Every case counts: Improving abusive head trauma detection in Alaska through Multi-source data linkage

INTRODUCTION: Serious abuse resulting in traumatic brain injury has severe health consequences.  Nearly 80% of abusive head trauma (AHT) occurs among children <2 years of age, with infants experiencing an incidence nearly 8 times that of 2 year olds.  This study describes the validation of the CDC Pediatric Abusive Head Trauma (PAHT) definitions when applied to a multi-source database at the state level, and provides a sensitive incidence estimate of AHT among children <2 years of age in Alaska.

METHODS: AHT cases among children in Alaska during 2005–2010 were identified by applying the PAHT coding schema to a multi-source database which included death records, Violent Death Reporting System (AK-VDRS), Maternal Infant Mortality Review – Child Death Review (MIMR-CDR), Alaska Trauma Registry (ATR), inpatient Hospital Discharge Database (HDD), and Medicaid claims. Using these data a statewide AHT annual incidence was calculated.  A validation study was conducted by comparison of captured cases with a subset of extracted cases from Alaska’s largest pediatric tertiary care center.

RESULTS: The databases with the highest case capture rates were the ATR and Medicaid systems, both at 51%, followed by HDD at 38%. Combined, the ATR, HDD, and Medicaid systems captured 91% of all AHT cases.  The linkage and use of the PAHT definitions yielded an estimated sensitivity of 91% and specificity of 98%.  During the study period we detected an annual average incidence of 34.4 cases per 100,000 children aged <2 years (95% CI 25.1, 46.1) and a case fatality proportion of 22% (10/45).  Among the AHT cases 82% were infants. Significant differences (p<0.05) in AHT were noted by age and race, but not sex.

CONCLUSIONS: In Alaska, applying the CDC PAHT definition to the multi-source database enabled us to capture 49% more AHT cases than any of the individual database used in this analysis alone.