The Future of Children

Children’s Exposure to Parental AODA

The prevalence of children’s exposure to parental AODA refers to the proportion of abused and neglected children who are affected by parental alcohol and other drug use at a given time. Estimates vary depending on the definition of AODA used to classify cases, the segment of the child population examined, and the method of data collection used to count the cases. Prevalence estimates are best generated through carefully conducted studies using uniform definitions that rely on samples of cases drawn at random or using some other statistically valid method of selection to generate an estimate within some margin of error, for example, plus or minus a few percentage points.

Because “substance abuse” is defined differently and measured more precisely by drug professionals than by ordinary folks, an important element of the estimation process is the definition of substance abuse that is used for classifying and counting. AODA is variously measured in terms of current use, lifetime use, abuse, or dependence. Current or lifetime use of illicit substances or large amounts of alcohol (often defined as four or more drinks in one day) is best measured using uniform screening questions such as those in the Composite International Diagnostic Interview-Short Form (CIDI-SF).⁴ In such diagnostic interviews, respondents are asked a series of questions such as, “In the past 12 months did you ever use… [insert name of substance]”? ⁵

Substance abuse and dependence are distinct concepts and refer to detrimental or debilitating use. They can be systematically measured with criteria specified in the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV).⁶ The manual lists seven potential dependency symptoms and suggests that dependence is indicated when at least three of the seven are present. The DSM-IV defines substance abuse in narrower terms, as a pattern of substance use that is “maladaptive” ⁷ without meeting the criteria for dependence. The manual specifies four [End Page 149] characteristic symptoms of substance abuse and specifies that at least one must be present to indicate a diagnosis of substance abuse.

The National Survey of Drug Use and Health (NSDUH; formerly known as the National Household Survey of Drug Abuse) conducts in-home surveys with probability samples of the population to estimate prevalence rates of alcohol and drug use within the past year. It uses DSM-based criteria to assess substance abuse and dependency. In 2002, the NSDUH found that among married women aged twenty-one to forty-nine living with children under the age of eighteen, 14.5 percent engaged in binge drinking and 4 percent used illicit drugs in the past month.⁸ The 2003 NSDUH found that among women aged eighteen to forty-nine, 5.5 percent abused or were dependent on alcohol or any illicit drug.⁹

These prevalence estimates suggest that between 6 million¹⁰ and 9 million¹¹ children live in households in which a caregiver abuses alcohol or drugs. These numbers far exceed the number of children who become involved in the child welfare system for any reason. Of the approximately 900,000 children with substantiated maltreatment allegations of any kind in 2005, about 300,000 (33 percent) were placed in foster care, leaving about 600,000 children with substantiated allegations at home with their parents.¹² Even if all of these substantiated cases with children in the home involved parental substance abuse, the number would conservatively reflect only about 10 percent of the estimated number of children living with a parent who abuses substances.

It is equally challenging to identify the prevalence of AODA among families already involved with the child welfare system.¹³ Just as substance abuse can be measured differently in general population studies, so can exposure to parental AODA in the child welfare population be defined and counted in a variety of ways. In the child welfare research literature, measures of AODA range from the impressions of state administrators elicited in phone surveys, to references in case files, to caregivers’ scores on standardized measures such as the CIDI-SF.¹⁴ As described below, when substance abuse is measured with standardized and validated measures, the resulting prevalence estimates tend to be lower than those of phone surveys and case records.

An added complication is that the child welfare population can also be defined in a variety of ways. The definitions range from the total number of children involved in CPS investigations to the fraction having a substantiated maltreatment report to the smaller number who are removed and placed into foster care. Prevalence rates vary not only across these different population groupings but also by geographical location and time period. Child welfare jurisdictions have different policies and norms regarding when substance abuse triggers child welfare [End Page 150] involvement, and those policies and norms change over time. Hence, even if the same child welfare subpopulations are assessed using the same substance abuse measures, prevalence rate estimates may vary depending on the specific location and time period examined.

In light of the range of possibilities, it is easy to see how specific choices of substance abuse definitions and child welfare subpopulations can affect prevalence estimates. The most reliable prevalence estimates come from studies that meet generally accepted criteria of sampling rigor and measurement precision. Studies with unspecified response rates, response rates of less than 50 percent, or those that use only impressions as an indicator of substance abuse tend to produce unreliable estimates. The best estimates derive from studies with well-defined indicators of substance abuse and clearly specified samples. The best studies will also differentiate between samples that focus on the smaller foster care subpopulation and those that focus on the larger population of abused and neglected children.

