Midwifery 82 (2020) 102597

Contents lists available at ScienceDirect

Midwifery
journal homepage: www.elsevier.com/locate/midw

Incidence of maternal near miss among women in labour admitted to

hospitals in Ethiopia
Ayele Geleto a,b,∗, Catherine Chojenta b, Tefera Taddele c, Deborah Loxton b
a
School of Public Health, College of Health and Medical Sciences, Haramaya University, Harar, Ethiopia
b
Research Centre for Generational Health and Ageing, School of Medicine and Public Health, Faculty of Health and Medicine, the University of Newcastle,
Newcatle, Australia
c
Health System and Reproductive Health Directorate, the Ethiopian Public Health Institute, Addis Ababa, Ethiopia

a r t i c l e i n f o a b s t r a c t

Article history: Objectives: To assess the incidence of maternal near miss and contributing factors among hospitals in
Received 22 June 2019 Ethiopia. The study also assessed the ability of hospitals to provide signal functions of emergency obstet-
Revised 18 October 2019
ric care and its regional distribution.
Accepted 2 December 2019
Design: A national dataset accessed from the Ethiopian Public Health Institute were analysed to assess
the incidence of maternal near miss and mortality index among women admitted to hospitals with ob-
Keywords:
stetric complications.
Maternal near miss
Maternal near miss incidence ratio Setting: Maternal health indicators including obstetric complications, maternal deaths and births con-
Mortality index ducted at all hospitals available in Ethiopia were included.
Ethiopia
Measurements: The maternal near miss incidence ratio, which is the number of near miss cases per
1,0 0 0 live births, and the mortality index were presented descriptively. Chi-squared test at p value ≤ 0.05
was used to assess the presence of significant regional differences of the provision of signal functions of
emergency obstetric care.
Results: In 2015, 78,195 women were admitted to hospitals with both the direct (68,002) and indirect
(10,193) causes of maternal mortality. Of women who experienced the direct causes, 435 died which
means there were 67,567 maternal near miss cases. In the same year, 323,824 live births were reported
in hospitals, making the crude maternal near miss incidence ratio of 20.8% (9.1-38.8%) and mortality in-
dex of 0.64% (435/68,002) for the direct causes of maternal mortality. A significant regional variation was
observed with regard to incidence of maternal near miss, mortality index and the provision of signal
functions of emergency obstetric care. Administration of parenteral antibiotics was the most frequently
practiced signal function of emergency obstetric care while blood transfusion was the least provided sig-
nal function.
Conclusions: In Ethiopian hospitals, the incidence of maternal near miss was unacceptably high. A signif-
icant regional variation was detected with regard to maternal near miss incidence ratio, mortality index
and the provision of signal functions of emergency obstetric care. The Ethiopian government needs to
work on equitable resource distribution and quality improvement initiatives in order to close the de-
tected regional variations.

Abbreviations: AIDS, Acquired Immunodeficiency Syndrome; APH, Antepartum Haemorrhage; CAFÉ, Computer Assisted Field Editing; CAPI, Computer Assisted Personal
Interviewing; CEmOC, Comprehensive Emergency Obstetric Care; EmOC, Emergency Obstetric Care; EPHI, Ethiopian Public Health Institute; HIV, Human Immunodeficiency
Virus; HREC, Health Research Ethics Committee; IFSS, Internet File Streaming System; MD, Maternal Death; MI, Mortality Index; MMR, Maternal Mortality Rate; MNM,
Maternal Near Miss; MNMIR, Maternal Near Miss Incidence Ratio; PPH, Postpartum Haemorrhage; WHO, World Health Organization; IV, Intravenous; MRP, Manual Removal
of Placenta; RRP, Removal of Retained Products; SGD, Sustainable Development Goal; SNNPR, Southern Nations, Nationalities and People Region.
∗
Corresponding author at: School of Public Health, College of Health and Medical Sciences, Haramaya University, Harar, Ethiopia.
E-mail addresses: [email protected] (A. Geleto), [email protected] (C. Chojenta), [email protected] (D. Loxton).

