The Journal of Behavioral Health Services & Research

, Volume 33, Issue 2, pp 225–243

A Comparison of Mothers with Co-occurring Disorders and Histories of Violence Living with or Separated from Minor Children


    • Center for Mental Health Services Research, Department of PsychiatryUniversity of Massachusetts Medical School
  • Norma Finkelstein
  • Valerie Williams
  • Jennifer Thom
  • Chanson Noether
  • Megan DeVilbiss
Regular Article

DOI: 10.1007/s11414-006-9015-5

Cite this article as:
Nicholson, J., Finkelstein, N., Williams, V. et al. JBHSR (2006) 33: 225. doi:10.1007/s11414-006-9015-5


Data from the Women with Co-occurring Disorders and Histories of Violence Study are used to examine characteristics distinguishing mothers currently providing care for all their minor children (n = 558) from mothers separated from one or more minor children (n = 1396). Mothers are described and compared on background characteristics and experiences, well-being and current functioning, situational context, and services used. Analyses control for number of children, race, and years of education. Mothers separated from children have more children, less education, have more often been homeless, in juvenile detention or jail, and have lower incomes than mothers living with all their children. Mothers separated from children have more extensive experiences of traumatic and stressful life events, and the groups differ in current functioning and patterns of services used. While cross-sectional data do not allow causal inferences, challenges faced by mothers have significant implications for policy and programs.


The issue of motherhood within the context of co-occurring mental illness, substance abuse, and violence has been ignored in the published research literature and by the services community.15 Analyses of national prevalence data from the National Comorbidity Survey (NCS)6,7 indicate that over two thirds of women with co-occurring mental health and substance use disorders are mothers.1,8 In addition, recent research indicates that individuals with co-occurring mental health and substance use disorders are more often victims of violence than those with either a mental health or substance use diagnosis alone.9 It is likely, therefore, that the majority of women with co-occurring disorders who are mothers have experienced violence either in childhood, as adults, or both.10 It is well known that parental mental illness and, particularly, substance abuse contribute significantly to the numbers of children in out-of-home placements.2,1113 Likewise, a history of abuse as a child may increase the potential for mothers to abuse their own children, resulting in the involvement of child-protective services and increasing the likelihood of mother–child separations.14 Mothers with mental illness and those with drug and alcohol problems identify loss of legal custody or disruptions in caregiving relationships with children as their most significant concern.3,15,16

Research efforts on mothers with mental illness and substance use disorders, parallel to services and interventions, remain largely segmented by field and diagnosis. Published studies tend to focus on small convenience samples of disadvantaged women with the most serious of disorders, typically receiving public sector services either in mental health or substance abuse treatment settings. Research reflects investigator-identified problems and deficits, rather than the voices of mothers themselves.2,8,15 There have been no studies, to the authors' knowledge, of mothers with co-occurring disorders and histories of violence and the relationship between their characteristics and experiences and maintaining caregiving responsibility for their children.

The Significance of Motherhood

Parenting is central to the lives of most mothers, whether living with mental illness or substance use disorders, and is a major source of identity and self-worth.2,5,15,17,18 Similar themes emerge in both the mental illness and substance abuse literatures, underscoring common issues for mothers with mental illness or substance use disorders, and suggesting a convergence of issues for women with co-occurring disorders. Motherhood is frequently seen by women with mental illness or who abuse substances as a “normalizing” experience, one that connects them with others, and provides a specific role and purpose.15,1820 Children may be a source of hope for the future. Women's attachments to their children may motivate them to seek help and participate in treatment. Difficulties in parenting, exacerbated by the lack of economic, familial, and social supports, as well as the stigma and discrimination often faced by women with mental health or substance use disorders, can result in anxiety, guilt, and negative self-images.2,15,17,2024

The Contribution of Experiences of Violence

Clearly, a history of violence as well as exposure to current violence can influence a woman's effectiveness as a parent and her experience of her children, particularly in terms of attachment.25,26 A woman's capacity to parent can be affected by multiple factors, including the extent and nature of the violence to which she has been exposed, her parenting role models, and her involvement in trauma-treatment and trauma-informed parent training.27 A woman's experiences of violence can increase the possibility of problematic parenting indirectly as well, through the impact of violence on current situations, such as homelessness or dangerous living situations, mental illness or substance abuse, and unsupportive relationships. These, in turn, may contribute to abusive or neglectful parenting.28 For example, mothers beaten by their partners have been found to have twice the child abuse rate of nonabused mothers.29 Another recent study found that involvement with a violent partner was a more significant predictor of child abuse than drug abuse alone, a finding not inconsistent with the literature on family violence.30

The Impact of Mother–Child Separations

There are no national data on the prevalence of custody loss or mother–child separations for mothers with mental illness, substance use disorders, or both, although it is thought to be common.8 In fact, fear of losing custody or contact with children is reported to be a deterrent to treatment participation for women.15 Also, no data are available regarding the prevalence of informal alternative caregiving arrangements, e.g., mothers voluntarily making arrangements with grandparents or other family members. Although smaller-scale studies, typically of clinic populations, report rates of “custody loss” of anywhere from 25 to 75% or higher,2,31 whether these figures represent formal, legal arrangements or informal relationships is often not specified.

