Mental Health and Substance Use Problems of Parents Involved With Child Welfare: Are Services Offered and Provided?

It is well documented that characteristics and problems of parents affect the quality of parenting and child outcomes ( 1 , 2 , 3 , 4 , 5 , 6 , 7 ). The detrimental effect on children when parenting is disrupted is perhaps nowhere more evident than in the child welfare system. Many parents served by public child welfare agencies experience personal and environmental stresses that interfere with parenting and subsequently place children at risk of poor outcomes. Researchers have consistently reported heightened rates of depression, alcohol dependence, and other stressors among parents who maltreat their children. A 1983 study found that maltreating parents were significantly more likely than other parents to have a mental disorder ( 8 ). Parents who reported that they beat their children or spanked hard enough to bruise were more likely to have alcohol abuse, depression, and antisocial personality disorder than parents who did not report doing so ( 9 ).

The rate of substance abuse was significantly higher among parents who reported physically abusive and neglectful behaviors compared with parents who did not report these behaviors ( 10 ). Similar results were found in another study; maternal alcohol and drug problems, maternal external locus of control, marital problems, and low paternal warmth and involvement were associated with child abuse and neglect ( 11 ). Other studies have shown that parental mental problems and substance abuse are associated with actual or potential child maltreatment ( 12 , 13 , 14 , 15 , 16 ).

Several studies have examined whether children in foster care and child welfare receive mental health care ( 17 , 18 , 19 , 20 , 21 , 22 , 23 ). Fewer have looked at parents' use of services. A study of physically abused children and their caregivers found that caregiver use of mental health care was positively associated with acknowledgment of the abuse ( 24 ). About half of the caregivers received mental health treatment. At six-month follow-up more child abuse victims than caregivers were receiving treatment. In a study of physically or sexually abused children and their parents, approximately half of the parents received individual or family therapy after the intake assessment, but very few used substance abuse treatment ( 25 ). Neither of these studies reported on parental problems or parents' need for services.

It is important to learn more about parents' problems and service use because meeting the needs of parents may enhance child outcomes. However, it is important to note that parental mental problems and substance use problems are not being assumed to cause child abuse and neglect. This article reports findings from an exploratory study that used national data to address several questions. When child welfare caseworkers identified caregivers with mental problems, were mental health services offered? If offered, were they provided? When child welfare caseworkers identified caregivers with substance use problems, were substance abuse treatment services offered, and if so, were they provided? What other problems were reported for caregivers with mental and substance use problems? Were services offered to specifically address the identified problems?

We hypothesized that there would be significant associations between caseworkers' identifying mental problems and offering mental health services, between caseworkers' identifying substance use problems and offering substance use services, and between caseworkers' identifying lack of parenting skills and offering parental training. We also examined whether caregivers with both substance use and mental problems were more likely than caregivers with only one of these problems to be offered services.

Methods

A secondary analysis of data from the 1994 National Study of Protective, Preventive and Reunification Services Delivered to Children and Their Families ( 26 ) was conducted. The purposes of the 1994 study were to describe the families served by the child welfare system and the services they received and compare the 1994 data with data from the 1977 National Study of Social Services to Children and their Families ( 27 ).

The 1994 National Study used a two-stage stratified probability sampling design. The first stage consisted of selecting counties from the 50 states and the District of Columbia. The second stage consisted of sampling families served by child welfare agencies in the selected counties. Counties were sorted into four categories by metropolitan size: large Metropolitan Statistical Area (MSA) counties with at least one city with 200,000 or more in total population, all remaining MSA counties plus other counties that have 25,000 or more in total population, non-MSA medium-size counties with at least 5,000 persons under the age of 22, and non-MSA small counties with at least 1,000 persons but under 5,000 persons under the age of 22. Within each of these categories the counties were sorted by the ten Health and Human Services regions, and within these regions the counties were sorted by the number of persons younger than age 22. Counties were systematically selected with probability proportionate to size. Substitute counties were selected when a county declined participation. Forty-four counties participated in the study; 41 were initially selected and three were substitutes.

