A maternal death is defined as the death of a woman occurring during pregnancy or within 42 days of its termination, regardless of the duration or location of the pregnancy, resulting from a cause related to or aggravated by the pregnancy, excluding accidental or incidental causes [1].

Globally, maternal mortality remains a major public health challenge. In 2020, approximately 287,000 women died from maternal causes [2]. Hemorrhage remains the predominant direct cause [2].

In sub-Saharan Africa, maternal mortality remains particularly alarming [3]-[5]. Several major explanatory factors have been identified, including limited access to quality obstetric care and insufficient prenatal monitoring; low coverage of prenatal consultations and the persistence of risky practices; unassisted childbirth by qualified personnel; and demographic and geographic variables, such as region or extreme maternal age [6]-[8]. In the Democratic Republic of Congo, the maternal mortality ratio is estimated at 760 deaths per 100,000 live births [9]. The surveillance system reveals significant regional disparities. Some provinces have critical ratios exceeding 300 deaths per 100,000 live births [10]. Hemorrhages, cardiovascular diseases, and structural weaknesses in the health system are cited as contributing factors [11].

Despite the severity of the overall situation, published scientific data on the factors contributing to maternal deaths remain scarce in Kasai Province. Yet, identifying these local factors is essential for guiding prevention interventions, improving the quality of obstetric care, and strengthening strategies to reduce maternal mortality.

The present study aimed to identify the factors explaining maternal deaths among women who experienced obstetric complications in the Kasai Province, in order to provide evidence that could support the decision-making of health authorities and contribute to the sustainable reduction of maternal mortality.

Table 1. Sociodemographic characteristics of the women interviewed in Kasai Province from 2022 to 2024.

A total of 708 women who experienced obstetric complications were included in the study. The majority were between 18 and 40 years old (91.53%), with a mean age of 29.4 ± 6.8 years. Nearly three-quarters of them lived in rural areas (71.61%). Primary education was the most common level (67.80%), followed by secondary education (30.51%) and, to a very small extent, university education (1.69%). More than half of the women were primiparous (59.89%), while 21.47% were multiparous (See Table 1).

Data from the period 2022-2024 reveal two distinct realities depending on the data collection source. The overall data for Kasai Province (DHIS2) indicates 873 maternal deaths and 730,934 live births recorded, representing a maternal mortality ratio of 119 per 100,000 live births (See Figure 1Figure 1). Meanwhile, data from the targeted survey (Healthcare Facilities & Community) reveal a high mortality rate in emergency situations: 177 maternal deaths out of 708 recorded obstetric complications, representing a mortality rate of 25% (i.e., 1 death for every 4 complications).

Figure 1. Frequency of maternal deaths in Kasai Province from 2022 to 2024. Source: (DHIS2) and our surveys.

All three levels of delay significantly increased the maternal mortality rate compared to the “close call” status (p < 0.05). Women who did not receive skilled personnel during delivery had more than twice the mortality rate (OR = 2.25; 95% CI: 1.56 - 3.25; p < 0.0001). Failure to attend prenatal care increased the mortality rate by 1.81 times (95% CI: 1.22 - 2.70; p = 0.0031). A delay of less than two years between pregnancies doubled the mortality rate (OR = 2.10; 95% CI: 1.48 - 2.97; p < 0.0001). Mothers over 40 years of age had more than twice the mortality rate compared to the 18 - 40 age group (OR = 2.21; 95% CI: 1.09 - 4.47; p = 0.0246). (See Table 2)

Place of origin, marital status, and gestational age did not show a statistically significant association with maternal mortality in this sample (p > 0.05).

Obstetric hemorrhage emerged as the leading direct cause of death in this study. Its prevalence was significantly higher among women who died (76.84%) than

Table 2. Determinants of maternal deaths among women interviewed in Kasai Province from 2022 to 2024.

among those who survived (35.03%), demonstrating a statistically significant association (p < 0.05). Other recorded obstetric complications were less frequent, including eclampsia, dystocia, and maternal infections. Furthermore, none of these showed a statistically significant link with maternal deaths in this sample. (See Table 3)

Table 3. Direct causes of maternal deaths among women interviewed in Kasai Province from 2022 to 2024.