Evidence meeting the above criteria suggests that caseworkers and investigators report substance abuse in about 11 to 14 percent of investigated cases¹⁵ and in 18 to 24 percent of cases with substantiated maltreatment.¹⁶ Of the cases that are opened for in-home services following a maltreatment investigation, 24 percent screen positive for alcohol abuse or illicit drug use in the past year.¹⁷ This figure is a nationwide average. In an urban sample with no specification about timing, 56 percent of such caregivers had a notation of illicit drug or alcohol abuse in their case files or self-reported as having engaged in drug or alcohol abuse.¹⁸

The prevalence of substance abuse runs higher for children taken into foster care, with estimates meeting the above criteria ranging from 50 to 79 percent among young children removed from parental custody.¹⁹ Although few studies meeting the specified criteria have assessed the prevalence of DSM-defined substance abuse or dependency in child welfare populations, those that do suggest that 4 percent of families having contact with the child welfare system²⁰ and 16 percent of families having a child in foster care²¹ meet DSM criteria for substance abuse or dependence. Comparing reports of prevalence of substance abuse or current use to more standardized measures of drug abuse and dependency suggests that approximately one-fourth of users of alcohol and other drugs who come to the attention of CPS authorities present serious enough problems to warrant a DSM designation.

Two key generalizations may be drawn from the research about the prevalence of children’s exposure to parental AODA. First, when detection methods and measures of substance abuse are more precise, prevalence estimates tend to be lower. Prevalence rates generated from impressions (from administrators, state liaisons, or caseworkers) or from wide-ranging references in case files (such as reports of past substance abuse or a past referral to substance abuse treatment) are substantially higher than are estimates generated through individual parent assessments or professional diagnosis. A clearer picture of links between substance abuse and child maltreatment will require greater attention to definitions of substance abuse and the timing and method of assessment. Second, the prevalence of parental substance abuse is lower among children who are subjects of a CPS investigation than among those who are indicated for maltreatment and [End Page 151] substantially lower than among those placed into foster care. These distinctions are important because, as noted, only about one-third of substantiated maltreatment allegations result in out-of-home care.²² Prevalence estimates derived from a foster care subpopulation should not be generalized to the larger child welfare populations of abused and neglected children.

Does Parental AODA Place Children at Increased Risk of Maltreatment?

Selective prevention, as distinct from universal prevention,²³ refers to interventions that target groups that exhibit above-average risks, such as children exposed to parental AODA. Several studies document a link between parental AODA and child maltreatment, particularly neglect.²⁴ However, establishing a causal relationship between parental substance abuse and child maltreatment is difficult. Most investigations of the link between substance abuse and child maltreatment start with a sample of parents involved with either child welfare or substance abuse services. For example, a sample of parents who have been found to abuse substances might be assessed for child maltreatment reports and the report rate may be compared with that of the general population or a matched comparison group without substance abuse problems. Sometimes such studies factor in other potential influences on child maltreatment, such as parental mental health or education. Such studies often find higher child maltreatment rates among parents in a substance abuse group than in the comparison group or, conversely, higher substance abuse rates among parents in a child welfare services group than in a comparison group.

Using similar methods, researchers have identified an association between parental substance abuse and child maltreatment as measured by scores on a child abuse potential index,²⁵ parental self-reports,²⁶ CPS reports,²⁷ and incidents of maltreatment noted in medical records.²⁸ In a rigorous study that is among the few prospective studies to assess the risk of child maltreatment among parents who abuse substances, Mark Chaffin and several colleagues²⁹ followed for one year parents from a community sample. The researchers compared parents identified as having a substance use disorder and parents without a substance use disorder in self-reports of child maltreatment. Parents with a substance use disorder were three times more likely than those without one to report the onset of child abuse or neglect within the one-year follow-up period. About 3 percent of parents with a substance abuse problem reported child abuse or neglect within the year compared with 1 percent of parents without a substance abuse problem. The researchers found that the influence of substance abuse on maltreatment was maintained even when the parents being compared were similar with respect to such characteristics as parental depression, obsessive-compulsive disorder, household size, age, race, marital status, and socioeconomic status.

