Literature Review

Effect_of_Child_Protective_Ser.pdf

Home >Social Science homework help >Literature Review

Effect of Child Protective Services System Factors on Child Maltreatment Rereporting

This study examined how child protective services (CPS) systems respond to initial and subsequent reports in the context of child maltreatment rereporting and to what extent CPS system factors are associated with the risk of rereporting after controlling for abuse type and child and family factors. This study followed 67,243 families who were reported to the CPS agencies in seven counties in Florida for child abuse and neglect over a period of 5.4 years and found that 14,218 families had one or more child maltreatment

rereports. Key findings include that CPS system factors were significantly different from initial report to subsequent reports. Five CPS system factors, reporting source, contact by CPS workers, investigative level at intake, postinvestigation services, and duration of CPS involvement were significantly associated with the risk of child maltreatment rereporting. Multivariate analyses found that CPS system factors were substantially dif- ferent for three categories of rereporting, unsubstantiated rere- ports, substantiated rereports, and recurrence reports. Interpretations and implications for practice are discussed.

Hwa-ok Bae Gyeongsang National University

Phyllis L. Solomon University of Pennsylvania

Richard J. Gelles University of Pennsylvania

Tammy White Greater Philadelphia Urban Affairs Coalition

33Child Welfare • Vol. 89, No. 3

Child protective services (CPS) systems are designed to identifyvictims of maltreatment and protect them from further victim- ization. One core goal of CPS investigation and intervention is to prevent reoccurrence of maltreatment. Despite the extensive services provided by CPS, large numbers of children experience subsequent maltreatment. A study from National Child Abuse and Neglect Data System (NCANDS) found that, between 1998 and 2002, approxi- mately one-third of 1.4 million children reported in nine states were rereported, and approximately 17% of victims were revictim- ized within five years (Fluke, Shusterman, Hollinshead, & Yuan, 2005). These large rereport and revictimization rates suggest that CPS agencies serve the same abused and neglected children repeatedly. This cycle of rereporting and recurrence results in significant costs for CPS agencies and creates a major burden on CPS workers, many of whom are already struggling with heavy caseloads. More importantly, repeat cases suggest a failure of CPS’ intended mission to adequately protect children from revictimization. Accordingly, identification of factors for child maltreatment rereporting may be the logical prereq- uisite for developing effective interventions for children and families.

Rereporting and Recurrence Rereporting possibly results in the subsequent identification of recur- rence (substantiated or indicated maltreatment). Way, Chung, Jonson-Reid, and Drake (2001) found that more than 10% of those whose index event was not substantiated were substantiated at the second report in a statewide administrative CPS data with 31,531 perpetrators of intrafamilial maltreatment. Drake, Jonson-Reid, Way, and Chung (2003) also found that 17.6% of the initially unsubstan- tiated neglect reports resulted in substantiated rereports. Similarly, Dorsey, Mustillo, Farmer, and Elbogen (2008) found that casework- ers assessed approximately one-third of the cases as having a high likelihood of maltreatment recurrence and one-fifth of the cases in the entire sample received subsequent reports of maltreatment. All of these findings confirm that rereporting leads to a significant rate of recurrence.

Vol. 89, No. 3Child Welfare

Ecological Model of Child Maltreatment Rereporting Among the theoretical perspectives employed to identify risk factors for child maltreatment, one widely used framework is the ecological model. The ecological model explains that child maltreatment is determined by a variety of factors operating through transactional processes at various levels (Belsky, 1993; Cash, 2001; Sidebotham, 2001). Thus, the ecological model enables one to combine the diver- sity of factors known to be relevant to the etiology of child maltreat- ment rereporting. Child and family factors related to the risk of child maltreatment rereporting include the child’s age (Fluke, Yuan, & Edwards, 1999), the child’s sex (Fryer & Miyoshi, 1994), the child or perpetrator’s race or ethnicity (Lipien & Forthofer, 2004; Way et al., 2001), the family size or number of children ( Johnson & L’Esperance, 1984; Wolock, Sherman, Feldman, & Metzger, 2001), and the fam- ily structure (Coohey, 2006; Fuller, Wells, & Cotton, 2001). In addi- tion to these child and family risk factors, the present study included CPS system factors into the research model to further explore the likelihood of child maltreatment rereporting from an enhanced eco- logical perspective.

CPS System Factors CPS systems involve complex processes of investigation and dis- position, and therefore, it is important to examine what system ele- ments may be associated with the likelihood of rereporting (Inkelas & Halfon, 1997). One key element that is theorized to be associ- ated with child maltreatment rereporting is a family’s continued exposure to the CPS system. More than any other single factor, prior abuse history or being known to the CPS system, was found to be a very strong predictor of rereporting in a number of studies (DePanfilis & Zuravin, 2002; Dorsey et al., 2008; Fuller et al., 2001; Inkelas & Halfon, 1997). Once the child and family are brought to the attention of the CPS system, they are subjected to a greater degree of observation and inspection. As such, this increased CPS surveillance may be related to a higher risk of child maltreatment

Child WelfareBae et al.

rereporting. The degree and extent of CPS surveillance can be measured by examining both the frontline of the child protection system (i.e., reporting and investigation) and the next gate of the system (i.e., disposition and service). The present study selected five CPS system factors as indicators of CPS surveillance: (1) report source, (2) contact by CPS workers, (3) investigative level at intake, (4) postinvestigation services, and (5) duration of CPS involvement. The researchers examined whether these CPS system factors were different from initial report to subsequent report(s) and to what extent the risk of child maltreatment rereporting was associated with these factors.

