Background

Despite improvements and worldwide attention on maternal mortality, it is still one of the top global health agendas, and there are many existing challenges to ending preventable maternal mortality, particularly in low and middle-income countries [1]. Successes in lowering maternal mortality during the Millennium Development Goal era have plateaued in the first five years (2016–2020) of the Sustainable Development Goals (SDG) [2]. If this progress is maintained, the Maternal Mortality Ratio (MMR) will be 222 by 2030, more than three times the SDG global target of 70 [2]. Globally, 287,000 maternal deaths occur each year, with Sub-Saharan Africa accounting for 70% of deaths [1].

Many women survive for every woman who dies, yet often experience long-lasting complications, such as adverse pregnancy outcomes, disability, and psychological complications [3, 4]. In 2004, the World Health Organization (WHO) highlighted the importance of moving beyond simply reporting deaths to create an understanding of why they occur and how they might be prevented [5]. Furthermore, in 2011, the concept of maternal near-miss emerged as a tool for assessing the quality of obstetric care [6]. Maternal near-miss (MNM) is defined by the WHO working group as a woman who nearly died but survived a life-threatening condition that occurred during pregnancy, childbirth, or within 42 days following childbirth due to getting the best evidence-based quality care or by chance [5, 7]. Its primary causes are hemorrhage, hypertensive disorders of pregnancy, postpartum sepsis, obstructed labor, uterine rupture, abortion, and anemia [1, 8, 9].

The near-miss approach is comprehensive and works on the concept of criterion-based clinical audit, which is considered a feasible and beneficial method of auditing the quality of maternal health care [10]. It assumes that women who survived life-threatening complications related to pregnancy and childbirth had many similarities with those who died [6]. The ultimate goal of the near-miss approach is to boost clinical practice and reduce preventable morbidity and mortality using the best evidence-based practices [5]. The approach enables health service delivery points to work on cases with a chance of survival, allowing for open discussion and removing fear of blame among clients and healthcare providers [11]. Furthermore, it has proven to be a valuable metric for evaluating the quality of safe motherhood programs in populations [6].

The global estimated figure of near-miss in 2022 was 18.67/1000, with continental variations; 3.10/1000 in Europe to 31.88/1000 LB in Africa [12]. Socioeconomic factors (age, education level, wealth status), obstetric (parity, gravidity, history of CS delivery), medical conditions (having chronic hypertension), and health system-related characteristics were associated with MNM [13,14,15,16,17].

Although small-scale studies regarding MNM have been conducted within African countries, they were limited to subnational levels [13, 16,17,18,19] and with a relatively small sample size (e.g. n = 183 [20]). Therefore, large-scale studies are scarce to estimate MMN prevalence and risk factors across the continent. Furthermore, a recently conducted systematic review and meta-analysis on the global prevalence of MNM have not identified its risk factors did not estimate the pooled primary (direct and indirect) causes of MNM and have limited detailed evidence to understand the unique intervention options relevant to Africa [12]. This evidence gap could be partly addressed by synthesizing and pooling estimates from existing country-level evidence via systematic review methods and meta-analysis.

Hence, the current study aimed to assess the magnitude of MNM, its primary causes, and its potential determinants in Africa. This study's findings could aid in identifying factors that contribute to maternal morbidity and death, which is necessary for designing targeted measures aimed at improving maternal health outcomes, aligned with SDG target 3.1: reducing maternal mortality below 70 per 100,000 live births [21]. Policymakers, healthcare providers, and other stakeholders working in maternal health can use these findings to inform evidence-based decision-making and implement interventions, ultimately improving maternal health outcomes through strengthening targeted service quality measures.

Methods and materials

Study design and reporting system

A systematic review and meta-analysis were performed by synthesizing peer-reviewed articles. Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) was used to report the findings [22] (Table S1).