The Chaffin study is rigorous and convincing. It offers the best type of evidence for demonstrating a link between substance abuse and child maltreatment. And similar patterns are found in repeated studies that control for other co-existing risk factors. Such studies, however, cannot rule out the possibility that other co-factors associated with substance abuse, such as parental depression, social isolation, or domestic violence, are more directly responsible for higher maltreatment rates. Targeting interventions on a “spurious” association between drug use and [End Page 152] maltreatment without attending to the underlying direct causes of both will be ineffectual. For example, researchers studying the effects of crack cocaine use during pregnancy found that the deleterious consequences originally attributed to substance abuse were actually related to the environments and associated hazards in drug users’ lives.³⁰

In the Illinois experiment on “recovery coach” services in promoting drug treatment and family reunification, among parents who were identified as having a substance abuse problem and having a child placed out of the home, substance abuse was the sole problem for only 8 percent. The vast majority of the parents experienced co-existing problems with mental health, housing, or domestic violence.³¹ The best studies attempt to control for these other risk factors, but even multiple-regression and matched-sample studies are challenged to control adequately for the myriad of social, environmental, and other variables that can “confound” the association between parental substance abuse and threats to child safety. Differences attributed to substance use can also arise from other unobserved factors that affect the detection or identification of substance use, maltreatment reporting (including self-reports), and the likelihood of child welfare involvement.

The role of substance abuse in increasing risks for child maltreatment will become clearer as researchers succeed in identifying exactly what it is that explains the link between parental substance abuse and child maltreatment. Researchers have proposed a range of potential explanations. For example, substance abuse may strain social support relationships, leading to social isolation and heightening the risks that family, friends, and neighbors will refrain from lending a hand or stepping in when child-rearing problems arise.³² Substance abuse may promote impulsivity or reduce parental capacity to control anger under stressful situations.³³ Substance abuse may also distract parents from meeting children’s needs or impair their ability to supervise them.³⁴ The links between parental substance abuse and child maltreatment surely warrant further study because different causal mechanisms call for different ways to conceptualize the problem and determine how to intervene. As one example, different substances may have different consequences for parenting and child safety. The ways in which a sedative, such as alcohol, impairs parenting or threatens child safety could be quite different from the ways in which a stimulant, such as methamphetamine, impairs parenting and threatens child safety. Perhaps child safety will be promoted most effectively by specifically targeted interventions for different types of substance abuse. Likewise, different mechanisms may explain different pathways to child neglect and physical abuse, or mechanisms may differ in different social or economic contexts.

Is It Possible to Target AODA Families for Treatment?

Indicated prevention³⁵ involves screening abuse and neglect cases for signs of parental substance abuse to promote sobriety and prevent the recurrence of maltreatment. To date, usual caseworker practices have not proved effective in identifying AODA problems among families in the child welfare system or in preventing subsequent maltreatment allegations once families are investigated for child maltreatment. An analysis using data collected on families reported for child maltreatment as part of the National Survey of Child and Adolescent Well-Being (NSCAW) found that among at-home caregivers who screened positive for past-year alcohol abuse or illicit drug use, only 18 [End Page 153] percent were identified by caseworkers as having a substance abuse problem. Among at-home caregivers meeting criteria for alcohol or drug dependency, caseworkers identified a substance abuse problem for only 39 percent.³⁶ Such findings are consistent with other research indicating that child welfare caseworkers are ill-equipped to identify substance abuse problems.³⁷

When substance abuse is indicated, evidence also casts doubt that CPS is effective in linking parents to substance abuse services and treatment. A study focusing on parents with substance abuse problems involved with child welfare services found that about half received substance abuse treatment; 23 percent were offered treatment but did not receive it; and 23 percent were not offered treatment.³⁸

Shares of parents completing treatment are similarly low. An Oregon-based study found that both before and after implementation of the Adoption and Safe Families Act of 1997, about one-third of mothers involved with the child welfare system who entered substance abuse treatment completed their first treatment episode; about half completed any treatment episode within a three-year observation window.³⁹ A more recent study found that among parents with substance abuse problems and children in foster care, only 22 percent completed treatment.⁴⁰

To upgrade identification of substance abuse problems and improve treatment access for parents in the child welfare system, service organizations in both child welfare and substance abuse treatment have increasingly adopted programs or policies that encourage or mandate inter-agency collaboration. For example, child welfare caseworkers are sometimes required to involve substance abuse treatment providers in service planning, or substance abuse treatment counselors may be required to enlist child welfare caseworkers in client engagement. Nevertheless, interagency collaboration in child welfare and substance abuse treatment has proven difficult to achieve.⁴¹ Organizational policies promoting collaboration have not always been sufficient to establish widespread changes in staff collaborative practices.⁴²