Report Source The likelihood of rereporting may be influenced by who initially and subsequently reports a case of suspected child maltreatment. Schools and law enforcement agencies are more likely than any other report source to report physical abuse; medical professionals are more likely to report sexual abuse; and neighbors and friends mostly report gen- eral neglect (Barth, Courtney, Berrick, & Albert, 1994; Giovannoni, 1995). Differential exposure to initial reporters may cause differ- ential rereporting rates. U.S. Department of Health and Human Services (USDHHS), National Center for Child Abuse and Neglect (NCCAN), and NCANDS (2004) data indicate that children reported by educational personnel were 25% more likely to be reported for a recurrence, while children reported by law enforce- ment or legal personnel were 9% less likely to be rereported than those initially reported by social services personnel. Second, families with specific social problems such as poverty are increasingly exposed to the CPS system (Drake, Jonson-Reid, & Sapokaite, 2006; Drake et al., 2003; Wolock et al., 2001), which may increase the likelihood of rereporting. Higher rates of rereporting were found for children with Medicaid mental health/substance-abuse treatment and special eligibility for emotional disturbance, whereas a lower rate of rereporting occurred among children with parents who were per- manently exited from their first spell on Aid to Families with Dependent Children (Drake et al., 2006).

Vol. 89, No. 3Child Welfare

Contact by CPS Workers During the initial stages of investigation CPS workers typically have contacts (e.g., home visits, phone calls) with the child and fam- ily reported for suspected maltreatment. The CPS workers have more frequent contacts with younger children and complex, large, and multiproblem families (Freeman, Levine, & Doueck, 1996; Jagannathan & Camasso, 1996). Since child neglect tends to be chronic, with parental behaviors unlikely to change rapidly, a CPS worker’s ongoing contact with the family may increase the likeli- hood of detecting the recurrence of neglect. The number of in- person visits to a family receiving CPS services affects the rates of recurrence (Cash, 2001; Johnson, 1994). Frequency of contact with a caseworker is positively associated with maltreatment recurrence (Fuller et al., 2001). This finding could be explained by the detec- tion of other maltreatment during the initial investigation or by the fact that maltreating families are more likely to be rereported to CPS following initial contact with the system (DePanfilis & Zuravin, 1998). However, the number of contacts or number of workers has not been found to predict recurrence of child maltreatment (DePanfilis & Zuravin, 2002).

Investigative Level at Intake Once a report is screened into the system, CPS workers assign an ini- tial level of risk ranging from low to high risk. Low risk level receives a low standard of investigation, whereas cases with high risk levels receive a high standard of investigation. Sexual abuse cases are typi- cally classified as a higher response priority for investigations com- pared to any other abuse types (Drake, 1995; Karski, 1999). Presence of injury is most often deemed to indicate abuse and thus investi- gated (Karski, 1999; Wells, Fluke, & Brown, 1995). Reports from law enforcement personnel are more likely to be judged emergen- cies for physical abuse and neglect, although reports from an anony- mous source and school personnel were more likely to be judged emergencies for sexual abuse (Drake, 1995). Age of child and certain family characteristics are associated with the investigative level at intake, with young children and high-risk families being more likely

Child WelfareBae et al.

to receive higher levels of investigation (Wells et al., 1995; Wells, Lyons, Doueck, Brown, & Thomas, 2004). The overall family risk rating at initial intake was related to maltreatment rereporting, with families judged as no risk and low risk occurring less frequently in the recurring sample, and families judged as moderate risk occur- ring more frequently in the recurring group (Fuller et al., 2001). Meanwhile, others indicate that agreement between caseworkers’ assessment of risk and rereport are low; only 16% of the cases assessed as low risk had a subsequent report of child maltreatment in the study by Dorsey et al. (2008). Therefore, investigative level at intake is likely to be more intensive at the second reporting, leading to the theory that rereporting will be associated with higher investiga- tive levels at intake.

Postinvestigation Services Postinvestigation services are provided to families to reduce the risk of future harm, offered on a voluntary basis by CPS agencies, or ordered by the courts (USDHHS, NCCAN, NCANDS, 2004). The likelihood of child maltreatment rereporting may vary by type of CPS services, as the type of service signifies the degree of CPS involvement. Family support services were found to be associated with a lower risk of recurrence (DePanfilis & Zuravin, 2002), whereas families receiving less intensive in-home services have a higher risk of rereporting compared to those not receiving services (Lipien & Forthofer, 2004). Drake et al. (2003) found that family preservation service cases had higher rates of recidivism, whereas those receiving traditional child welfare services were no different from the not- served group in the rates of recidivism. Children are more likely to be replaced on an emergency basis during or after an investigation of abuse. More controlling services, such as foster care placement and court-ordered custody, will decrease the risk of rereporting because of the removal of the children from families at risk of future mal- treatment. Other researchers found that foster placements with a relative were associated with lower risk of recurrent maltreatment reports than no services (Connell, Vanderploeg, Katz, Caron, Saunders, & Tebes, 2009; Lipien & Forthofer, 2004), and foster care services

Vol. 89, No. 3Child Welfare

significantly decreased the risk of rereporting of neglect cases ( Jonson- Reid, Drake, Chung, & Way, 2003).

