Thursday, January 3, 2013

Prevention and the Child Protection System

----------------------- Page 1-----------------------


Prevention and the Child Protection System

Jane Waldfogel


The nation’s child protection system (CPS) has historically focused on preventing maltreatment

in high-risk families, whose children have already been maltreated. But, as Jane Waldfogel

explains, it has also begun developing prevention procedures for children at lower risk—those

who are referred to CPS but whose cases do not meet the criteria for ongoing services.

Preventive services delivered by CPS to high-risk families, says Waldfogel, typically include

case management and supervision. The families may also receive one or more other preventive

services, including individual and family counseling, respite care, parenting education, hous-

ing assistance, substance abuse treatment, child care, and home visits. Researchers generally

find little evidence, however, that these services reduce the risk of subsequent maltreatment,

although there is some promising evidence on the role of child care. Many families receive few

services beyond periodic visits by usually overburdened caseworkers, and the services they do

receive are often poor in quality.

Preventive services for lower-risk families often focus on increasing parents’ understanding

of the developmental stages of childhood and on improving their child-rearing competencies.

The evidence base on the effectiveness of these services remains thin. Most research focuses

on home-visiting and parent education programs. Studies of home visiting have provided some

promising evidence. Little is as yet known about the effects of parent education.

Waldfogel concludes that researchers have much more to learn about what services CPS agen-

cies should expand to do a better job of preventing maltreatment. Some families, especially

those with mental health, substance abuse, and domestic violence problems, are at especially

high risk, which suggests that more effective treatment services for such parents could help.

Very young children, too, are at high risk, suggesting a potentially important role for child

care—one area where the evidence base is reasonably strong in pointing to a potential preven-

tive role. Although preventive services for the lower-risk cases not open for services with CPS

are much more widespread today than in the past, analysts must explore what CPS agencies can

do in this area too to ensure that they are delivering effective services.


Jane Waldfogel is a professor of social work and public affairs at the Columbia University School of Social Work.

                                                                            VOL. 19 / NO. 2 / FALL 2009     195

----------------------- Page 2-----------------------

Jane Waldfogel

                  very state in the United States         to be screened in, substantiated, or kept open

                  has a public child protection           for ongoing protective services with CPS but

                  system (commonly known                  whose children nevertheless are at risk of

                  by the acronym CPS) that                becoming victims of abuse or neglect. Such


                  receives and responds to                services may be delivered by the CPS agency

reports of child abuse and neglect. Funding               (with the case kept open on a voluntary or

for CPS agencies comes from federal, state,               preventive basis) but are more commonly

and sometimes county or local sources.                    delivered by community-based agencies.

Although these state systems vary consider-               Indeed, since the reauthorization of the

ably, they do share some common elements.                 federal Child Abuse Prevention and

In particular, all CPS agencies have staff and            Treatment Act (CAPTA) in 2003, CPS

procedures in place to respond to reports of              agencies have been required to develop

suspected child abuse and neglect, with some              procedures to refer children in lower-risk

agencies also accepting other types of refer-             families to community-based agencies or

rals or applications for services. Although               voluntary preventive services.

CPS agencies work in partnership with other

state agencies as well as community-based                 In this article I examine the effectiveness of

agencies, some core functions—in particular,              both types of prevention efforts. For those

receiving and responding to reports of abuse              focusing on families whose cases are opened

or neglect—are carried out mainly by CPS                  for ongoing services with CPS, I describe the

agency staff, while other functions—such as               services provided, explore their effectiveness

services for families or foster or group care—            in preventing repeat maltreatment, and ask

may be contracted out or purchased from                   whether other approaches might do a better

other agencies.                                           job. For efforts focused on lower-risk families

                                                          whose cases are not opened or kept open

Historically, the child protection system                 for services by CPS, I consider what types

has focused most of its limited resources                 of services are provided and to what types of

on preventing maltreatment and promoting                  families, how widespread the services are,

permanency and well-being among children                  how the services are funded and delivered,

who are identified as having already been the             and how effective they are in preventing

victims of abuse or neglect. A sizable share              maltreatment. I conclude with suggestions

(more than a third) of families who come to               for further research and policy.

the attention of CPS are screened out at the

time of the initial referral, while others have           Prevention Efforts for Cases

their cases closed after an investigation. The            Opened for Ongoing Services

cases that receive services from CPS on an                with CPS

ongoing basis constitute a minority of those              Figure 1 illustrates the flow of families (and

referred—a minority made up of families                   children) into the CPS system, using data

who are judged to be at highest risk.                     from the most recent report on child mal-

                                                          treatment issued by the U.S. Department of

States and localities, however, also invest                                                           1

                                                          Health and Human Services (DHHS).  Of

some resources into services to prevent                   the 6 million children (representing some 3.3

maltreatment among lower-risk families—                   million families) reported to CPS agencies

families whose cases do not meet the criteria             nationwide in 2006, about 60 percent were


----------------------- Page 3-----------------------

                                                                                Prevention and the Child Protection System

Figure 1. Pathways for Children Reported to CPS in 2006

                                                       Cases of children reported to CPS

                                                                       (6 million)

                                                  60 percent                                 40 percent

                                                 screened in                                screened out

                                                 (3.5 million)                               (2.4 million)

                                  30 percent                  70 percent

                                 substantiated             not substantiated

                                  (1.0 million)               (2.5 million)

                      60 percent              40 percent not                    30 percent               70 percent not

                   open for services         open for services               open for services          open for services

                      (600,000)                  (400,000)                       (750,000)                (1.75 million)

     37 percent placed              63 percent                 13 percent placed              87 percent

       out-of-home                   in-home                      out-of-home                  in-home

