Maternal death and delays in accessing emergency obstetric care in Mozambique

Table 1 below presents the women’s demographic characteristics, previous obstetric history, facility characteristics, timing and cause of death (direct or indirect). Also this table presents the distribution of proportion of delays within the selected characteristics.

Most women (62.1%) were between 20 and 34 years old at the time of their death. The median age was 25 years and only 18.4% were nulliparous. Most women arrived at the health facility with an obstetric complication (78.4%), and 53.1% were referred from another health facility. Delay II (reaching the health facility where she died) was observed in 40.4% of deaths and delay III (receiving treatment after arrival at the health facility) was registered in 14.2% of cases, 13.9 faced the both type of delays and in 18.5% of cases no evidence of delay was identified. Direct obstetric complications were registered as the main cause of death for most cases (73.7%).

Around 10% of women developed complications after admission in the health facilities. Out of this group, 53% of deaths occurred at the tertiary hospitals while 35% at the health Centers and 12% at the Central Hospitals. The main causes of death in this group were post-partum hemorrhage which accounted for 28% followed by uterine rupture responsible of 14% of deaths. Uterine rupture was the first cause at the health centers level.

Table 2 summarizes the causes of death in relation to the type of delay. Amongst the direct causes of death, hemorrhage was observed in 60.6% of cases, followed by sepsis in 20.8%, while HIV/AIDS was frequent amongst the indirect causes, observed in 40.6% of the cases.

Around 40.4% of women suffered delay type II but not delay type III, 13.9% experienced both delay type II and type III and 18.5% did not experienced any type of delay at the referral facility.

The modelling results in Table 3 show a significant association between the main cause of death and delay type III after controlling for other risk factors. Pregnant women who died of a direct pregnancy complication were twice more likely to have experienced a delay type III than pregnant women who died due to indirect causes. However, no significant association between delay type II and cause of death was observed.

Experience of delay type II was strongly and positively associated with having been referred due to complications, and with care-seeking at central hospitals compared to seeking care at a health center. The odds of delay type II among women who were referred were roughly two times that of women not referred.

A negative correlation was estimated between the probability of experiencing both delays type II and III. Delay type III is less likely to occur in a woman who experienced delay type II and vice-versa. A test on the overall effect of risk factors based on contrasts between model parameters showed that the significant effects on Table 3 were different between Delay II and Delay III.

This study evaluated the associations between two types of delays for a women experiencing pregnancy complications and the cause of death: delays type II, reaching a facility with capacity to manage pregnancy related complications and delays type III, delay in receiving appropriate treatment. This study focused on type II and III delays and therefore only provided the analysis of the deaths that occurred within health facilities.

The second delay was shown to be more frequent in this setting when compared with the third delay regardless of cause of death (direct or indirect). It was observed in more than 60% of women who died. The delay type II has been described in other studies [18‐20]. Patient referral from a peripheral facility was strongly associated with the delay type II. This finding suggests that part of the delay in reaching the appropriate facility to treat the complication may occur at the peripheral health facility. At these facilities we often find junior and less qualified staff with critical limitations with regard to decision-making (diagnosis and treatment) in the face of an obstetric emergency. Another source of delays is related to the referral process management which is characterized by frequent fuel stock out and lack of ambulance maintenance or even inexistence of one functioning in the District. This slows down the decision-making to transfer patients and results in an inefficient functioning referral system. Referral has been noted elsewhere in Mozambique to be a risk factor for stillbirth outcomes [21]. A combination of factors such as the perception of poor quality of care and the lack of availability of skilled personnel have been identified in other studies to influence the choice of where to deliver and ultimately contributes to the delay for women to reach the health facility [22‐25]. Elsewhere lack of transportation, prolonged travel time and seeking care at more than one facility were also identified as contributing factors to the delay type II [20].

Delay type II was identified and potentially contributed to one third of facility based maternal deaths. Women who died of acute direct obstetric complications were more likely to have faced delay type III. This might be interpreted as a lack of adequate emergency response from health services. Appropriate and timely complication management requires specific midwifery and specialist skills from the health providers as well as adequate availability of EmOC services [26‐28]. Thus, availability of qualified services providers in right quantity must be combined with the availability of supplies for emergency care (e.g. drugs, blood), and the structural functioning of the health facilities such as the referral system and the functionality of the operating room [6, 27, 29‐31].

We identified a negative relationship between delay II and delay III. A woman experiencing a delay in reaching the health facility may trigger a quick response from the health system when she arrives at the health facility and is less likely to experience a third delay. This correlation has been noted elsewhere [11]. This emphasizes the importance of ensuring the availability of services to identify and effectively manage the emergency resulting from pregnancy complications. Measures to improve delay type II tend to contribute to a reduction in delay III as well [32].

In a study in Brazil of maternal near miss cases the occurrence of any delay increased the severity of maternal outcome, particularly leading to maternal death [5]. Other authors have found an association between delays and severe pregnancy outcomes and maternal death [33].

In our study, direct obstetric causes were the most important causes of maternal deaths and hemorrhage the most frequent among this group followed by sepsis. HIV/AIDS was the most common cause of death among indirect causes. This finding aligns with the pattern of causes of maternal deaths in the developing world described elsewhere [5, 34].

In this national study 44% of the reviewed cases went directly from home to a health facility where they died. Another study conducted in a more urbanized setting in Mozambique found that 29% of near miss cases went directly from home to the health facility evaluated [14]. It is important to note that around 10% of the women who died were admitted to health facilities without any complication. Data from this group of women showed hemorrhage and obstructed labor as responsible of 42% of deaths. This results may indicate inadequate labor and delivery management including the issue of referral system discussed previously. Women still face important risk of dying even arriving in the maternities in healthier status [11]. The health facilities in this setting suffer from the same challenges described for the developing world in relation to human resources and the overall health services provision. A comprehensive effort to improve health system responsiveness is desirable for better maternal health outcomes.

Investment is needed to strengthen referral linkages and secure hospital and health center readiness to rapidly diagnose and manage pregnancy related complications, and thus, lessen the impact of delays II and III, and ultimately reduce institutional maternal mortality. Further studies that include women who survive will be important to understand the broader impact of the delays on maternal health outcomes in this setting.