Search strategies

This study considered studies published before October 31, 2023. Searches were performed from October 1–31, 2023 using electronic databases, namely PubMed/Medline, Scopus, Web of Science, Directory of Open Access Journals, and Google Scholar. Medical subject heading (MeSH) with Boolean operators (AND and OR) and truncation were employed to connect the keywords: maternal near miss, maternal morbidity, risk factors and Africa. A search strategy used for PubMed was: ((((((((epidemiology [All Fields]) OR (prevalence[All Fields])) OR (level[All Fields])) OR (magnitude[All Fields])) OR (proportion[All Fields])) OR (incidence[All Fields])) AND (((((((((maternal near miss[All Fields]) OR (maternal near-miss[All Fields])) OR (severe maternal outcome*[All Fields])) OR (pregnancy complication*[All Fields])) OR (life-threatening condition*[All Fields])) OR (maternal morbidit*[All Fields])) OR (Severe maternal complication*[All Fields])) OR (maternal mortality[All Fields])) OR (maternal death[All Fields]))) AND ((((determinant*[All Fields]) OR (factor*[All Fields])) OR (predictor*[All Fields])) OR (Associated factor*[All Fields]))) AND ((Africa*[All Fields]) OR (Sub-Saharan Africa*[All Fields])) Search strategies used across the database with their example are presented in the supplementary material (Table S2).

Eligibility criteria and study selection

The systematic review and meta-analysis used the mnemonic Condition, Context, and Population (CoCoPop) for question formulation method [23].

Articles were included if they met the following inclusion criteria.

  1. 1.

    Condition (Co): Assessed the magnitude and/or determinants of MNM

  2. 2.

    Context(Co): Conducted in Africa

  3. 3.

    Population: All women who were pregnant, gave birth, or were within postpartum periods (42 days).

  4. 4.

    Study type: Observational (cross-sectional, case–control, and cohort) studies that reported the prevalence of MNM, its causes or determinants.

The scope of the review was limited to quantitative peer-reviewed published studies in the English language. The most complete and up-to-date study was included in case of duplicate studies sourced from the same data. Case reports, case series, commentaries, conference abstracts, letters to editors, technical reports, qualitative studies, and other opinion publications were excluded.

Study selection, and data extraction

All retrieved studies were imported into the EndNote X7 library and checked for duplication. After removing duplicate studies, two independent reviewers (AH and YS) screened all articles for eligibility by looking at the title, abstract, and full text. A third reviewer (LL) independently assessed 20% of the excluded papers and collected the screened articles; any disagreements were resolved through discussion. Two authors (AH and YS) extracted the data by using Microsoft excel 2013 spreadsheet, which includes the author’s name, publication year, study year, study design, country, region, data collection technique, sample size, response rate, prevalence of MNM, each cause of MNM, and determinants.

Quality assessment

The quality of the articles was assessed using the Joanna Briggs Institute (JBI) Critical Appraisal Checklist [24]. Two reviewers (AH and YS) independently rated the quality of the studies. The tool considers eight parameters, each with equal weight: (1) well-stated inclusion and exclusion criteria (2) a detailed description of study subjects and setting (3) measurement of exposures validly and reliably, (4) has well-stated objective with standard criteria used for measurement of the condition, (5) proper identification of confounders, (6) strategies to deal with confounders were well-stated (7), measurement of outcome validly and reliably and (8) use of appropriate statistical analysis. The evaluators rated the study a '1' if it met each specific parameter and a '0' if it did not (no or unclear). A composite index was computed and those studies with a score of ≥ 6 were included in the final analysis (SRMA) [25] (Table S3).

Outcome measurement

MNM was assessed using the WHO MNM criteria and computed as the total number of MNM cases per total number of live births. MNM is defined as a woman admitted to health facilities with at least one of the following severe maternal complications: hypertensive disorders of pregnancy (severe preeclampsia or eclampsia), severe postpartum hemorrhage, uterine rupture, sepsis or severe systemic infection, or severe complications of abortion, but she survived [6]. Determinants of MNM were estimated using a pooled AOR with corresponding 95% CIs.