One such collaborative approach is a “cooperative interagency relationship” implemented in Montgomery County, Maryland, during the late 1990s. The collaboration between county child welfare and substance abuse services involved information sharing, cross-training and internal supports, new service standards to assure quality, and new protocols and standards for assessment, referral, and follow-up. A key aspect of the effort was the co-location of a substance abuse specialist at the county’s central child welfare office. The substance abuse liaison consulted with child welfare staff on substance abuse cases, helped intervene with substance abuse cases, [End Page 154] and facilitated substance abuse referrals for child welfare clients. After three years, evaluation measures indicated that child welfare workers had increased their consultation with and involvement of substance abuse specialists in their cases.⁴³

Another intervention emphasizing interagency collaboration is the Engaging Moms Program, which promotes treatment entry and engagement among low-income mothers who used crack cocaine.⁴⁴ In one evaluation, mothers of infants were randomly assigned to Engaging Moms or to regular services. The evaluation found that mothers in Engaging Moms were more likely than those receiving regular services both to enter treatment (88 percent, as against 46 percent) and to stay in treatment for at least four weeks (67 percent, as against 38 percent). After 90 days, however, rates for the two groups had become more similar (39 percent of the Engaging Moms group were still in treatment, compared with 35 percent of the regular services group). Whether the Engaging Moms Program, which was run by university researchers, could be transferred to community practice settings is uncertain, but the evaluation illustrates the program’s promise for promoting treatment entry and short-term retention while underscoring the challenges associated with long-term treatment retention among mothers of young children.

As states and localities work to promote collaboration among child welfare and substance abuse services, evidence suggests that adopting organizational policies or rules regarding collaboration may result in uneven implementation among front-line staff.⁴⁵ Given individual influences on the implementation of organizational dictates, states and localities adopting pro-collaboration policies and programs should communicate their goals effectively and convince front-line staff of their value.

How Effective Is Substance Abuse Treatment in Preventing Maltreatment Recurrence?

Concerted efforts to link clients with treatment sometimes fall short of the goal of preventing subsequent maltreatment, either because of problems with program attendance or because of the nature of the services provided. Barbara Rittner and Cheryl Davenport Dozier⁴⁶ studied a sample of children with maltreatment allegations who either remained at home under court supervision or were placed with relatives. In about half the cases, a caregiver was mandated by the courts to attend substance abuse treatment. After rating the caregivers for treatment compliance and tracking the cases for eighteen months, the researchers found no correlation between caregivers’ treatment compliance and subsequent child maltreatment. In the researchers’ view, the findings raise questions about whether mandated treatment can prevent subsequent maltreatment and whether the treatment is of sufficient quality to help parents. Reflecting on the study findings, the researchers speculate that child welfare caseworkers may rely too heavily on indications of caregiver treatment compliance and give too little attention to family functioning and other indicators of child safety.

In an investigation with related findings,⁴⁷ researchers studied an urban sample of children following an initial CPS report of maltreatment. All the children in the sample were living in families that received public assistance. Those in families that also received Medicaid-funded substance abuse or mental health services before the first CPS report were about 50 percent more likely to [End Page 155] have a subsequent maltreatment report within seven years than were children in families that had not received the services. The study findings suggest an increased risk of maltreatment among families with substance abuse or mental health problems even when compared with other families involved with child welfare services. The findings also raise questions about the effectiveness of substance abuse and mental health services in preventing child maltreatment.

An evaluation of a treatment service program for women who used drugs during pregnancy lends support to the argument that treatment compliance, per se, may not be enough to promote child safety.⁴⁸ The evaluation found that program attendance was not related to subsequent maltreatment reports—mothers who attended more sessions were about as likely to have subsequent maltreatment reports as mothers who attended fewer sessions—but completion of treatment goals reduced chances of a subsequent report. That is, mothers who attained treatment goals were less likely than those who simply attended treatment sessions to have a subsequent maltreatment report. The authors argue that full and “genuine” engagement in treatment may be associated with child safety.

Uncertainties about whether substance abuse treatment services can prevent subsequent maltreatment are also reinforced by a series of studies using data from the National Study of Child and Adolescent Well-Being (NSCAW) involving children reported to CPS who remained at home.⁴⁹ Aware that the apparent benefits of treatment can often reflect the characteristics of the clients who access, enter, and attend treatment rather than the net effects of the services received, researchers matched caregivers according to characteristics that indicated a need for substance abuse treatment using propensity score methods. Among in-home caregivers matched on need for treatment, those who received treatment services were more likely than those who did not to incur a subsequent maltreatment report within the next eighteen months. In addition, children of the in-home caregivers who received treatment had lower well-being scores than children of caregivers who did not receive treatment. Questions raised by such perplexing findings are further discussed below.