The transfer of legal custody to others may not always result in the loss of caregiving responsibility or contact. For example, mothers may be involved with their children on a regular basis, providing care or visiting, when children are in the legal custody of relatives or extended family.2,15,32,33 Likewise, alternative or additional caregiving relationships may develop informally, without legal involvement or sanction, particularly in times of emergency, when respite is warranted, or when women enter inpatient or residential treatment facilities or are incarcerated.

There are times when voluntarily relinquishing custody or caregiving responsibility, or involuntarily separating children from their parents may be in everyone's best interest. Indeed, federal policy dictates that the safety and well-being of children are society's priorities, and mandates actions within certain time frames to ensure children's safety and well-being.34 Both maternal mental illness and substance use have been linked to the elevated risk of child abuse and neglect,29,3539 developmental and mental health problems in children,8,40,41 school failure,42 future substance use and abuse,43,44 and delinquency and crime.45 State laws vary as to whether maternal substance abuse is regarded as an “automatic” reason for removing children from a parent's home. However, in a 1997 national survey of state public child welfare agencies conducted by the Child Welfare League of America, parental chemical dependency was a contributing factor in the out-of-home placement of at least 53% of the children and youth in state custody.46

Mother–child separations, particularly those that are unwanted or unnecessary, are likely to affect women's treatment and recovery. Connections are fundamental to women's psychological growth and healing.17 Separation from a child may be experienced as a “disconnection,” which may exacerbate a mother's difficulties. Women may turn to alcohol or other drugs to relieve the anxiety and pain caused by this disconnection.17 Mothers with mental illness may literally decompensate when relationships with children are disrupted. In a British study, the majority of mothers with mental illness prior to separation from their children spent more time in the psychiatric hospital after separation from their children than before.47 Mothers who involuntarily lose or voluntarily relinquish custody of their children, or who give up caregiving responsibility for them, are often seen by others and feel themselves to be “failures.”48 Mothers are likely to feel intense shame and ambivalence after the removal of a child, as well as feelings of sadness and loss that last a lifetime.15,49

The Women, Co-occurring Disorders and Violence Study

Data in the study reported here are obtained from the larger Women, Co-occurring Disorders and Violence Study (WCDVS), a nationwide, federally funded, longitudinal study of women with co-occurring mental health and substance use disorders and histories of violence. The study was conducted at nine program sites across the USA, with six programs located on the East Coast, two in California, and one in Colorado. The WCDVS is the first large-scale, comprehensive effort to characterize and describe this population, as well as to examine the effects of innovative, integrated, comprehensive, and trauma-informed services on both short- and long-term outcomes for women across the domains of mental health, substance use, and trauma symptoms. A unique and innovative aspect of the WCDVS is the active and meaningful role played by women with co-occurring disorders and histories of violence themselves, “consumer/survivor/recovering women” or “C/S/Rs,” in the design and implementation of the research and interpretation of findings, ensuring that women's voices were heard throughout the entire research process.

Although the WCDVS was not designed as a study of motherhood, it provides a unique opportunity to describe a large sample of a previously ignored but significant group of mothers, and to begin to address the research question of what distinguishes mothers who continue to provide care for their minor children from those who are separated from their minor children, regardless of legal custody status. Variables reflect mothers' (1) background characteristics and experiences, (2) well-being and current functioning, (3) situational context, and (4) services used in the past 3 months. There is little or no literature to inform the development of specific research hypotheses regarding the relationship between mothers' characteristics and experiences and mother–child separations, and the cross-sectional nature of the data does not permit the testing of causal relationships. It is reasonable to suggest, however, that mothers who currently have caregiving responsibility for their children are likely to be those with greater individual, situational, and social resources and supports; who have less severe histories of abuse and trauma; and who report better well-being and current functioning.

Whereas the focus of the larger WCDVS was the testing of an innovative intervention approach over time, the wealth of data obtained at baseline allows for the description of this rarely studied group of women and the comparison of the subgroups of interest noted above. Findings will provide the essential first step in the development of policies and programs specifically designed to meet the needs and build on the strengths of mothers and children whose lives are shaped by these issues.


Study design

The overall WCDVS design was quasi-experimental, with nonrandom group assignment to intervention or comparison conditions based on naturally occurring circumstances, typically portal program of entry into services.50 In some cases where multiple intervention and comparison programs existed in a given site, intervention and comparison programs were further matched based on direct program comparisons. Although some differences did exist among programs, when aggregate comparisons were conducted, the profiles of women in the WCDVS sample at baseline were remarkably similar between intervention and comparison conditions.51 A total of 2729 women from across the USA participated in the WCDVS.