         (220,000)                   (380,000)                      (100,000)                 (650,000)

Source: U.S. Department of Health and Human Services, Child Maltreatment 2006 (Washington: U.S. Government Printing Office, 2008).

screened in for investigation or assessment                       of children—roughly 650,000—was provided

and about 30 percent of those cases (roughly                      with in-home services by CPS as a result of

20 percent of the families originally reported)                   their cases having been reported and investi-

were ultimately substantiated for abuse or                        gated but not substantiated by CPS (again,

neglect. The majority of families whose cases                     that number excludes children whose cases

are substantiated (about 60 percent in 2006)                      were opened for services before 2006). At

go on to receive post-investigation services,                     first glance it may seem surprising that more

whose main focus is on preventing further                         unsubstantiated than substantiated cases

maltreatment, whether the family remains                          were kept open for in-home services. But so

intact (about two-thirds of cases) or the child                   many more cases are unsubstantiated than

is placed out-of-home with kin, in foster care,                   are substantiated that even though the

or in group care (just over a third of cases).                    unsubstantiated cases receive services at a

                                                                  lower rate, the total number receiving

As figure 1 shows, some 380,000 children                          services is larger. It is also important to note

were provided with in-home services in 2006                       that some children whose cases are not

as a result of their cases having been                            substantiated have in fact been maltreated.

reported, investigated, and substantiated by                      Following the differential response systems

CPS that year (that number excludes chil-                         put in place over the past decade by many

dren whose cases were opened for services                         states, some CPS agencies now provide a

before 2006 and who continue to receive                           family “assessment,” in place of an investiga-

services from CPS). An even larger number                         tion, for low- and moderate-risk cases. In

                                                                                      VOL. 19 / NO. 2 / FALL 2009         197

----------------------- Page 4-----------------------

Jane Waldfogel

these assessments the focus is on developing                one-third of the children were re-reported

a service plan for the family, rather than                  within five years. Children who received

identifying a perpetrator and producing a                   post-investigation services were more likely

substantiation decision.2                                   to be re-reported than those who did not

                                                            receive services. This finding applied alike to

The services delivered to intact families                   children whose cases had and had not been

typically include case management and                       substantiated (and in fact was more pro-

supervision by a CPS worker (or perhaps a                   nounced for those who had not been substan-

worker from an agency under contract with                                     5

                                                            tiated initially).  Similarly, analyses of data

CPS), often supplemented by one or more                     on roughly 3,000 children from the National

other preventive services. The specific                     Survey of Child and Adolescent Well-Being

services delivered to any given family depend               (NSCAW), a nationally representative

on the family’s assessed need, the willingness              sample of children reported to CPS, find

of family members to engage in and accept                   that nearly a quarter of the children whose

particular services, and the availability of                cases were opened for in-home services were

services in their area. According to DHHS,                  re-reported within eighteen months, and that

post-investigation services may include                     children were more likely to be re-reported if

“individual counseling, case management,                    their families received parenting services.6

family-based services (services provided to

the entire family such as counseling or family              Such findings are the opposite of what one

support), [and other] in-home services” as                  would expect if post-investigation services

well as “foster care services, and court                    were effective at preventing maltreatment.

services.” Intact families may also receive                 But the findings may be misleading for

what DHHS categorizes as preventive                         several reasons. One problem is selection

services, which may include “respite care,                  bias. If CPS systems are operating efficiently,

parenting education, housing assistance,                    the families who receive services should be

substance abuse treatment, daycare, home                    the ones whose children are at highest risk

visits, individual and family counseling, and               of maltreatment and hence whose cases

home maker help.” 3                                         are at highest risk of being re-reported or

                                                            re-substantiated. Estimates that do not take

Researchers know remarkably little about                    selection bias into account may erroneously

how effective post-investigation and preven-                interpret a recurrence of maltreatment after

tive services are in stopping maltreatment                  service receipt as an effect of service receipt.

among the families whose cases are opened                   Another potential source of bias is the “sur-

for services with CPS. Although a few studies                                   7

                                                            veillance effect.”   Clients whose cases are

have found that maltreatment is less likely to              opened for services may be at higher risk

recur in open cases that receive services than              of being reported because they have more

in those that do not, most studies find that,               frequent contact with CPS workers and ser-

if anything, families that receive services are             vice providers rather than because they have

more likely to be re-reported and substanti-                higher levels of maltreatment.


ated subsequently.  For example, analyses

of data on 1.4 million children from nine                   Because existing research is not designed to

states from the National Child Abuse and                    address these two potential sources of bias,

Neglect Data System (NCANDS) find that                      it is not possible to conclude that the links it


----------------------- Page 5-----------------------

                                                                          Prevention and the Child Protection System

finds between service delivery and height-                   suggesting that developing and delivering

ened risk of reporting or substantiation are                 more effective treatment services for such

causal. But neither does the research provide               parents (as discussed in other articles in this

much evidence that services provided by CPS                 volume) could help prevent further

reduce the risk of subsequent maltreatment.                  maltreatment.10

Why are CPS services for families in open                   Young children are also at high risk for

cases not more effective in promoting child                  repeated maltreatment. For example, both

safety and preventing future maltreatment?                   the NCANDS and NSCAW studies discussed

Recent analyses of data from the National                    above found that the risk of re-reporting was

Survey of Child and Adolescent Well-Being                   highest for the youngest children (in par-