Do Substance Abuse Interventions Promote Family Reunification?

Failure to engage parents in drug recovery services or to prevent the recurrence of maltreatment will usually precipitate the children’s removal from parental custody and placement into foster care. In these circumstances, attention turns to encouraging or compelling parents to attain sobriety or total abstinence so that the children can safely be restored to their care. The shock of child removal is thought to provide a sufficient incentive for parents to engage in treatment⁵⁰ to avoid permanent separation from their children through continued state custody or termination of parental rights.

A statewide long-term study of substance-abusing mothers in Oregon⁵¹ found that the more quickly mothers entered treatment and the more time they spent in treatment, the fewer days their children spent in foster care. Also, children of mothers who completed at least one treatment episode were more likely to be reunified with their parents than were children whose mothers did not complete treatment.

In an effort to boost reunification rates among children taken from substance-involved parents, the Illinois Department of Children [End Page 156] and Family Services secured federal permission to fund a randomized controlled trial of a state-funded enhanced services program that previous quasi-experimental findings suggested showed promise. The Illinois demonstration was initially implemented in Cook County (which includes the city of Chicago) in April 2000. The demonstration randomly assigned Illinois Performance-Based Contracting agencies to treatment and comparison conditions. Parents were referred on a rotational basis to these agencies and subsequently screened for drug abuse problems. Eligible parents assigned to the comparison condition received the standard substance abuse services. Those assigned to the treatment condition received the standard services plus a package of enhanced services coordinated by a “recovery coach.” The recovery coach worked with the parents, child welfare caseworker, and AODA treatment agency to remove barriers to drug treatment, engage the parents in services, provide outreach to re-engage the parent if necessary, and provide ongoing support to the parent and family throughout the permanency planning process.

The final results from the independent evaluation⁵² showed that assignment of a recovery coach only marginally increased parental participation in drug treatment (84 percent versus 77 percent, not significant) but that 43 percent of the treatment group managed to complete at least one level of treatment compared with 23 percent of caregivers in the comparison group. The higher rate of completion in the treatment group helped to boost the difference in reunification rates between the treatment and comparison groups by a small but statistically significant difference of 3.9 percentage points (15.5 percent versus 11.6 percent). Although this difference was compelling enough for federal officials to grant Illinois a five-year extension to expand the demonstration to downstate regions, the failure of the sizable difference in treatment completion rates to carry over to a larger difference in reunification rates prompted a closer look at some possible explanations for the shortfall.

An investigation by Jeanne Marsh and several colleagues⁵³ found that although completing at least one level of treatment helped to boost reunification rates, only 18 percent of participants in the Illinois demonstration completed all levels of treatment. Furthermore, besides substance abuse, participants faced other serious problems, such as domestic violence, housing, and mental illness. Only 8 percent of participants had no other problem besides substance abuse; 30 percent had at least one other problem; 35 percent had two other problems; and 27 percent had three or more. Parents whose only problem was substance abuse achieved a 21 percent reunification rate, while parents with one or more other problems achieved only an 11 percent rate. Reunification rates were highest among the 5 percent of participants who completed mental health treatment (41 percent) and next highest among the 10 percent of participants who solved their housing problems (12 percent). Of the 18 percent of participants who completed all levels of drug treatment, only 25 percent regained custody of their children. The authors concluded that a service integration model designed to increase access to substance abuse treatment will not successfully promote reunification unless outreach and retention services can ensure client progress in the three co-occurring problem areas as well as in completing substance abuse treatment.

In another area, preventing subsequent substance-exposed infant (SEI) reports, [End Page 157] assignment of a recovery coach was linked with a reduced likelihood of recurrence. At baseline, 69 percent of parents randomly assigned to the treatment group had previously delivered an infant reported for intrauterine substance exposure compared with 70 percent in the comparison group. After at least eighteen months of follow-up, 21 percent of parents assigned to the comparison group experienced a subsequent SEI report compared with 15 percent in the treatment group.⁵⁴ Prior SEI reports were most strongly associated with the hazards of subsequent SEI reports. Parents with prior SEI reports were seven times more likely than those without reports to experience the birth of a child reported for intrauterine substance exposure. Parents randomly assigned to the comparison group were 1.4 times more likely than those assigned to the recovery coach treatment to have a subsequent SEI report. Despite the lowered risk in the treatment group, the fact that 15 percent of mothers assigned a recovery coach experienced a subsequent SEI report further compounds the permanency planning dilemma—whether to continue investing in the uncertain outcomes of drug recovery and family reunification or to cut the process short by terminating parental rights and proceeding with adoption or other planned permanency arrangements such as legal guardianship and long-term placement with extended kin.