https://doi.org/10.1016/j.midw.2019.102597
0266-6138/© 2019 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license.
(http://creativecommons.org/licenses/by-nc-nd/4.0/)

Downloaded for Anonymous User (n/a) at National Library of Indonesia from ClinicalKey.com/nursing by Elsevier on April 14, 2020.
For personal use only. No other uses without permission. Copyright ©2020. Elsevier Inc. All rights reserved.
2 A. Geleto, C. Chojenta and T. Taddele et al. / Midwifery 82 (2020) 102597

Implications for practice: The Ethiopian government needs to practice evidence-based maternal health
strategies, including capacity building of the regional hospitals in order to improve the distribution of
resources and quality of maternal health.
© 2019 The Author(s). Published by Elsevier Ltd.
This is an open access article under the CC BY-NC-ND license.
(http://creativecommons.org/licenses/by-nc-nd/4.0/)

benefits because MNM happens more frequently and has similar

Background causes with maternal deaths (Chhabra 2014; Kalhan et al., 2017).
In developing countries, where MMR has shown a declining trend
In developing countries, including Ethiopia, maternal mortality over the last two decades with the advent of different targeted
and morbidity continue to be unacceptably high. The risk of ma- interventions (WHO, UNICEF et al., 2015) and is usually under-
ternal death in this region of the world is 1 in 41 live births as reported due to high proportion of home delivery (EDHS 2016),
opposed to 1 in 3300 live births in developed nations. For every studying MNM shows the real figure of the quality of obstetric
woman who dies, 20 more women experience acute and chronic care. Studying MNM also allows rigorous quantitative analysis of
obstetric complications from the direct causes of maternal mortal- factors leading to maternal mortality (Kasahun and Wako 2018).
ity (Firoz et al., 2013; WHO, UNICEF et al., 2015). The incidence of MNM varies across world regions. It is highly
The high Maternal Mortality Ratio (MMR) in developing coun- aggregated in middle and low income countries with the high-
tries was recognized and included in the Sustainable Develop- est prevalence among African and Asian countries (Tunçalp et al.,
ment Goals (SDGs) where target was set to reduce MMR to be- 2012). According to the results of a systematic review conducted
low 70 per 10 0,0 0 0 live births for all countries by the year 2030 by the WHO, showed that the prevalence of MNM ranges be-
(IAEG 2016). Identifying and targeting the major causes of mater- tween 0.80 and 8.23% among studies which used clinical crite-
nal deaths would contribute to the realization of the SDG’s target ria and 0.01–2.99% among studies that used management criteria
of reducing maternal mortality. Ethiopia, one of the countries with (Say et al., 2004). A recent systematic review also indicate that in
the highest MMR in the world (Hogan et al., 2010), is hoping to sub-Saharan African countries, the median MNM ratio was 24.2 per
benefit from this global plan. 10 0 0 live births (Tura et al., 2019).
Maternal near miss (MNM) is a situation where a woman nearly In Ethiopia, the MNM incidence ratio (MNMIR) varies among
dies from complications of pregnancy, childbirth or within 42 days different regions of the country. The few available literature shows
of termination of the pregnancy, but survives either due to the care that MNMIR was lowest in Addis Ababa, the capital city of the
she receives or due to chance (Say et al., 2009). In order to elimi- country (8.01 per 10 0 0 live births) (Liyew et al., 2017). While MN-
nate confusions of using different definitions of MNM, the World MIR was 17 per 10 0 0 live births in the Harari region (Tura et al.,
Health Organization (WHO) developed a set of criteria that are 2018), the highest MNMIR (23.3%) was reported in the Amhara re-
used to precisely define MNM. According to the WHO MNM cri- gion (Abate and Dile, 2015). The underlying causes of the majority
teria, a woman who presents with any of the conditions listed in of MNM were hypertensive disorders and obstetric haemorrhage.
Table 1 that could occur during pregnancy and childbirth but sur- However, studies about MNM are scarce and the available litera-
vived the complication is considered as a MNM case (Say et al., ture are not representative of the country as they were conducted
2009). in a small number of health facilities of some regions. The avail-
Although MMR in developing countries remains quite high, in- able literature also did not include the case of private hospitals
vestigating MNM has become the best measure of quality of ob- (Liyew et al., 2017; Tura et al., 2018).
stetric care (Pattinson and Hall 2003; World Wealth Organiza- Two categories of obstetric services commonly known as ‘sig-
tion 2004). Studying MNM rather than maternal deaths has further nal functions of Emergency Obstetric Care (EmOC) have been

Table 1
World health organization maternal near miss definition criteria (Say et al., 2009).

