Maternal mortality is a sentinel public health indicator essential to the measurement of health care quality both nationally and internationally; yet, the United States has not published an ?official? U.S. maternal mortality rate since 2007. This has created an information deficit at a time when more international attention has been focused on maternal mortality than ever before. For example, the United Nations Millennium Development Goal 5a was to reduce maternal mortality by 75% worldwide between 1990 and 2015. Maternal mortality reduction also figures prominently in the newer United Nations Sustainable Development Goals. Recently, major problems in the collection and coding of U.S. maternal mortality data have been identified, which make current estimates of maternal mortality levels and trends highly unreliable. This study analyzes the 2015-16 cause of death literal data (actual words written in the cause-of-death section of the death certificate) to identify and correct problems in data collection and coding of maternal deaths. Specifically, we will determine which records are not maternal deaths and represent errors in data collection and coding by correctly excluding incidental causes of death according to the World Health Organization maternal mortality definition. For records with non-specific causes of death that are appropriately classified as maternal deaths, we will develop supplementary coding methods which allow the cause of death data to be coded to specific organ systems and disease pathways, thus increasing the specificity of reported maternal mortality data. Newly developed maternal mortality estimates will be compared to those from the previous inaccurately coded data to assess the degree of bias in the previously coded data. In the second part of the project, we will use the more accurate and detailed maternal mortality data developed in Aim 1 to analyze maternal mortality patterns by socio- demographic and health variables and cause of death. Bivariate and multivariate methods will examine associations between these variables and the maternal mortality rate. Major contributions of the proposed research include: 1) a better understanding of maternal mortality data quality; 2) development of improved methods for analyzing maternal mortality data; 3) concrete recommendations to improve data quality needed to restart production of U.S. maternal mortality rates; and 4) more accurate estimation of U.S. maternal mortality rates and disparities by socio-demographic characteristics, cause of death, and region. Producing accurate maternal mortality rates is essential for national and international reporting, and for monitoring progress toward meeting the United Nation?s Sustainable Development Goals. The identification of specific and potentially preventable causes of maternal death will lead to efficient targeting of prevention efforts towards the most problematic causes of death. More detailed knowledge of maternal mortality disparities (for example, by maternal age, race/ethnicity, or region) will lead to the more accurate identification of at-risk populations, essential to effective targeting of prevention programs.
Recent studies have identified gross inaccuracies in United States maternal mortality data, due to problems related to both case ascertainment, and data coding. We investigate these inaccuracies, through examination of the cause-of-death literal text (actual words written in the cause-of-death section of the death certificate), and develop supplementary analytical methods to produce more accurate U.S. maternal mortality estimates by characteristics.
