, and all program staff were invited to take part in this study. The data presented here were gathered from all but one current and former adult tenant (N = 8), their resident children (N = 12 children), program staff (N = 3), the program administrator (N = 1), and the prosecutors who originally conceptualized the program (N = 5, including the DA at the time and 4 assistant DAs). Women also reported information about their non-resident children (N = 8). The purpose of the broad inclusion criteria was to gather the range of stakeholder experiences developing, working in, and living in this program, each of which was thought to play a potentially important role in program culture and health impact. The only current or former resident who did not participate was removed from the program for repeatedly violating curfew and visiting rules. She could not be located for recruitment. As noted above in the program description, the criminal courts determined program entry; it was not in the control of the researcher.

Participating women can be characterized as single mothers of color with very low levels of education (see Table 1 for participant characteristics). Women tended to bring some but not all of their children and to bring their youngest children. The mean age on entry of resident children was seven years younger than their non-resident siblings. Top criminal charges were evenly split between drug, property, weapons, and violent (reckless endangerment, assault) crimes. This was the first felony charge for all but one woman. Of interest in terms of women’s social networks as sources of support and alternate caregiving during an incarceration, partners or family members were often directly involved in the charges that resulted in program placement. By the end of this study, six women successfully completed court requirements. No residents were re-arrested in the local jurisdiction over the three-year period from the beginning of the program to the end of this study.

Table 1.

Characteristics of Program Residents

	M (SD)	Range
Length of stay until program completion, months (n = 6)	15.2 (5.13)	7.03 – 21.32
Mothers (N = 8)
Age on entry	30.92 (7.59)	20.04 – 40.74
Race/Ethnicity
Black, non-Latino	37.5%
Black, Latino	12.5%
White, Latino	50%
Marital status
Single, never married	100%
Education (Highest level completed)
Some high school	86%
Top criminal charge
Crimes against persons	25%
Drug Possession or Sales	25%
Property	25%
Weapons Possession	25%
Partner or family member directly involved in charge	50%
Children (N = 20)
Number of children per mother, total	2 (1)	1–4
Number of children per mother, resident (n = 12)	1 (1)	1–3
Age on entry, resident children	5.0 (4.31)	43 – 13.00
Age on entry, non-resident children	12.3 (5.45)	2.43 – 17.22

Methods

An ethnographic approach was used to explore the historical development of and community within this SH ATI program and their interactions with the health of tenant women and their children. Ethnographic methods are suited to assessing a shared experience, uncovering interactions between program culture and impact, and documenting innovative areas with limited prior inquiry (Spradley, 1979). They also allow the gathering of in-depth information from the perspective of those most affected to address human problems and change current systems (Madison, 2010; Thomas, 1992).

Data collection took place between January and August 2011. Data were collected by multiple methods, including unstructured and semi-structured interviews, and participant observation. Interviews and observation focused on the specific aims of this study but remained flexible enough to capture unexpected findings and relationships among variables. I conducted all interviews with current and former program tenants, their children, and program staff, as well as all of the observations. Another trained researcher conducted all interviews with the prosecutors and program administrator. Having a separate interviewer speak with legal authorities helped maintain the strict separation between potentially damning information collected from women currently under mandate and those who could affect their legal fates. Unstructured interviews with resident adults, children, and staff occurred throughout data collection, at first to establish rapport, then with time, to confirm or discomfirm emerging findings. Semi-structured interviews were conducted once with all adult participants. These occurred in a private setting without children present. The length of semi-structured interviews ranged from approximately 20 minutes to two and a half hours, with the average length being approximately 45 minutes.

Semi-structured interviews with tenant women began with broad questions eliciting life on a typical day in the program and their experiences of living and mothering there. Women were also asked to describe their paths into the program and their court mandates. The health and caregiving histories of all of their children and their own decision-making around which child or children to bring into the program were also explored. Finally, women were asked to contrast life and mothering in this program with any time spent in other institutional and non-institutional living situations. Staff questions focused on their roles and responsibilities and the experience of working there. Prosecutors and the program admininistrators were queried on the historical development of the program, the process by which women gain entry, and their decision-making around staffing and service provision.

Participant observation was conducted approximately six hours per week over the full eight months of data collection. This included extensive time in common areas, participants’ apartments, and when invited, at special events hosted by individual women or program staff. Most observation occurred on weekdays and included the time when children returned home from school. This timing maximized the observation of interactions between adult tenants, children, and staff. Observation was guided by a framework in which the following major dimensions of the social situation were noted: who was observed, their actions and interactions, their apparent goals, the emotional tenor of the interaction, and the outcomes. Extensive field notes written at the end of each observation day documented observations and unstructured interviews.