Duration of CPS Involvement CPS involvement includes the investigation period (from case intake and to case disposition) and the intervention period (from case dis- position to case closing). While the increased likelihood of rere- porting may be associated with increased CPS surveillance, the risk for rereporting is higher during the early period of the CPS inves- tigation and then becomes lower as the duration of services is extended. In other words, the risk for rereporting is not evenly dis- tributed across the time of CPS involvement. Studies using survival analysis confirm that the risk of child maltreatment recurrence was greatest in the immediate period such as within the first month following an initial report (DePanfilis & Zuravin, 1999; Fryer & Miyoshi, 1994), thus indicating the likelihood of rereporting in the early period of CPS involvement. Cases in which the family received no services within the first 60 days following case opening was approximately twice as likely to recur than cases that received at least one service during this time (Fuller et al., 2001). Further, other stud- ies found that longer durations of services are associated with lower risks of child maltreatment rereporting (Marshall & English, 1999; Johnson & L’Esperance, 1984).

This study, therefore, hypothesizes that the risk of child mal- treatment rereporting will be associated with report source, inves- tigative level at intake, contact by CPS workers, postinvestigation services, and duration of CPS involvement after controlling for abuse type and child and family factors.

Method The data source is the Florida Department of Children and Families Florida Abuse Hotline Information System (FAHIS). The data available for this study included all FAHIS records starting from January 1, 1996, to May 31, 2002, in seven Florida counties (Broward, Hillsborough, Lee, Manatee, Palm Beach, Pasco, and Pinellas).

Child WelfareBae et al.

This study merged four datasets: (1) report dataset, (2) reporter dataset, (3) dependent dataset, and (4) child-caretaker pair dataset from FAHIS by selecting the family as the unit of analysis (Kinnevy, Cohen, Huang, Gelles, Bae, Fusco, & Dichter, 2003). Since this study used the family as the unit of analysis rather than the child, a ran- dom selection of one child per family was conducted for those families, where more than one child was reported on the same day. Consequently, multiple reports for more than one child in a given family on the same day were not counted as a rereport. All variables were based on the randomly selected child.

This study regarded a report without a prior CPS report in the preceding one year of the study period as the first report. Studies using survival analysis have found that the greatest risk of recurrence is within the first month to four months following an initial incident; hence, a majority of rereporting tends to cluster in the early period of CPS involvement (DePanfilis & Zuravin, 1999; Fryer & Miyoshi, 1994; Lipien & Forthofer, 2004). Accordingly, this study regarded a report of a family without a prior CPS report in the preceding one year as the starting point of the rereporting interval. Therefore, this study defines the rereporting period as an additional report of phys- ical abuse, sexual abuse, or neglect of any child in the family that was reported to a CPS agency during a 5.4-year period from the start of the first report or index report for purposes of this study.

The final dataset included 67,243 families who were tracked until May 31, 2002, to assess whether the family had any rereport of child maltreatment. Among them, 14,218 families (21.1%) were rereported one or more times, with a total of 21,219 rereports. Each rereport for a family was counted as a separate rereport whereby CPS had under- gone the procedures of intake, investigation, and disposition for the case. The 21,129 rereports were grouped into three categories accord- ing to prior and subsequent substantiation sequence: (1) unsubstan- tiated rereports, which included unsubstantiated-to-unsubstantiated reports (35.2%) and substantiated-to-unsubstantiated reports (22.4%); (2) substantiated rereports, which included unsubstantiated-to-sub- stantiated reports (21.5%); and (3) recurrence reports, which included substantiated-to-substantiated reports (20.9%; Fluke et al., 2005).

Vol. 89, No. 3Child Welfare

Results Univariate Results The total sample was composed of 67,243 families. The majority of families (58%) were initially reported for neglect, and more than half (58%) of the initial reports were deemed unsubstantiated. The majority of the children (61%) were white, and almost three- quarters were aged 12 and younger. Descriptive statistics, including abuse type and child and family characteristics of the sample, are presented in Table 1.

Bivariate Results Bivariate Chi square analyses were conducted on the subsample of families (n � 14,218) who had rereports to explore differences in the five CPS system factors among the three categories of rereporting (Table 2).

Report Source Overall, more than half of the unsubstantiated rereports were reported by mandatory reporters at both initial report and subse- quent reports, contrary to substantiated rereports, where the pre- dominant number of families was made by nonmandatory reporters. Recurrence reports were least likely to be made by nonmandatory reporters. Among mandatory reporters, educational personnel were more likely to identify unsubstantiated rereports and substantiated rereports, whereas legal personnel were more likely to identify recur- rence reports both at initial report and at subsequent reports. Across all categories of rereporting, social services personnel were more likely to, and medical personnel and legal personnel were less likely to iden- tify all categories of rereports at subsequent reports.

Investigative Level The majority of families were investigated at a medium intensity level both at initial report and subsequent reports for all categories of rereporting.

Child WelfareBae et al.

