Development of an Early Warning System to Prevent Severe Maternal Outcomes in Humanitarian Settings

Introduction: Obstetric Early Warning Systems (OEWS) improve maternal morbidity outcomes across various settings by enhancing emergency obstetric care and facilitating early intervention. However, there is limited research analyzing the effectiveness and implementation of OEWS in low-resource, low-complexity care settings, and none to date in humanitarian settings. This study aims to develop and validate a Modified Early Obstetric Warning System (MEOWS) to predict severe maternal outcomes in refugee camp health centers in Cox’s Bazar, Bangladesh. Methods: A retrospective case-control methodology was used, and de-identified obstetrics electronic medical records (EMR) from October 2017 to June 2024 were obtained via the camp’s EMR system, NIROG, at two sites: Kutupalong and Balukhali. Data was split into developmental and validation datasets. Primary predictor variables were systolic blood pressure, diastolic blood pressure, and heart rate; secondary predictors included age, BMI, gravida, gestational age at birth, and elevated fasting blood glucose (FBG). Severe maternal outcomes were the primary outcome variable based on ACOG’s Obstetrics Care Consensus and WHO’s near-miss criteria. A stepwise logistic regression was used to create the model. Results: Among 446 patients with complete data, 36 experienced SMOs. Controls were randomly selected and included at a 4:1 control-to-case ratio. At a 5% level of significance, only elevated FBG was associated with SMOs. Both the developmental and validation datasets revealed a fourfold higher likelihood of SMOs in patients with elevated FBG at 80% and 90% confidence, respectively. However, 43% of those with elevated FBG did not have a diabetes diagnosis. Conclusion: These findings suggest that elevated FBG should be included in obstetrics early warning systems for the camp health centers and similar settings, as FBG is not regularly included in standard obstetrics early warning systems. Furthermore, the lack of associated diabetes diagnosis suggests a possible area for patient care improvement within the camp health centers in Cox’s Bazar.