Data Analysis

Data were analyzed using the inductive thematic approach outlined by Madison (2012). I was immersed in these data from the beginning of the project, such that data collection and the emerging analysis informed one another throughout the course of the study. For example, by comparing information across interviews and observations as documented in the field notes, I was able to ask follow-up questions to clarify inconsistencies. All semi-structured interviews were audio recorded and transcribed verbatim by a professional transcriptionist. I then reviewed each transcript for accuracy while listenening to the recorded interviews, and reread the field notes for that day to re-create the context in which the data were collected. Program development, life, and health impact were the areas of particular focus in this study. Keeping this in mind, after all data had been collected, they were clustered around these broad categories. Exploratory coding was then applied to data within each cluster, such that lists of themes that could best describe that cluster were developed from line by line analysis of the free-flowing text. The identified themes were then compared and contrasted for overlap and marked distinctions, leading to a final reduction of the data into the presented themes. Narrative text from within each theme was then chosen to give voice to participant experiences.

Validity

The following ethnographic strategies recommended by Stewart (1998) were used to maximize validity. I conducted weekly observation for eight months. This extended observation facilitated the collection of a large amount of data to determine patterns and disconfirm or reorient observations. It also decreased the potential effects of reactivity over time as my visits became normative, and allowed me, as an outsider, to build strong relationships with families and program staff. Through the strength of these relationships, women continued to willingly and independently share their lives with me throughout the course of data collection. Multiple informants and data sources were also used to triangulate findings. A data trail was maintained to address threats to objectivity. Audio-recording, professional transcription, and an additional review were used to ensure the accuracy of semi-structured interview transcripts. Finally, extensive field notes were written immediately upon exit from participant observation, be it at the program site or elsewhere.

Ethical Issues

Institutional Review Board approval and a Certificate of Confidentiality were obtained for all described study activities, including observation and informal interviewing of children. Pseudonyms are used in this article to further ensure the confidentiality of study participants. Naturalistic observation occurred only in common areas, unless I was invited into a family’s apartment. The physical layout of this site supported the ability of participants to maintain privacy and avoid observation when they so desired. Each family had their own fully separate apartment with a locking front door. Women and children who did not wish to be observed on any given day when the researcher was present could quickly and easily remove themselves.

Adult current and former program tenants received $25 and an age-appropriate book for each of their resident children upon recruitment. Women did not receive additional monetary incentives for continued participation. Staff, the administrator, and the prosecutors did not receive a monetary incentive for participation.

I became an accepted member of this social scene in the role of peripheral-member-researcher. I was unable to assume a more active role given the context of the setting as an ATI, and potentially because of my demographic differences with the group as a white person with a high level of formal education (Adler & Adler, 1987). Women and staff knew about my clinical experience as a nurse, but they did not appear to see my role in the house as care provider. I never declined to answer health questions, which were mostly oriented around whether they or their child should seek care for a particular issue, but was infrequently asked to do so and never asked to provide health care. Participants were informed a month in advance of the planned ending of participant observation. As a final thank you gift, a family portrait was offered. I remain in touch with participants who initiate continued contact. Since data collection has ended, I have returned to the site upon invitation to celebrate holidays and birthdays.

Results

Three major themes are presented with narrative evidence to illustrate the path this program took to fruition, life within the program, and how these impacted the health of resident women and their children: “The Cycle,” “This is My Home,” and “This Doesn’t Go with That.” Within this structure the very different perspectives of the stakeholders included in this study provide a comprehensive view of how early conceptions led to the programs greatest strength of keeping women with at least some of their children, while also contributing to its greatest weakness, the unaddressed health and social needs of resident families.

The Cycle

For the three prosecutors who originally conceptualized this program, family preservation was the key to breaking “the cycle,” their term for when the child of a formerly incarcerated parent also gets in trouble with the law. In their collective long experience as prosecutors they had seen great increases in the incarceration of women, something they believed to be a prime contributor to intergenerational incarceration. The DA summed the group’s thinking up this way: “the concept was if we could, in an apartment setting, keep the child with the mother, who might be prison bound for problems like drug addiction, maybe we could break that cycle.” Another prosecutor described how her distress over witnessing intergenerational incarceration spurred her to action, “I think the philosophy is really about trying to break the cycle, trying to help these kids and trying to not have what happened to their mothers happen to them. Their fathers are in jail, and you know that their kids are going to jail if they go to jail. You just know it, so you’re desperately trying to figure it out.” This group of prosecutors’ had determined that their own prior approaches to illegal behavior in the form of mandatory incarcerations were ineffective and potentially harmful, and that they wanted to try something different.

