Abstract
BACKGROUND: The safety of out-of-hospital birth in the United States remains contested. A neglected methodological issue is the selective nonreporting of 5-min Apgar scores, which may conceal adverse outcomes and bias safety comparisons. This study examined whether Apgar score missingness differs systematically by birth setting and whether such "informative missingness" alters risk estimates. METHODS: We conducted a population-based analysis of 3,066,021 term, normal-birthweight, midwife-attended singleton births in the United States (2016-2023). Birth settings included hospitals, freestanding birth centers, and planned home births. Missing 5-min Apgar scores were quantified, and deterministic sensitivity analyses modeled the impact of varying assumptions about unrecorded low scores (<4 and <7). Hospital births served as the reference group. FINDINGS: Five-minute Apgar scores were missing in 0.13% of hospital, 1.9% of birth-center, and 3.1% of home births. Severe compromise (Apgar <4) occurred in 0.17%, 0.20%, and 0.26%, respectively. When half of missing scores were imputed as <4, adjusted odds of severe compromise increased to 7.7 for home and 4.9 for birth-center births vs. hospitals. INTERPRETATION: This study evaluates documentation integrity of US births. Selective nonreporting of 5-min Apgar scores at out-of-hospital births introduces major bias, distorting apparent safety of out-of-hospital births. Complete and enforceable outcome reporting is essential for scientific validity and ethically sound informed consent. FUNDING: None declared.