Women recruited for the study met the following criteria:
  • at least 18 years of age;

  • a DSM-IV Axis I mental disorder and/or Axis II personality disorder;

  • a DSM-IV substance use disorder;

  • a history of interpersonal violence, e.g., physical and/or sexual abuse;

  • at least two prior episodes of mental health or substance abuse treatment;

  • English or Spanish as their primary language

Either the mental disorder or the substance abuse disorder had to be current (within the past 30 days), with the other occurring within the past 5 years.


The vast majority of the women participating in the larger WCDVS study (n = 2729) were mothers (about 87%; n = 2366), i.e., answered “yes” to the question, “Have you ever had any children?” and had given birth or had adopted children, stepchildren, or foster children. Almost 6% of the women were pregnant at the time of the baseline interview. Detailed descriptive data on the total WCDVS sample of women are available in a previous paper.51 All women were recruited from the same program locations regardless of motherhood/parenting status. Detailed information regarding the eligibility and recruitment of mothers vs. nonmothers or mothers providing care for children vs. those separated from children was not collected.

Mothers' responses to two interview items were used to group them into categories: “How many of your children are under the age of 18?” and “Of those under 18, how many live with you? With their father? With another relative? In foster care? With someone else?” Because the focus of this study is a comparison of mothers who are current caregivers for their children with mothers separated from their children, mothers whose children were all over the age of 18 were excluded, as parents are typically not legally responsible for children over the age of 18 who are considered adults themselves. Pregnant women with no other living children were excluded as well. The remaining mothers in the current sample (n = 1954) were then grouped into two categories: (1) mothers with minor children, all of whom were living with them (n = 558); and (2) mothers with minor children, who were separated from at least one child, i.e., at least one child was living elsewhere (n = 1396). Of mothers separated from at least one child, about three-quarters report that all their children under the age of 18 are living apart from them. Children under the age of 18 not living with their mothers are living with their fathers (25%) or other relatives (43%), someone unrelated to them (8%), or are in foster care (24%).

Interviewing procedures

Study procedures were reviewed and approved by Institutional Review Boards in all participating sites. Trained research interviewers obtained informed consent from study participants before they joined the study. A federal Certificate of Confidentiality was obtained for the study to provide participants as much protection as possible and to encourage them to respond openly to study questions. Participants were interviewed in a standardized format using a cross-site protocol developed for the WCDVS by researchers, providers, and consumer/survivor/recovering women who served as paid consultants and staff members for the study. Women were interviewed in locations they identified as private, comfortable, and safe, often in their homes. Test–retest reliability for all survey items was assessed via the readministration of the interview to 20 participants per site within 14 days of the first administration and is detailed in a previous publication.52


Variables relevant to the research question posed were selected for the current analyses. The WCDVS baseline interview involved the use and/or adaptation of established measures, as well as the creation of interview items and measures specifically for purposes of the WCDVS. A detailed description of the measures, as well as the corresponding item and scale development and adaptations, may be found elsewhere.52 Self-report measures were used in the WCDVS for all individual-level variables; this decision reflected the philosophical underpinnings of the study and the prioritizing of respect for the knowledge and credibility of study participants.62 All measures had satisfactory reliabilities.

Background characteristics and experiences

(1) Basic demographics, e.g., age, number of children, ethnicity/race (Hispanic, white/Caucasian, black/African American, other, multiracial, nonspecified), education (number of years), relationship status (married/living with significant other, widowed, divorced, separated, and never married), and, as measured by the Life Stressor Checklist-Revised (LSC-R)53, which was modified for the WCDVS,54 items that: (2) solicit information on lifetime occurrence of serious money problems (“Have you ever had serious money problems, e.g., not enough money for food, clothing, or rent?” with responses of “yes,” “no,” “refused,” and “don't know”), homelessness (“Have you ever been homeless?”), and ever having been sent to jail or juvenile detention (“Have you ever been sent to jail or juvenile detention?); (3) reflect childhood circumstances, e.g., “Were you ever put in foster care or put up for adoption as a child?” and “Were your parents separated or divorced when you were a child?”; and (4) assess lifetime and current experiences of stressful events and interpersonal abuse.

Lifetime exposure to stressful events, lifetime frequency of interpersonal abuse, current exposure to other stressful events, and current exposure to interpersonal abuse were computed variables. Lifetime exposure to stressful events (“Have you ever...”) is the sum of “yes” responses to 31 items (possible range = 0–31) reflecting serious disaster, accident, incarceration, family disruptions, money problems, illnesses, disabilities, death, violence, abuse, neglect, assault, restraint, discrimination, harassment, and death of a child. Lifetime frequency of interpersonal abuse (If “yes” response to “Have you ever...,” then “How often has this happened?”) is the sum of frequency responses (“once” = 1, “a few times” = 2, “a lot” = 3) to 12 probes (possible range = 0–36) regarding having witnessed violence, having been abused and neglected, stalked or threatened, forcibly restrained, and having forced sexual encounters. Current exposure to other stressful events (“Has this happened in the past 6 months?”) is the sum of “yes” responses to 20 items (possible range = 0–20) reflecting life stressors other than interpersonal abuse, e.g., serious disaster, accident, incarceration, etc. Current exposure to interpersonal abuse (“Has this happened in the past 6 months?”) is the sum of “yes” responses to 8 items (possible range = 0–8) reflecting being the target of emotional and physical abuse and neglect, stalking or threats of serious harm, forcible restraint, and forced sexual encounters.