Click for larger view
View full resolution

Figure 1.

Birth Cohorts of Substance-Exposed Infant (SEI) Reports, Indicated Reports, Protective Custody Taken, and Foster Care Placements per Thousand Births in Illinois, Fiscal Years 1985–2007

Substance-Exposed Infants: The Case of Illinois

As noted, two decades ago Illinois became one of the first states to make the presence of illegal drugs in newborns prima facie evidence of abuse and neglect. It enacted legislation that expanded the definition of abused or neglected minor to include newborns whose blood, urine, or meconium contained any amount of a controlled substance or its metabolites. The mandate helped to fuel a rise in the number of SEI reports that peaked at 20 per thousand births in fiscal year 1994 (see figure 1). More than 90 percent of reported SEI cases were subsequently indicated for maltreatment because a positive toxicology report meets the credible evidence standard that abuse or neglect has occurred. The proportion of substance-exposed infants who were taken immediately into protective [End Page 158] custody (PC) lagged behind the steep rise in reports and hit its highest point in 1999 with 41 percent of reports triggering the state’s removal of the infant at birth. Currently the proportion of protective custodies hovers around 33 percent of SEI reports. The risk of removal, however, does not end with the child’s birth. Substance-exposed infants run a high risk of being placed in foster care throughout their early childhood.

Figure 1 also charts the foster care rates as of March 30, 2008, among successive cohorts of children born substance-exposed from fiscal years 1985 to 2007. The rate of foster care was highest among the cohort of children born in fiscal year 1994. Of the 2 percent of infants reported as substance-exposed during that year, the proportion that was later taken into foster care for any reason reached 50 percent as of March 2008. Among all birth cohorts, the removal proportion hit a high of 56 percent among children born substance-exposed during fiscal year 1999. Since that time, the proportion has stabilized at around 50 percent for recent birth cohorts.

There was some debate in Illinois over whether the drop in SEI rates after fiscal year 1994 mirrored a decline in maternal drug abuse or instead simply reflected changes in hospital surveillance practices. In Illinois, children are not universally screened at birth for substance exposure. Each hospital differs in its protocols as to what risk factors—for example, no prenatal care, past drug use, low birth weight—warrant ordering a drug test. As a result, concerns arose that publicly funded, inner-city hospitals were using protocols that resulted in more drug testing than the protocols used by privately insured, suburban hospitals, thus bringing African American infants disproportionately to the attention of CPS. For example, approximately 59 percent of Illinois infants born in 1995 were non-Hispanic whites and 20 percent were African Americans. In that same year, approximately 12 percent of SEI reports involved non-Hispanic white infants while 83 percent involved African American infants. These figures translate into a disproportionality ratio of twenty SEI reports on black infants for every one report on a white infant. The disproportionality ratio was the same when black infants were compared with Hispanic infants.

By 2002, the disproportionality ratio in Illinois had fallen to seven SEI reports on black infants for every one report on a non-Hispanic white infant. The entire decline in racial disproportionality was explainable by the 64 percent drop in black SEI rates from 65.9 per thousand births in 1995 to 23.9 per thousand births in 2002. During the same period, Hispanic SEI rates also fell by 61 percent, from 3.2 per thousand births in 1995 to 1.2 per thousand births in 2002. In contrast, SEI rates rose slightly among non-Hispanic white infants, from 3.2 to 3.5 per thousand births. While it cannot be discounted that the large SEI decline among African Americans reflected an actual drop-off in the prevalence of parental drug abuse from its epidemic levels in the early 1990s, the concomitant decline among Hispanics but not among majority whites suggests that changes in drug surveillance practices, particularly in the inner city, may have also figured in the SEI decline.

During the years when SEI reports were climbing in Illinois, child welfare advocates and drug professionals were calling for the expansion of drug treatment programs for women and children. After the fall-off in report rates, attention turned to treatment retention and the completion of services. The [End Page 159] shift in focus from program availability to service completion reflected both the aforementioned decline in SEI levels as well as new insights gained from the tighter collaboration between drug and child welfare professionals in the state.