The passion for this project seemed in part also to emanate from a feeling of culpability for not recognizing or addressing the family effects of maternal incarceration sooner. When speaking of the evolution of their thinking, narratives took on a confessional quality. One prosecutor said, “I had never thought about that issue before, which sounds ridiculous. Seventeen years at that point, doing this kind work; never had thought about it, never thought about the kids.” This jurisdiction had no prior formal programming in the criminal courts for families, or specifically for women with children. They were known for the wide use of alternative sentencing programs, like drug courts, but a formal evaluation of the local drug court, a program in which they took great pride, showed women did not do as well as men. Clients were sentenced to long periods of residential drug treatment. Some women reported choosing jail time to limit the length of separation from their children. Working on a new paradigm of keeping women with their children felt much better to prosecutors than the status quo. A prosecutor with 21 years of experience said, “to see the children being able to stay with their moms is a real shot in the arm for me as a career prosecutor who spent so much time with kids that came from disenfranchised homes and had to go into foster care… I actually feel like a rock star when I go to court, because for the first time in my career I have defense attorneys that are smiling at me. They love me.”

The cycle also refers to political and funding cycles that affect which interventions are possible at any given time. Actively working to develop a diversion program for women facing felony charges and their children was, and possibly still is, a novel idea for prosecutors. This was especially true in 1999 at the height of the U.S. prison boom. In their view, this novelty at a time when others were focused on incarceration worked against them, in that they struggled for over seven years to get others behind it. With legal colleagues in and out of their office, the prosecutors described a deep level of discomfort for a diversion program like this one. For one, allowing women with felony charges to stay with their children, was not perceived as enough punishment. One prosecutor outlined her approach with colleagues this way, “We’re changing hearts and minds. It’s the best way to try to help the kids and, look—it’s not like the woman’s won the lottery.” Secondly, colleagues expressed to them that mothering in CJ-involved women was not something worthy of support. One prosecutor who had spoken with people across the country over the years said, “you would often hear, and you knew, that there was this feeling that these women didn’t deserve their children.”

They also struggled to find funding support. The long-time DA of this jurisdiction, a powerful politician at the time, had personally approached other politicians and funders about this project. While people told his less powerful colleagues it would never work, they gave him positive feedback but not funding. He said, “I couldn’t convince anyone. I went to people in the business world; I went to people in the political world. And everyone said it was a great idea, but the money’s not there. The money’s always there to build prisons.” The CJ focus at that time was on programs to support reentry, the period of time after incarceration. “If we called it a reentry program, we could have gotten the money,” another prosecutor said. The double whammy of taking away the incarceration altogether and adding in children was much less palatable.

But while rare, there’s was not the only jurisdiction experimenting with family diversion options. Two of the prosecutors toured four programs across the US. They knew that they wanted a program to keep women with their children but still were not clear on what it should look like, who should be let in, how many children they could bring, or what services should be provided. They described being most struck that programs limited the number of children to one per woman and set age cutoffs in toddlerhood. One said, “I felt very strongly that you know if we are interested in keeping the family together, we can’t come up with these false numbers.” With these visits, the program began to take more shape from their original broad idea into a place for women to bring multiple children with no age limits.

In the end the final program structure as family SH was not the result of thoughtful planning but of a chance meeting and funding priorities in the areas of homelessness prevention, not CJ. The prosecutors had reached out to multiple local drug treatment providers over the course of seven years, thinking that this was the best service provider group with which to partner. From one of these meeting, they were introduced to the administrator of a local nonprofit that had just received funding for family SH but had not yet found families for the spaces. “It was somewhat serendipity, and spontaneous, and being in the right place at the right time, and the right pot of money was there,” the program administrator said. And yet, the strings attached to this funding were not ideal. It stipulated that a family be technically homeless, as defined above, to gain entry, and that program families could remain in program apartments as long as they would like. In interviews with prosecutors over two years after the program opened, dissonance remained around how long families could stay. One prosecutor said, “the program is for a defined period of time, and then when they’re stabilized they can move on, and of course the apartment can be taken up by the next family.” This contrasted with the terms of the major funding support and what tenant women came to understand when they moved in, but after years of waiting the prosecutors jumped at this chance to finally open the program. As one prosecutor said, “We told her that was great. And you know what? I asked no questions.” Within a year of the meeting, the first tenant accepted a plea bargain, received a suspended sentence for reckless endangerment, and moved into this family SH program with her infant.