Individual items from the LSC-R reflect ever having been physically abused and neglected; ever having sex in exchange for money or material goods; ever having been robbed, mugged, or physically attacked by a stranger; and ever having been touched or made to touch in a sexual way, with responses of “yes,” “no,” “refused,” or “don't know.” Good validity for the LSC-R has been demonstrated in diverse populations of women and in several languages.54 In the test–retest sample of 186 women, the lifetime summary variables and scales for current exposure in the WCDVS version of the LSC-R demonstrated acceptable test–retest reliability, with intraclass correlation coefficients ranging from 0.77 to 0.88; the kappas for individual items included in these analyses ranged from 0.52 to 0.85.54


(1) Perceived physical health, as measured by a single item from the Short Form 12 (SF-12)55 and rated from “1 = poor” to “5 = excellent.” The SF-12 has been used across a wide range of populations for varying purposes and has good demonstrated reliability and validity.55 (2) An item adapted from the 2000 National Health Interview Survey to collect information on any serious physical illnesses and/or disabilities, to which respondents reply “no” or “yes.” (3) Mental health symptom severity, i.e., the global symptom severity rating of psychological distress as measured by the Global Severity Index of the Brief Symptom Inventory (BSI).56 Higher scores on individual BSI items reflect responses indicating greater distress (“1” = “not at all” to “5” = “extremely”) regarding problems such as “nervousness or shakiness inside,” “faintness or dizziness,” etc. The BSI has exhibited good reliability, both in terms of internal consistency (Cronbach alphas of 0.71–0.85 for the primary symptom dimensions) and test–retest reliability (intraclass correlation coefficients of 0.80–0.90 for the global indices).56 (4) Trauma-related symptom severity as measured by the total score of the 17-item Posttraumatic Stress Disorder (PTSD) Symptom Scale-Interview Version (PSS-I),57 i.e., the sum of respondent ratings, from “1 = not at all” to “4 = almost always,” of the extent to which they are bothered by each of the specified PTSD symptoms, e.g., “having upsetting thoughts or images...,” “having bad dreams or nightmares...,” etc. Good reliability and concurrent validity have been demonstrated for the PSS.58,59 (5) Current alcohol severity and current illicit drug use severity, assessed using the Addiction Severity Index (ASI),60 modified slightly in consultation with its author for use in the WCDVS. The Alcohol Composite Score and Drug Composite Score were calculated to measure problem severity in the past 30 days; higher scores reflect greater substance use problem severity. Several studies have established the reliability and validity of the ASI across subgroups of people divided by age, race, sex, or primary drug problem.61,62 The ASI is appropriate for alcohol- and drug-dependent adults, multiple substance abusers, psychiatrically ill substance abusers, the homeless, and prison populations. The ASI composite scores are widely recognized as valid and reliable measures of current substance use.63

Current functioning

(1) Two items, felt to be most closely related to parenting, regarding the degree to which respondents have been “experiencing difficulty in the area of managing day-to-day life” and “household responsibilities” during the past week, drawn from the Behavior and Symptom Identification Scale (BASIS-32).64 Respondent ratings are recoded from “1 = extreme difficulty” to “5 = no difficulty,” so that higher scores reflect lesser reported difficulty. The BASIS-32 has well-established reliability and validity.65 (2) Perceived parenting effectiveness, assessed by four items developed for use in the WCDVS, e.g., “I feel confident in my ability to help my child(ren) grow and develop,” “I feel my family life is under control,” “I believe I can solve problems with my child(ren) when they happen,” and “I feel I am a good parent,” rated from “1 = strongly disagree” to “5 = strongly agree,” so that higher scores indicate higher levels of perceived effectiveness. (3) Current employment status: employed (full or part time) or unemployed, i.e., looking for work, disabled, volunteering, or retired. (4) Average monthly income in the 30 days prior to the interview, including wages, SSI, SSDI, or disability, benefits, and entitlements, child support or alimony, income or money from a spouse, partner, or family member, and retirement income.

Situational context

(1) Current living arrangement: shelter, street, institution, housed or other; six safety-related items developed for the WCDVS including: (2) how safe the woman feels where she is living now, rated from “1 = not at all safe” to “5 = extremely safe”; and (3) an overall rating of personal safety, which is the average of a woman's responses to five items regarding feeling unsafe with people (two items), in places (two items), and alone (one item) rated from “1 = almost always” to “4 = not at all.”