Statistical analyses

Higgins I-square (I2) statistics and Cochran’s test were used to examine the presence of statistical heterogeneity across the included studies. Accordingly, considerable heterogeneity [I2 = 99.5%, p < 0.001] was detected, and the pooled prevalence of MNM and each severe maternal complication was estimated using a random-effects model with the DerSimonian-Laird method [26]. Furthermore, the adjusted odds ratio (AOR) and 95% CIs were extracted, and the pooled estimates were computed using a random- or fixed-effect model based on their level of heterogeneity. Forest plots were used to present a visual summary of data. In addition, subgroup analyses were performed based on region and study year.

A univariate meta-regression analysis with sample size, publication years, and study years as factors was performed to identify probable sources of heterogeneity among the studies [27]. Visual and statistical methods were used to check for publication biases. A funnel plot was used during the visual inspection, with a symmetrical and large inverted funnel used as a proxy for low publishing bias. In addition, statistical methods such as Egger's and Begg's tests were used to support visual assessment, p-value of < 0.05 suggests the possibility of publication bias. A random-effects model was used for the sensitivity analysis to examine the impact of a single study on the overall pooled prevalence of MNM.

Results

Study selection

Of 5698 retrieved studies, 4821 were duplicates (Fig. 1). Subsequently, 877 studies were reviewed by their titles and abstracts, with 189 articles meeting the full-text eligibility criteria. Sixty-five studies were included in this systematic review and meta-analysis. Most of the full-text reviewed articles were excluded (n = 124) due to not having insufficient data (n = 83), followed by failing to clearly state the outcome of interest (n = 26) (Fig. 1).

Fig. 1
figure 1

PRISMA flow diagram describing the selection of studies for systematic review and meta-analysis

Full size image

Characteristics of included studies

In 65 studies, nearly one million (N = 968,555) participants were included, with the sample size in individual studies ranged from 183 [20] to 323,824 [28] women (Table 1). Nearly three-fourths (n = 47) of the studies were cross-sectional, and the remainder were case–control (n = 10) or cohort (n = 8) studies. The studies’ publication period spans from 2011 to 2023. Half of the studies (n = 33) were conducted by record review only. The majority of the studies were carried out in the East Africa (n = 43) and West Africa (n = 11) regions (Table 1).

Table 1 Descriptive summary of studies included in systematic review and meta-analysis of the prevalence of MNM and its determinants in Africa, 2008–2021
Full size table

The pooled estimate of MNM in Africa

The pooled estimate of MNM in Africa was 73.64/1000 Live births (95% CI: 69.17, 78.11) The I2 test statistic (I2 = 99.50%; p < 0.001) revealed that there was significant variation between the included studies (Fig. 2).

Fig. 2
figure 2

Forest plot showing the pooled estimates of MNMR in Africa, 2008–2021. The pooled prevalence of severe maternal complications among near-miss cases

Full size image

Subgroup analyses

Subgroup analyses by region, country, and study year were performed to examine the sources of variation in the pooled prevalence of MNM. East and West African regions have a higher pooled prevalence of MNM (114.82/1000LB (95% CI: 104.94, 123.59) and 78.34/1000LB (95% CI: 67.23, 89.46) respectively. In contrast, the Northern (10.40, 95% CI: 3.15, 17.64) and Southern (11.20, 95% CI: 7.5, 14.9) African regions had the lowest prevalence (Fig. 3).

Fig. 3
figure 3

Sub-group analysis for the pooled prevalence of MNMR by regions of Africa, 2008–2021

Full size image

For studies conducted before or during the Millennium Development Goals and during the SDG, the pooled prevalence was 81.42/1000 (95% CI: 73.70–89.14) and 70.36/1000 (95% CI: 64.56–76.16), respectively (Fig. 4).