This is My Home

The unmarked, salmon-colored row house sits quietly on a residential block. With the exception of a boarded up house across the street, the building and the block have curb appeal. The neighboring properties are other well-maintained brick row houses and large apartment buildings, some of which are also managed by social service agencies. In front of this building, there are flowers in a window box on the first floor in the spring and summer, and the sidewalk is the first on the block to be shoveled in the winter. As you walk in the building, there are no guard desks, metal detectors, or signs with rules. Instead there is a bulletin board with pictures of famous African American women, a small bright yellow office, a narrow staircase, and the closed doors of individual apartments. When the back door is propped open to let in the breeze, you see a backyard with a swing set and a garden with vegetables and flowers.

Reliance on peer-based staffing further contributed to the non-institutional feel. Aretha, the case manager, was a professional-ex, someone who used her own history of CJ involvement to help others (Maruna, 2001). After serving 11 years in a maximum-security prison, she completed a Bachelors degree in social work, then started a Master’s in social work during the course of this study. This job was her second position within an agency serving CJ-involved people. Her approach with tenant women was influenced less by her formal studies than by instincts honed on the streets, her experience of being treated with respect “and allowed to be a woman” during her own incarceration, and her memory of the pain of being separated from her daughter. In describing the program and her role in it, Aretha said, “It’s not just supportive housing. We become almost part of a family. I’m like their mother, their sister, their aunt. I’m like a father to some of the children. It’s more intimate.” Tina, the housing manager, was living in SH herself when she got this position. She too saw a broader role for herself in the program than attending to maintenance issues. “I make sure maintenance is taken care of, but the main thing is taking care of the ladies. I mentor them; I make sure their everyday functions are calm and quiet. I also take care of the kids, you know, mentoring them, see how they’re doing, how well they get along in school.” Both the case and housing manager took their formal job descriptions and ran them through the filter of their own experiences of CJ involvement and housing instability. Women and children sought proximity to them. Tenants approached them with respect and used them as social, but not always professional, support. Children treated them like older aunts, seeking their attention, approval, and affection. Even when enforcing curfew and visiting rules or in their roles as reporters to the court liaison, tenants never appeared to view staff as correctional actors.

Inside the five apartments in the building that house tenants (the sixth is reserved for the live-in housing manager) there is a familiar domestic scene. Women and children show off their apartments with pride. In all, they took great care to set up their spaces with a mix of simple, program-provided furniture and their own decorations, some purchased on layaway, handed down between tenants, or made by their children. Walls and refrigerator doors are adorned with drawings on construction paper and macoroni art. Older non-resident siblings stare down from candid pictures on the walls. Stuffed animals, action figures, and coloring books lay on the floors and tables. The rooms of younger children are bright with cartoon character–inspired toys and bedding. Shirley picked Tinkerbell for her young girls. The older kids chose movie posters for themselves. Jennifer’s boys, who said their favorite thing to do is go to the movies with their mother, chose Batman, whereas Aricele’s daughter chose Twilight. To families living here this is not a correctional space but a home.

Women consistently used the word home to describe their living situations while in the program. The supervisory aspects were acknowledged, but placed secondary to staying together with children and having separate apartments. Marisa’s quote was emblematic, “They gave us our space. We had our own apartments. We had rules to obey and curfew, but I didn’t really care about that. I was with my daughter. It was homey. It was my home.” Women did not need to have all of their children with them to consider it home. Jennifer, who lived with two of her three children, said, “It’s not even time that I’m doing here because as you can see, this is my home. The only thing is, it’s just a little bit of rules that we got to live by.” Shirley was the only tenant who denied that the program was her home, but her narrative and behavior suggested proprietary feelings over the apartment. She alternately described where she and two of her children lived as “my apartment” and “my own space.” She bristled that according to the rules a staff member could come into her apartment without her consent, even if it was just to address maintenance concerns. She also threw her children large birthday parties, to which she invited family, friends, and her most respected community leaders, including her pastor and the principal of her oldest daughter’s school. In these ways Shirley demonstrated that while she did not choose this space or the rules within, this program had become a home for her and two of her children.

