BACKGROUND: Maternal deaths are a serious public health concern that has tremendous family and societal impact. Georgia’s maternal mortality ratio (MMR) has increased over the last decade; from 18.0 per 100,000 live births in 2002 to 43.6 in 2011. Furthermore, racial/ethnic disparities exist; in 2002-2011, the MMR was four times higher in Black, non-Hispanics (39.2) than their White counterparts (9.7). The staggering trend and the underlying racial/ethnic disparities served as the impetus for the creation of a statewide maternal mortality review committee (MMRC) in 2012. The mission of the Georgia (GA) MMRC is to identify pregnancy-associated deaths, review those caused by pregnancy complications and other selected deaths, and identify problems contributing to these deaths and interventions that may reduce these deaths. However, there are inherent barriers that make the efficient identification and review of maternal deaths difficult. The timely and accurate identification of maternal death cases are crucial to the successive GA MMRC review of cases, recommendations and implementation of interventions. As the GA MMRC has developed, the case identification processes have been evaluated to improve efficiency.
METHODS: Three empirical (mandatory reports, vital records death registration pregnancy checkbox and International Classification of Diseases, 10th Revision, cause-of-death codes in the O series) and two analytical (deterministic and probabilistic linkages) data-gathering strategies were employed to identify maternal deaths. Strategies were evaluated according to timeliness, complexity, time involved, accuracy and comparability over time.
RESULTS: Between 2012 and 2013, 266 possible maternal death cases were identified through combined strategies. Upon abstraction by the GA MMRC, 80 (30%) possible cases were found to have no evidence of pregnancy (false positives); empirical strategies yielded over 90% of false positives. However, empirical strategies (mandatory reports and pregnancy checkbox) resulted in more timely case identification; specifically, mandatory reports were the timeliest (within 30 days) method. Analytical strategies yielded fewer (10%) false positive cases and higher comparability over time; however, these methods were more complex, required more time, and less timely.
CONCLUSIONS: Multiple strategies are necessary for comprehensive maternal death case identification. However, with the execution of strategies accuracy and timeliness are inversely related. Quality assurance efforts, including a feedback loop for vital records certifiers, are needed to improve the accuracy of strategies ranked highest for timeliness.