(NSCAW) and its companion survey, Caring                     ticular, infants and toddlers) and decreased

for Children in Child Welfare (CCCW),                        sharply with age. That pattern suggests a

provide some clues. One possible explanation                potentially important role for services such as

is that many families receive few services                   child care. Although research on how child

beyond periodic visits by usually over-burdened              care functions within CPS is limited, the


caseworkers.  Another possible explanation is               broader evidence base on child care suggests

that services are poor in quality and insuffi-

                                                             that it could be important in reducing the risk

cient in quantity. For example, although                     of maltreatment.

rigorous research has proved several parent

training programs effective, fewer than half of

                                                             Child care has long been a core service

families whose cases are opened for services

                                                            provided to open CPS cases with the explicit

receive any parent training at all. Those who                                                                   11

                                                            intent of helping to prevent maltreatment.

do get training typically receive only fifteen or

                                                            The Alaska CPS agency, for instance, explains

fewer hours of training from a program that

                                                             that “protective day care services provide day

has not been proven effective. Nor is the

                                                             care to children of families where the chil-

training they receive monitored to ensure that

                                            9                dren are at risk of being abused or neglected.

it is being implemented as intended.

                                                            The services are designed to lessen that risk

Given the poor overall track record of today’s              by providing child care relief, offering

preventive services, the question arises                     support to both the child and parents,

whether other types of services are or could                 monitoring for occurring and reoccurring

be more effective in reducing the risk of                    maltreatment, and providing role models to

maltreatment. To date, however, evidence on                  families.”12 Such care is also expected to

that question is quite limited.                              enhance the development of children who

                                                             might otherwise be at risk for poor outcomes.

One indirect way to answer the question is to               The Illinois CPS agency, for instance, says:

extrapolate from the characteristics of                      “Day care services are provided to high-risk

families whose children are known to be at                   families whose children are in open … cases;

high risk of recurring maltreatment. For                     they are used to prevent and reduce parental

instance, studies have found that families in                stress that may lead to child abuse or neglect.

which parents have substance abuse, domes-                  The services also help children to develop

tic violence, or mental health problems are                 properly and enable families to remain

more likely than others to be re-reported,                   together.” 13

                                                                               VOL. 19 / NO. 2 / FALL 2009      199

----------------------- Page 6-----------------------

Jane Waldfogel

The developmental benefits of child care                  based programs but also fell (5 percent)

are well documented. High-quality care                    among those in home-based programs.

has been shown to improve the cognitive

development of disadvantaged children and                 Similarly, a random-assignment study of the

may also improve their social functioning.14              Infant Health and Development Program

Researchers have not yet conducted formal                 (IHDP), an early child care program for

evaluations of whether child care prevents                low-birth-weight children, found reduced

maltreatment among families whose cases                   spanking by mothers in the previous week,

are open with CPS.15 But studies of Head                  although the effect was confined to boys.19

Start and other child care programs sug-

gest that child care services can help reduce

maltreatment.                                             Studies of Head Start and

Head Start, a compensatory early education                other child care programs

program for low-income children, has been                 suggest that child care

in operation since 1965 and now serves nearly

1 million preschool-aged children annually                services can help reduce

(including about 62,000 children under age                maltreatment.

three in the Early Head Start program, begun

in 1994).16 Head Start was recently the

subject of a randomized study that evaluated,             Also suggestive of a potentially protective role

among other outcomes, its effect on parent-               of Head Start and other formal child care is

ing and discipline. The findings indicated that           evidence from an observational study of chil-

parents of three-year-olds who had been                   dren from the Early Childhood Longitudinal

randomly assigned to Head Start were less                 Study-Kindergarten (ECLS-K) cohort, a large

likely than control group parents to report               nationally representative sample of children

spanking their child in the previous week and             entering kindergarten in the fall of 1998.20 In

also reported spanking less frequently, with              that study, parents of disadvantaged children

particularly pronounced effects for teen                  who had attended Head Start before kinder-

mothers (though there were no significant                 garten were more likely to report that they

effects for parents of four-year-olds).17                 never used spanking, and also reported less

Although using spanking as a marker for                   domestic violence in their home, than parents

potential child maltreatment requires                     of children who had not attended child care.

caution, these findings are nevertheless                  Parents whose children had attended Head

promising.                                                Start or other center-based child care were

                                                          also more likely to say they would not use

Another randomized study found that Early                 spanking in a hypothetical situation. The

Head Start improved parenting and reduced                 study’s authors speculated that having a child

spanking by both mothers and fathers.18                   attend Head Start or other center-based child

Parents of children assigned to Early Head                care may have reduced parents’ use of physi-

Start were less likely than control group par-            cal discipline by relieving parental stress,

ents to have spanked their child in the previ-            by exposing parents to alternative forms of

ous week. The share of mothers spanking fell              discipline, and by making the children more

most (10 percent) among children in center-               visible to potential reporters (for example,


----------------------- Page 7-----------------------

                                                                         Prevention and the Child Protection System

child care providers) who would be aware if                 reported that only 2 percent of caregivers

they were being maltreated.                                 were uncooperative at initial contact, as com-

                                                            pared with 44 percent of those in investiga-

As noted, measuring the effects of child                    tion track cases.

care on spanking is not the same as measur-

ing its effects on maltreatment. One quasi-                 Minnesota is exceptional in that funding from

experimental evaluation of the Chicago                      the McKnight Foundation allowed it to

Child-Parent Centers, however, addresses                    expand services to low-risk families. Families

maltreatment directly. The study found that                 receiving the alternative response were more

children in the program, which provides care                likely to have their cases opened for services

to children from disadvantaged neighbor-                    (36 percent vs. 15 percent). They were more

hoods during the two years before kinder-                   likely to receive not only the types of services,

garten, had only half as many court petitions               such as counseling, that are traditionally

related to maltreatment as did children in                  prescribed and paid for by CPS, but also

similar neighborhoods that did not have the                 services, such as assistance with employment,


program.                                                    welfare programs, and child care, from other

                                                            community resources not funded by CPS.

