Child Protective Service Disparities and Serious Mental Illnesses: Results From a National Survey

Data came from the Truven Health Analytics’ PULSE Survey, conducted in the United States. PULSE is the largest privately funded national survey covering a variety of health care topics. The PULSE survey involves multimodal sampling conducted via telephone (landline and cell) as well as the Internet. The survey involves contacting a geographically stratified random sample of the U.S. population every month, approximately 7,400 individuals, by phone (landline and mobile; 65%) or online (35%). The telephone and online surveys are conducted by a leading market research firm. Landline phone calls consist mainly of random-digit dialing, but a small percentage (no more than 10%) target people who responded to the previous year’s survey. Cell numbers are obtained from a multinational survey firm, cover the full set of phone exchanges assigned in the United States, and are optimized to enhance representation of results. The online survey is conducted with a prominent technology provider that sends out e-mails with links to the survey, with methods intended to achieve a sample that is nationally representative. Individuals participating in the survey may receive rewards in the form of points that may be redeemed for prizes, gift certificates, or cash. Since 2014, the PULSE survey topics have included medical tourism, the Ebola virus, birthing decisions, high-deductible health plans and their impact on utilization, vaccinations, and health care and flexible spending accounts. Their samples are representative of the overall U.S. population in terms of geography, race, ethnicity, income, marital status, employment, and education. A number of reports and news articles on various health-related topics have been published by using the Truven PULSE data.

Truven allows custom topics to be added to the PULSE survey for a fee. To conduct this study, questions were added to identify persons with a lifetime history of a serious mental illness and ask about parenting and CPS involvement. The presence of a lifetime history of a serious mental illness was based on the 1992 ADAMHA Reorganization Act (P.L. 102–321) and was defined as a “diagnosable mental, behavioral, or emotional disorder [that] has resulted in functional impairment which substantially interferes with or limits one or more major life activities.” After consultation with several psychiatric epidemiologists about brief approaches to assessing whether a serious mental illness is present, the following yes-no sequence of questions was developed and added to the Truven survey: “Have you ever been told by a psychiatrist or other mental health professional that you have major depression, bipolar disorder, manic depression, schizophrenia, or schizoaffective disorder?” “Have you ever been hospitalized for this mental health or emotional problem?” and “Has this mental health or emotional problem substantially interfered with or limited your ability to participate in any major life activities such as work, school, recreation, social activities, religious activities, family relationships, or caring for yourself?” Respondents who answered yes to all three questions were determined to have a lifetime history of a serious mental illness. Another study using this approach resulted in estimates that were comparable to those found by other methods (22).

To assess parenting status, respondents were asked, “Do you now or have you ever had any children?” Those who answered yes were then asked the following questions regarding involvement with CPS: “Has CPS ever contacted you with concerns about your parenting of your children?” If so, they were asked, “Did any contact by CPS result in your receiving in-home services?” “Did any contact with CPS result in your receiving out-of-home services?” and “Did any contact with CPS result in a change of living arrangements for your children?”

The sample of parents with serious mental illness was generated by using data from 42,761 individuals between the ages of 18 and 65 who were contacted by Truven between September 2014 and December 2015. Demographic data for the population from which the sample was obtained, including age, gender, race, education, income and marital status, can be found in a previous publication (22). A diagram outlining how the sample of parents with a serious mental illness was determined is shown in Figure 1.

The general population sample was generated by using Truven data from a single month (September 2014) because it was expected that a large sample of parents could be obtained in a single month. Of 4,661 individuals between ages 18 and 65 who were contacted that month, 3,765 (81%) responded no to the diagnosis question, meaning they would not meet the lifetime serious mental illness criteria. A diagram outlining how the general population of parents was determined is shown in Figure 2.

The proportions of individuals who were parents, who were contacted by CPS, and who had various types of CPS involvement were calculated for both groups. Chi-square tests were used to assess whether any group differences were statistically significant. PROC GENMOD in SAS was used to compute relative risks of CPS involvement and change in living arrangements; these computations were made by using a robust error variance (23, 24).

Demographic characteristics of the parenting samples are provided in Table 1. The groups did not differ significantly by gender, but parents with a serious mental illness were less likely to be white, were younger, and were less likely to be married, to have more than a high school education, and to have a household income above $50,000 at the time of the survey. There was no statistical difference between the two groups in the likelihood of being a parent (serious mental illness, N=1,682, 69%; general population, N=2,656, 71%).