In 1997, the independent evaluators of a joint initiative between the Illinois Department of Children and Family Services and the Illinois Department of Alcohol and Substance Abuse were forced to drop the intended “no treatment” comparison group from their quasi-experimental study because they unexpectedly discovered that nearly three-quarters of their intended control group had in reality received some kind of substance abuse treatment.⁵⁵ In their peer-reviewed article,⁵⁶ the authors instead focused on the differences between women who received regular treatment services and those who received enhanced treatment services that provided special outreach and case management services as well as transportation and child care services to lower the barriers that prevent mothers from succeeding in treatment. The results of the evaluation linked participation in the enhanced services program with lower self-reported drug use but, surprisingly, linked better access to transportation and child services with higher use. The authors concluded that clearly something else besides access to services made the enhanced service program more effective.⁵⁷

Also in 1997, early results from the Illinois Performance-Based Contracting Initiative showed providers were far less successful in achieving permanence for children by reunification than they were by adoption or guardianship.⁵⁸ Analysis of permanency outcomes showed that reunification rates were particularly low among children born substance-exposed. Of the 1,859 substance-exposed infants in fiscal year 1994 who were ever removed, less than one-fifth (18 percent) were reunified with birth parents, whereas two-thirds were adopted (65 percent) and one-tenth were taken into subsidized guardianship. Also of concern were the racial disparities in family preservation and reunification patterns. Of all SEI reports in fiscal year 1994, only 55 percent of black infants were retained in or ever returned to parental custody compared with 71 percent of non-Hispanic white infants and 73 percent of Hispanic infants.

Might Other Interventions Better Address the Risk of Child Maltreatment?

In the spring of 2008, the Chicago Tribune ran a story about a recent graduate of Morehouse College under the headline: “Proof Positive of Flawed Data.” It told the story of a Rhodes Scholarship finalist who was born substance-exposed at the start of the SEI epidemic in Chicago in 1986, “among a wave of inner-city babies exposed to crack in their mother’s womb, children written off by much of society as a lost generation doomed to failure.” ⁵⁹ The article asserted that the drug panic was fueled by flawed data that warned of neurologically damaged and socially handicapped children that would soon flood the nation’s schools and, later on, its prisons.

More recent opinion has backed away from such dire predictions. Much of the earlier work failed to consider the myriad of adverse social, environmental, and other factors that confound the association between parental substance use and impaired childhood growth and development. Barry Lester was among the first researchers to note that early studies of substance-exposed infants overestimated [End Page 160] the effects of cocaine exposure by attributing to cocaine adverse effects that were probably related to other influences such as multiple-drug use, poverty, or cigarette smoking.⁶⁰ The challenges associated with identifying specific effects of prenatal cocaine exposure, along with the wide-ranging findings of research on the topic, led a group of leading researchers, including Lester, to argue publicly that no particular set of symptoms supports the popular notion of a “crack baby” syndrome.⁶¹ They asked the media to stop using the stigmatizing term.⁶²

Recently, however, Lester has noted that some well-designed studies that control for a range of influences are identifying some apparent effects of prenatal cocaine exposure that may even increase over time.⁶³ The studies suggest that prenatal cocaine exposure may have neurological effects that become visible only when “higher level demands are placed on the child’s cognitive abilities.” ⁶⁴ Lester argues that just as it was initially a mistake to overstate the effects of prenatal cocaine exposure, it would also be a mistake to overlook potential effects that are still largely unknown and warrant further research.

A recent study in Atlanta, Georgia, helps to isolate the effects of prenatal cocaine exposure from the effects of the caregiving environment.⁶⁵ The researchers compared cocaine-exposed infants who remained with their mothers and cocaine-exposed infants placed with alternative caregivers. At two years old, despite having more risk factors at birth, the toddlers with non-parental caregivers had more positive cognitive-language and social-emotional outcomes than did the toddlers living with their parents. Outcomes for the cocaine-exposed toddlers with non-parental caregivers were even slightly more positive than for other toddlers in the study who had not been exposed to cocaine and remained with their mothers. The results underscore the importance of a nurturing caregiving environment for children’s well-being and illustrate that efforts to identify and isolate effects of prenatal cocaine exposure must account for the caregiving context.

In the absence of a definitive link between intrauterine substance exposure and developmental harm, it is difficult to justify categorizing such exposure as a form of child abuse and neglect in its own right. At the same time, it would be imprudent to back off entirely from drug screening at birth. Although some of the higher association of intrauterine substance exposure with subsequent maltreatment is clearly self-referential—that is, drug addicts are more likely to be indicated for future child maltreatment than non-addicts simply because ingestion of illicit substances during pregnancy is itself a reportable allegation—an indicated SEI report is still a useful marker of future risk.⁶⁶ SEI reports are correlated with mental illness, domestic violence, poverty, homelessness, and other disadvantages that may be more directly associated with child maltreatment. The major inadequacy with existing hospital surveillance practices is that screening is done selectively in such a way that puts African American infants at disproportionate risk of CPS detection and involvement.