Another potentially promising approach to

prevention is “differential response,” which,               At the one-year follow-up, families in

as noted, entails greater CPS flexibility in                Minnesota’s alternative response group

responding to allegations of abuse. States are              reported less financial stress and stress associ-

increasingly coming to believe that they can                ated with relationships with other adults, as

effect more lasting change in lower-risk cases              well as fewer problems with drug abuse and

by providing services that are engaging for                 less domestic violence. Effects on other out-

families and attentive to their needs rather                comes for the children and families, however,

than by using a more traditional adversarial                were few.

investigative response.22 What does the evi-

dence show?                                                 It should be noted that the study does not

                                                            establish which of the Minnesota results

A recent review of the as-yet limited research              were due to the added funding. Most states

base suggests the promise of a differential                 using differential response have not had

response approach in preventing future                      extra resources. And the reforms in those

maltreatment.23 The strongest evidence                      other states, while yielding some promising

comes from a random-assignment study in                     evidence, have not been subject to a random-

Minnesota that found that cases assigned to                 assignment evaluation.

the alternative response track were less likely

to be re-reported subsequently than cases                   In addition to altering service delivery for

assigned to the investigative track, a finding              cases opened with CPS, differential response

that was linked to the alternative response                 reforms also increase the likelihood that CPS

track’s provision of increased services to fami-            will refer to community-based agencies the

lies.24 The evaluation and an accompanying                  cases that are not opened. An explicit part of

process study provided many indications that                the alternative assessment approach is

families were more engaged. For example,                    working with families to identify their service

workers delivering an alternative response                  needs and to make appropriate referrals.

                                                                              VOL. 19 / NO. 2 / FALL 2009      201

----------------------- Page 8-----------------------

Jane Waldfogel

Some differential response models also                     referrals” because they do not need to be

explicitly set out a preventive track for reports          referred by CPS to be served and funded).

that should be handled by community-based

agencies instead of CPS right from the outset.             The federal Department of Health and

A further impetus to such referrals was the                Human Services, in its annual report on child

2003 Child Abuse Prevention and Treatment                  maltreatment, distinguishes between children

Act (CAPTA) requirement that states develop                receiving preventive services and those

the ability to refer children who are not at               receiving post-investigative services. The

imminent risk of harm to community organi-                 distinction perhaps suggests that their data on

zations or voluntary child protective services.            children receiving preventive services mainly

Both differential response and the new                     capture children from the above groups—

CAPTA requirement, then, are likely to have                children receiving preventive services funded

increased the number of lower-risk families               by CPS even though their cases are not open

receiving some kind of preventive services                 for services with CPS (while post-investigative

from community-based agencies, without                     services would refer to children whose cases

being open for services with CPS. I turn to               were substantiated and kept open for ser-

this group of families next.                              vices). In 2006, state CPS agencies reported a

                                                           total of 3.8 million children receiving preven-

Prevention Efforts for Lower-Risk                          tive services.25 Some of these children were

Families Not Opened or Kept                                referred to CPS in 2006; others were referred

Open for Services with CPS                                 earlier; and still others were served without

Figure 1 highlights (in italics) three groups of           having been referred to CPS at all (the

children in lower-risk cases not opened or                 so-called “open referrals”).

kept open for services with CPS. The three

groups are: the 2.4 million children annually             According to DHHS, preventive services

reported to CPS but screened out; the                      “are designed to increase parents’ and other

roughly 1.75 million children annually whose               caregivers’ understanding of the developmen-

cases are reported to CPS and screened in                  tal stages of childhood and to improve their

but not substantiated and not kept open for                child-rearing competencies.” As noted, exam-

services with CPS; and the roughly 400,000                ples of preventive services include “respite

children annually whose cases are substanti-               care, parenting education, housing assistance,

ated but not kept open for services with CPS.              substance abuse treatment, daycare, home

Some of these children receive preventive                 visits, individual and family counseling, and

services from community-based agencies                     home maker help.” 26

(which may or may not be funded by CPS),

but data are not available on precisely how                Funding for preventive services for lower-risk

many children from each group do so.                       cases comes from several different sources.27

Another group—not shown in the figure—                     The most common source reported by states

that receives preventive services from                     in 2006—covering nearly 30 percent of

community-based agencies consists of                       children receiving preventive services

children who are not reported to CPS but                   nationwide—was Promoting Safe and Stable

whose families apply voluntarily or are                    Families funding under Title IV-B of the

advised to do so by someone in the commu-                  Social Security Act. The second most com-

nity (these cases are sometimes called “open               mon source—covering nearly 20 percent


----------------------- Page 9-----------------------

                                                                           Prevention and the Child Protection System

Table 1. Federal Funding for Preventive Services for Children Whose Cases Are Not Open with CPS,


Source                                                                                         Amount

Promoting Safe and Stable Families (Title IV-B of the Social Security Act)                     $250 million

Social Services Block Grant (Title XX of the Social Security Act)                              $340 million

Community-Based Child Abuse Prevention (Title II of the Child Abuse Prevention and Treatment Act) $  42 million

Source: Author’s calculations based on data in 2004 and 2008 Green Book.

nationally—was the Social Services Block                      of the $1.7 billion allocated in 2006, or about

Grant (SSBG) under Title XX of the Social                     $340 million, was devoted to preventive ser-

Security Act. Community-Based Child Abuse                     vices (about 13 percent was devoted to child

