A Public Health Approach to Preventing
Maltreatment-Related Child Deaths
(Originally published August 2011)
Two recent studies by Emily Putnam – Hornstein and (in one study) Barbara Needell suggest the potential of a public health approach to prevention of child abuse / neglect related fatalities. Putnam – Hornstein linked data for 4.3 million children born in California between 1999 and 2003 with Confidential Death Master Files from California’s Department of Public Health (CDPH) and CPS records from California’s statewide child welfare information database maintained by the Center for Social Services Research at the University of California at Berkeley.
“A probabilistic linkage methodology was employed to establish linkages between records in all three files,” Putnam – Hornstein states. This methodology identified 4,317,321 live births between 1999 and 2006, 596,962 children with a CPS record during these same years, 25,496 deaths which occurred before children’s fifth birthday and 1,917 child deaths of children younger than age 5 due to intentional or unintentional injuries. Approximately 14% of CPS records were not matched to a birth record, and 1.5% of death records were unmatched to a birth record. All children reported to CPS (including screened out cases) were included in the analysis. Children reported to CPS on or after the date of a fatal injury were recoded as “non – CPS” since there had been no opportunity for CPS to intervene prior to a child’s death. Multivariate models were used to control for socio-demographic variables such as child’s sex (male vs. female), child’s health (risk present vs. no risk), birth payment method (public insurance vs. private insurance), race / ethnicity (Black vs. White, Hispanic vs. White, Asian / Pacific Islanders vs. White), maternal age at birth (24 years or less vs. 25 years or older), maternal education at birth (high school degree or less vs. some college or college degree), established paternity on the birth record (missing vs. established), and birth order (later born vs. first born).
Less than 1% of children reported to CPS died of injuries prior to age 5; child deaths from unintentional injuries far outnumbered deaths from intentional injuries.
20.7% of children who died from injuries (intentional and unintentional) prior to age 5 had been reported to CPS prior to death; a third of children who died from intentional injuries had prior CPS reports compared to 11.9% of all children born in California from 1999- 2006.
Paternity was missing in almost one fifth of fatal injuries of young children and one third of intentional injuries compared to 9.4% of 4.3 million children born in California from 1999- 2006.
Almost half (49.6%) of mothers of children with fatal injuries were 24 years old or less at the deceased child’s birth; two thirds of mothers of children who died from intentional injuries were 24 years old or less at the deceased child’s birth compared to about one third of the mothers of all children born in California from 1999-2006.
72.7% of mothers of children with fatal injuries had a high school education or less; almost four fifths (79.8%) of mothers of children who died of intentional injuries had a high school education or less compared to 57.3% of mothers of all
children born in California from 1999- 2006.
13.2% of children who died from injuries were Black; and 21.7% of children who died from intentional injuries were Black compared to 6.1% of all children born in California from 1999- 2006. Asian Pacific Islander and Hispanic children were underrepresented among injury fatality victims.
Birth payment was labeled Public for 56.3% of children who died of injuries and almost two thirds (64.2%) of children who died of intentional injuries compared to 43.5% of all children born in California from 1999- 2006.
Multivariate Analysis – after controlling for baseline and family characteristics
Children who had been reported to CPS were fatally injured at a rate 2.5 times as great as unreported children.
Children reported to CPS died of unintentional injuries at twice the rate of unreported children.
Children reported to CPS died of intentional injuries at almost six times (5.86) the rate of unreported children.
Putnam – Hornstein comments that “a prior, nonfatal allegation of maltreatment is an
independent risk factor for injury death during the first 5 years of life in California,” in fact “In the fully adjusted model, an earlier (CPS) report emerged as the strongest predictor of all manners of injury death.” Putnam- Hornstein asserts that “the findings
of the current study lend support to researchers… (who) have called for CPS to be pursued under a broader, public health – oriented agenda, focused on the prevention of all manners of injury..”
How a public health approach to preventing maltreatment related injuries might work is suggested by a study recently published by Putnam – Hornstein and Needell. This study linked child welfare data to birth records for children born in California in 2002; . Sixteen percent of child welfare records could not be linked to a birth record. All children reported to CPS before the age of 5 were included in the study. Twelve variables included in the birth record were selected for analysis:
1. Child’s sex
2. Birth weight
3. Prenatal Care
4. Birth abnormality
5. Maternal Birth place
6. Maternal Race / Ethnicity
7. Maternal Age
8. Maternal Education
9. Abortion History
10. Father Information
11. Total Children Born to Mother
12. Medi-Cal Status
Over half a million children (531,035) were born alive in California in 2002, and 74,182 children (13.9%) were referred to CPS due to alleged abuse or neglect before their fifth birthday.