Of the 1,665 parents with a serious mental illness who answered the question about CPS contact, 597 (36%) reported contact. Of the 580 parents with a serious mental illness who answered all three questions about specific CPS services, 452 (78%) received in-home services, 456 (79%) received out-of-home services, and 435 (75%) had a change in living arrangements. Of the 2,645 parents in the general population who answered the question about CPS contact, 120 (5%) reported contact. Twenty-four (20%) received in-home services, 22 (18%) received out-of-home services, and 27 (23%) had a change in living arrangements.

The rate of CPS contact was nearly eight times greater for parents with a serious mental illness compared with parents in the general population (relative risk [RR]=7.90, 95% confidence interval [CI]=6.56–9.52, z=21.76, p<0.001) and did not change substantially after the analyses controlled for gender and race (RR=7.94, 95% CI=6.60–9.57, z=21.88, p<0.001). Fathers were more likely than mothers to have CPS involvement (RR=1.53, 95% CI=1.36–1.72, z=7.05, p<0.001), but there were no significant differences in CPS involvement rates for whites and nonwhites. Within-group analyses showed that gender and race were not associated with CPS involvement in the general population, but among parents with serious mental illness, fathers (N=299, 47%) were significantly more likely than mothers to have CPS involvement (N=298, 29%) (χ²=57.50, df=1, p<.001).

Among all parents, those with a serious mental illness were much more likely to receive CPS services (RR=20.00). For parents with a serious mental illness, the risk of receiving in-home and out-of-home services was 29.8 and 32.8 times greater, respectively, compared with the risk for parents in the general population, whereas the risk of having a change in living arrangements was 25.6 times higher. No differences by race were found. Fathers were more likely than mothers to experience a change in their children’s living arrangement (RR=2.05, 95% CI=1.75–2.40, z=8.84, p<0.001), but the difference was due to outcomes among parents with serious mental illness. Among those parents, fathers were significantly more likely (N=242 of 623, 39%) than mothers (N=189 of 1,025, 18%) to have a change in living arrangements (χ²=83.54, df=1, p<0.001).

Parents without a serious mental illnesses were much more likely to receive no CPS services (RR=2.58, 95% CI=2.01–3.31, z=7.42, p<0.001). Parents with a serious mental illness were 3.8 and 4.2 times more likely to receive in-home and out-of-home services, respectively, and 3.3 times more likely to report a change in living arrangements.

After we controlled for gender and race, the adjusted RR of having a change in living arrangements following CPS contact was greater among parents in the serious mental illness group compared with parents in the general population (RR=3.18, 95% CI=2.28–4.45, z=6.79, p<0.001). Among all parents with CPS contact, fathers were more likely than mothers to have a change in living arrangements (RR=1.29, 95% CI=1.17–1.43, z=5.22, p<0.001), and whites had marginally lower rates of change in living arrangements than nonwhites (RR=.91, 95% CI=.82–1.00, z=−1.95, p=0.051). Among parents with CPS contact in the serious mental illness sample, fathers (N=242 of 289, 84%) were significantly more likely than mothers (N=189 of 291, 65%) to have a change in living arrangements (χ²=26.81, df=1, p<0.001).

The fact that parents with serious mental illnesses experience greater CPS contact and custody loss does not necessarily mean that they are more likely to abuse and neglect their children or that a mental health diagnosis is a sole or direct contributing factor. Research rarely disentangles factors that are associated with greater CPS involvement and custody loss, such as poverty and unemployment (28–30), unaffordable housing, inaccessible health care, community violence, social isolation, substance abuse, and criminal involvement (28, 30–32), which are more prevalent among persons with serious mental illness (33).

Parents with serious mental illnesses may also experience overt bias from CPS workers, judges, and clinicians that also might account for higher rates of CPS involvement. Parenting with a significant impairment, whether it is physical, sensory, cognitive, developmental, or psychiatric, places these populations at risk of CPS involvement. Overt discrimination toward parents with disabilities is well documented, including a sterilization law influenced by the eugenics movement that was applied to individuals with disabilities, “incurable” illnesses, and epilepsy as well as to criminals and orphans. This law was adapted by Virginia and upheld by the Supreme Court in Buck v. Bell, establishing the precedent that resulted in over 60,000 procedures performed on those with health impairments and disabilities (34–36). Although eugenics has since been discredited, Buck v. Bell has yet to be overturned and has been used as precedent as recently as 2001 by a federal appeals court (37). Federal and state statutes continue to limit parental rights by either allowing a mental illness as grounds to terminate parental rights in 36 states (37) or expediting the process to terminate parental rights, and some states allow for a mental illness as grounds to not provide reasonable efforts to reunify a family (38). In numerous cases, parenting competency has been questioned on the basis of presence of a parental health condition or disability (37).