Downloaded for Anonymous User (n/a) at National Library of Indonesia from ClinicalKey.com/nursing by Elsevier on April 14, 2020.
For personal use only. No other uses without permission. Copyright ©2020. Elsevier Inc. All rights reserved.
 A. Geleto, C. Chojenta and T. Taddele et al. / Midwifery 82 (2020) 102597 3

recommended as critical lifesaving care for women who experi- maternal and neonatal health indicators from all health facilities
enced obstetric complications. The first category is ‘signal func- found in the country. In the EPHI’s survey, all functional health
tions’ of Basic Emergency Obstetric Care (BEmOC) which include facilities available in Ethiopia including hospitals, health centers
administration of parenteral antibiotics, parenteral anticonvulsants, and clinics, which had attended births in the last 12 months of
parenteral uterotonics, removal of retained products, manual re- the data collection period (N = 3804), were included. Although
moval of the placenta and assisted vaginal delivery. The second data about maternal and neonatal health indicators were collected
category is ‘signal functions’ of Comprehensive Emergency Obstet- from all public and private health facilities, the current analysis
ric Care (CEmOC), which include the provision of caesarean section was conducted using only hospitals’ data. Altogether, there were
and blood transfusion in addition to all the signal functions of BE- 293 hospitals in Ethiopia and all of them were included into the
mOC (UNPF 2009). current study. Hospitals’ maternal data were used in this analysis
The government of Ethiopia is striving to reduce MMR through as women who have experienced obstetric complications are most
implementation of different targeted interventions. Several actions likely referred to hospitals to receive specialized treatment.
were taken to improve access and utilization of maternal health
services and all maternity health services were made cost free Study participants
at all public health facilities (Federal Ministry of Health Ethiopia
2015). Despite all of these efforts, the maternal morbidity and mor- Altogether, 293 hospitals (58 private and 235 public hospitals)
tality remain unacceptably high (EDHS 2016). Therefore, the mag- were included in the national survey conducted by the EPHI. One
nitude and possible causes of MNM need to be well-assessed using hundred and sixty primary hospitals, 103 general hospitals and 30
representative data. Thus, the main aim of the current study was specialized hospitals were included from all regions and city ad-
to assess the incidence of MNM and contributing factors among ministrations of the country. Data of women who sustained all
women in labour who are admitted to hospitals in Ethiopia. The types of obstetric complications were included in to the current
ability of the hospitals in the provision of signal functions of emer- analysis.
gency obstetric care and its regional distributions were also ac-
cessed. Inclusion criteria

Methods The data about type and magnitude of obstetric complications,

maternal deaths and all live births conducted in 2015 at all avail-
Study area and period able hospitals were included. Data about maternal health from all
public and private hospitals found in Ethiopia were included in the
Ethiopia is the second most populous country in Africa, after analysis.
Nigeria. According to the elaboration of the latest United Nations
data (2019), the current population of Ethiopia is 112,921,560 as Data collection process
of October 2019 (Worldometers 2019). Currently, the total fertility
rate of Ethiopia is 4.6, the contraceptive prevalence rate for cur- During the EPHI survey, several maternal and neonatal health
rently married women is 36%, and the unmet need for family plan- indicators were collected from hospitals during a single facility
ning is 22%. The proportion of pregnant women who received ante- visit. The data about performance of the Emergency Obstetric and
natal care from a skilled provider was 62% while only 32% of preg- Newborn Care (EmONC) signal functions were retrospectively col-
nant mothers had at least four antenatal care visits during their lected from all hospitals using a standardized questionnaire. The
pregnancy. Only 17% of women were reported as meeting the Safe reference period for the data collection was 1st January to 31st
Motherhood Programme’s recommendation of receiving a postnatal December 2015 inclusive and the data were collected from May to
care check within two days of delivery (CSA and ICF 2016). December 2016. A standardized tool was used to collect the data
For this study, we accessed and analysed a national dataset from the hospitals’ registers and records. The data were collected
from the Ethiopian Public Health Institute (EPHI). The EPHI is a using Computer Assisted Field Editing (CAFE) and Computer As-
large public health research institute conducting various researches sisted Personal Interviewing (CAPI). The cleaned data were sent to
on several priority public health issues. In 2015, the EPHI con- the central office of EPHI every day using an Internet File Stream-
ducted a national cross-sectional survey and collected data on ing System (IFSS). All data management was done using CSPro 6.1

