FROM THE CHILD ABUSE RECOGNITION EXPERIENCE STUDY (CARES): PATIENT AND FAMILY FACTORS IN PRIMARY CARE PRACTITIONERS' (PCPS) DECISION TO REPORT CHILD ABUSE (CA)
E. Flaherty, R. Sege, J. Griffith, N. Dhepyasuwan, CARES Study Group, A Report for PROS, NMAPedsNet.. Pediatrics, Children's Memorial H., Chicago, IL; I. Clin. Res. & Health Policy, Tufts-New England Medical Ctr., Boston, MA; AAP, Elk Grove, IL; Washington, DC.

BACKGROUND: Previous studies have suggested that poverty, race and ethnicity influence PCP CA reporting, but these factors have not been examined in a large prospective study.

OBJECTIVE: Describe patient and family variables associated with PCP's decision to report suspected CA to Child Protective Services (CPS).

DESIGN/METHODS: 434 PCPs in CARES collected data about 15,003 child injury office visits. PCP level of suspicion (1-5 scale; 1=no suspicion) was recorded. Visits with no suspicion were excluded, resulting in 1685 eligible visits. PCPs described child demographics, familiarity with patient, if referred for suspected CA, injury severity/type, child risk factors (eg, injury not compatible with history, delay in care), family risk factors (eg, drug/ alcohol abuse, prior involvement with CPS), and if reported to CPS. Generalized Estimating Equations (GEE) were used to assess factors affecting PCP decisions to report suspected CA. Separate stepwise models were constructed to select most important factors among: demographics, familiarity, injury, and child and family risk factors. Variables included in these models were then included as candidate factors in a stepwise GEE model.

RESULTS: PCPs reported 95 patients (5.6%) to CPS. Independent factors associated with PCP reporting included: child injury not compatible with history (odds ratio [OR] 23.6, 95% confidence interval [CI] 11.5-48.5); patient referred to PCP for suspected CA (OR 10.4, 95%CI 4.1-26.5); number of family risk factors (for an increase of 1 between 0 and 2; OR 4.2, 95%CI 2.8-6.3); serious injury (OR 2.6, 95%CI 1.4-5.0); any child risk factor present other than injury not compatible (O.R. 2.4, 95%CI 1.2-4.6); and presence of a laceration (OR 0.3, 95%CI 0.1-0.7). In the final multivariate model, patient insurance type (a proxy for SES), race and ethnicity did not independently influence reporting.

CONCLUSIONS: When deciding about reporting to CPS, PCPs put greatest weight on whether an injury is compatible with the history and consider known patient and family factors. These factors appear to outweigh poverty, race and ethnicity considerations.