Fig. 4
figure 4

Sub-group analysis for the pooled prevalence of MNMR by study year in Africa, 2008–2021

Full size image

The pooled prevalence of severe maternal complications among near-miss cases

The primary causes for being a near-miss case were severe postpartum haemorrhage (36.15%) (Fig. 5) and severe hypertension (27.52%) (Fig. 6). Severe anemia (18.88%) (Fig. 7), uterine rupture (13.89%) (Fig. 8), sepsis (11.62%) (Fig. 9), and septic abortion (8.34%) (Fig. 10) were also common severe maternal complications among the near-miss cases in Africa.

Fig. 5
figure 5

Forest plot showing the pooled prevalence of severe postpartum hemorrhage among near-miss cases in Africa, 2008–2021

Full size image
Fig. 6
figure 6

Forest plot showing the pooled estimates of severe forms of hypertension among near-miss cases in Africa, 2008–2021

Full size image
Fig. 7
figure 7

Forest plot showing the pooled estimates of severe anemia among near-miss cases in Africa, 2008–2021

Full size image
Fig. 8
figure 8

Forest plot showing the pooled estimates of uterine rupture among near-miss cases in Africa, 2008–2021

Full size image
Fig. 9
figure 9

Forest plot showing the pooled estimates of sepsis among near-miss cases in Africa, 2008–2021

Full size image
Fig. 10
figure 10

Forest plot showing the pooled estimates of abortion among near-miss cases in Africa, 2008–2021

Full size image

Heterogeneity and publication bias

To determine the likely cause of heterogeneity, a univariate meta-regression analysis was performed using publication year, study year, and sample size. The sample size (p = 0.0074) substantially explained the heterogeneity, but significant heterogeneity was not observed by the study year (p = 0.421) or the publication year (p = 0.321) (Table 2).

Table 2 A univariate meta-regression analysis of factors affecting between-study heterogeneity, 2023
Full size table

A funnel plot was used to examine publication bias visually, and the vast majority of studies were under an inverted funnel, indicating that publication bias was unlikely (Fig. 11). Furthermore, Egger's regression (p = 0.11) and adjusted Beggs rank correlation test (p = 0.11) did not show significant publication bias.

Fig. 11
figure 11

Funnel plot displaying publication bias of studies reporting the MNM in Africa, 2022

Full size image

Sensitivity analysis

A sensitivity analysis using a random-effects model was carried out to detect the impact of a single study on the total meta-analysis estimate. There was no evidence that a single study had an effect on the overall prevalence of MNM (Fig. 12).

Fig. 12
figure 12

Sensitivity analysis for the pooled prevalence of MNM in Africa, 2008–2021

Full size image

Determinants of MNM in Africa

Nineteen variables were extracted from the included studies to identify determinants of MNM (S4 Excel). The risk of MNM was higher among women with advanced age, living in rural areas, low educational achievement, reported low ANC uptake, living far from a health facility, reported delay to access health service, and have previous history of CS or pre-existing medical condition (Table 3).

Table 3 Results of meta-analysis for significant determinants of MNM in Africa, 2008–2021
Full size table

The effect of age on being a near-miss case was identified in four studies [16, 17, 35, 83], with the pooled risk of being a near-miss case was 2.03 times higher among women aged 30 years and above than women aged < 30 years [AOR = 2.03; 95%CI: 1.65, 2.40)]. From pooled estimates of seven studies, being a rural resident was associated with MNM [17, 30, 32, 40, 77, 83, 84]; women from rural areas were 2.06 times more likely to be near-miss cases than urban counterparts [AOR = 2.06; 95%CI: 1.50, 2.61)]. Using the data of seven studies [17, 39, 79,80,81,82,83], the overall likelihood of MNM was 1.82 times higher among women with no formal education [AOR = 1.82; 95%CI: 1.36, 2.28]. Thirteen studies [4, 13, 17, 30, 31, 35, 37, 39, 79,80,81, 83, 84] were selected to assess the pooled association between not receiving ANC and MNM, and women who did not receive ANC were 1.80 times more likely to become near miss cases than women who did receive ANC [AOR = 1.80; 95%CI: 1.64, 1.97]. A pooled estimate from ten studies [4, 13, 16, 35, 56, 71, 78, 79, 81, 82] revealed that women with a previous history of CS were 4.35 times more likely to have MNM than their counterparts[AOR = 4.35; 95%CI: 3.44, 5.26]. All three (1st, 2nd, and 3rd) delays were significantly associated with MNM. The odds of MNM were 2.51 [AOR = 2.51; 95% CI: 1.79, 3.23], 2.12[AOR = 2.12; 95% CI: 1.42, 2.82], and [AOR = 3.38; 95% CI: 1.21, 5.55] times higher among women who experienced 1st, 2nd and 3rd delays respectively. Long distance to health facilities and low monthly income were also identified as significant predictors of MNM in Africa (Table 3).