Prevention (CBCAP) grants under Title II of                   welfare services other than foster care, with

the Child Abuse Prevention and Treatment                      another 8 percent devoted to child care).31

Act (CAPTA) covered roughly 15 percent,                       With regard to the CBCAP program, here we

while funds from the Basic State Grant under                  can assume that most (if not all) of the total

Title I of CAPTA covered just over 5 percent.                 $42 million available in 2006 went to preven-

Other federal or state programs funded the                    tive services, because that is the main focus

remaining 30 percent of preventive services                   of the program. (These estimates are summa-

for children.28 States vary considerably in the               rized in table 1.)

funding sources they use. New York, for

example, relied on SSBG funding for 85                        Little information is available about spending

percent of its preventive services in 2006,                   on specific types of preventive service

while Texas relied exclusively on Promoting                   programs, such as respite care and parent

Safe and Stable Families funding.                             education. One exception is home-visiting

                                                              programs, which have been a subject of

DHHS does not track total dollars spent                       increased interest in Congress and which

on these preventive services for lower-risk                   received an additional $10 million in federal

families, but it is possible to create some                   funding in 2008, under an initiative designed

rough estimates using other data.29 Thus,                     to expand support for empirically validated

of the $410 million appropriated in 2006                      models of home visiting such as the Nurse-

for the Promoting Safe and Stable Families                    Family Partnership.32

program (the single largest source of funding

for preventive services nationally, as noted), a              The above data on spending for prevention

reasonable estimate is that about 60 percent,                 refer only to federal funding and do not

or roughly $250 million, went for preven-                     include funding from state and local sources.

tive services such as family support and                      Federal dollars represent only half the funds

prevention and family preservation (with the                  spent on overall child welfare services and a

remainder going for other services such as                    much smaller share of funding for preventive

reunification and adoption planning).30 With                  services, which are more likely than other

regard to the SSBG (the second largest fund-                  types of child welfare services to rely on state

ing source for preventive services nationally),               and local funding.33 In 2004, states spent a

program data indicate that roughly one-fifth                  total of $9 billion on child welfare services,

                                                                                VOL. 19 / NO. 2 / FALL 2009       203

----------------------- Page 10-----------------------

Jane Waldfogel

while localities spent at least $2.5 billion.34           significant effects on protective and risk

Most of these state and local dollars, however,           factors for child abuse and neglect. Two other

went for services such as foster care, with               programs were reported to be effective,

only a small portion going for preventive                 although they lacked a random-assignment

services.                                                 evaluation. Both deliver augmented parent-

                                                          ing and family support services in child care

Although prevention programs have                         settings. One is the Circle of Security parent-

expanded rapidly and now exist in all fifty               ing program in Head Start and Early Head

states, researchers still know little about their         Start in Spokane, Washington; the other is

effectiveness. In 2003, a review conducted                the Families and Centers Empowered

by DHHS noted that most of the research                   Together (FACET) family support program

focused on just two types of prevention                   in child care centers in high-risk neighbor-

programs—home visiting and parent educa-                  hoods in Wilmington, Delaware. Given the

tion.35 The evidence base on home visiting                promising evidence on the role of child care

programs, as discussed in other articles in this          in preventing maltreatment reviewed above,

volume, is promising. Although not all home               these programs—which explicitly aim to

visiting programs have been demonstrated                  increase the protective role of child care

to be effective, randomized evaluations of                settings—are potentially promising and worth

the Nurse-Family Partnership program have                 close attention.

found decreased rates of child maltreat-

ment among the group randomly assigned

to receive home visits. Regarding parent

education programs, perhaps the most com-                 Although prevention

monly provided type of prevention services,               programs have expanded

the DHHS review concluded: “The record

is neither rich nor, on the whole, particularly           rapidly and now exist in

compelling. However, a few studies have                   all fifty states, researchers

demonstrated positive findings. Many of the

existing studies in this area rely on outcomes            still know little about their

that do not include actual maltreatment                   effectiveness.

reports, but focus on short-term gains in

knowledge, skills, or abilities. Thus, taken as

a whole, little is known about the impact of

these programs on child maltreatment in the               The DHHS review also highlights two

long term.” 36                                            essential characteristics of effective preven-

                                                          tion programs—of whatever type. The first is

When the same DHHS review invited                         that the program be delivered in sufficient

nominations for effective programs, only                  dosage. In the prevention area, as in other

one—the University of Maryland’s Family                   areas of social policy, successful programs are

Connections program for at-risk families with             often implemented with less intensity or for a

children aged five to eleven—met their two                shorter time than the original model specifies,

standards for effectiveness: having been                  thus diluting the effectiveness of the program

evaluated by a study using a random-                      and leading to disappointing results. The

assignment design and having demonstrated                 second essential characteristic is the ability of


----------------------- Page 11-----------------------

                                                                        Prevention and the Child Protection System

frontline staff to engage with families to                 existing research sheds little light on what

encourage them to agree to participate in                  types of services might be most effective in

services and to continue participating. But                meeting that goal. As other analysts have

engaging families is also extremely difficult              noted, CPS agencies provide “a somewhat

because many of the target families are                    haphazard set of services that aim to help

socially isolated and may distrust helping                 abusive families and their children … [with]

professionals, however well-intentioned.                   a shortage of effective intervention programs

Thus, recruiting and training effective                    to provide needed services [and] a dearth of

prevention staff is a common challenge.                    prevention services.”38

Looking Ahead: Suggestions for                             Program data—and common sense—suggest

Further Research and Policy                                that any intervention that aims to prevent