Table 2
Operational definitions of the technical terms used throughout the paper.

Operational definitions

 Major direct causes of maternal deaths: in this study, the following were considered as major direct causes of maternal deaths: Antepartum Haemorrhage
(APH), Postpartum Haemorrhage (PPH), retained placenta, prolonged/obstructed labour, ruptured uterus, severe eclampsia/preeclampsia, complications of
abortion and ectopic pregnancy.
 Indirect causes of maternal deaths: are potentially lethal complications, which could occur during pregnancy, childbirth and immediate postpartum
period but not induced by the pregnancy. These conditions included malaria, anaemia, hepatitis, HIV/AIDS, and other non-specific life-threatening conditions.
 Maternal Near Miss (MNM): refers to a woman who nearly died but survived a complication that occurred during pregnancy, childbirth or within 42 days
of termination of pregnancy. It was estimated by subtracting the number of women with life-threatening conditions minus number of women who died of
the conditions.
 Maternal death (MD): is the death of a woman while pregnant or within 42 days of termination of pregnancy.
 Live birth (LB): refers to the complete expulsion or extraction from its mother of a product of conception, irrespective of the duration of the pregnancy,
which, after such separation, breathes or shows any other evidence of life.
 Women with life-threatening conditions (WLTC): refers to all women either who qualified as having maternal near miss or who died
(WLTC = MNM + MD)
 Maternal Near Miss Incidence Ratio (MNMIR): refers to the number of maternal near miss cases per 1000 live births; the numerator being the magnitude
of MNM and the denominator is live births conducted at the hospitals.
 Mortality Index (MI): refers to the number of maternal deaths divided by the number of women who experienced life-threatening conditions, expressed
as percent. The higher the mortality index, the more women with life-threatening conditions die (low quality of care), whereas the lower the index, the
fewer women with life-threatening conditions die (better quality of care). [MI = MD/(MNM+MD)] (Say et al., 2009).

Downloaded for Anonymous User (n/a) at National Library of Indonesia from ClinicalKey.com/nursing by Elsevier on April 14, 2020.
For personal use only. No other uses without permission. Copyright ©2020. Elsevier Inc. All rights reserved.
4 A. Geleto, C. Chojenta and T. Taddele et al. / Midwifery 82 (2020) 102597