Agencies provided two sampling frames for the second stage of sampling. One included all families with open cases on March 1, 1994. This was to provide point-in-time data to compare with the 1977 national study. The second frame consisted of families with cases open at some time between March 1, 1993, and February 28, 1994, but whose case was closed as of March 1, 1994. About 40 open cases and ten closed cases from each county were targeted for inclusion in the sample. The sample consisted of 2,109 families, 443 with closed records and 1,666 with open records. The total sample represented the range of families seen in child welfare, including those who enter and exit quickly and those who remain in the system for a longer period.

Data collection for the 1994 national study consisted of a one-time structured telephone interview with the caseworkers of the 2,109 families. Caseworkers referred to case records as needed to answer the questions. To help them prepare for the interview, caseworkers received cards in advance that showed the response options. The duration of most interviews was 30 to 50 minutes. Interviews were conducted in the summer of 1994. The analysis described in this article was certified as exempt from full review by an institutional review board.

Caseworkers were read a list of problems and asked whether the primary caregiver experienced that problem. Likewise, caseworkers were read a list of services and asked whether the service was offered to the primary caregiver. If the service was offered, caseworkers were asked if it had been provided. The primary caregiver was the person responsible for the daily care of the child. If the child was in foster care, then the primary caregiver was the person responsible for the child's care before foster care or the person with whom the child was expected to be reunified. The mother was considered the primary caregiver when a mother and father (or other relative) were jointly responsible for the child's care.

Analysis of the 1994 national study data required the use of weights for the estimates to accurately reflect the national population. The data set contains sample weights and replicate weights. Sample weights adjusted for nonresponses and replicate weights accounted for the nonindependence of observations resulting from the multistage clustering design. Replicate weights were developed with jackknife replication ( 26 ). WesVarPC software, version 4.0 ( 28 ), was used for the analyses.

The amount of missing data for caregiver problems ranged from 10.2% to 14.5%. Missing data mainly concerned the situation in which a primary caregiver was not identified, which was the case for 8.1% of families with children in long-term foster care. "Don't know" and "not ascertained" codes accounted for the remaining missing data. Missing data for whether services were offered ranged from 2.4% to 8.9%, and missing data for whether services were provided ranged from none to 3.5%. Percentages and frequencies were calculated for the first three research questions; chi squares tested for the fourth research question. Weighted percentages and frequencies are reported throughout.

Results

Caseworkers identified 35.0% (N= 390,897) of caregivers with mental health or substance use problems; 16.8% (N=188,214) had mental problems, and 23.3% (N=259,648) had substance use problems (alcohol or drug problems). The two groups (those with mental problems and those with substance use problems) were not exclusive. Only 4.6% (N=56,966) were identified as having both problems.

Among caregivers with mental problems, alleged child abuse and neglect was the main reason for opening a case for 72.6% (N= 136,560) of the families, whereas for 22.5% (N= 42,274) the main reason was other family problems (data missing for 5.0% of families). Among caregivers with substance use problems, 69.1% (N=179,485) of cases were for alleged abuse and neglect and 21.0% (N=54,751) for other family problems (data missing for 9.8% of families). Of caregivers with mental problems, 39.9% (N=75,087) had a child in foster care during the study period, as did 47.2% (N=122,604) of caregivers with substance use problems. (Data were missing for 8.5% of caregivers with mental problems and 6.0% of caregivers with substance use problems.)

