This study investigates maternal mortality in 59 health facilities using the PBF approach in the province of Maniema in the DRC.
The method used considered data collected in two periods (five years and one year).
The five-year data were collected from all health facilities under study, allowing the determination of the number of pregnant women who gave birth, their maternal mortality rate, the time of occurrence, and delays justifying these deaths. Additional data were unavailable because of limitations linked to health centers’ incomplete files (24). This state of affairs had been discussed at the Seventy-second session of the World Health Organization for the eleventh revision of the International Classification of Diseases, during which the member states acknowledged that despite the efforts made, the system of notification, registration, and declaration of maternal deaths and obstetrical causes remains limited at the community level and in primary healthcare facilities at the first level (25).
Because of the aforementioned limitations, we conducted a second data collection for one year exclusively from reference health structures, including HGR, CSR, and maternity units. This step made it possible to generate complete, specific data, with a great possibility of conducting all the analyses.
This approach was developed by Reinke and colleagues (26), who justified using this procedure considering that data collected over a long period were considered incomplete or unrepresentative. In this respect, Ntambue states that the maternal death registration system in developed countries considered as efficient seems utopian in developing countries where the national health information system is weak (27).
However, the prominent results of this study reveal that the mortality rate in the health facilities studied (i.e., CS, CSR, HGR, and maternity units) was 620 deaths per 100,000 live births and that the main risk factor was first delay. The results of this study are discussed as follows.
Maternal mortality ratio
The maternal mortality rate in CS, CSR, HGR, and maternity units was 620 per 100,000 live births.
These results do not fulfill the United Nations Sustainable Development Goals (SDGs), which aims to reduce the global maternal mortality rate to ≤70 deaths per 100,000 live births by 2030 and for no country to have a maternal mortality rate more than twice the world’s average (28).
However, considering the World Health Organization principle that no woman should die due to childbirth, we observed that despite the implementation of the PBF strategy by the DRC, the endeavors in this area remain limited. This observation is supported by a maternal death rate that is three times higher than the rate estimated at the global level in 2017, which was 211 maternal deaths per 100,000 live births (29). This rate is twice as high as that of developing countries, estimated at 415 maternal deaths per 100,000 live births (30). By contrast, it is similar to that of sub-Saharan Africa, estimated at 542 deaths per 100,000 live births in 2017 (28).
This maternal mortality rate is 60 times higher than that of Europe, with 10 deaths per 100,000 live births, and 80 times higher than that of Australia, with 7 deaths per 100,000 live births (30), which are industrialized countries.
This high prevalence of maternal mortality rates in our research setting is sufficient justification for the relevance of our study in the Doctoral research program.
Maternal mortality rate in the reference hospital setting
The mortality rate in reference hospitals (i.e., HGR, CSR, and maternity units) is 1300 maternal deaths per 100,000 live births. Our results are lower than those found by Ntoimo (2,085 per 100,000 live births) (31), Ousmane (1962 per 100,000 live births in the Labé Regional Hospital in Guinea) (32), and Kamga (1,538.9/100,000 live births in three university hospitals in Yaoundé) (33). However, our results are higher than those found by Sissoko (201.87 per 100,000 live births) (34) and Geleto (149 per 100,000 live births in Ethiopian hospitals) (35). This diversity of results confirms that maternal mortality in hospitals varies according to the context (36). Thus, it is difficult to compare the results of industrialized countries with those of African countries or of provinces that have more hospitals with those that have fewer hospitals.
Furthermore, this high rate of maternal death in referral facilities is explained by the fact that most parturients spend more time at first-level facilities and are only referred to a referral facility when obstetric complications appear (31-33).
Regarding our results, our view and that of the aforementioned researchers are the same. First, the referral health structures in our study settings are designated to receive and manage obstetric referrals from the first-level primary healthcare facilities (e.g., dispensary, health post, and health center); second, these facilities confront serious obstetric complications; and third, in the PBF approach, all referrals from first- to second-level health facilities are purchased through a third-party payment supported by the government (9).