Vol. 89, No. 3Child Welfare

Variables Number of Families Percent Substantiation

Substantiated 39,217 41.7 Not substantiated 28,026 58.3

Abuse type Neglect 39,240 58.4 Physical abuse 22,244 33.1 Sexual abuse 5,759 8.6

Age of child �2 10,363 15.4 2– 5 12,126 18.0 6– 12 26,397 39.3 13– 18 18,357 27.3 (Mean) (8.0)

Sex of child Girls 34,155 50.8 Boys 33,088 49.2

Race/ethnicity Blacks 19,749 29.4 Latinos 5,336 7.9 Whites 41,058 61.1 Others 1,100 1.6

Family structure Both parents 27,056 40.2 Stepparents 7,081 10.5 Mother only or with other 26,943 40.1 Father only or with other 3,501 5.2 Relatives and others 2,662 4.0

Number of dependents 1 23,923 35.6 2 21,954 32.7 3� 21,366 31.8 (Mean) (2.2)

Total 67,243 100.0

Table 1 Description of the Sample

Child WelfareBae et al.

U ns ub st an ti at ed

Su b st an ti at ed

R ec ur re nc e

R er ep o rt

R ep o rt

(n = 12 ,1 66 )

(n = 4, 54 0)

(n = 4, 42 3)

In it ia l

Su b se q ue nt

In it ia l

Su b se q ue nt

In it ia l

Su b se q ue nt

R ep o rt

R ep o rt s

R ep o rt

R ep o rt s

R ep o rt

R ep o rt s

(% )

Re po

rt s ou

rc e

� 2

� 49

.2 **

* �

2 �

23 .1

** *

� 2

� 46

.1 **

* M

ed ic al

8. 8

6. 8

7. 8

6. 9

12 .1

7. 8

Ed uc

at io

na l

16 .8

17 .5

16 .9

17 .0

18 .3

16 .3

So ci al

s er

vi ce

s 11

.1 14

.9 10

.7 14

.0 10

.8 15

.0 Le

ga l

14 .1

13 .1

10 .1

8. 0

23 .1

20 .3

D ay

c ar

e an

d fo

st er

c ar

e 1.

2 1.

4 1.

3 2.

5 1.

4 2.

2 N on

m an

da to

ry 48

.1 46

.3 53

.1 51

.6 34

.4 38

In ve

st ig

at iv

e le

ve l

� 2�

11 8.

1* **

� 2�

36 .7

** *

� 2�

51 .6

** *

H ig

h 10

.4 11

.1 8.

7 8.

6 10

.2 9.

1 M

ed iu

m 79

.5 84

.5 81

.5 86

.8 80

.9 87

.7 Lo

w 10

.1 4.

3 9.

9 4.

7 8.

9 3.

Ta b

le 2

C PS

S ys

te m

C ha

ra ct

er is tic

s Re

sp on

di ng

t o

In iti

al R

ep or

t an

d Su

bs eq

ue nt

R ep

or ts

Co nt

in ue

d on

n ex

t p ag

Vol. 89, No. 3Child Welfare

Ta b

le 2

c on

t. U ns ub st an ti at ed

Su b st an ti at ed

R ec ur re nc e

R er ep o rt

R ep o rt

(n = 12 ,1 66 )

(n = 4, 54 0)

(n = 4, 42 3)

In it ia l

Su b se q ue nt

In it ia l

Su b se q ue nt

In it ia l

Su b se q ue nt

R ep o rt

R ep o rt s

R ep o rt

R ep o rt s

R ep o rt

R ep o rt s

(% )

N um

be r of

c on

ta ct

s �

2 �

13 .7

** �

2 �

3. 5

� 2

� 17

.0 **

* 0

2. 5

2. 7

2. 4

3. 2

2. 7

3. 2

1 71

.3 69

.3 75

.1 73

.2 63

.9 61

.9 2

20 .0

20 .1

18 .3

18 .9

22 .7

20 .2

3 an

d m

or e

6. 2

7. 9

4. 2

4. 7

10 .8

14 .7

Po st

in ve

st ig

at io

n se

rv ic es

� 2

� 87

.4 **

* �

2 �

27 .2

** *

� 2

� 25

.2 **

* C ou

rt s er

vi ce

s 5.

0 8.

1 0.

8 1.

7 12

.4 16

.4 C PS

s er

vi ce

s 24

.1 28

.4 22

.7 28

.5 30

.8 34

.1 O th

er p

ro gr

am 20

.9 17

.4 18

.0 15

.8 28

.1 24

.5 N o

se rv

ic es

50 .0

46 .1

58 .5

54 .0

28 .8

25 .0

To ta

l 10

0. 0

10 0.

0 10

0. 0

10 0.

0 10

0. 0

10 0.

0 In

vo lv

em en

t pe

rio d

(m on

th )

8. 7

7. 5

6. 9

6. 2

11 .6

9. 4

*� 0.

** �

0. 01

*� 0.

00 1

Number of Contacts At both initial report and at subsequent report, the majority of fam- ilies were contacted once by CPS across all three categories of rere- porting. Families with recurrence reports significantly received more frequent contacts with CPS relative to those with unsubstantiated rereport and substantiated rereport.