It is now widely accepted that CPS has an                  maltreatment must be intensive, and its

important role to play in preventing maltreat-             frontline staff must be able to engage with

ment not just among the relatively high-risk               families. But beyond that, researchers have

cases opened for services, but also among the              much more to learn about what types of

lower-risk families who come to its attention              services should be expanded if CPS agencies

but do not meet the thresholds for case open-              are to do a better job of preventing maltreat-

ing or continuing service delivery. Failing to             ment among their open cases. The demo-

prevent maltreatment among open cases is a                 graphics of recurrence suggest that some

signal that CPS intervention has failed in its             families, especially those with mental health,

primary role of promoting child safety and                 substance abuse, and domestic violence prob-

well-being among the most vulnerable group                 lems, are at higher risk than others, pointing

of children. And failing to refer lower-risk               to issues that services will need to address

families for effective preventive services                 effectively if they are to reduce the risk of

represents a missed opportunity to intervene               maltreatment. The demographics of recur-

before the risk of maltreatment escalates into             rence also point to young children as being

full-blown abuse or neglect, saving children               particularly at risk, suggesting a potentially

needless suffering while also saving CPS                   important role for such services as child care.

and other agencies the costs that would be                  Indeed, child care is one area where the

entailed by a subsequent report, investiga-                evidence base is reasonably strong in pointing

tion, and ongoing service delivery.                        to a potential preventive role. This is certainly

                                                           an area where further experimentation would

How well are CPS agencies doing at pre-                    be worthwhile.

vention? We know from the federal Child

and Family Services Reviews that in 2005,                  With regard to the lower-risk cases not

6.6 percent of open CPS cases nationally                   open for services with CPS but referred to

experienced a new incident of substantiated                preventive services, the good news is that

maltreatment within six months of being                    such services seem to be much more wide-

opened.37 That rate, although somewhat                     spread today than in the past, reflecting the

lower than it was a few years previously, still            expanded availability of federal and other

exceeds the 6 percent target set by the Child              funds as well as the increased recognition

and Family Service Reviews, and state CPS                  that a one-size-fits-all investigative response

agencies are actively trying to lower it. But              will not meet the needs of all families

                                                                              VOL. 19 / NO. 2 / FALL 2009     205

----------------------- Page 12-----------------------

Jane Waldfogel

referred to CPS. Nevertheless, challenges                  be on the threshold of an exciting new era

remain. Analysts have much to learn about                  in the provision of prevention programs. To

what CPS agencies can do to support and                    take fullest advantage of the opportunities

monitor preventive programs to ensure that                 this expansion of interest is likely to offer, it

they are delivering effective services.39 They             is worth keeping a few principles in mind.

also have much to learn about coordinating                 The first is that if studies are to yield reliable

services across the many types of community                evidence documenting that programs suc-

agencies that may play a role in prevention.40             cessfully prevent maltreatment, they must

                                                           use randomized designs whenever possible

Although the evidence base on preventive                   and must measure maltreatment outcomes.

programs for lower-risk families remains                   The second is that policy makers must keep

fairly thin, with a few exceptions such as the             in mind the lessons learned from past efforts,

results from randomized studies of the Nurse-              in particular, the importance of dosage and

Family Partnership program, programs and                   family engagement. As tempting as it may

evaluations in this area are expanding rapidly.            be to cut corners and save dollars, there is

Both DHHS and the federal Centers for                      no substitute for systematically implement-

Disease Control and Prevention are actively                ing and evaluating promising interventions.

reviewing program effectiveness and spur-                  If not, we could well find ourselves a decade

ring states to commission and participate in               from now with no more evidence on preven-

program evaluations. It seems the nation may               tion in CPS than we have today.


----------------------- Page 13-----------------------

                                                                           Prevention and the Child Protection System


1. All statistics in this paragraph are from U.S. Department of Health and Human Services, Administration on

    Children, Youth, and Families, Child Maltreatment 2006 (Washington: U.S. Government Printing Office,

    2008) ( [accessed July 29, 2008]).

2.  For an overview of alternative response systems, see Jane Waldfogel, “Differential Response,” in Community

    Prevention of Child Maltreatment, edited by Kenneth Dodge (New York: Guilford Press, 2009).

3.  U.S. Department of Health and Human Services, Child Maltreatment 2006 (see note 1), p. 83.

4.  These studies are reviewed by John D. Fluke and Dana Hollinshead, “Child Maltreatment Recurrence,”

    report prepared for the National Resource Center on Child Maltreatment (Duluth, Ga.: NRCCM, 2003)

    ( [accessed April 1, 2009]), and by John D. Fluke and

    others, “Reporting and Recurrence of Child Maltreatment: Findings from NCANDS,” report prepared for

    the U.S. Department of Health and Human Services, Office of the Assistant Secretary for Planning and

    Evaluation (DHHS, 2005) ( [accessed August 1, 2008]). See also Jessica Kahn, “Child

    Welfare Recidivism,” doctoral dissertation, Columbia University School of Social Work, 2006. These

    reviews cite only a few studies that find that families who received services had a lower likelihood of being

    re-reported. See Brett Drake and others, “Substantiation and Recidivism,” Child Maltreatment 4, no. 4

    (2003): 297–307; M. J. Camasso and R. Jagannathan, “Modeling the Reliability and Predictive Validity of

    Risk Assessment in Child Protective Services,” Children and Youth Services Review 22, no. 11/12 (2000):

    873–96; T. L. Fuller, S. J. Wells, and E. E. Cotton, “Predictors of Maltreatment Recurrence at Two

    Milestones in the Life of a Case,” Children and Youth Services Review 23, no. 1 (2001): 49–78; and Diane

    DePanfilis and Susan J. Zuravin, “The Effect of Services on the Recurrence of Child Maltreatment,” Child

    Abuse and Neglect 26, no. 2 (2002): 187–205.