Explanatory factors for in-hospital maternal deaths: Risk markers
The results of this study show that in-hospital maternal death is significantly associated with the age group of the parturient, specifically, the extreme ages (≤19 years and ≥40 years). After adjustment for independent variables, in-hospital maternal death remained significantly associated with maternal deaths under 20 years of age (p < 0.001).
These results corroborate those of Salem and friends (37), who found high maternal mortality in parturients aged ≤20 and ≥35 years. In this context, the authors explained that in Tunisia, pregnancy outside the age range of 19–34 years is a risk factor for maternal and fetal morbidity and mortality (37). However, the high mortality observed among adolescents may support the idea of the Europe PMC Funders group that “special attention should be paid to girls under 20 years of age, as they are more at risk than other categories” (38). Furthermore, our results contradict those of Mouté and Zinvi (39), who suggested that the groups aged 15–19 and 40–44 years were least exposed to maternal deaths, which seems to be confounded by the physiological immaturity and the occurrence of complications in childbirth at extreme ages (39).
In the context of our study, maternal mortality in the group aged ≤20 years can be explained, in addition to the aforementioned elements, by sociocultural constraints that force early marriage. Furthermore, the ignorance of young girls could increase the risk of death of parturients aged ≤20 years. Thus, this argument sufficiently justifies the interest in good sensitization regarding the prevention of early marriage.
Although the results of our study show that the educational level of the deceased women does not constitute as a factor of mortality in the parturient (p = 0.054), it cannot be eliminated considering the abundant literature on the preponderance of this variable in maternal death at maternity hospitals (40). On this subject, Baldé explained that a high frequency of of deceased women dying were among those who had low level of education (41). Similarly, Ousmane study demonstrated that in 69.23% of maternal deaths, the deceased had no education (32). Based on the aforementioned literature, the low level of education seems to be an important indicator that justifies maternal mortality in this study. In this regard, meh stated that education level is fundamental in explaining the behavior of individuals in a society (40). Accordingly, low educational level among women leads to low or no use of modern healthcare. Further studies are necessary to confirm the exclusion of this variable in maternal mortality.
Regarding marital status, the results of this study show that this variable is a protective factor for maternal death (p > 0.724 f). This finding is contrary to that of Kamga, who showed that the majority of parturients who died were single (75%) (33); according to them, marriage is the ideal setting for sexual activity and procreation; consequently, unmarried women are more likely to die during childbirth. However, because marriage takes place early in this context (as demonstrated in ref. 32 and 42), this sociocultural consideration leads women of childbearing age to enter into an early marriage and have children at a very early age, thus increasing the risk of maternal death.
The results of this study also reveal that the woman’s occupation is not a risk factor in parturient death; however, the differences observed are not statistically significant at the 95% confidence level (p > 0.268 f). This result is contrary to that of Ousmane (32) who found that maternal mortality was prevalent among homemakers. However, in our study, attention could be given to women farmers who often conduct heavy fieldwork, thus obliging pregnant farmers to carry large loads that can increase the risk of obstetrical complications.
The results of this study establish no link between the variable religion and the risk of maternal death (p ˃ 0.298 f). However, medical literature has demonstrated that religious beliefs significantly influence access to maternal healthcare. For example, certain religious practices and requirements that disrupt social life are harmful (43). Researchers have argued that religious constraints limit access to maternal healthcare (43). Thus, this argument supports the hypothesis that religion is a risk factor for death. This idea was supported by Ariyo et al., who demonstrated that Muslim women were 52% more likely to experience maternal deaths (OR: 1.52; CI: 95%: 1.10–2.11) than Christian women (44).