Postinvestigation Services More than half of the families with substantiated rereports did not receive postinvestigation services, whereas a majority of families with recurrence reports received postinvestigation services at both initial report and subsequent reports. Across all three categories of rere- porting, families were more likely to receive court-ordered services and general CPS services at subsequent reports.

Involvement Period Families with recurrence reports had longer periods of CPS involve- ment compared to those with unsubstantiated or substantiated rere- ports. Across all three categories of rereporting, the involvement period was shorter following a subsequent report than at initial report.

In summary, the CPS system factors differently responded to initial report and subsequent reports. Overall, the CPS system was likely to provide more intensive surveillance to subsequent reports than to initial reports, and families were more intensely investigated, more frequently contacted, and more likely to receive some postin- vestigation services at subsequent reports across all three categories of rereporting.

Multivariate Results Nonproportional Cox regression analyses included 67,243 families and their 88,372 reports and rereports to examine the impact of the five CPS system factors on the risk of unsubstantiated rereport, sub- stantiated rereport, and recurrence report, when controlling for abuse type and child and family factors (Table 3). The analyses incorporated one time-dependent covariate (duration of CPS involvement) that

Child WelfareBae et al.

Vol. 89, No. 3Child Welfare

U ns ub st an ti at ed

Su b st an ti at ed

R ec ur re nc e

R er ep o rt

R ep o rt

Pa ra m et er

H az ar d

Pa ra m et er

H az ar d

Pa ra m et er

H az ar d

Es ti m at e

R at io

Es ti m at e

R at io

Es ti m at e

R at io

A bu

se t yp

e Ph

ys ic al

a bu

se �

0. 18

33 **

* 0.

83 30

� 0.

17 60

** *

0. 83

90 �

0. 43

51 **

* 0.

64 70

Se xu

al a

bu se

� 0.

27 25

** *

0. 76

10 �

0. 37

34 **

* 0.

66 80

� 0.

36 79

** *

0. 69

20 N eg

le ct

(r ef

.) C hi

ld a

nd fa

m ily

fa ct

or s

C hi

ld ’s a

ge �

0. 02

31 **

* 0.

97 70

� 0.

03 52

** *

0. 96

50 �

0. 03

15 **

* 0.

96 90

C hi

ld ’s s ex

G irl

0. 01

16 1.

01 20

0. 04

63 1.

04 70

� 0.

01 57

0. 98

40 Bo

y (r ef

.) Ra

ce /e

th ni

ci ty

La tin

o �

0. 18

96 **

* 0.

82 70

� 0.

33 07

** *

0. 71

80 �

0. 33

66 **

* 0.

71 40

W hi

te s

0. 10

20 **

* 1.

10 70

� 0.

00 69

0. 99

30 �

0. 09

20 *

0. 91

20 O th

er �

0. 44

63 **

* 0.

64 00

� 0.

45 00

** 0.

63 80

� 0.

58 23

** *

0. 55

90 Bl

ac k

(r ef

.) Fa

m ily

s tr uc

tu re

St ep

pa re

nt (s

) 0.

07 11

* 1.

07 40

0. 06

90 1.

07 10

0. 20

61 **

* 1.

22 90

M ot

he r on

ly o

r w

ith o

th er

0. 21

04 **

* 1.

23 40

0. 29

93 **

* 1.

34 90

0. 30

90 **

* 1.

36 20

Fa th

er o

nl y

or w

ith o

th er

0. 25

14 **

* 1.

28 60

0. 20

63 **

1. 22

90 0.

24 12

** 1.

27 30

O th

er r el

at iv

e 0.

16 80

** *

1. 18

30 0.

21 23

** 1.

23 70

0. 03

74 1.

03 80

Bo th

p ar

en ts

(r ef

Ta b

le 3

N on

pr op

or tio

na l H

az ar

ds M

od el

fo r th

e Ri

sk o

f R er

ep or

tin g

Child WelfareBae et al.

N um

be r of

d ep

en de

nt s

0. 18

58 **

* 1.

20 40

0. 16

78 **

* 1.

18 30

0. 15

72 **

* 1.

17 00

C PS

s ys

te m

fa ct

or s

Re po

rt s ou

rc e

M ed

ic al

� 0.

23 80

** *

0. 78

80 �

0. 30

85 **

* 0.

73 50

0. 15

71 **

1. 17

00 Ed

uc at

io na

l �

0. 02

28 0.

97 70

� 0.

07 64

0. 92

60 0.

39 45

** *

1. 48

40 So

ci al

0. 03

97 1.

04 10

� 0.

01 39

0. 98

60 0.

27 66

** *

1. 31

90 Le

ga l

� 0.

24 27

** *

0. 78

50 �

0. 63

42 **

* 0.

53 00

0. 46

31 **

* 1.

58 90

C hi

ld c ar

e an

d fo

st er

c ar

e �

0. 19

26 *

0. 82

50 0.

02 96

1. 03

00 0.

27 02

* 1.

31 00

N on

m an

da to

ry (r

ef .)

In ve

st ig

at iv

e le

ve l

0. 02

58 1.

02 60

0. 08

85 *

1. 09

20 0.

11 26

** 1.

11 90

N um

be r of

c on

ta ct

s 0.

02 49

** *

1. 02

50 0.

01 13

1. 01

10 0.

