RESUMO
BACKGROUND: The predictors of postpartum severe maternal morbidity and mortality have not been well-described using patient-level data. OBJECTIVE: This study aimed to evaluate the application of maternal early warning criteria in the postpartum period and generate a preliminary predictive model for severe maternal morbidity and mortality occurring after delivery hospitalization discharge until 42 days postpartum. STUDY DESIGN: A retrospective case-control study was conducted from January 2013 to September 2020. Cases were identified from electronic medical records using the International Classification of Diseases, Tenth Revision codes for Centers for Disease Control and Prevention-defined severe maternal morbidity. Patients meeting the criteria for severe maternal morbidity and mortality from delivery hospitalization discharge until 42 days postpartum were matched for delivery hospital and year with the controls in an approximate 1:2 fashion. The objective was to identify the demographic and clinical risk factors during the antepartum through postpartum periods for postpartum severe maternal morbidity and mortality. Multivariable logistic regression was performed to estimate the risks, and a receiver operating characteristic curve was derived to evaluate the model. RESULTS: Ninety cases of postpartum severe maternal morbidity and mortality that occurred following delivery hospitalization discharge were identified. These were matched with 175 controls. Women with postpartum severe maternal morbidity and mortality had more postpartum assessments (mean: 1.7 vs 1.4, P=.005) and a higher frequency of maternal early warning criteria (58% [52/90] vs 2% [3/175]; P<.001) preceding the diagnosis of severe maternal morbidity and mortality than controls. Black women had higher odds of postpartum severe maternal morbidity and mortality than White women (odds ratio, 1.93; 95% confidence interval, 1.14-3.27). Women with maternal early warning criteria during postpartum assessments were more likely to experience subsequent postpartum severe maternal morbidity and mortality (odds ratio, 67.2; 95% confidence interval, 21.3-211.6) than women with no maternal early warning criteria. Although the point estimate was different in Black women (odds ratio, 161.8; 95% confidence interval, 8.9 to >999) than White women (odds ratio, 47.9; 95% confidence interval, 13.8-167.1), the effect modification between the maternal early warning criteria and race was not statistically significant (P=.93). In a multivariable model, race, body mass index, cesarean delivery, and maternal early warning criteria at postpartum assessments were associated with subsequent severe maternal morbidity and mortality, with an area under the curve of 0.905 (95% confidence interval, 0.864-0.946). CONCLUSION: Maternal early warning criteria are associated with increased odds of postpartum severe maternal morbidity and mortality. A straightforward model that includes race, body mass index, cesarean delivery, and presence of maternal early warning criteria appears to be a promising tool to identify those at risk for postpartum severe maternal morbidity and mortality following delivery hospitalization discharge. This is an important first step in improving the ability to recognize and respond to conditions preceding postpartum severe maternal morbidity. These findings should be validated in a prospective cohort.