Services used

These items request information about the services a woman has used in the past 3 months. The respondent is asked to respond “yes” or “no” to a sequence of 17 items developed for the WCDVS beginning, “In the past three months did you receive...” or “In the past three months did you stay at...,” about a range of service options including hospital, emergency room, detoxification services, residential treatment, homeless or domestic violence shelter, jail/correctional facility, individual and group counseling (residential and nonresidential), case management (residential and nonresidential), medical clinic, medications, and peer support/self-help. In addition, respondents are asked whether any other providers helped them in the past 3 months (“yes” or “no”), and whether they had any unmet service needs in the past 3 months (“yes” or “no”). Service use responses could reflect treatment for a physical complaint, violence/abuse/trauma, psychological problem, psychiatric medication check, alcohol/other drug abuse, parenting support, legal or housing assistance, educational or vocational/job assistance, or entitlements. Responses reflect whether a particular service option was received or not, rather than the volume of services used.


With each domain of variables, the two groups of mothers with co-occurring disorders and histories of violence are described and compared. Comparisons of groups on categorical measures were made using chi-square tests. Differences between groups on continuous measures were assessed using analysis of variance. p values were adjusted for background characteristics that exhibited a significant bivariate relationship with the group variable. These characteristics included total number of children less than 18 years of age, race, and years of education. Adjustment was carried out using SAS® PROC GLM. The adjusted p values generated from the generalized linear models were corrected for multiple comparisons using a modified Bonferroni procedure.66 [Note: It is important to compare the number of significant results to the number expected to be false positives (wrongly judged as significant) given the total number of tests conducted. Twenty-seven of the 56 adjusted tests conducted yielded results significant at or below the 0.05 level after applying a modified Bonferroni correction. This value exceeds by nine times the figure of 5% that would be expected because of chance alone.]


Background characteristics and experiences

Women in both groups are similar in age, with the average being about 35 years, and race/ethnicity, when adjusted for number of children, race, and years of education (Table 1). Mothers separated from children have significantly more children on average (between two and three) than do mothers living with all their children (about two). Mothers living with all their children are likely to have significantly more education than mothers separated from children, although the differences between the means of 11.7 and 11.1 years, respectively, may not have real-world significance. In fact, over 70% of the mothers living with their children and over 78% of the mothers separated from children completed 12 or fewer years of education. Equivalent percents of women in both groups are married or living with significant others (about 40%), whereas the majority in both groups are without current partners, i.e., either widowed, divorced, or separated from partners (about 30%), or never married (about 30%).
Table 1

Background characteristics and experiences


Mothers living with all childrena (n = 558)

Mothers separated from ≥ 1 childa (n = 1396)


Adjusted pb

Background characteristics

 Age in years, mean (SD)

34.3 (7.6)

34.6 (7.1)



 Number of children, mean (SD)

1.9 (1.2)

2.7 (1.6)



 Hispanic ethnicityd (%)





 Race (%)










  Other racec









  None specified




 Years of education, mean (SD)

11.7 (2.2)

11.1 (2.2)



 Current relationship status (%)


  Married/living with significant other




  Widowed, divorced, separated





  Never married




Experiences (lifetime and current)

 Ever had serious money problems (%)





 Ever homeless (%)





 Ever sent to jail/juvenile detention (%)





 Ever put in foster care or up for adoption (%)





 Mother's parents separated/divorced (%)





 Traumatic life events


  Lifetime exposure to stressful events, mean (SD)

15.5 (4.7)

16.8 (4.4)



  Lifetime frequency of interpersonal abuse, mean (SD)

14.8 (7.3)

16.4 (7.4)



  Current exposure to other stressful events, mean (SD)

2.8 (1.9)

3.5 (2.1)



  Current exposure to interpersonal abuse, mean (SD)

1.0 (1.3)

1.3 (1.7)



  Ever physically abused (%)





  Ever physically neglected (%)





  Ever had sex when not wanted in exchange for money or material goods (%)





  Ever robbed, mugged, or physically attacked by a stranger (%)





  Ever touched or made to touch in sexual way because felt forced (%)





aUnadjusted values.

bp value adjusted for total number of children, race, and years of education.

cCategory includes respondents who identified two or more races.

dHispanic ethnicity was measured independent of race.

ep < 0.02 after modified Bonferroni correction.

fp < 0.05 after modified Bonferroni correction.

Mothers separated from children are significantly more likely to have ever been homeless (78% compared with 58% of mothers living with all their children) or to have been in juvenile detention or jail (76% compared with 60% of mothers living with all their children). There were no significant differences between groups in percentage having childhood experiences of foster care or adoption or the percentage with parents who separated or divorced. Mothers separated from children report significantly greater lifetime and current exposure to stressful and interpersonally abusive events. They are significantly more likely to have ever been physically abused (87% compared with 82% of mothers living with all their children), to have ever had unwanted sex in exchange for money or material goods (61% compared with 48.5% of mothers living with all their children), to have ever been robbed, mugged, or physically attacked by a stranger (47% compared with 40% of mothers living with all their children), and to have ever been touched or been forced to touch someone else in a sexual way (almost 70% compared with about 63% of mothers living with all their children).