04 34

** *

1. 04

40 Po

st in

ve st

ig at

io n

se rv

ic es

C ou

rt s er

vi ce

0. 15

22 *

1. 16

40 �

0. 88

93 **

* 0.

04 11

1. 23

66 **

* 3.

44 20

C PS

s er

vi ce

s 0.

11 95

** *

1. 12

70 0.

06 93

1. 07

20 0.

77 13

** *

2. 16

30 O th

er p

ro gr

am 0.

10 00

** 1.

10 50

� 0.

04 87

0. 95

20 0.

83 90

** *

2. 31

40 N o

se rv

ic e

(r ef

.) In

vo lv

em en

t pe

rio d

(m on

th )

0. 00

21 1.

00 20

� 0.

00 68

** 0.

99 30

0. 01

05 **

* 1.

01 10

N um

be r of

e ve

nt 11

,4 28

4, 47

5 3,

81 0

Lo g

lik el

ih oo

d ra

tio 30

20 .4

80 0*

*� 0.

** �

0. 01

*� 0.

00 1

changes the hazard at different rates for families and has different effects on rereporting (Allison, 1995).

Unsubstantiated Rereports Abuse type and child and family factors. Results indicate that phys- ical abuse or sexual abuse was less likely to be associated with an unsubstantiated rereport than neglect. The child’s age was signifi- cantly associated with an unsubstantiated rereport, but the child’s gender was not. Family race or ethnicity, structure, and size were sig- nificantly associated with an unsubstantiated rereport. Compared to black families, Latino and other families were significantly less likely, but white families were more likely, to have an unsubstantiated rere- port. All types of family structure than both-parent families were more likely to have an unsubstantiated rereport. The greater the number of dependents, the greater the likelihood of unsubstantiated rereports.

CPS system factors. When controlling for abuse type and child and family factors, families with unsubstantiated rereports were significantly less likely to be made by medical personnel, legal per- sonnel, and child care and foster care personnel than those by non- mandatory reporters. Statistically, the risk ratios of unsubstantiated rereport for families reported by medical personnel and legal per- sonnel were 21.2 and 21.5% lower than those reported by non- mandatory reporters (e�0.2380�1, e�0.2427�1). Investigative level was not significantly associated with the likelihood of unsubstantiated rereporting. Families with unsubstantiated rereports were more likely to have frequent contacts of CPS workers and were more likely to receive postinvestigation services. Duration of CPS involvement was not significantly associated with the likelihood of unsubstantiated rereporting.

Substantiated Rereports Abuse type and child and family factors. Nonproportional Cox regression analyses on substantiated rereporting produced very few different results compared to unsubstantiated rereporting. For exam- ple, white families and stepparent families lost statistical significance,

Vol. 89, No. 3Child Welfare

but other child and family factors did not substantially differ from the ratios found in the examination of unsubstantiated rereports.

CPS system factors. Report sources risk ratios were compara- ble to the ratios found by the unsubstantiated rereport analyses. Investigative level was significantly associated with the likelihood of substantiated rereporting. An increase in the investigative level from initial report to rereport (from low to medium, or from medium to high) was significantly and positively associated with unsubstanti- ated rereport, meaning that the family was investigated at a higher level. In contrast, number of contacts by CPS workers, and CPS serv- ices and other programs lost statistical significance. Families with substantiated rereports were 58.9% less likely to receive court serv- ices (e�0.8893�1). A longer duration of CPS involvement was nega- tively associated with a substantiated rereport. In other words, each one month increase of the involvement period decreased the risk of a family having a substantiated rereport by 0.7% (e�0.0068�1).

Recurrence Reports Abuse type and child and family factors. Nonproportional Cox regression analyses on recurrence reports produced similar results found by the unsubstantiated rereport analyses. Abuse type, child’s age, and family race or ethnicity, structure, and size significantly predicted the risk of a family having a recurrence report. White families were less likely to have a recurrence report compared to black families.

CPS system factors. All of the CPS system factors were signifi- cantly associated with the likelihood of recurrence report. Families with recurrence reports were more likely to be made by mandatory reporters than by nonmandatory reporters. Families with recurrence reports were more likely to be intensively investigated, more fre- quently contacted by CPS workers, more likely to receive any serv- ices, and more likely to be involved in the CPS system longer.

In summary, multivariate analyses indicate that abuse type and child and family factors were not much different for the three cate- gories of unsubstantiated rereports, substantiated rereports, and recurrence reports. Across three categories of rereporting, large,

Child WelfareBae et al.

single-parent families reported for neglect were more likely to be rereported to the CPS system.

The system factors significantly differed, however, by categories of rereporting. Medical personnel and legal personnel were less likely to be associated with the risk of unsubstantiated rereports and sub- stantiated rereports, but were more likely to be associated with the risk of recurrence reports. Higher investigative level positively pre- dicted the risk of substantiated rereports and recurrence reports. Number of contacts by CPS workers and all types of postinvestiga- tion services positively predicted the risk of unsubstantiated rereports and recurrence reports. Duration of CPS involvement negatively pre- dicted the risk of substantiated rereport but positively predicted the risk of recurrence reports.