More recently, following a rise in discrimination-based complaints from parents with disabilities, the Department of Health and Human Services (DHHS) and Department of Justice (DOJ) released a joint technical assistance statement in which they reported, “In the course of their civil rights enforcement activities, OCR [DHHS Office for Civil Rights] and DOJ have found that child welfare agencies and courts vary in the extent to which they have implemented policies, practices, and procedures to prevent discrimination against parents and prospective parents with disabilities in the child welfare system” (39). They went on to offer technical assistance for state and local child welfare agencies and the courts regarding requirements of Title II of the Americans with Disabilities Act (ADA) and Section 504 of the Rehabilitation Act.

Prejudice and discrimination toward parents with a serious mental illness are likely to occur throughout the child welfare system, and more work is needed to help child welfare and courts do a better job of supporting the rights of parents with disabilities. As a result of such biases, it is not surprising that parents with serious mental illnesses may be less likely to seek assistance or support from others because it increases their engagement with mandated reporters. Increased surveillance by mandated reporters likely raises the probability of both increased CPS contacts and loss of parental custody experienced by parents with serious mental illnesses through two mechanisms (20, 40). First, increased surveillance leads to greater detection of actual abuse or neglect among parents with serious mental illnesses, but lower surveillance of parents without a serious mental illness results in less detection of actual abuse and neglect that they are involved in, creating a disparity. Second, increased surveillance combined with prejudicial beliefs about the parenting abilities of individuals with a serious mental illness may lead to greater false positives in reporting that, when combined with biases present elsewhere in the child welfare system, may also contribute to the disparities. The end result is that parents with serious mental illnesses may be less inclined to engage with any type of services, including mental health services, as a result of fears about the effects of increased surveillance.

This is problematic because effective parenting programs and supports within a mental health service context exist (41, 42), including online programs (43), and should be utilized. Supported parenting programs may meet the needs and concerns of parents, improve self-esteem and motivation for treatment, and enhance parenting abilities and experiences (27). Adapting existing mental health services to better address the role of parenting and support parents and families would be more cost-effective than a reliance on child welfare systems, given that CPS involvement is associated with higher economic costs and poorer developmental outcomes (12).

Still, these parenting programs are few and far between, and findings suggest that most practitioners are poorly equipped to address issues related to parenting (1, 44). Mental health professionals rarely attend to whether or not someone is a parent and do not inquire about CPS involvement, plausibly because they are not adequately prepared in their professional training to focus on parenting issues as part of treatment or familial reunification. Furthermore, evidence points to poor communication and collaborative treatment planning between mental health practitioners and CPS workers, leading to fragmented service delivery (7).

Additional work is needed to inform other pertinent `systems about the issues surrounding parenting, mental illness, and CPS involvement. Increased opportunities for education and training in these topics are particularly important among professionals who work in child welfare, legal systems, and courts as well as attorneys and educators. These efforts should focus on increasing awareness and understanding of the prejudice pertaining to parents with a serious mental illness and developing and sharing strategies for working with the unique challenges faced by these families.

Surveys tend to have an overrepresentation of women and more educated respondents (45), as was the case here. Our general estimates of CPS involvement for those with a serious mental illness may, therefore, be an underestimate, given our finding that men with a serious mental illness may experience more CPS involvement.

The presence of a serious mental illness was assessed without a structured diagnostic interview. This approach produced estimates of serious mental illness that were very comparable to previous and concurrent findings (22). There was limited opportunity to use demographic characteristics to control for age, education, income (i.e., poverty), employment, family size, child characteristics, and other factors known to be associated with CPS involvement. Comparing groups that had CPS contact somewhat controlled for these factors. Embarrassment and other social desirability factors may have led to an underreporting of CPS involvement and types of involvement, although this would have likely affected both populations, suggesting that our RR estimates would be minimally affected. Participants’ recall of types of CPS involvement or understanding of the various types of involvement about which they were asked may also produce some error, but again, these factors would not affect RR estimates.