Comparison of Incidence of Severe Hyperbilirubinemia Between HFHSF Newborns and the Newman et al Study2
| Variables | Newman et al Study | HFH:HFHSF Group | Relative Risk* (RR) After Controlling for Main Effect (Newman et al Study as Reference Group) | |||
|---|---|---|---|---|---|---|
| Total Newborns(n = 51 387)(%) | Newborns With TSB ≥ 20 mg/dL(n = 1002)(%) | Total Newborns(n = 3498)(%) | Newborns With TSB ≥ 20 mg/dL(n = 21)(%) | RR | 95% CI | |
| Total cohort | 100 | 2.0 | 100 | 0.6 | ||
| Gestational age in weeks | 0.26 | 0.17–0.44 | ||||
| 35 | — | — | 1.7 | 1.7 | ||
| 36 | 2.9 | 5.2 | 4.0 | 2.2 | ||
| 37 | 5.9 | 5.7 | 7.6 | 1.9 | ||
| 38 | 13 | 3.2 | 19.6 | 0.3 | ||
| 39 | 24.2 | 1.7 | 26.8 | 0.2 | ||
| 40 | 36.8 | 1.4 | 26.9 | 0.4 | ||
| 41 | 14.3 | 0.7 | 12.0 | 1.0 | ||
| ≥42 | 2.9 | 0.6 | 1.4 | 0.0 | ||
| Gender | 0.31 | 0.20–0.48 | ||||
| Male | 51.5 | 2.3 | 50.4 | 0.7 | ||
| Female | 48.5 | 1.6 | 49.6 | 0.5 | ||
| Maternal race | 0.49 | 0.31–0.76 | ||||
| White | 52.7 | 1.7 | 14.4 | 1.2 | ||
| Black | 18.5 | 1.0 | 67.2 | 0.4 | ||
| Latino | 15.8 | 1.7 | 10.4 | 0.8 | ||
| Asian | 9.3 | 3.9 | 2.1 | 1.3 | ||
| Other/unknown | 3.7 | 1.9 | 5.8 | 0.5 | ||
| Maternal age in years | 0.33 | 0.21–0.50 | ||||
| <20 | 8.5 | 1.3 | 13.1 | 0.0 | ||
| 20–24.9 | 17.2 | 1.5 | 27.6 | 0.6 | ||
| 25–29.9 | 28.2 | 2.0 | 26.3 | 0.1 | ||
| 30–34.9 | 28.5 | 2.3 | 20.1 | 1.1 | ||
| ≥35 | 17.7 | 2.1 | 12.9 | 1.3 | ||
| Feeding type† | 0.49 | 0.32–0.76 | ||||
| Exclusive breastfeeding | 65.5 | 2.7 | 30.2 | 0.8 | ||
| Partial breastfeeding | 11.4 | 0.9 | 24.9 | 0.8 | ||
| Formula feeding | 23.1 | 0.4 | 45.0 | 0.2 | ||
The distribution of the two samples was significantly different in gestational age (P < .0001), gender (P = .018), maternal race (P < .0001), and maternal age (P < .0001).
↵* This is the overall relative risk of developing an age-specific TSB exceeding AAP criteria at HFHS versus the hospitals in the Newman et al study. Each variable was controlled for separately because cross-tabulation summary tables were not available. For example, after controlling for maternal race, the regression coefficient for the HFHSF group in the Poisson regression analysis is −0.722, indicating that the relative risk for developing severe hyperbilirubinemia is twofold lower in HFHSF newborns than in the Newman et al study (relative risk for the HFHSF group: Newman et al study = e−0.722 = 0.49).
↵† Applied feeding distribution reported in a nested-case control study that used the same cohort for two different studies.12