Most (96.6%, N=181,778) of the caregivers with mental problems were birth mothers, as were 92.0% (N= 238,544) of caregivers with substance use problems. The mean±SE age of caregivers with mental problems was 33.6±.91, similar to the age of caregivers with substance use problems (32.5±.63). Of caregivers with mental problems, 74.8% (N=140,797) were white and non-Hispanic, 15.5% (N= 29,161) were black and non-Hispanic, 3.1% (N=5,896) were Hispanic, 1.5% (N=2,858) were Asian or Pacific Islander, and less than 1.0% (N=303) were American Indian or Alaskan Native. Race was coded as other, unknown, or not ascertained for the remaining 4.9%. Among caregivers with substance use problems, 57.6% (N= 149,624) were white and non-Hispanic, 34.6% (N=89.776) were black and non-Hispanic, and 4.3% (N=11,039) were Hispanic. Less than 1.0% (N= 2,944) were American Indian or Alaskan Native or Asian or Pacific Islander. The remaining 2.0% were of other or unknown race or ethnicity, or the data were not ascertained.

As shown in Table 1 , 77.9% of caregivers with mental problems were offered mental health services. Outpatient treatment was offered to 69.2%, whereas psychological assessments and inpatient treatment were offered to 30.1% and 8.1% of caregivers, respectively. Mental health services were provided to 84.0% of the caregivers who were offered them.

Enlarge table

Of caregivers with substance use problems, 58.1% were offered outpatient substance abuse treatment and 23.9% were offered inpatient treatment. Of those with substance use problems, 65.7% were offered substance use services and 67.5% of those received them ( Table 1 ).

Table 2 lists additional problems of caregivers with mental health or substance use problems. Caseworkers identified 60.1% of the caregivers as lacking parenting skills. Other frequent problems included parent-child conflict, lack of income, lack of education and job skills, and housing problems.

Enlarge table

The final research question explored associations between caseworkers' identifying problems and offering services. Caregivers with mental problems were significantly more likely than caregivers without such problems to be offered mental health services (N=1,242,539, χ² =286.25, df=1, p<.001). Still, mental health services were offered to 26.8% of caregivers not identified as having mental problems. A significant association was found between having substance use problems and being offered substance abuse treatment (N= 1,242,539, χ² =870.57, df=1, p<.001). Only 3.0% of caregivers not identified as having substance use problems were offered treatment services. Significant associations were also found between caseworkers' identifying lack of parenting skills and offering in-home parent training or parenting classes (for caregivers with mental problems, N=188,214, χ² =63.99, df= 1, p<.001; for caregivers with substance use problems, N=259,648, χ² = 71.80, df=1, p≤.001). Of caregivers with mental problems who lacked parenting skills, 67.0% (N=73,740) were offered parent training or classes, as were 69.7% (N=102,304) of caregivers with substance use problems and a lack of parenting skills. However, even when parenting skill was not an identified problem, 24.4% (N=18,022) of parents with mental problems and 32.1% (N=33,156) of parents with substance use problems were offered parent training or classes.

Of caregivers with co-occurring mental health and substance use problems, 82.2% (N=45,722) were offered mental health services, compared with 76.0% (N=97,389) with mental problems only. This was not a statistically significant difference. However, substance abuse treatment services were significantly more likely to be offered to caregivers with both problems than to caregivers with a substance use problem only (N=1,242,538, χ² =17.47, df=1, p<.001). Substance abuse treatment services were offered to 82.0% of caregivers with both problems and to 65.7% of caregivers with a substance use problem. All caregivers with both problems were offered at least one service; 80.3% (N=45,723) were offered both services and 19.7% (N=11,244) were offered either.

Discussion

We found a higher rate of service utilization among caregivers in child welfare than what is reported for the general population, where less than one-third of adults with a mental disorder receive treatment ( 29 ). Child welfare caseworkers may facilitate use of services among caregivers, serving as "gatekeepers" for adults as well as children ( 30 ). However, the findings suggest that caregivers with substance use problems may not receive needed services. Fewer caregivers with substance use problems were offered and provided services than were caregivers with mental problems. Caregivers with both substance use and mental problems were significantly more likely to be offered substance abuse services than were caregivers with substance use but not mental problems.