Children also made the house their home. Daily life had a familiar domestic rhythm built around children’s schedules. Mothers and children walked to school or day care together in the morning and back in the afternoon. In the summer, there was day camp from which the school-aged children return sweaty from playing, and the houses’ one adolescent returned tired from work as a paid camp counselor. In the afternoon, children took over the house. The central hallway was a gym and a dance floor, and the stairwell was a place to do homework. When the weather was nice, the backyard filled with preschoolers to preteens running around or playing with the stray cat that dropped by for scraps. The little ones scooted on ride-on toys or got pushed on the swings. The older children played tag or threw a basketball into a small plastic hoop.

One-third of resident children were below the age of two years upon their entry, and thus could have a limited understanding of why they moved into the home. School-aged children, mothers felt, were also unaware of why they had moved or that they were living in an ATI, even when they were aware of their mothers’ CJ involvement by virtue of witnessing their arrests or visiting them in jail. To prevent possible disclosure to children who were not aware of the program’s focus, I did not specifically ask their thoughts on why they moved into the building, and they did not independently mention it to me during the course of our play. Their behavior and their mothers’ reports suggested they too felt a sense of ownership over the space and their apartments. Teresita, speaking from the perspective of her six and seven year olds, said, “Wow! Now me and my mommy got our own apartment.’ They don’t know nothing. They just think, ‘my mommy got her own house. We got our own apartment,’ and they just like it.”

This feeling of home contrasted with prior unstable or crowded living situations. Half of the women described years of precarious housing on the streets, in shelters, and in the homes of others who were not able to take them in for extended periods. Aricele, who had the longest and most serious history of drug dependence, had been living on the streets without either of her children for many years. This was the first stable housing situation she had lived in as an adult. Jada had lived in a series of shelters and the home of her abusive partner’s mother. Shirley, while pregnant with her youngest, slept on someone’s floor with her toddler before she decided to enter the shelter system. Jennifer and her two school-aged children slept in the assisted living facility of an elderly relative with dementia, putting the woman’s housing in jeopardy. Of the four women who had been stably housed, only Marisa had had on the books employment and her own market rate apartment. Receiving Section 8 shortly after becoming an adult had kept CeCe and her children stably housed. Her housing assistance had been immediately terminated, however, upon her incarceration for drug possession. For the youngest two women in the house, Teresita and Jacinta, this was their first time living apart from an older female relative who had provided much of their care and the care of their children.

The home’s safety and calm were also a relief after years of trauma and loss. These were not the natural losses of elderly relatives that one can expect in the normal course of life. They were brutal and unnatural. For example, Jada and Aricele had lost children; Marisa lost her young brother; and Shirley had also lost her mother to murder during her own childhood. Losses and traumas were shared or witnessed by children. For example, Aricele and Shirley had lost former partners, while their children lost fathers. Children in three families were exposed to violence in their prior living situations. In describing the night before her arrest on the burglary charges for which she came to this program, Jennifer said, “I was so scared of him and of me almost dying that night. My kids witnessed everything. It traumatized my kids. I was all bruised up, all beat up, broken nose, stab wounds.” There was no indication in any family that violence was directed at the children, but nevertheless they were exposed in some cases, like the one above, to extreme violence directed at a loved one. Women’s stories were carried barely under the surface, emerging without direct questioning in private interviews, and in informal one-on-one and group conversations between residents. The experiences remained raw, unprocessed, and unresolved. Sharing stories created strong ties among the women. At these times, other residents and the case and house managers were a ready source of understanding and informal support, but women declined to reach out to the part time counselor about their trauma.

Women brought these experiences into a program about which they knew little. They all described jumping at “the opportunity” to stay together or reunite with their children, even as they worried what form the program might take, and how it might affect their children. Marisa thought, “it was going to be like a little jail. I thought it was gonna be a bunch of ghetto girls, and a bunch of fighting.” Teresita, who said she had been “mommy and daddy” to her oldest two children since their births, expressed feeling fearful that, “I’m not going to be able to come outside, I’m going to be locked up with my kids. Like when they first told me about it, I was happy that I could be with my kids, but then again, I didn’t want my kids to go through that, to be locked up in one place and they can’t go outside and have any freedom.” CeCe and Jennifer had been in residential drug treatment and felt that the program was going to be similar.

Women’s willingness to bring their children with them into an unknown institutional program highlighted a shared desperation to remain together in some families or prevent further dissolution in others. All of the women but one were separated from their resident children during jail stays prior to program entry. These ranged from three days to four months. For the five youngest children in the program, this time represented their only caregiving separations. Shirley’s children, who had uninvolved fathers and few family members in the US, were the only ones to go into non-kinship foster care during her incarceration. The other children stayed with fathers, aunts, and grandmothers until they could be reunited upon their mother’s release. Aricele and her teenage daughter were not separated by her incarceration because they had not lived together since the child’s infancy due to Aricele’s long history of opiate dependence and mental illness. Theirs was the only reunification in the home after a period of extended separation.