No CPS Report CPS Report
Low Birth Weight 6.1% 8.6%
Prenatal Care 88% 76%
Teenage Mother 8% 18%
Mother Younger than 25 30% 50%
US Born Mother 50% 30%
No Paternity Information 5.5% 17.6%
Medi –Cal Coverage 39% 65%
Percentage of children who were referred to CPS before their fifth birthday by race / ethnicity
34% of Native American children
30% of Black Children
14% of Hispanic children
13% of White children, and
5% of Asian / Pacific Islander children .
Multivariate Analysis – after adjusting for all other birth variables:
Birth to a mother who immigrated to the U.S. was a strong protective factor.
For children third or higher in the birth order, the risk of a CPS referral was doubled.
The associations between low educational achievement and CPS referral “were notably dampened for children born on Medi –Cal (i.e., poor children) and amplified for children who are not (on Medi- Cal at birth, i.e., non – poor children).”
Both Black and Native American children were significantly more likely than White children to have been referred to CPS by age 5, though Black children with Medi – Cal coverage at birth were less likely than White children to have been referred to CPS.
Lack of prenatal care doubled the risk of CPS referral for children not covered by Medi- Cal; children covered by Medi – Cal with no prenatal coverage were 1.5 times more likely to be referred to CPS compared to poor children whose mothers had received prenatal care.
A maternal history of one or more abortions was a significant risk factor for CPS referral.
These authors maintain “that objective data collected at birth can be used to identify those children in a given birth cohort who are at greatest risk of future CPS contact.” And they continue ,”if we were to utilize just a handful of risk factors for CPS contact (prenatal care that began after the first trimester, missing father information, high school degree or less for mothers, three or more children in the family, Medi- Cal coverage for U.S. born mothers) to classify as “high risk” any child born with three or more of these risk factors, we could identify 50% of children referred to CPS before the age of five from just 15% of the total birth cohort.”
Putnam – Hornstein’s and Needell’s view is that services targeted at high risk families as identified by a handful of risk factors immediately following a child’s birth or prior to a child’s birth would be a cost effective approach to prevention at a time of diminished resources and state governments’ budget cuts. Their analysis of a California birth cohort indicates the plausibility of this approach to prevention. However, they do not propose specific services or programs or comment on the empirical support for prevention. Studies of Nurse Family Partnership home visitation program and the Chicago Parent Child Centers, the two programs with the strongest evidence that they reduce child maltreatment, have found that much of these programs’ preventive effects on maltreatment rates are not evident until the 15 year follow up. The largely delayed effects of these programs on maltreatment rates and out- of home placements has not been adequately explained. It is both fascinating and encouraging that prevention programs that serve families for years can have a range of positive influences on family and child outcomes that extend for a couple of decades or more. Nevertheless, the question remains of what to do on behalf of high risk children beginning prior to or at birth that would protect children from all manner of injuries, including fatal injuries, during the first years of life.
What risk assessment tools do best is target families for services, but if services are ineffective, nothing much is gained by targeting families for special attention. This issue has plagued the use of risk assessment tools in child protection, and may do the same in a large scale public health approach to prevention absent bigger investments in research of prevention programs.
There are, however, some lessons that can be learned from past mistakes in child welfare, for example, the federal government’s “all in” bet on family preservation programs in the 1990s.
Policymakers and practitioners should not be looking for a single program or two to meet the needs of diverse populations.
Despite their limitations and cost, there is no substitute for long term experimental and quasi- experimental studies.
Programs that address co - occurring substance abuse / mental health disorders prior to a CPS referral and include persistent outreach to troubled families are urgently needed. Ditto for domestic violence interventions for both victims and batterers.
Allowing almost 10% of U.S. children to grow up in severely poor (with incomes less than $10,000 per year) families guarantees that pervasive and intractable child neglect symptomatic of family breakdown will be an ongoing child welfare dilemma.
Policymakers and practitioners should also ponder the power of social attitudes to reduce maltreatment rates over a few decades. Arguably, the decline in sexual abuse and severe physical abuse of children during the past 15-18 years has not been the result of prevention programs, but of ‘zero tolerance’ social attitudes expressed in a large variety of ways, including CPS decision making, police behavior, tort actions and intense social disapproval of these forms of child maltreatment. Once a society or community has a collective intent to prevent specific forms of maltreatment and or other early adverse experiences, the effects of prevention services and programs will be greatly enhanced.
Putnam – Hornstein, Emily, “Report of Maltreatment as a Risk Factor for Injury Death: A Prospective Birth Cohort Study,” Child Maltreatment , 2011.
Putnam – Hornstein, Emily, Needell, Barbara, “Predictors of Child Protective Service Contact between Birth and Age Five: An Examination of California’s 2002 Birth Cohort,” Children and Youth Services Review,”