Indirect causes of maternal death revealed significant differences between the group of women who died and the group of survivors, particularly for malaria and anemia. Malaria was the most frequent condition, significantly affecting more women who died (93.79%) than those who survived (79.66%) (p < 0.05). Conversely, anemia of pregnancy was observed in 3.95% of women who died compared to 10.55% of those who survived, also showing a statistically significant difference (p < 0.05). All other conditions (asthma, heart disease, pneumonia, tuberculosis, HIV/AIDS) remained rare and showed no statistically significant association with maternal death (See Table 4).

Table 4. Indirect causes of maternal deaths among women surveyed in Kasai Province from 2022 to 2024.

After adjustment, the multivariate analysis shows four main factors independently and significantly associated with maternal mortality: The absence of skilled assistance at delivery multiplies the odds of death by more than two (AOR = 2.3558; 95% CI [1.55 - 3.57]; p = 0.0001). An interpregnancy interval of less than two years triples the odds of maternal death (aOR = 3.3039; 95% CI [1.99 - 5.48]; p = 0.0000), obstetric hemorrhage nearly sevenfold (aOR = 6.7209; 95% CI [4.25 - 10.62]; p = 0.0000), and multiparity nearly halves the odds of maternal death due to a professional care deficit (aOR = 0.6590; 95% CI [0.49 - 0.89]; p = 0.0071). The other variables studied did not show a significant association with maternal mortality after adjustment. (See Table 5)

The profile of the 708 women analyzed highlights a context of marked social vulnerability. A low level of education was observed among the women, with 67.80% having not progressed beyond primary school. There is a predominantly rural population, with 71.61% of the respondents residing in remote areas, zones characterized by geographical isolation and limited access to healthcare. The majority of the women were between 18 and 40 years old (91.53%) and were first-time mothers (59.89%) (Table 1 and Table 2).

A discrepancy exists between the provincial ratio of the health information system (DHIS2: 119 deaths per 100,000 live births) and the hospital mortality rate observed in the study (25%, or 1 death for every 4 obstetric emergencies). This difference is explained by systemic underreporting within the DHIS2, linked to deaths occurring at home (outside of the healthcare system), gaps in the completeness of reports, and the absence of systematic audits. The 25% case fatality rate reflects the extreme severity of the cases treated and the inability of hospital facilities to manage major obstetric emergencies. These results are consistent with data from the Demographic and Health Survey in the Democratic Republic of Congo [9], which confirm the persistence of a high level of maternal mortality in the country.

Multivariate analysis identified four independent determinants of maternal death. Obstetric hemorrhage (ORa = 6.72; 95% CI [4.25 - 10.62]) (Table 3 and Table 5) was the leading direct cause of death (76.84% of women who died). The mortality rate from an obstetric emergency is nearly seven times higher in the presence of uncontrollable bleeding. This finding reflects the lack of availability of safe blood products, the absence of emergency kits, and the delayed administration of uterotonics (e.g., oxytocin, misoprostol). This prevalence is widely documented internationally, in both low-resource and high-income countries [2][12]. Comparable results have been reported in Abidjan, where hemorrhages accounted for approximately one-third of direct causes [13].

In contrast, eclampsia and maternal infections were not significantly associated with deaths in our study. This finding differs from some previous research [14] and could be explained either by the low frequency of these conditions during the study period or by diagnostic difficulties in peripheral settings, which often lack appropriate laboratory equipment.

A short interpregnancy interval (<2 years) triples the mortality rate (aOR = 3.30; 95% CI [1.99 - 5.48]). This leads to “maternal exhaustion syndrome,” characterized by uncorrected nutritional deficiencies and an increased risk of uterine atony, severe anemia, and placental complications [15] (Table 5).

