Urban pediatric services and high risk environments often go hand in hand.“High risk” is often taken at face value without exploring community differences. Factors contributing to risk in families may vary, needing different community-based pediatric intervention. Caregivers of young children were interviewed in 2 studies in 2 neighboring cities. Study 1 (S1) included 311 families recruited by zip code while waiting in the resident continuity clinic at an academic medical center in a low-income, inner city neighborhood. Study 2 (S2) included 142 families of low/moderate income in a smaller city identified by census tract, birth records and day care waiting lists. As part of 2 larger studies, subjects responded to the CRISYS, a 50-item survey of life stressors and the CES-D scale, a 16-item measure of depression. Community comparison by t-test used the 10 content domains of the CRISYS.

The percentage of depressed caregivers (CES-D > 15) was similar in both samples (S1:34%; S2:32%) with comparable CES-D means (S1:13.0 ± 11.3; S2:13.6 ± 11.1, t-value 0.5, p = 0.6). The mean number of life stressors experienced in the 6 mo reporting period was different (S1:8.8± 4.8;S2:6.3 ± 4.4;t-value 5.4,p = 0.001). In both S1 and S2, the sum of stressors predicted CES-D scores(S1:R2 = 0.22, p < 0.001; S2:R2 = 0.26, p < 0.001). Subjects in S1 showed significantly higher rates of life stress than S2 in the following areas: finances (t-value = 3.4, p < 0.01); personal health (t-value = 6.3, p < 0.001); family/friend health (t-value = 3.7, p < 0.001); community safety (t-value = 6.4, p < 0.001). Content domains with similar rates of endorsement in the 2 samples were legal and career issues, interpersonal relationships, family violence, housing and relationships with authority.

Using the 10 content domains as predictors of depression revealed 2 very different patterns for the 2 samples. In S1, 6 of 10 content domains contributed to depression (R2 = 0.29, p < 0.001); in S2, 3 domains predicted depression (R2 = 0.32, p < 0.001). Only family violence predicted depression in both samples. A high rate of maternal depression in both samples supports the families' characterization as “at risk,” but the antecedents and thus the intervention would vary. Further analyses examine within-community differences.