data programming software. Although several maternal and neona- the complications. The indirect causes of maternal death, which are
tal data were collected during the survey, for this paper, we used potentially lethal complications including malaria, anaemia, hep-
the data about the number of live births conducted in hospitals, atitis, HIV/AIDS, and other nonspecific life threatening conditions,
number of women who experienced obstetric complications and accounted for 10,193 (13.0%) of the complications. There were 481
number of maternal deaths recorded in hospitals. maternal deaths recorded in Ethiopian hospitals in the same year.
Healthcare professionals with at least a bachelor’s degree were The majority (N = 435) of these deaths were due to direct causes
recruited to conduct data collection. The data collectors were while 46 of the deaths were attributed to indirect causes. Pro-
trained on interview techniques, survey tools and field procedures. longed/obstructed labour (20.3%) was the leading cause of morbid-
The survey was conducted under close supervision of the techni- ity while hypertensive disorder (25.2%) was the leading cause of
cal working group, which included experts from different partners mortality (Table 3).
and professional associations. In addition, regional coordinators su- Fig. 1 shows how the provision of EmOC signal functions were
pervised the data collection process and conducted spot-checking distributed among hospitals. Overall, administration of parenteral
to ensure the quality of the data. Several terms and phrases used antibiotics was the most frequently practiced EmOC signal func-
throughout this paper are operational defind hereunder (Table 2). tion as it was provided by 282 (96.2%) of the hospitals. Administra-
tion of parenteral anticonvulsant (N = 278; 94.8%) was the second
Data analysis most frequently provided signal function. The signal functions of
CEmOC were the least provided services as only 189 (64.5%) and
The data were accessed in IBM SPSS Statistics for Windows, 237 (80.8%) of the hospitals provide blood transfusion and cae-
version 24 (IBM Corp., Armonk, N.Y., USA) from the EPHI central sarean section respectively.
server. Then the dataset was exported to Stata version 15 for anal- Gambella region has only one hospital and it provided all EmOC
ysis. Descriptive statistics including means, percentages, frequency signal functions. Harari region (100%) and Addis Ababa city admin-
tables and ratios were performed to describe facility specific char- istration (90.9%) were the most frequent providers of caesarean
acteristics, and the magnitude of MNM and the MI. The causes of section while Somali (90%) and Oromia (82.2%) regions were the
both the direct and indirect MNM and their case fatality rates were most likely to provide blood transfusion. There were statistically
also analysed. The ability of hospitals in the provision of the sig- significant regional difference in the provision of parenteral utero-
nal functions of EmOC were analysed with due consideration to tonic (p = 0.012), manual removal of placenta (p = 0.003), removal
regions. The MNMIR and MI were presented separately for each of retained product (p = 0.007) and blood transfusion (p = 0.001).
region and city administration for the purpose of comparison. A However, the difference among regions with regard to the provi-
Chi-squared test was performed to examine the presence of sig- sion of the remaining EmOC signal functions was not statistically
nificant regional differences of the provision of signal functions of significant (Table 4).
EmOC. Maternal Near Miss Incidence Ratio refers to the total num-
ber of MNM per 10 0 0 live births. Mortality index is the ratio of Maternal near miss indicators
maternal deaths to the total number of women who sustained life-
threatening complications. Mortality index indicates the quality of Of the 68,002 women who experienced the direct obstetric
obstetric care, as MI > 1 is interpreted as excess mortality and MI complications (40,080 major direct and 27,922 ‘other direct’ com-
< 1 means fewer deaths than expected (Vandecruysa et al., 2002). plications) 435 died, which means there were 67,567 MNM cases.
Overall, 323,824 livebirths were reported in hospitals, making the
Results MNMIR due to direct obstetric causes 20.8%. The highest MNMIR
was reported in Benshangul Gumuze (38.8%) and the lowest was
Magnitude of obstetric complications and maternal deaths observed in Gambella region (9.1%). With regard to MI due to di-
rect obstetric causes, although an acceptable level was observed
In 2015, 78,195 women were admitted to hospitals across nationally, significant regional variations were observed and it was
Ethiopia due to both the direct and indirect causes of maternal highest in Gabella region (3.82%). Excess mortality was observed in
mortality. Slightly more than half, (N = 40,080; 51.3%) of these Afar, Gambella, Harari and Somali regions as mortality indexes in
complications were due to the major direct causes of maternal these regions exceed one (Table 5).
deaths. The ‘other direct causes’ including premature rupture of Additionally, of the 10,193 cases of indirect causes of mater-
membrane, post-term labour, cord prolapse, breech presentation, nal deaths, 46 women died. HIV/AIDS 6249 (61%) was the leading
and other possible problems that are not considered as a major cause of the indirect maternal morbidity while severe anaemia 19
direct cause of maternal mortality accounted for 27,922 (35.7%) of (41%) was the leading indirect cause of maternal death (Fig. 2). The

Table 3
The magnitude of obstetric complications and severe maternal outcomes among hospitals in
Ethiopia, 2015.