Regarding the lack of skilled assistance, it has been noted that giving birth without the help of trained personnel (midwife, physician) more than doubles the mortality rate (aOR = 2.36; 95% CI [1.55 - 3.57]) (Table 5). The failure to detect dystocia or residual bleeding early transforms a manageable complication into a fatal one. The lack of skilled assistance during childbirth has been identified as a key factor in several African contexts, notably in Senegal [4].

The analysis shows a protective effect of multiparity compared to pauciparity (ORa = 0.6590; 95% CI [0.49 - 0.89]). This counterintuitive result may reflect a survival bias (Healthy Mother Effect): women who have experienced several pregnancies without major complications have developed more effective use of healthcare services or exhibit less vulnerable obstetric physiology than primiparous women who are confronted from the outset with unrecognized emergencies. Previous research shows an association between high multiparity and maternal deaths [16].

The bivariate analysis reveals a very high prevalence of malaria among the women who died (93.79% vs. 79.66%, p < 0.0001). Malaria caused by Plasmodium falciparum promotes hemolysis and maternal-fetal distress. This condition was thus the main indirect cause of death, which corroborates the data reported in Cameroon [5].

Clinical anemia was reported in 3.95% of deaths compared to 10.55% of survivors (p = 0.0076). This contrasts sharply with the findings of many previous studies that have linked this condition to maternal deaths [5][11]. This paradoxical reversal is probably due to imputation or documentation bias: in patients who died rapidly from fulminant hemorrhage, the diagnosis of underlying chronic anemia was often not recorded in the medical file, unlike in stabilized patients (survivors) who underwent a complete blood count.

Regarding the three-delay model and the reality on the ground, qualitative data confirmed that mortality in Kasai stems from dysfunction at each stage of care: The first delay (decision to seek medical attention) is exacerbated by low levels of education, women’s financial dependence on their partners, and late access to prenatal consultations (often initiated in the third trimester). The second delay (access to care) is compounded by the isolation of rural areas, the cost of transportation, and the time it takes to transfer patients from health centers to general referral hospitals. The third delay (hospital care) is marked by drug shortages, the absence of functioning blood banks, a lack of expertise in resuscitation equipment, and staff demotivation (cemented by free maternity care not compensated by effective state subsidies). This systemic approach aligns with the findings of Bohoussou et al. [13], who emphasized the combined involvement of individual, institutional, and professional factors. These observations reinforce the importance of a maternal mortality surveillance system that integrates these three levels of analysis.

These results can be extrapolated to other similar contexts, as the organization of the health system and the structural constraints of Kasai Province are comparable to those of other provinces in the DRC and many resource-limited countries.

Despite the relevance of its results, this study had some limitations:

1) Data collection relied on routine registers, exposing the study to a risk of information bias (variable completeness and rigor among healthcare providers). This bias was mitigated, however, by the use of cross-referenced secondary sources (partograms and death review forms).

2) The lack of detail regarding the time to treatment and the actual quality of care at the time of the complication prevented a detailed analysis of the impact of the “three delays” (decision, access, and treatment). Triangulation with community member testimonies helped to mitigate this limitation.

3) The scarcity of published studies specific to Kasai Province limited direct comparisons with previous local data, a constraint partially offset by the use of national surveys (DHS).

4) The difficulty in clinically differentiating malaria from other febrile syndromes may have influenced the attribution of indirect causes. Nevertheless, the application of a strict operational definition (fever associated with a positive laboratory test) has strengthened the validity of the diagnoses made.

Maternal mortality remains a major public health problem in Kasai Province. This study highlighted the persistence of largely preventable causes, revealing shortcomings both at the community level and within the health system.

The data collected revealed several critical factors. Three delays in delivery, lack of skilled birth attendants, low attendance at antenatal care services, delayed access to healthcare, maternal age over 40, and interpregnancy intervals of less than two years (closely spaced pregnancies) were the risk factors associated with maternal deaths. Direct obstetric causes were largely dominated by postpartum hemorrhage. Indirect causes were primarily related to gestational malaria (See Table 6).

Table 6.To sustainably reverse the trend of maternal mortality in Kasai, interventions must jointly target the three identified levers.