Similarly, incarceration was not the cause of the extended separation in other older resident and non-resident children. Shirley’s sister kept her two oldest when she moved to the city to seek employment. At the time of this study they were nearing adulthood, and she was not planning to resume care for them. All other extended separations were related to women’s drug dependence. Jennifer’s mother had cared for her oldest child from infancy while she “ran the streets.” She subsequently had two children who needed care when she went for inpatient drug treatment. These children stayed with their father during her 18-month stay. Her story highlights the lack of safe alternate caregivers in some women’s lives. “My mother couldn’t take my kids. And you know this was the hardest thing that I did was give my kids to their father. My kids went through hell. He was getting public assistance for them. And the kids weren’t eating properly. They weren’t dressing them.” After a custody battle, Jennifer reunited with her two children and was able to bring them into this program. CeCe’s story was different. Her youngest child could live in the drug treatment facility with her, but the program was not open to her two preteens. As a result they moved three hours away to live with their father. CeCe then entered this program because it allowed older children. In coordinating their move back to the city, CeCe’s children told her they would prefer to stay with their father.

Despite not residing in the home, older children were a presence there. Women and younger children visited siblings who lived in the local area as frequently as every weekend. They used Skype and social networking to keep in touch with children who lived farther away. From their new home bases, CeCe, Aricele, Jennifer, and Shirley, strengthened collaborative parenting relationship with their children’s caregivers, even though they never planned to regain custody. In CeCe’s case, she made peace with being unable to reunite with her older children. She said, “Time changes, but it was by the grace of God that I let them go and live with their father. Cuz [their smaller town] is different than down here. These kids get into everything, and these kids follow one another and do all kinds of stuff, and I do not want my kids involved in all that. There used to be a lot of bickering between us [her and her ex-husband], but we got over that, past that. He tells me if something is going on with one of them, and tells me to talk to them because they listen to me for some reason, and they give them a hard time.”

One-third of women in this study chose to stay in their apartments after they were allowed to leave, resulting in only two of five apartments being used for new ATI tenants. All women interviewed prior to their completion used the term “moving on” to describe their dreams for after the program. In these dreams they would find another clean, safe, affordable apartment close to their family members and without a curfew or visiting rules. While women independently desired this move, the narrative framework of “moving on” originated with housing staff who introduced the concept with women for both positive and negative reasons. On the positive side, they were responding to women’s broader desires to start a new life after their mandates by helping them secure external housing. Negatively, staff exerted pressure on women to let someone else have their spots. Aretha described her approach, “I tell them, you know, it’s time to move on. I encourage it. I let them know, listen, I’m not trying to put you out or put you into the street, or I’m not saying that you’re not welcome here, that we don’t want you here, but the fact is that this is ATI program and to continue staying here means that you still will have to abide by the rules.”

As their mandates neared or passed, women expressed ambivalence about moving. They described weighing safety and low risk of family separation against increased freedom and a break from program rules. This occurred in the context of poverty and a difficult urban housing market, yet some declined opportunites to move. Aricele had been there for over a year after being allowed to move. Shirley and Teresita, similarly, expressed no plans to leave. For women who chose to leave their post-mandate housing patterns were similar to the circumstances in which they lived prior to this CJ contact. Cece, whose housing had been supported by Section 8 that was lost upon her drug felony charge, moved into an SH apartment at another site. She reported she had no plans to move prior to being offered the new apartment. Jada, as she had in the past, moved in with her new partner’s family, and Jacinta moved back in with her mother and older child. Marisa, was the only tenant to sustain on the books employment throughout her stay and to seamlessly move into market rate housing with her child’s father.

This Didn’t Go with That

The organic and somewhat informal way that the program evolved got it launched after a seven-year struggle. The program succeeded with this small group in keeping women out of carceral settings and in safe housing with their youngest children. In other fundamental ways, however, the program failed to meet its own goals. At the level of the program, the SH structure was both a strength and a potentially fatal weakness. Women and children accepted their program apartments as their homes. As more women asserted their tenancy rights after completion of the mandate, fewer families could move in. In the end this meant that the main program goal of diverting women with children from incarceration could not be met, as only two apartments were open to ATI clients at the end of this study.