Well-being and current functioning

When other variables are controlled for, there are no significant differences between groups of women on perceived physical health, reports of physical illness or disability, mental health symptom severity, trauma-related symptom severity, current alcohol severity, or current illicit drug use severity (Table 2). Mothers living with all their children report significantly greater difficulty managing day-to-day life and household responsibilities than mothers separated from children. Mothers separated from children report significantly lower scores on the four items measuring perceived parenting effectiveness, suggesting they feel less confident, less in control, not as good at solving problems, and less like good parents than mothers living with all their children. Mothers separated from children have significantly lower monthly incomes than mothers living with all their children.
Table 2

Well being and current functioning


Mothers living with all childrena (n = 558)

Mothers separated from ≥ 1 childa (n = 1396)


Adjusted pb

Physical/psychological well-being

 Perceived physical health, mean (SD)

2.7 (1.1)

2.8 (1.1)



 Physical illness/disability (%)





 Mental health symptom severity, mean (SD)

1.4 (0.8)

1.3 (0.8)



 Trauma-related symptom severity, mean (SD)

23.2 (11.6)

24.3 (11.8)



 Current alcohol severity, mean (SD)

0.21 (0.3)

0.20 (0.3)



 Current illicit drug use severity, mean (SD)

0.16 (0.2)

0.17 (0.2)



Current functioning

 Managing day-to-day lifec, mean (SD)

3.6 (1.2)

3.9 (1.2)



 Household responsibilitiesc, mean (SD)

3.9 (1.2)

4.2 (1.1)



 Perceived parenting effectivenesse, mean (SD)



4.4 (0.8)

4.2 (1.0)



  Under control

3.3 (1.1)

2.9 (1.2)



  Problem-solving ability

4.0 (0.9)

3.9 (1.0)



  Good parent

4.2 (0.9)

4.0 (1.1)



 Currently employed (%)





 Average monthly income (30 days prior to interview), mean (SD)

$1028 (1195)

$659 (895)



aUnadjusted values.

bp value adjusted for total number of children, race, and years of education.

cHigher scores indicate lesser reported difficulty.

dp < 0.02 after modified Bonferroni correction.

eHigher scores indicate higher levels of perceived effectiveness.

Situational context

The vast majority of mothers in both groups are currently housed (Table 3). While both groups of mothers feel equally safe in their current living situations, mothers separated from children report significantly lower ratings of personal safety than mothers living with all their children when number of children, race, and years of education are controlled for.
Table 3

Situational context


Mothers living with all childrena (n = 558)

Mothers separated from ≥ 1 childa (n = 1396)


Adjusted pb

Situational context

 Current living arrangement (%)


  Shelter or street














 How safe in current living situation, mean (SD)

3.8 (1.0)

3.9 (1.0)



 Personal safetyc, mean (SD)

3.1 (0.8)

3.0 (0.9)



aUnadjusted values.

bp value adjusted for total number of children, race, and years of education.

cHigher scores indicate higher levels of perceived safety.

dp < 0.05 after modified Bonferroni correction.

Services used

There are significant differences between groups in types of services used in the previous three months (Table 4). Mothers separated from children are significantly more likely to report using residential-, shelter-, or correctional facility-based or detoxification services than are mothers living with all their children. Mothers living with all their children are significantly more likely to report using outpatient services, e.g., individual and group counseling in nonresidential settings. Mothers separated from children are significantly more likely to report using peer support or self-help services than mothers living with children.
Table 4

Services used in the past 3 months


Mothers living with all childrena (n = 558)

Mothers separated from ≥ 1 childb (n = 1396)


Adjusted pb

Treated/admitted to a psychiatric ward/hospital (%)





Emergency room (%)





Detoxification services (%)





Residential treatment services (%)





Individual counseling (residential) (%)





Group counseling (residential) (%)





Case management (residential) (%)





Other services (residential) (%)





In a homeless or domestic violence shelter (%)





In a jail/correctional facility (%)





Individual counseling (nonresidential) (%)





Group counseling (nonresidential) (%)





Case management (nonresidential) (%)





Medical clinic/doctor's office services (%)





Medications for emotional/psychological problems (%)





Medications for substance abuse problems (%)





Peer support/self-help (%)





Any other providers that helped (%)





Unmet service needs (%)





aUnadjusted values.

bp value adjusted for total number of children, race, and years of education.

cp < 0.02 after modified Bonferroni correction.


The vast majority of women with co-occurring disorders and histories of violence participating in the WCDVS are mothers (about 87%), confirming that it is much more likely that these women will become mothers than not. This percentage exceeds the percentage of women in the NCS who meet lifetime criteria for co-occurring psychiatric and substance use disorders who are mothers (67.2%) and the percentage meeting criteria for PTSD who are mothers (72.8%).1 The NCS sample comprises women who may or may not be in treatment. The higher percentage of mothers in the WCDVS sample, all of whom were recruited in treatment settings, may reflect the greater motivation of women who are mothers to enter treatment, whether that motivation is “internally” driven or externally induced by family members, child-protective workers, or treatment providers.