By contrast, in our study, from a demographic view, Muslim religion is predominant in southern Maniema province (Catholics, Protestants, Revivalists, and Kimbaguists were also observed). This cultural diversity also explains the diversity of models of perceptions of the use of modern healthcare, because each religion has a set of traditional health system practices specific to its culture. From a health perspective, the predominance of Muslim women may indicate the lack of social support from Muslim husbands to their wives, which may influence the low utilization of health services. This topic, Al-Mujtaba and al., (53), think in their study on "Assessment of religious influences on the use of maternal health services among Muslim and Christian women", carried out at the center -north of Nigeria, although stated in a Hadith that "A woman should only travel with a Dhu-Mahram (her husband or a man with whom this woman cannot marry at all according to Islamic Jurisprudence), and no man can visit her, except in the presence of a Dhu-Mahram, religion does not seem to see any influence on the choice of attendance of maternal health services. But they believe, on the other hand, that husbands who do not make the necessary arrangements for their wives to attend health services in real time, may contribute to misuse of services (53).
This non-involvement of Muslim husbands would expose parturients to obstetrical complications that could lead to maternal death. On this point, Iliyasu and his colleagues (54), in their study carried out in Ungogo, a community in northern Nigeria, found in their study that only 32.1% of men agreed to accompany their wives for maternity care (54). Hence, special attention should be given to this issue in our study setting.
The results of this study show that in-hospital maternal death is significantly associated with parity. Thus, the risk of dying increases in primiparous and large multiparous women.
These results corroborate those of Yambare and Yambare and et al. (46) who presented the high risk of maternal death in primiparous and large multiparous women. As in medical literature, the authors justified their results by referring to the physiological immaturity of the deceased primiparous parturients and the occurrence of complications during delivery in large multiparous women (45).
Causes of maternal death
In relation to the causes of maternal death, in order of importance, the obstetric causes diagnosed on the admission of parturients to the hospital were dystocia (18%), hemorrhage (10.4%), infections (1.8%), and uterine rupture (1.1%). However, after adjustment for independent variables, the obstetrical causes that were significantly associated with maternal death were those found frequently in the literature, mainly in the African region, namely, hemorrhage, uterine rupture, infection, and mechanical dystocia.
Our results are comparable to those of other studies conducted in African countries, especially those conducted by Kamga, Ntoima, Ousmane, Mbeva, and Yambare. Considering their results, these researchers have demonstrated that obstetric complications are either treated late or inappropriately by healthcare personnel (31-33,46,47).
In relation to our study environment, our results demonstrated two main types of delayed decision making, namely, that of the parturient or the family to go to the hospital in the event of obstetric complications and that of healthcare personnel at the first level of the health pyramid (health center) to transfer parturients in time in the event of obstetric complications.
Delays as a major risk factor for maternal deaths.
In this regard, Thaddeus and Maine and Actis Danna and colleagues have suggested that women who made decisions in time to go to the place of delivery at the onset of danger signs, women who arrived at the health center or hospital in time to give birth, and women whose hospital admission was adequate are more likely to be saved from obstetric complications than those who experience all three delays (19,42,48).
Indeed, the most important risk factor revealed in our study is delay, which could be delay related to the decision by the parturient or the family to go to the place of delivery (first delay), delay related to the arrival at a health center or hospital (second delay), and delay related to hospital management (third delay). Our results show that the cause of death of more than one third of parturients was related to at least one of the three types of delay, and the first delay had the highest proportion (46%).
Our results contradict those of Mbeva, who reported that the highest proportion of maternal deaths was related to the third type of delay (i.e., hospital management; 49%) (47); they believed that healthcare providers are incompetent in managing obstetrical emergencies in the interaction between members of the healthcare team, management of inputs including blood for transfusion, identification and resolution of obstetrical problems in real time, and use of equipment (47).
Based on of our results, a qualitative study could further our understanding of the reasons for these delays in our study environment.
However, before discussing the results of this study, the method used is discussed.
Limitations and strengths of this study.
The study has some limitations. First, the data were collected in two stages, the first being that of five years where the data were collected in all the health structures (CS, CSR, HGR and maternity). And the second phase is that of a year when the data collections were made in the reference health structures (HGR, CSR and maternities). Second, the data collected during the first step did not allow us to perform specific analyzes that could allow us to associate the dependent variables with our variable of interest. However, the data collected during the second phase, exclusively in the reference health structures (HGR, CSR and maternities), which allowed us to generate complete data and carry out more detailed analyzes, as well as the studied population ( nearly 100,000 pregnant women), could justify the power of this study.