5.  Fluke and others, “Reporting and Recurrence of Child Maltreatment” (see note 4). The study also found

    that among children who had initially been substantiated, about 17 percent were the subject of another

    substantiated investigation over the next five years. Nationally, data compiled for the Child and Family

    Services Reviews indicate that in 2005, 6.6 percent of substantiated victims were the subject of another

    substantiated investigation in the next six months, an improvement over the rate of 7.5 percent in 2002;

    see U.S. Department of Health and Human Services, Administration for Children and Families, “Child

    Welfare Outcomes 2002–2005: Report to Congress” (DHHS, 2008) (

    cwo05/chapters/executive.htm [accessed September 12, 2008]).

6.  Patricia Kohl and Richard Barth, “Child Maltreatment Recurrence among Children Remaining In-Home:

    Predictors of Re-Reports,” in Child Protection: Using Research to Improve Policy and Practice, edited by

    Ron Haskins, Fred Wulczyn, and Mary Bruce Webb (Washington: Brookings Institution Press, 2007).

7.  The “surveillance effect” is discussed on p. 13 of Fluke and Hollinshead, “Child Maltreatment Recurrence”

    (see note 4).

8.  Ron Haskins, Fred Wulczyn, and Mary Bruce Webb, “Using High-Quality Research to Improve Child

    Protection Practice: An Overview,” in Child Protection: Using Research to Improve Policy and Practice,

    edited by Haskins, Wulczyn, and Webb (see note 6).

9.  Michael Hurlburt and others, “Building on Strengths: Current Status and Opportunities for Improvement

    of Parent Training for Families in Child Welfare,” in Child Protection: Using Research to Improve Policy

    and Practice, edited by Haskins, Wulczyn, and Webb (see note 6).

                                                                                VOL. 19 / NO. 2 / FALL 2009       207

----------------------- Page 14-----------------------

Jane Waldfogel

10.  See reviews by Fluke and Hollinshead, “Child Maltreatment Recurrence” (see note 4), and Fluke and

    others, “Reporting and Recurrence of Child Maltreatment (see note 4); and Nick Hindley, Paul G.

     Ramchandani, and David P. H. Jones, “Risk Factors for Recurrence of Maltreatment: A Systematic

     Review,” Archives of Disease in Childhood 91, no. 9 (2006): 744–52.

11.  See, for example, Martha G. Roditti, “Child Day Care: A Key Building Block of Family Support and Family

     Preservation Programs,” in Child Day Care, edited by Bruce Hershfield and Karen Selman (Edison, N.J.:

    Transaction Publishers, 1997).

12.  State of Alaska, Office of Children’s Services (OCS), “OCS Family Preservation” (OCS, 2008) (www.hss. [accessed July 10, 2008]).

13. Illinois Department of Children and Family Services, “Day Care and Early Childhood”(DCFS, 2008)

     ( [accessed July 10, 2008]).

14.  Regarding cognitive development, see, for example, Margaret O’Brien Caughy, Janet A. DiPietro, and

     Donna M. Strobino, “Day-Care Participation as a Protective Factor in the Cognitive Development of Low-

     Income Children,” Child Development 65, no. 2 (1994): 457–71. Regarding social development, see, for

     example, Sylvana Cote and others, “The Role of Maternal Education and Nonmaternal Care Services in the

     Prevention of Children’s Physical Aggression Problems,” Archives of General Psychiatry 64, no. 11 (2007):


15. Although a small-scale study (of twenty-two children) found that infants placed into protective day care

    were more likely than other infants to be removed from their families subsequently, this appears to be

     an isolated finding. See Patricia M. Crittenden, “The Effect of Mandatory Protective Daycare on Mutual

    Attachment in Maltreating Mother-Infant Dyads,” Child Abuse and Neglect 7, no. 3 (1983): 297–300.

16. Information on Head Start from the U.S. House of Representatives, Committee on Ways and Means, 2008

     Green Book ( [accessed August 1, 2008]).

17.  U.S. Department of Health and Human Services, Administration for Children and Families, “Head Start

     Impact Study: First Year Findings” (Washington: DHHS, 2005) (

    impact_study [accessed August 6, 2008]).

18. John M. Love and others, “Making a Difference in the Lives of Infants and Toddlers and Their Families:

    The Impacts of Early Head Start. Final Technical Report” (Princeton, N.J.: Mathematica Policy Research,


19. Judith R. Smith and Jeanne Brooks-Gunn, “Correlates and Consequences of Mothers’ Harsh Discipline

    with Young Children,” Archives of Pediatric and Adolescent Medicine 151 (1997): 777–86.

20.  Katherine Magnuson and Jane Waldfogel, “Pre-School Enrollment and Parents’ Use of Physical Discipline,”

    Infant and Child Development 14, no. 2 (2005): 177–98.

21.  Arthur J. Reynolds and D. Robertson, “School-Based Early Intervention and Later Child Maltreatment in

    the Chicago Longitudinal Study,” Child Development 74 (2003): 3–26.

22.  Differential response reforms in many states have complicated efforts to measure the effectiveness of

     services provided by CPS in preventing future maltreatment, because states now differ sharply in how

    they define reports and substantiated cases. For a discussion of the origins and rationale for differential

     response, see Jane Waldfogel, The Future of Child Protection: Breaking the Cycle of Abuse and Neglect


----------------------- Page 15-----------------------

                                                                                 Prevention and the Child Protection System

     (Harvard University Press, 1998), and Jane Waldfogel, “The Future of Child Protection Revisited,” in

     Child Welfare Research: Advances for Practice and Policy, edited by Duncan Lindsey and Aron Shlonsky

     (Oxford University Press, 2008). For a brief overview, see U.S. Department of Health and Human Services,

     Administration for Children and Families, “Differential Response to Reports of Child Abuse and Neglect,”

     an issue brief prepared for the Child Welfare Information Gateway (DHHS, 2008) (

     [accessed August 1, 2008]).