The number of mothers separated from all their minor children is three times that of mothers still living with at least one of their minor children and about twice as many as those mothers continuing to live with all their children under the age of 18. It is important to note that over two-thirds of the children separated from their mothers are living with fathers or other relatives, compared with only about one quarter who are in foster care. The data do not tell us whether these alternative living arrangements are made voluntarily by mothers themselves, or whether these mothers have come to the attention of the child welfare system, and children are placed in kinship care by child protection workers.

Mothers currently separated from some or all of their children have histories severely affected by disadvantage, i.e., a range of psychosocial, economic, legal, and other life challenges including homelessness and criminal justice involvement. While this is not surprising, causal relationships between mother–child separations and the above life circumstances cannot be determined in this study as the data are collected at one point in time. Most likely, these relationships are not linear, but circular and multidirectional. For example, homelessness can lead to loss of children, but loss of children can also lead to homelessness through the loss of income supports and lack of access via ineligibility to family shelters or emergency housing. Likewise, criminal justice involvement frequently results in loss of children.67 Alternatively, it may be important to explore whether and how the despair, shame, anger, and helplessness caused by losing children may lead to criminal justice involvement, possibly via the route of substance abuse or other high-risk behaviors.

Mothers separated from children report significantly greater exposure to stressful life events and sexual and physical abuse. Stressful and abusive life events, including childhood and current violence and abuse, may be strongly related to actual loss of children through the involvement of child-protective services. Adults whose parenting role models in childhood were abusive may be more likely to become abusive as parents, increasing the likelihood that their children will be placed in alternative situations.25

Mothers currently separated from children describe themselves less confident and less in control as parents and as having poorer problem-solving ability than mothers living with all their children. The cross-sectional nature of the WCDV study muddies interpretation of the data regarding perceived parenting effectiveness and separation from children. The findings may reflect conditions that contributed to separations from children or may be the consequence of these separations. Interestingly, although mothers living with all their children feel more confident and in control, they report greater difficulty in managing day-to-day life and household responsibilities than mothers separated from children. This finding is consistent with reports of mothers with mental illness who have difficulty balancing caring for children and managing their illnesses with other demands of daily life.15

In general, more mothers living with their children report using outpatient or community-based services, whereas more mothers separated from children, in general, report using residential or institution-based services, e.g., in homeless shelters or correctional facilities. Given the lack of significant differences between groups in current mental health and trauma symptoms, as well as alcohol and illicit drug use, the differences in patterns of services used over the past 3 months are extremely interesting and suggest issues of access, focus, and networks. It may be that mothers were separated from children because they entered programs that would not allow children. On the other hand, mothers currently separated from children may have had more severe symptoms in the past, necessitating residential or inpatient treatment, and perhaps were functioning better at the time of the interview, i.e., at levels similar to mothers using outpatient services, in response to treatment.

The WCDVS sample is larger than those in previous studies of mothers with mental illness or substance use disorders. Women met stringent recruitment criteria for mental illness, substance abuse, histories of violence, and treatment episodes. Since women were recruited at treatment sites, they are not necessarily representative of the general population of women with co-occurring disorders who may not be in treatment. However, data are from baseline interviews with women and represent their characteristics and situations at the beginning of their involvement in the study and prior to receiving specialized treatment services.

Whereas the issue of the reliability of recall of traumatic experiences is somewhat controversial, and may be affected by many factors and subsequent experiences, it is important to note that women in both groups reported extremely high lifetime frequencies of a variety of stressful and abusive events. Rates of lifetime trauma exposure in the general population found in the NCS, as measured by modified versions of the DSM-III-R PTSD module from the Diagnostic Interview Schedule and of the Composite International Diagnostic Interview, are significantly lower.68 For example, about 7% of women in the NCS report ever having experienced a physical attack; over 80% of the WCDVS mothers report ever having been physically abused, and over 40% report ever having been robbed, mugged, or physically attacked by a stranger. About 12% of the women in the NCS report ever having experienced sexual assault other than rape, and about 9% report ever having experienced a rape. Over half of the WCDVS mothers report ever having sex when not wanted in exchange for money or material goods, and approximately two-thirds report ever feeling forced to touch or be touched in a sexual way.

Whereas interview topics may have been extremely sensitive in nature, interview items were, in select cases, written by and, in all cases, reviewed and approved by women with similar backgrounds and experiences to achieve the greatest relevance and maximum validity. Considerable strength derives from the active participation of women with co-occurring disorders and histories of violence in the design and implementation of this study, as well as in the interpretation and dissemination of findings.