At the level of individual tenant women, unmet mental health needs represented a serious and ongoing concern. The program lacked on-site health services, health care coordination, or an established connection to women’s off-site psychiatric and drug treatment providers. Case management was oriented toward helping women complete court requirements, which in all women included some amount of drug treatment. The case manager prioritized and frequently reviewed adherence to correctional supervision and treatment mandates, yet in the two cases discussed below adherence to mandates appeared to be the problematic issue. Staff were unprepared to deal with the complexity of multi-system involvement or to question conflicting treatment goals between psychiatric and drug treatment providers.

Aricele and Jennifer’s experiences illustrate gaps at the intersection of CJ, drug treatment, and mental health systems that the program was unable to address, leading to the two women stopping their psychiatric medications against medical advice. Aricele said, “The Suboxone didn’t go with this medication, and this didn’t go with that, so I stopped it and just never went back.” After stopping her opioid replacement medication, Aricele briefly relapsed. She was re-admitted to the program after a short admission into residential drug treatment. Her daughter spent the time at her grandmother’s house. Jennifer was instructed by drug treatment personnel at the center to which she was mandated to discontinue prescribed psychiatric therapies or face remand. “My psychiatrist had finally found medications that worked well for me. Then it came to my [drug treatment] program. They was like, ‘you can’t take this. Your toxicology is going to come back positive.’ So right now I’m not seeing a doctor, and I’m not under medications. So it’s like I don’t want to be codependent also on meds. I’ve been off meds for quite a while.” She was instead mandated to attend anger management groups. Both women completed the program in the end.

The program was created for a population of mothers with historical and ongoing parenting vulnerabilities, but provided no formal intervention in this area. Five of the eight women in the home came in with active child welfare cases. When asked what services the program needed, CeCe immediately responded, “some help with the kids. Because some people are just getting back their kids and they don’t know how to cope dealing with their kids.” Relational issues commonly observed between women and their children included harsh communication, inconsistent discipline, differential treatment of siblings, and role reversal. For example, Jennifer, whose children had repeatedly witnessed her being assaulted by their father, took pride in them being protective of her. She described them as “my little body guards.” Women with child welfare cases reported receiving didactic parenting education classes but denied any relational interventions.

Aretha and Tina role modeled positive communication and discipline when children misbehaved in the office or halls. Neither had formal training in this area, and the program had not adopted an intervention model. They were hesitant to exert a stronger influence for fear of offending the women. Aretha said, “people have these different ideas on how to raise their children, you know, so you don’t want to tell them what to do, you kind of want to give them and drop them hints, and kind of give them suggestions in a way that, you know, well, for a child this age, it might be healthier if…you know, or something like that. And, when I was younger this worked for me…you know, whatever, to just try to encourage them to do better by their children.”

While children were the stated reason for its inception, this program had no services specifically for them, apart from a summer day camp. There was also no collaboration between the program and any pediatric health provider. Children presented with and developed a range of health needs during their time in program. Half of the children had witnessed domestic and community violence prior to their arrivals, but none had received therapy. Some of the health conditions were endemic to urban, low-income areas. For example, Shirley’s youngest had a speech delay related to lead poisoning that occurred during a shelter stay, and four of the children in the home had asthma. Shirley’s oldest, at the age of five, had unfilled baby bottle caries so extensive that the dentist threatened to report her to child welfare. Jada’s son was slowly being evaluated for potentially serious genetic disorder causing global developmental delays.

Discussion

This SH ATI program provided women facing incarceration with a safe, home-like space in which to complete court requirements while caring for their minor children. Positive results were tempered by unmet needs for trauma-informed mental health and pediatric care. Allowing families to remain in their apartments after program completion kept them together and stably housed, but greatly limited access to the program for other CJ involved women and their children. To my knowledge, this study is the first to describe SH for women charged with felonies and their children. It is also unique in obtaining the perspectives of multiple stakeholders and using participant observation. It adds to a strong body of research supporting gender-responsive, health-focused CJ programming and responds to calls for research on less-restrictive options for CJ-involved women with children (Belknap, 2014; Haney, 2013).

Given the non-institutional model, broad eligibility criteria for women and children, and limited available data on other programs, situating these results within the available literature presents a challenge. The shared work to create normalcy and the relative freedom of resident families contrasted with the restrictive environments of community programs described elsewhere (Haney, 2013). Allowing families to remain in their apartments after completion of CJ requirements was also unique. Primary funding source may be a key determinant of this difference. Whereas this program received SH funding, others were funded through departments of correction. Required program elements, such as maintaining a locked environment, onsite security personnel, and drug testing, may vary greatly between funders with a public health or social service versus a correctional mission.