Discussion

The pooled prevalence of MNM was 73.77/1000 live births, which varied significantly across the regions and study periods. The risk of MNM was higher among women with advanced age, living in rural areas, low educational achievement, reported low ANC uptake, living far from a health facility, reported delay to access health service, and have previous history of CS or pre-existing medical condition.

The current finding of MNM in Africa (73.77/1000 live births) was considerably higher than the global estimate (18.67/1000LB) [12]. This could be attributed to a lack of access to adequate healthcare services, road infrastructure and transportation access limitations, ill-equipped health facilities, socioeconomic inequities, low educational achievement and high fertility rate, all of which are prevalent across the continent [85,86,87]. The pooled prevalence of MNM was higher in the East and West African regions. Compared to the northern and southern sub-regions of Africa, these two regions are known for poor healthcare infrastructure [88, 89], low skilled birth attendance rates [90], poverty and lack of education, a high rate of harmful traditional practices such as female genital mutilation [91], and political and social instability, all of which contribute to poor maternal health outcomes.

Furthermore, there has been a decrease in prevalence of MNM since 2015 (during the SDG era) compared to that before 2015 (during the MDG era). This, might be attributed to the implementation of SDG goal 3: ensuring healthy lives and promoting well-being for all. In particular, Goal 3.1 focuses on the global reduction of the maternal mortality ratio through great investment and effort to address complications that contribute to MNM [92]. In addition, governments emphasize the significance of establishing robust and resilient health systems during the SDG by providing skilled maternal health services such as prenatal, skilled delivery and postnatal services, which are vital for preventing and managing problems that can lead to MNM [93, 94]. Moreover, it could be attributed to technological breakthroughs and enhanced healthcare interventions, increasing global awareness and advocacy for maternal health, and a focus on women's empowerment.

Women who did not receive adequate ANC had a higher likelihood of being near-miss cases, which is consistent with the previous studies [95,96,97]. Timely and adequate ANC entails regular check-ups and monitoring of maternal and fetal health, along with counselling about danger signs and the need to obtain healthcare when needed [98]. In addition, ANC provides preventive services (vaccination, iron and folic acid supplementation, and mother-to-child HIV transmission prevention) as well as screening for risk factors such as hypertension and diabetes [98, 99]. If these check-ups, counselling, preventive services, and screening are not provided as part of regular ANC follow-ups, these problems may go unnoticed and untreated, increasing the likelihood of a near miss. Moreover, ANC is often linked to planning for skilled birth attendance, as part of the birth preparedness and complication readiness (BPCR) plan [100]. Thus, a lack of ANC could lower the likelihood of accessing skilled delivery services, increase the risk of complications during childbirth, and limit access to emergency obstetric care, all of which increase the risk of severe maternal outcomes. Thus, efforts should be made to ensure universal access to ANC for a positive pregnancy experience by addressing barriers to accessing healthcare services for pregnant women, improving the healthcare system, and promoting educational campaigns to improve maternal and neonatal outcomes.