Discussion The present study found that the five CPS factors tested were sig- nificantly associated with the risk of child maltreatment rereporting after controlling for abuse type and child and family factors. The results of the multivariate analyses were quite consistent with those from the bivariate analyses. These findings suggest that the risk of rereporting is to some extent explained by the CPS system factors. These results indicate, however, that the risk ratios for these system factors were different for unsubstantiated rereports, substantiated rereports, and recurrence reports.

There are two plausible explanations. One is that the CPS sys- tem differently responds to substantiation status of initial report and subsequent reports. More importantly, substantiation of initial report seems to be critical in differentiating the degree of response by the CPS system. The CPS system seems to provide more intensive sur- veillance to children and families of the initially substantiated cases. For example, families with substantiated rereports were least likely to be identified by mandatory reporters and were less likely to receive contacts by CPS workers and some postinvestigation services. Among the families with unsubstantiated rereports, a significant proportion of families had substantiated initial reports (22.4%). If

Vol. 89, No. 3Child Welfare

the unsubstantiated rereports were categorized into two groups according to the substantiation status of initial report, there may be greater clarity in interpreting the findings.

The second explanation is that the three categories of rereporting may represent different characteristics of the children and families involved. In other analyses of the same dataset, researchers found that families with recurrence reports were more likely to have multiple rereporting compared to those with unsubstantiated rereports and substantiated reports, while families with unsubstantiated rereports were more likely to have single rereporting (� � 42012.42, p � .001). Families with recurrence reports inherently had different character- istics; younger age of child victims, single mother or stepparent, and large family significantly increased the odds of multiple child mal- treatment recurrence relative to single recurrence and no recurrence (Bae, Solomon, & Gelles, 2009).

Given the design of this study, researchers cannot say defini- tively that the CPS system factors cause the risk of rereporting. It is impossible to establish a causal relationship between explanatory variables (CPS system factors) and dependent variables (categories of rereporting) using state administrative data with a retrospective research design. However, it can be suggested that the CPS system factors respond differently to initial report and subsequent reports; further, researchers can conjecture that CPS system factors are likely associated with the risk of child maltreatment rereporting. Thus, the future risk of child maltreatment rereporting can be predicted. For example, if the initially unsubstantiated report was made by medical personnel or legal personnel, it was associated with decreas- ing likelihood of rereporting. Sexual abuse and emergency cases were not likely rereported, as sexual abuse was mostly reported by medical personnel and removal to foster care or custody was dealt by legal personnel.

Implications for Practice Study results indicate that the likelihood of a family being rereported is greater for families that have higher degrees of exposure to and

Child WelfareBae et al.

involvement with CPS agencies. While this may indicate that a sys- tem designed to protect children from continued maltreatment is falling short of its duties, results actually can be interpreted more pos- itively. In fact, if children known to the CPS system are under more intense scrutiny and thus more likely to be rereported, then one inter- pretation is that the CPS system is being extra diligent in monitor- ing children under its care.

One caveat, however, is that each rereport requires a repeated procedure of intake, investigation, and disposition. This repetition of CPS procedures, therefore, can result in significant costs for CPS agencies and can create additional burden on CPS workers who may already be struggling with heavy caseloads. A key implication resulting from this study is that CPS agencies may need to provide additional training to its workers or create or modify policies that explicitly outline when rereporting is indicated. Given that this study found that 58% of the 21,129 rereports were unsubstantiated, it would behoove CPS agencies to carefully weigh the benefits of rereporting against the human and capital costs associated with rere- porting. More intensive input of both time and energy into identi- fying and substantiating initial reports may possibly prevent the occurrence of subsequent reports.

Findings from this study suggest additional research is needed to explore the extent to which other CPS system factors not included here may be associated with the risk of rereporting. It is evident that the CPS system plays a crucial and significant role in the likelihood of a child being rereported for maltreatment, and the more the sys- tem factors associated with rereporting are understood, the more effi- cient the rereporting system can be.

References

Allison, P. D. (1995). Survival analysis using the SAS system: A practical guide. Cary, NC: SAS Institute.

Bae, H., Solomon, P. L., & Gelles, R. J. (2009). Multiple child maltreatment recurrence rel- ative to single recurrence and no recurrence. Children and Youth Services Review, 31(6), 617– 624.

Vol. 89, No. 3Child Welfare

Barth, R. P., Courtney, M., Berrick, J. D., & Albert, V. (1994). From child abuse to permanency planning. New York: Aldine de Gruyter.

Belsky, J. (1993). Etiology of child maltreatment: A developmental-ecological analysis. Psychological Bulletin, 114, 413– 434.

Cash, S. J. (2001). Risk assessment in child welfare: The art and science. Children and Youth Services Review, 23(11), 811– 830.

Connell, C. M., Vanderploeg, J. J., Katz, K. H., Caron, C., Saunders, L., & Tebes, J. K. (2009). Maltreatment following reunification: Predictors of subsequent child protective services contact after children return home. Child Abuse & Neglect, 33, 218– 228.

Coohey, C. (2006). Physically abusive fathers and risk assessment. Child Abuse & Neglect, 30, 467– 480.

DePanfilis, D., & Zuravin, S. J. (1998). Rates, patterns, and frequency of child maltreatment recurrences among families known to CPS. Child Maltreatment, 3(1), 27– 42.

DePanfilis, D., & Zuravin, S. J. (1999). Epidemiology of child maltreatment recurrences. Social Service Review, 73, 218– 239.

