Evaluation of Ohio's Maternal Mortality Surveillance System: Looking Back, Moving Forward

BACKGROUND: Maternal mortality has increased in Ohio and nationally, highlighting the importance of surveillance.  Re-established in 2010, the Ohio Pregnancy Associated Mortality Review (PAMR) is a passive system that uses the pregnancy status checkbox and O-codes on death certificates to ascertain pregnancy-associated deaths (deaths during or within one year after pregnancy).  PAMR compiles information from reporting sources and facilitates a review of circumstances surrounding deaths to identify opportunities for quality improvement and public health intervention to reduce mortality and morbidity.  A multidisciplinary committee meets triennially and has reviewed deaths occurring in 2008-2010.  We aimed to evaluate the operation of this relatively new system and provide recommendations to improve quality, efficiency and usefulness.

METHODS: CDC’s Updated Guidelines for Evaluating Public Health Surveillance Systems were followed. To evaluate usefulness and acceptability, Ohio licensed hospitals with a maternity unit (N=112) were surveyed electronically (74.1% response rate).  Data quality and timeliness were assessed by qualitative information obtained through key informant interviews, observation of a committee review meeting, and document review. 

RESULTS: PAMR has proven useful by providing estimates of maternal mortality and identifying associated factors. However, 46% of surveyed hospitals were unaware of PAMR prior to survey completion.  Limited data analyses were performed with spreadsheets.  Therefore, a lack of widespread dissemination of findings has minimized the system’s usefulness in bringing about systematic change and improving health outcomes.  Ninety-three percent of hospitals had concerns about transmitting records to PAMR; concerns included staffing, confidentiality, and legal authority.  The overall positive predictive value (PPV=74%) of current ascertainment methods indicates that about a quarter of identified cases are false positives, leading to a time-consuming case confirmation process and decreased system efficiency. Sensitivity of the checkbox is unknown.  Timeliness is reduced by slow case ascertainment and up to a six month delay in receipt of records from reporting sources.

CONCLUSIONS: PAMR’s usefulness and acceptability may be improved by broader analysis, dissemination of findings, and stakeholder engagement. Sensitivity of case ascertainment may be improved by data linkages. Efficiency and timeliness may be increased by improved checkbox accuracy, direct access of PAMR staff to vital statistics records, stakeholder engagement to speed receipt of records, and the adoption of a comprehensive data system.  Ohio’s participation in the Association of Maternal and Child Health Program’s (AMCHP) Every Mother Initiative — a three-year, six-state Action Learning Collaborative — provides framework and support for accomplishing improvements such as these, enhancing PAMR’s capacity to reduce maternal mortality and morbidity in Ohio.