23.  Waldfogel, “Differential Response” (see note 2).

24.  The Minnesota results are reported in Anthony L. Loman and Gary L. Siegel, Minnesota Alternative

     Response Evaluation: Final Report (St. Louis: Institute of Applied Research, 2004) (

     [accessed July 24, 2006]); Anthony L. Loman and Gary L. Siegel, “Alternative Response in Minnesota:

     Findings of the Program Evaluation,” Protecting Children 20, no. 2–3 (2005): 79–92; and Anthony L.

     Loman and Gary L. Siegel, “Extended Follow-Up Study of Minnesota’s Family Assessment Response:

     Final Report” (St. Louis: Institute of Applied Research, 2006) ( [accessed September 18,

     2007]). Results from Minnesota as well as other states are reviewed in Waldfogel, “Differential Response”

     (see note 2).

25.  U.S. Department of Health and Human Services, Child Maltreatment 2006 (see note 1).

26.  Ibid., p. 83.

27.  All statistics in this paragraph are from U.S. Department of Health and Human Services, Child Maltreatment

     2006 (see note 1).

28.  These other sources of funding are quite varied and include other federal agencies such as the Centers

     for Disease Control and Prevention, the Maternal and Child Health Bureau, and the U.S. Department of

     Justice, as well as a variety of state and private funding sources.

29.  In particular, I rely on estimates from various editions of the Green Book, published at regular intervals by

     the U.S. House of Representatives, Committee on Ways and Means. As of this writing, the 2008 version of

     the Green Book was being published in stages. For some sections, the 2008 version is available, while for

     others, the latest release was the 2004 version. See also Emilie Stoltzfus, “Child Welfare Issues in the 110th

     Congress,” CRS Report for Congress RL34388 (Congressional Research Service, 2008) (http://opencrs. [accessed January 15, 2009]); and Emilie Stoltzfus, “Child Welfare: Recent and Proposed Federal

     Funding,” CRS Report for Congress RL34121 (Congressional Research Service, 2007) (http://opencrs.cdt.

     org [accessed January 15, 2009]).

30.  Data from the U.S. House of Representatives, Committee on Ways and Means, 2004 Green Book, Section

     11—Child Protection, Foster Care, and Adoption Assistance ( [accessed

     January 15, 2009]).

31.  Data from the U.S. House of Representatives, Committee on Ways and Means, 2008 Green Book,

     Section 10—Title XX Social Services Block Grant Program (

     asp?section=2168 [accessed January 15, 2009]).

32.  See Stoltzfus, “Child Welfare Issues in the 110th Congress” (see note 29), and Stoltzfus, “Child Welfare:

     Recent and Proposed Federal Funding” (see note 29).

33.  In 2005, federal funds were 49 percent of total child welfare spending, with state funds making up 39

     percent and local funds making up 12 percent; see Cynthia Andrews Scarcella and others, “The Cost

                                                                                       VOL. 19 / NO. 2 / FALL 2009          209

----------------------- Page 16-----------------------

Jane Waldfogel

     of Protecting Vulnerable Children, V: Understanding State Variation in Child Welfare Financing”

     (Washington: Urban Institute, 2006).

34.  Ibid.

35.  David Thomas and others, “Emerging Practices in the Prevention of Child Abuse and Neglect,” report

     prepared for the U.S. Department of Health and Human Services, Children’s Bureau Office on Child Abuse

     and Neglect (DHHS, 2003) ( [accessed July

     28, 2008]). The federal Centers for Disease Control and Prevention (CDC) are also involved in reviewing

     the effectiveness of prevention programs; see, for example, Centers for Disease Control and Prevention,

     National Center for Injury Prevention and Control, “Using Evidence-Based Parenting Programs to Advance

     CDC Efforts in Child Maltreatment Prevention” (CDC, 2004) (

     ChildMalT-Briefing.pdf [accessed August 3, 2008]).

36.  Quote from p. 15 of Thomas and others, “Emerging Practices in the Prevention of Child Abuse and

     Neglect” (see note 35).

37.  See U.S. Department of Health and Human Services, Administration for Children and Families, “Child

     Welfare Outcomes 2002–2005: Report to Congress” (DHHS, 2008) (

     cwo05/chapters/executive.htm [accessed September 12, 2008]).

38.  Quote from p. 2 of Haskins, Wulczyn, and Webb, “Using High-Quality Research to Improve Child

     Protection Practice” (see note 8).

39.  See discussion in Fred Wulczyn, “A Community’s Concern,” Child Welfare Watch 14 (Summer 2007): 29–30.

40.  The need for coordination arises, in large part, because children at risk for maltreatment often have mul-

     tiple needs and thus require services that cut across agencies. See Roger Bullock and Michael Little, “The

     Contribution of Children’s Services to the Protection of Children” (Dartington, England: Dartington Social

     Research Unit, 2002) (; and Nick Axford and Michael Little, Refocusing Children’s

     Services towards Prevention: Lessons from the Literature (London: Department for Education and Skills

     Research Report RR10, 2004) (


----------------------- Page 17-----------------------

Prevention and the Child Protection System

         VOL. 19 / NO. 2 / FALL 2009                                       211

----------------------- Page 18-----------------------