Universal screening of all births for substance exposure may be one way to address the inequities in the current process, but targeting illicit substances for special attention may serve only to reify the belief that drug treatment, recovery, and abstinence mark out the best route for ensuring child safety and justifying family reunification. Attending to [End Page 161] this one visible manifestation of an underlying complex of family and personal problems can give the false impression that complying with treatment regimes and demonstrating prolonged abstinence are sufficient for deciding when to move forward with reunification plans. But the best evidence to date suggests that successful completion of drug treatment is no better a predictor of future maltreatment risk than non-completion.⁶⁷ Caseworkers and judges seem to have learned this lesson from their own experience because only one-quarter of participants who successfully completed drug treatment in the Illinois AODA demonstration were eventually reunified with their children.

Conversely, parental failures to comply with treatment plans and to demonstrate abstinence may be imperfect indicators of their capacity to parent their children at a minimally adequate level. The best evidence to date suggests that parents of substance-exposed infants pose no greater risk to the safety of their children than parents of other children taken into child protective custody.⁶⁸ Caseworkers and judges may thus want to consider implementing reunification plans some time after parents engage successfully in treatment but before they demonstrate total abstinence from future drug use. Perhaps the best course of action is to take the spotlight off of parental drug abuse and treatment completion and shine it instead on other co-factors, such as mental illness, domestic violence, and homelessness, that may be more directly implicated in causing harm to a child. A shift of attention from substance abuse to other risk factors could have the additional benefit of reducing stigma and the conflict parents may face if they fear that admitting substance abuse or asking for help with an addiction will lead to loss of child custody.

Although clearly more can be done to improve the integration of services to address the myriad of family and personal problems, such as mental illness, domestic violence, and homelessness, that, along with substance abuse, impair parenting, at some point in the intervention process attention needs to turn to the permanency needs and well-being of the child. Even though the young man profiled in the Chicago Tribune story was one of the 50 percent of substance-exposed infants who were never taken into foster care, by his own account life was not easy for him: “Mom would get drunk and hit me. I had to call the cops and send her to the drunk tank a couple of times.” ⁶⁹ Things finally turned around when his aunt, a Chicago Public Schools administrator, took him into her home at age fourteen: “My aunt’s house was a place of peace. She gave me a place that allowed me to grow. She had books everywhere, even in the bathroom.” ⁷⁰

Both personal accounts and the best research evidence indicate that finding a safe and lasting home for children born substance-exposed is critical to their healthy development and well-being. As of December 2007, however, only 39 percent of children assigned to the treatment group under the Illinois AODA demonstration had exited from foster care, compared with 36 percent in the comparison group. Not only does this small, albeit statistically significant, difference raise concerns about the advisability of heavily investing in recovery coach services, it raises additional questions about the permanency needs of the remaining 61 to 64 percent of drug-involved children who are still in foster care. Because the average age of children born substance-exposed who are removed from parental custody is less than three, it should not be too challenging to find them permanent homes with relatives either as [End Page 162] guardians or as adoptive parents or with foster parents who are willing to become their adoptive parents. Although it is unwise to set too firm guidelines, it strikes us as sensible to set a six-month timetable for parents to engage in treatment and twelve to eighteen months to show sufficient progress in all identified problem areas (presuming that both engagement and progress are determined with fair and valid measures). Thereafter, permanency plans should be expedited to place the child under the permanent guardianship of a relative caregiver or in the adoptive home of a relative, foster parent, or other suitable family. As regards the birth of another substance-exposed infant, it seems reasonable, assuming the availability of services, to initiate alternative permanency plans for all of the children unless the parent demonstrates sufficient progress in all problem areas within six months of the latest child’s birth.

In light of the difficulty of isolating the direct effects of prenatal substance abuse and the most recent evidence that some detrimental effects of intrauterine substance exposure on child development may increase over time, the newest empirical findings on the efficacy of Illinois’ recovery coach model in decreasing births of substance-exposed infants helps to bolster the case for improved treatment and service coordination regardless of whether intrauterine substance exposure is considered a form of child maltreatment in its own right. Preventing another potential risk to future child well-being, even if parental substance abuse and intrauterine substance exposure prove not to be determinative of child maltreatment directly, seems well worth the cost of investing in parental recovery from substance abuse and dependence. Such efforts, however, should not substitute for a comprehensive approach that addresses the myriad of social and economic risks to child well-being beyond the harms associated with parental substance abuse. [End Page 163]