Consistent with other studies, women remained or reunited with their youngest children. Tight age criteria separated families in some programs, but across studies women also reported custody losses prior to CJ contact (Barkauskas, et al., 2002; Brennan, 2008; Kubiak, et al., 2004; Lichtenwalter, et al., 2010). The attributions for prior custody loss made by women in this study fit with research in incarcerated mothers, which shows a pathway to CJ contact through trauma, mental illness, and drug dependence (Brennan, et al., 2012).

The core belief that preventing incarceration-related mother–child separation would lead to positive outcomes was also consistent across studies. While other studies also highlighted complex relational needs (Brennan, 2008), only the attachment intervention described by Cassidy and colleagues (2010) systematically addressed the effects of prior trauma on caregiving. This intervention was limited to mothers of infants, though there is a Circle of Security protocol for toddlers and preschoolers (Marvin, Cooper, Hoffman, & Powell, 2002). Other programs, relied on potentially less effective parenting classes (Brennan, 2008; Wiewel & Mosley, 2006; Lichtenwalter, et al., 2010; Siefert & Pimlott, 2001). Haney (2013) cautioned that parenting interventions within CJ settings can also be used to further punish or shame women, though this was not found in the present study.

Wide variation is seen across studies in on-site health services. Some, like this program, appeared to provide no health services (Lichtenwalter, et al., 2010), while prenatal care in particular was the main focus for the WIAR program (Barkauskas, et al., 2002). These data suggest that limited on-site services, lack of established connections to high quality external health providers, and communication challenges between treatment providers created health threats, and affected all residents through additional stress in the program milieu. The experiences of women with co-occurring disorders (COD), in particular, call attention to the need for integrated treatment in community corrections. A large study assessing COD treatment in community corrections settings across the US recently documented drug court staff unfamiliarity with mental health issues and negative beliefs about psychotropic medication use by people with a history of drug dependence (Peters, Kremling, Bekman, & Caudy, 2012).

Health service gaps could be addressed in ways tailored to the small size of the program. Off-site services decrease program costs and help ensure continuity of care if residents move. Development of strong partnerships with high-quality mental health, drug treatment, and pediatric primary care providers may ease coordination challenges. Regular visits by a public health nurse or family nurse practitioner could provide systematic screening, initial treatment, referrals to appropriate care, and continued coordination. The communal setting also supports the use of group interventions for trauma, stress, and parenting.

Further research is critical to guiding the development of mother-child ATI programs. Questions remain regarding intervention models, optimal program size, and staffing skill mix. CJ laws and policies, research ethics, and small program sizes create methodological challenges to comparing co-residing families with those separated due to maternal incarceration or offered other diversion options. Innovative methods may be needed to approximate differences using non-equivalent comparison groups.

Limitations

This project reflects an in-depth, qualitative look at one urban mother–child ATI that used an SH model. While this study, like all ethnographies, cannot claim generalizability, it does provide insights applicable to similar contexts. Similarities and differences of the women in this study to other CJ-involved families within and outside the area are unclear. Prosecutors, the courts, and even women themselves upon taking the proffered plea agreement determined who entered this program. Women who participated in this program represented a small portion of women in this large jurisdiction who met program eligibility criteria. Those who received a different sentence were not approached for recruitment. Relatedly, this research did not address the process by which the DA’s office offered the program to some women but not others. This work may have been enriched by observation of the larger network of social situations in which resident women were involved, including the courts, the third party agency that monitors compliance with court mandates, mandated treatment centers, women’s families and other social contacts. While women and staff provided extensive verbal information on their experiences in and with those institutions, I was only able to observe within the program itself. Finally, the overall young age of resident children limited my ability to better incorporate their unique perspectives. For this reason, much of the information included here was derived from participant observation or interviews with their mothers.

Conclusion

Overall, these results give preliminary support for the expansion and replication of SH as an ATI for felony-charged mothers with minor children, particularly in light of the documented negative health effects associated with maternal incarceration and the small number of family ATI programs. Keeping families together is only the beginning. A trauma-informed approach and attention to family health needs is needed to realize long-term intergenerational goals. In this first report of a unique SH ATI, it has been demonstrated that a small number of participants can complete court mandates, provide an appropriate home environment for minor children, and live safely with each other. Allowing CJ-involved mothers and their children to live together in ATI programs is an innovative approach to addressing inequities, preventing mother–child separations, and avoiding the stresses of incarceration and release.