Findings from the National Survey of Child and Adolescent Well-Being also indicated that caregivers with substance use problems were less likely to receive services than caregivers with mental health and substance use problems ( 31 ). Findings about caregivers parallel those of Garland and her colleagues ( 32 ) on youths. They found that the extent of unmet need was greatest for youths with substance use disorders. Youths with psychiatric disorders (with or without a substance use disorder) were most likely to receive treatment. A number of reasons could explain why caregivers with substance use problems were not offered services, including a dearth of substance abuse treatment programs for women ( 33 ). Moreover, child welfare caseworkers are frequently not trained on how to intervene with parents with substance use problems ( 34 ). More research, with standardized measures of need and service use, is needed to learn whether these findings would be replicated in other studies.

The findings on matching services to identified problems are mixed. Mental health, substance use, and parenting services were more likely to be offered when caseworkers identified these problems than when they did not identify them. However, some caregivers with problems were not offered services, and some caregivers without problems were offered services. Mental health services and parent training were offered to a substantial percentage of caregivers not identified as having these problems. More research is needed to understand why child welfare caseworkers would offer mental health and parenting services to caregivers who do not seem to need them. There are many gaps in knowledge about the correlates of caregiver use of services in the child welfare system. The child welfare research agenda should include efforts to identify the factors that influence whether caregivers are offered and provided services and whether and how caregivers' service use is associated with child, parent, and family functioning.

Our study has several limitations. Information was not available on caregivers' functioning level or coping strategies or whether their mental health and substance use problems affected their parenting; all likely affected whether caseworkers offered services. It is not known what procedures were used for the assessment of mental health, substance use, and other problems, but standard procedures apparently were not used across sites. The data do not shed light on why services were not offered to some caregivers with problems or why some services offered were not provided. It may be that caregivers were already receiving services before involvement with child welfare and so were not offered services. Information was not available on the process of how caseworkers offered services or what providing a service really means. Caseworkers may have interpreted the question differently, with some counting one contact as a provided service and others discounting that contact as a provided service. The number of contacts with the service provider, the treatment duration, and other relevant information were not available. The same caseworkers who worked with the families provided the study data, which may have led to systematic bias in reporting whether services were offered and provided.

It is also not known how changes in child welfare services and policy since 1994 have affected whether services are offered and provided to caregivers. For example, 1994 legislation resulted in the implementation of child and family service reviews that include outcomes relative to appropriate service referrals and provision ( 35 ). Given how slowly change occurs in large systems, the extent of social problems that continue to exist, and the gaps identified by the child and family service reviews ( 35 ), the findings are likely still applicable to current child welfare practice.

Conclusions

Research is needed to increase knowledge about caregiver problems and whether and how service use varies for different problems. For example, caregivers with substance use problems were less likely to be offered treatment than were caregivers with both substance use and mental problems. Preliminary analysis not presented here showed that having a child placed in foster care during the study period increased the likelihood of services being offered for caregivers with substance use problems but not for caregivers with mental problems. This finding suggests that factors in addition to identified caregiver problems influenced whether they were offered services.

Prospective studies are needed to understand the process of service utilization for caregivers in child welfare. This process includes caseworker assessments, treatment recommendations and referrals, and caregivers' accessing and using services. Research that aims to understand caseworker and caregiver experiences with the service system would increase knowledge about service use barriers and facilitators. More research is needed to understand the impact of parent mental health and substance use on parenting behaviors and on children's mental health and well-being. We also need to learn more about whether and how caregiver service use is related to parenting skills, parent-child interaction, and child well-being and safety.

Acknowledgments and disclosures

Data were made available by the National Data Archive on Child Abuse and Neglect, Cornell University. Data from the National Study of Protective, Preventive, and Reunification Services Delivered to Children and their Families, 1994, were originally collected by Westat, Inc. Funding for the National Study was provided by ACF contract no. 105-90-1601. Neither the collector of the original data, the funder, the National Data Archive, Cornell University, or its agents or employees bear any responsibility for the analyses or interpretations presented here. The secondary analysis reported here was supported by grant 5K01-MH-068473 from the National Institute of Mental Health to the first author.

The authors report no competing interests.