Evaluation of the Completeness of the American Hospital Association Database for Studying Maternal Risk-Appropriate Care

Background In a prior study we presented a method to identify a hospital level of maternal care using the American Hospital Association Annual Survey Database. However incomplete survey data prevented assessment of hospital level of maternal care for 558 hospitals, which may limit use of the data to study state policies to increase use of maternal risk-appropriate care. The purpose of this study is to determine the extent to which this sample represents the universe of maternity hospitals and the ability to measure hospital level of maternal care for each state; this measurement is necessary to study maternal risk-appropriate care. Methods We compared the number of births reported in the 2021 American Hospital Association Annual Survey Database to the number of births in the 2020 National Vital Statistics Data. The completeness of the sample was measured as the ratio of the number of births reported by hospitals in the state to the number of birth certificates reporting hospital births. Data was considered complete for a state if the ratio was within one standard deviation of the national average of completeness. Completeness of survey data was measured as the proportion of hospitals whose level of maternal care could be evaluated. Results Hospitals in the database were the location for 97.6% of all live hospital births in 2020. The average state level of completeness was 0.99 (SD 0.10); four state ratios were outside one standard deviation of the national average of completeness and one state ratio was outside two standard deviations. We measured hospital level of maternal care for 1,923 (69.1%) of maternity hospitals which were the location for 77.2% of all hospital live births. The average state completeness was 0.75 (SD 0.21); six states were outside one standard deviation and six were outside two standard deviations. We identified wide variation in completeness between states. Conclusions This study demonstrated that the American Hospital Association Database can be used to describe the distribution of maternity hospitals in the United States. The completeness of level of maternal care data for several states allows use of the database to study maternal risk-appropriate care. Improving data completeness would improve the ability to study maternal risk-appropriate care.