The range of domains of interest in federally funded, multi-site studies such as the WCDVS tends to be wide, in part because these studies are often the first step in developing an in-depth understanding of a treatment population and testing innovative interventions. Initial findings and limitations in interpretation suggest directions for future research and ultimate testing of the model. Clearly, a multivariate, multiple-respondent, longitudinal design is required to disentangle the causal relationships between proximal and distal factors and conditions of motherhood, i.e., protective factors that enhance the likelihood that mothers with co-occurring disorders will successfully provide care for their children, or risk factors that contribute to the likelihood of mother–child separations, voluntary or coerced. More thorough understanding of the process by which women make decisions regarding parenting and the care of their children, and the impact of parenting on women's functioning, service use, and healing and recovery, will allow for the development of interventions targeted to successful outcomes for both women and children.

Implications for Behavioral Health

The research literature and services community point to the neglect of parenting status in treatment planning or the development and testing of services for women with mental illness and substance use disorders.6974 Young children are described as a barrier to both treatment engagement and retention,69,7577 although children have also been described as motivating women to seek treatment.15 The negative attitudes of family members and providers, inadequate transportation and child care, threats of custody loss, and women's enormous guilt and shame over “failing” as mothers have consistently been cited as major barriers to treatment for parenting women.33,7881 Recent policy developments, including the Federal Adoption and Safe Families Act34 (ASFA) and welfare reform, have exacerbated treatment access issues for women, either by shortening the time allowed to participate in and benefit from services to re-unify with one's children or risk the move to permanency planning, or by impacting a parent's access to income supports, thereby perhaps limiting access to reimbursement for treatment services through Medicaid.12

Unfortunately, both mental health and addiction treatment programs, including those specifically developed for women, frequently do not incorporate the critical issues involved with motherhood and caring for children into service design and delivery,78 nor do they typically focus on issues of loss for mothers or their children, who may lose contact with mothers as well as with siblings who may be removed to alternative placements. Given the prevalence of family disruption among mothers in this study (71%), and the percentage of participating mothers reporting separation from all their children (over 55%), both women and children are extremely likely to experience significant losses.

Comprehensive community treatment services and supports can optimize mothers' functioning and prevent the multiple losses so common for these women, as well as support positive relationships with children, partners, and other family members and enhance children's development and well-being.5,9 However, appropriate family-centered, integrated, and comprehensive interventions that can address the full range of service needs for both a mother and her children are too often not available or empirically untested.4,77,78,82 Funding streams or mechanisms may limit treatment participation to eligible individual adult or child family members, but not both; services for mothers and children may be fragmented and provided in different locations.72

In addition, the stigma associated with mental illness and substance abuse and, in particular, the negative attitudes of family members and providers regarding the desire or efforts of women with these issues to parent, may contribute to mothers not asking for help with parenting, not seeking treatment, and saying little about their families while in treatment.1,17,18,32 Mothers identify additional reasons they refrain from seeking treatment, including the following: beliefs that they must put the needs of their children first and that drugs enhance their ability to cope with parenting; denial or minimization of the impact of their illness on family functioning; the lack of gender-specific treatment options particularly for mothers with their children; the threat of legal intervention; the belief that they are “terrible” people and “bad” mothers; and the need to avoid confronting their pervasive feelings of guilt about perceived failures.3,17,67,83,84 It is not clear whether mothers “coerced” into treatment by child welfare professionals or others by the threat of removal of children or termination of parental rights, are motivated to participate more fully and derive greater benefit from services, or are disempowered by fear, anxiety, and feelings of failure and, hence, derive less benefit if relevant, effective treatment is even available.69

Policies and programs should be put into place to promote the integration of a full range of family services and supports. True “family-centered care,” in which a family defines their needs, the focus is on strengths, interventions are broadly conceived, families have choices, and services and systems are flexible and responsive, may be the most relevant and effective approach for women with co-occurring mental illness and substance use disorders, as well as for their children.85

Mothers are frequently overwhelmed with the multiple tasks facing them, especially if they are raising children with limited financial and emotional resources while also dealing with mental illness, addiction, and the impact of violence and trauma on their own and their children's lives. Women with co-occurring disorders who succeed in raising their children do so despite extraordinary challenges. These women and their successes have not been studied. Effective prevention and treatment strategies for women with co-occurring disorders and their families will require: the study of mothers who are successful in raising their children, as well as children who are resilient; review and reformulation of current policies to encourage rather than impede families in their efforts; and the development and testing of innovative treatment models to support women's healing and recovery while enhancing the development and well-being of their children.


This study was funded under Guidance for Applicants (GFA) No. TI 00-003 entitled Cooperative Agreement to Study Women with Alcohol, Drug Abuse and Mental Health (ADM Disorders who have Histories of Violence: Phase II) from the Department of Health and Human Services, Public Health Service, Substance Abuse and Mental Health Services Administration's three centers: Center for Substance Abuse Treatment, Center for Mental Health Services, and Center for Substance Abuse Prevention (March 2000).

Copyright information

© National Council for Community Behavioral Healthcare 2006