Complications Number of women affected N (%) Maternal Death N (%)

Antepartum Haemorrhage (APH) 4462 (5.70) 23 (4.8)

Postpartum Haemorrhage (PPH) 2811 (3.60) 81 (16.9)
Retained placenta 2201 (2.81) 4 (0.8)
Prolonged/obstructed labour 15,875 (20.30) 25 (5.2)
Rupture of uterus 1516 (1.92) 41 (8.5)
Postpartum sepsis 1460 (1.87) 17 (3.5)
Severe Eclampsia/Preeclampsia 7912 (10.12) 121 (25.2)
Abortion complications 2042 (2.61) 6 (1.3)
Ectopic pregnancy 1801 (2.30) 0 (0.0)
Other direct complications 27,922 (35.71) 117 (24.2)
Indirect causes 10,193 (13.03) 46 (9.6)
Total 78,195 481

Downloaded for Anonymous User (n/a) at National Library of Indonesia from ClinicalKey.com/nursing by Elsevier on April 14, 2020.
For personal use only. No other uses without permission. Copyright ©2020. Elsevier Inc. All rights reserved.
 A. Geleto, C. Chojenta and T. Taddele et al. / Midwifery 82 (2020) 102597 5

Fig. 1. Distribution of hospitals in Ethiopia by their ability in provision of emergency obstetric care signal functions, 2015.

Table 4
Regional distribution of hospitals in Ethiopia by the provision of EmOC signal functions, 2015.

Total IV Antibiotic IV uterotonic IV anticonvulsant AVD MRP RRP Blood transfusion

Interventions hospitals N (%) N (%) N (%) N (%) N (%) N (%) N (%) CS N (%)

AA 33 33 (100) 29 (87.8) 26 (78.7) 23 (69.7) 19 (57.5) 19 (57.5) 24 (72.7) 30 (90.9)

Afar 6 6 (100) 4 (66.7) 5 (83.3) 4 (66.7) 5 (83.3) 5 (83.3) 2 (33.3) 5 (83.3)
Amhara 56 54 (96.4) 52 (96.3) 46 (82.1) 50 (89.3) 48 (85.7) 42 (75) 37 (66.1) 44 (78.5)
BG 3 3 (100) 3 (100) 2 (66.7) 2 (66.7) 2 (66.7) 3 (100) 2 (66.7) 2 (66.7)
DD 6 5 (83.3) 5 (83.3) 3 (50) 4 (66.7) 3 (500 2 (33.3) 4 (66.7) 5 (83.3)
Gambella 1 1 (100) 1 (100) 1 (100) 1 (100) 1 (100) 1 (100 1 (100) 1 (100)
Harari 6 6 (100) 6 (100) 3 (50) 6 (100) 4 (66.7) 4 (66.7) 4 (66.7) 6 (100)
Oromia 73 72 (98.6) 73 (100) 64 (87.6) 66 (90.4) 67 (91.7) 66 (90.4) 60 (82.2) 65 (89)
SNNPR 60 54 (90) 56 (93.3) 46 (76.7) 51 (85) 50 (83.3) 46 (76.7) 27 (45) 42 (70)
Somali 10 10 (100) 10 (100) 8 (80) 8 (80) 9 (90) 9 (90) 9 (90) 7 (70)
Tigray 39 38 (97.4) 39 (100) 33 (84.6) 34 (87.2) 34 (87.2) 31 (79.4) 19 (48.7) 30 (76.9)
Total 293 282 (96.2) 278 (94.8) 237 (80.8) 249 (84.9) 242 (82.5) 228 (77.8) 189 (64.5) 237 (80.8)
(X2 test) 33 p = 0.232 p = 0.012 p = 0.32 p = 0.17 p = 0.003 P = 0.007 p = 0.001 P = 0.206

IV=Intravenous; AVD=Assisted Vaginal Delivery; MRP=Manual Removal of Placenta; RRP=Removal of Retained Products; CS=Caesarean Section, AA=Addis Ababa,
BG=Benishangul Gumuz, DD=Dire Dawa, SNNPR=Sothern Nations, Nationalities and People Region.