The current findings regarding the higher risk of MNM among women with three delays of service use were supported by previous studies [101,102,103]. These three delays refer to a framework used in maternal health to identify and address factors contributing to MNM [104]. An expectant mother who experiences the first delay (delay at home), the second delay (delay on the road to the health facility), and the third delay (delay at the health facility) could experience greater difficulties by delaying timely care during pregnancy and childbirth [103,104,105,106]. The possible reasons behind those delays are being unaware of danger signs, delayed decision-making, lack of transportation, and ill-equipped health system. Thus, African governments need to work together to address all three delays through community education, better infrastructure construction, and improved care quality.

Women with a history of Caesarean section were at a higher risk of experiencing MNM, which is in line with previous studies conducted in Brazil [96], India [107], and Thailand [108]. Caesarean section (CS) is a life-saving intervention for the fetus, mother, or both at the time of life-threatening conditions such as obstructed labor, fetal distress, and obstetric hemorrhage [108]. However, deliveries after previous CS have been reported to have a higher risk of adverse pregnancy outcomes. This could be because scar tissue from previous CS can complicate subsequent deliveries by causing uterine rupture and antepartum hemorrhage (due to placenta previa and placenta accreta) [109,110,111]. This study implies that when evaluating the clinical grounds for CS, healthcare providers ought to weigh its potential risk over its benefits (especially in the case of elective CS) and may consider alternative birthing options when appropriate. On the other hand, healthcare personnel should pay special attention to women with a history of CS during prenatal and intrapartum care.

Similarly, women with pre-existing medical conditions had a higher risk of developing MNM, in line with similar studies [4, 101, 112, 113]. This might be due to chronic medical conditions, such as hypertension or diabetes, which can lead to life-threatening complications during pregnancy, such as preeclampsia, gestational diabetes, or worsening of an existing medical condition [112, 113]. In addition, these medical disorders might impair the immune system [114], leaving pregnant women more susceptible to infections, which, if not treated effectively and promptly, can lead to severe maternal outcomes.

Background characteristics, such as lack of formal education, rural residence, low monthly income, and distance from health facilities, were also identified as significant predictors of MNM. Previous studies have supported these findings [115, 116]. A possible explanation could be that those women have limited access to healthcare services and may need to travel far to reach health facilities, which might result in delays in receiving essential maternity care [13]. Furthermore, they may have limited access to maternal healthcare, which might result in delayed detection and management of complications that lead to MNM. Thus, a comprehensive approach is needed to ensure universal access to maternal healthcare for women in hard-to-reach areas by improving healthcare infrastructure and promoting community awareness.

This study has both strengths and limitations. This is the first systematic review and meta-analysis in Africa to examine the pooled prevalence of MNM and its contributing factors. In addition, the number and the quality of articles that have been meta-analysis are high, reflecting a comprehensive view of MNM. Furthermore, this study revealed primary severe maternal problems that resulted in MNM. Thus, the findings could be used as input for stakeholders in Africa who work on reducing maternal mortality and morbidities. However, the findings should be interpreted in light of the following limitations. First, since the vast majority of the included studies were hospital-based and the data collection techniques relied on record review, the findings may not be generalizable to near-misses that were not present at service delivery points. Furthermore, as the majority of the articles were from Eastern, Western, and Southern African regions, this may raise the issue of generalizability.

Conclusion

The prevalence of MNM was 73.77/1000 live births, with higher rates reported in eastern, western, and middle African countries. The risk of MNM increased among women living in rural areas, possessing low income, not attended formal education, not received ANC, living far from health facilities, reported three delays in seeking health service, have a previous history of CS, and had pre-existing medical conditions. A comprehensive approach is needed to strengthen and ensure universal access to education and maternal health services, especially ANC, to women in hard-to-reach areas by improving healthcare infrastructure and promoting community awareness. Stakeholders should work together to tackle all three delays through community education and awareness campaigns, improve access to road infrastructure and transportation, and improve the quality of care provided at service delivery points.