Related Experiment Video
Updated: Jul 31, 2025
Inverse Probability of Treatment Weighting Propensity Score using the Military Health System Data Repository and National Death Index
Published on: January 8, 2020
Replication and validation of a statewide linkage method to estimate incidence proportion of child maltreatment
Abigail Newby-Kew1, Lynn M Marshall1, Suzanne Zane2
1Oregon Health Science University and Portland State University School of Public Health, 1810 SW 5th Ave, Portland, OR.
Purpose:
To study familial factors associated with child maltreatment in a birth population, Alaska piloted a mixed-design method that linked child welfare data with the Pregnancy Risk Assessment Monitoring System (PRAMS). We replicated this approach in Oregon and validated it in both states.
Methods:
We linked vital records, child welfare, and PRAMS data to create two 2009 birth cohorts for each state: one based on vital records (full birth cohort) and one on PRAMS (stratified random sample). For each cohort, we estimated the incidence proportions (IP) of child maltreatment before age 9 years and compared those estimated using PRAMS with those observed using the full birth cohort.
Results:
The Oregon PRAMS cohort estimated that 28.7% (95% CI: 24.0, 33.4), 20.9% (17.1, 24.7), and 8.3% (6.0, 10.5) of children experienced an alleged, investigated, and substantiated maltreatment, respectively, versus 32.0%, 25.0%, and 9.9% from the birth cohort. The corresponding Alaska estimates were 29.1% (26.1, 32.0), 22.6% (19.9, 25.2), and 8.3% (6.7, 9.9) of children from the PRAMS cohort versus 29.1%, 23.5%, and 9.1% in the birth cohort.
Conclusions:
The IP of child maltreatment in two states was accurately estimated with PRAMS cohorts. Researchers can study a comprehensive set of factors that may influence child maltreatment by incorporating PRAMS into birth cohort linkages.
