Determinants of maternal near miss among women in public hospital maternity wards in Northern Ethiopia: A facility based case-control study

Background: In Ethiopia, 20,000 women die each year from complications related to pregnancy, childbirth and post-partum. For every woman that dies, 20 more experience injury, infection, disease, or disability. “Maternal near miss” (MNM), defined by the World Health Organization (WHO) as a woman who nearly dies, but survives a complication during pregnancy, childbirth or within 42 days of a termination, is a proxy indicator of maternal mortality and quality of obstetric care. In Ethiopia, few studies have examined MNM. This study aims to identify determinants of MNM among a small population of women in Tigray, Ethiopia. Methods: Unmatched case-control study was conducted in hospitals in Tigray Region, Northern Ethiopia, from January 30-March 30, 2016. The sample included 103 cases and 205 controls recruited from among women seeking obstetric care at six (6) public hospitals. Clients with life-threatening obstetric complications, including hemorrhage, hypertensive diseases of pregnancy, dystocia, infection, and anemia or clinical signs of severe anemia (in women without hemorrhage) were taken as cases and those with normal obstetric outcomes were controls. Cases were selected based on proportion to size allocation while systematic sampling was employed for controls. Binary and multiple variable logistic regression (“odds ratio”) analyses were calculated at 95% CI. Results: Roughly 90% of cases and controls were married and 25% experienced their first pregnancy before the age of 16 years. About two-thirds of controls and 45.6% of cases had gestational ages between 37–41 weeks. Among cases, severe obstetric hemorrhage (44.7%), hypertensive disorders (38.8%), dystocia (17.5%), sepsis (9.7%) and severe anemia (2.9%) were leading causes of MNM. Histories of chronic maternal medical problems like hypertension, diabetes were reported in 55.3% of cases and 33.2% of controls. Women with no formal education [AOR = 3.2;95%CI:1.24, 8.12], being less than 16 years of age at first pregnancy [AOR = 2.5;95%CI:1.12,5.63], induced labor[AOR = 3.0; 95%CI:1.44, 6.17], history of cesarean section[AOR = 4.6; 95% CI: 1.98, 7.61] or chronic medical disorder[AOR = 3.5;95%CI:1.78, 6.93], and women who traveled more than 60 minutes before reaching their final place of care[AOR = 2.8;95% CI: 1.19,6.35] had higher odds of experiencing MNM. Conclusions: Macro-developments like increasing road and health facility access as well as expanding education will all help reduce MNM. Work should be continued to educate women and providers about common predictors of MNM like history of C-section and chronic illness as well as teenage pregnancy. These efforts should be carried out at the facility, community, and individual levels. Targeted follow-up with women with history of chronic disease and C-section could also help reduce MNM. The study was conducted in six (6) hospitals in Tigray, Ethiopia, from January 30 to March 30, 2016. Hospitals were randomly selected from 16 public hospitals in the region[14]. The study was a facility-based, unmatched case control design. Sample size was estimated using a double population proportion formula based on a study from Morocco that showed hypertensive disease contributing the most to MNM [15]. Based on the Morocco study, we hypothesized the proportion of chronic hypertension to be double in cases (63.9%) and controls (47%) at a 95% confidence level and 80% power of the test, with a 1:2 ratio for cases and controls. Final sample size was 308, of which 103 were cases and 205 controls. We considered MNM as a condition meeting any of the five disease-specific criteria proposed by Filippi [16]. In sampled hospitals, using medical notes, any woman diagnosed with at least one of the following complications was considered as a case: severe obstetric hemorrhage leading to shock; hypertensive diseases of pregnancy, including eclampsia and severe preeclampsia; dystocia, including uterine rupture and impending rupture; infections, including hyper- or hypothermia or a clear source of infection and clinical signs of shock, and; anemia, including low hemoglobin (<6 g/dl) or clinical signs of severe anemia in women without hemorrhage. Women not meeting the above criteria were considered as controls. Cases were sequentially recruited whereas controls were selected through systematic sampling. Data was collected using a structured questionnaire, administered in-person by nurse midwives. Socio-demographic characteristics, obstetric history, and knowledge of pregnancy-related danger signs were collected. Questionnaire was based on tools validated by the World Health Organization (WHO) and in different literature and adapted to include context-specific factors [11–13, 15, 17]. Questionnaire was prepared in English, translated to Tigrigna, and back-translated to English separately by two individuals to ensure consistency. Data was collected by 12 nurse midwives with experience in obstetric care. Data collection was supervised and data checked for consistency and completeness. Incomplete and unclear questionnaires were returned to interviewers to be completed. Data was entered, cleaned and analyzed using SPSS 20. Data was cleaned by running frequencies, cross-tabulation and sorting cases. Bar graphs and frequencies were used to represent results of categorical variables. Bivariate and multivariate logistic regression analyses were used to determine the association of independent variables with the dependent variable. Variables with p<0.25 in bivariate analysis were entered into a multivariate logistic regression model. Odds ratios with 95% confidence were computed to identify the presence and strength of associations, and statistical significance was declared if p<0.05 was found. The final model was checked using the Hosmer–Lemeshow goodness of fit test. Co-founders, interaction and multi-collinearity were checked to minimize bias. Study protocol was approved by the Institutional Research Review Board of Mekelle University’s College of Health Sciences and Community Services Ethical Review Committee. Permission was obtained from Tigray Regional Health Bureau and participating hospitals. Informed verbal consent was obtained from participants prior to enrollment in the study. Participation in the study was voluntary and participants were informed of the right to withdraw from the study. Data collection was conducted confidentially and data de-identified, de-linked and stored in a secure location.