DePanfilis, D., & Zuravin, S. J. (2002). The effect of services on the recurrence of child mal- treatment. Child Abuse & Neglect, 26, 187– 205.

Dorsey, S., Mustillo, S. A., Farmer, E. M. Z., & Elbogen, E. (2008). Caseworker assessments of risk for recurrent maltreatment: Association with case-specific risk factors and rere- ports. Child Abuse & Neglect, 32, 277– 391.

Drake, B. (1995). Associations between reporter type and assessment outcomes in child pro- tective referrals. Children and Youth Services Review, 17, 503– 522.

Drake, B., Jonson-Reid, M., & Sapokaite, L. (2006). Rereporting of child maltreatment: Does participation in other public sector services moderate the likelihood of a second mal- treatment report? Child Abuse & Neglect, 30, 1201– 1226.

Drake, B., Jonson-Reid, M., Way, I., & Chung, S. (2003). Substantiation and recidivism. Child Maltreatment, 8(4), 248– 260.

Fluke, J., Shusterman, G., Hollinshead, D., & Yuan, Y. (2005). Rereporting and recurrence of child maltreatment: Findings from NCANDS. Washington, DC: U.S. Department of Health and Human Services, Office of the Assistant Secretary for Planning and Evaluation.

Child WelfareBae et al.

Fluke, J. D., Yuan, Y., & Edwards, M. (1999). Recurrence of maltreatment: An application of the national child abuse and neglect data system (NCANDS). Child Abuse & Neglect, 23, 633– 650.

Freeman, J. B., Levine, M., & Doueck, H. (1996). Child age and caseworker attention in child protective services investigations. Child Abuse & Neglect, 20(10), 907– 921.

Fryer, G. E., & Miyoshi, T. J. (1994). A survival analysis of the revictimization of children: The case of Colorado. Child Abuse & Neglect, 18, 1063– 1071.

Fuller, T. L., Wells, S. J., & Cotton, E. E. (2001). Predictors of maltreatment recurrence at two milestones in the life of a case. Children and Youth Services Review, 23, 49– 78.

Giovannoni, J. M. (1995). Reports of child maltreatment from mandated and non-mandated reporters. Children and Youth Services Review, 17, 487– 501.

Inkelas, M., & Halfon, N. (1997). Recidivism in child protective services. Children and Youth Services Review, 19, 139– 161.

Jagannathan, R., & Camasso, M. J. (1996). Risk assessment in child protective services: A canonical analysis of the case management function. Child Abuse & Neglect, 20, 599– 612.

Johnson, W. (1994). Effect of number of in-person visits on rates of child maltreatment recurrence. Paper presented at the Second Annual Roundtable on Outcome Measures in Child Welfare Services, San Antonio, TX.

Johnson, W., & L’Esperance, J. (1984). Predicting the recurrence of child abuse. Social Work Research and Abstracts, 29, 21– 26.

Jonson-Reid, M., Drake, B., Chung, S., & Way, I. (2003). Cross-type recidivism among child maltreatment victims and perpetrators. Child Abuse & Neglect, 27, 899– 917.

Karski, R. L. (1999). Key decisions in child protective services: Report investigation and court referral. Children and Youth Services Review, 21, 643– 656.

Kinnevy, S., Cohen, B., Huang, V., Gelles, R. J., Bae, H., Fusco, R., & Dichter, M. (2003). Evaluation of the transfer of responsibility for child protective investigations to law enforcement in Florida: An analysis of Manatee, Pasco, and Pinellas Counties. Philadelphia: University of Pennsylvania, Center for Research on Youth and Social Policy.

Lipien, L., & Forthofer, M. S. (2004). An event history analysis of recurrent child maltreat- ment reports in Florida. Child Abuse & Neglect, 28, 947– 966.

Vol. 89, No. 3Child Welfare

Marshall, D. B., & English, D. J. (1999). Survival analysis of risk factors for recidivism in child abuse and neglect. Child Maltreatment, 4, 287– 296.

Sidebotham, P. (2001). An ecological approach to child abuse: A creative use of scientific models in research and practice. Child Abuse Review, 10, 97– 112.

U.S. Department of Health and Human Services, National Center for Child Abuse and Neglect, National Child Abuse and Neglect Data System. (2004). Child maltreatment 2004: Reports from the states to the National Center on Child Abuse and Neglect. Washington, DC: U.S. Government Printing Office.

Way, I., Chung, S., Jonson-Reid, M., & Drake, B. (2001). Maltreatment perpetrators: A 54-month analysis of recidivism. Child Abuse & Neglect, 25, 1093– 1108.

Wells, S. J., Fluke, J. D., & Brown, C. H. (1995). The decision to investigate: Child protec- tion practice in 12 local agencies. Children and Youth Services Review, 17, 523– 546.

Wells, S. J., Lyons, P., Doueck, H. J., Brown, C. H., & Thomas, J. (2004). Ecological fac- tors and screening in child protective services. Children and Youth Services Review, 26, 981– 997.

Wolock, I., Sherman, P., Feldman, L. H., & Metzger, B. (2001). Child abuse and neglect refer- ral patterns: A longitudinal study. Children and Youth Services Review, 23(1), 21– 47.

Child WelfareBae et al.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.