Table 5
Regional distributions of incidence of maternal near miss among hospitals in Ethiopia, 2015.

Regions WLTC (N) Total number (N) of Near miss cases (N) Total livebirths MNMIR/100 Livebirths MI (%)
maternal deaths in hospitals (N)

Direct Indirect Direct Indirect Direct Indirect Direct Indirect Direct Indirect

Addis Ababa 14,453 2698 39 5 14,414 2693 45,049 31.9 5.9 0.27 0.18
Afar 358 187 5 4 353 183 1643 21.4 11.1 1.40 2.18
Amhara 11,263 1687 89 6 11,174 1681 43,899 25.4 3.8 0.79 0.35
Benish. Gumz 1302 268 4 0 1298 268 3338 38.8 8.0 0.31 0
Dire Dawa 1437 133 8 0 1429 133 5330 26.8 2.4 0.55 0
Gambella 157 2 6 0 151 2 1667 9.1 0.1 3.82 0
Harari 1384 112 15 0 1369 112 4823 28.3 2.3 1.10 0
Oromia 17,084 2031 128 14 16,956 114 97,105 17.4 0.1 0.75 12.28
SNNPR 12,942 1179 82 9 12,860 73 68,422 18.8 0.1 0.63 12.32
Somali 1710 548 29 5 1681 24 8181 20.5 0.2 1.69 20.83
Tigray 5912 1348 30 3 5882 27 44,367 13.2 0.06 0.51 11.11
Total 68,002 10,193 435 46 67,567 10,147 323,824 20.8 3.1 0.64 0.45

MNMIR: Maternal Near Miss Incidence Ratio, WLTC: Women with Life Threatening Complications, MI: Mortality Index.

MNMIR due to the indirect causes of maternal mortality was high- Discussion
est in Afar region (11.1%) followed by Benishangul Gumuz region
(8.0%). The mortality index for indirect maternal morbidity was Causes of maternal near miss
highest in Somali region (20.83%) followed by SNNPR (12.32%) and
Oromia (12.28%). However, the national level of mortality index In this study, hypertensive disorder of pregnancy was the
for indirect causes (0.45%) was within the recommended thresh- leading cause of MNM followed by obstetric haemorrhage. The
old (Table 5). current findings are supported by findings of several studies

Downloaded for Anonymous User (n/a) at National Library of Indonesia from ClinicalKey.com/nursing by Elsevier on April 14, 2020.
For personal use only. No other uses without permission. Copyright ©2020. Elsevier Inc. All rights reserved.
6 A. Geleto, C. Chojenta and T. Taddele et al. / Midwifery 82 (2020) 102597

Fig. 2. Morbidity and mortality due to indirect maternal complications among hospitals in Ethiopia, 2015.

conducted in high and middle-income countries (Say et al., 2009; A significant regional difference exists in the provision of EmOC
Donati et al., 2012; Ghazal-Aswad et al., 2013) and other African signal functions. A full coverage of EmOC signal functions was ob-
studies (Adeoye et al., 2013; Nelissen et al., 2013; Rulisa et al., served in Gambella region. This could be due to the availability
2015) including Ethiopia (Liyew et al., 2017; Tura et al., 2018). of a hospital, which facilitated improved investment of resources
For women who experienced hypertension and bleeding, treatment on the only hospital available in the region. Harari region and Ad-
with drugs might not be considered (Getaneh and Kumbi 2010) dis Ababa city administration were the most frequent providers of
or there might be delayed initiation of management (Ridge et al., caesarean section. The two regions are urban centers with better
2010; Jackson et al., 2017). In some conditions, anticonvulsants and health services coverage. It is a usual finding that EmOC facilities
antihypertensive drugs might not be available at the health facili- in sub-Saharan Africa are concentrated in capitals and urban areas
ties (Gaym et al., 2011) and this challenged initiation of early treat- (Banke-Thomas et al., 2019). The presence of regional difference in
ment. It might also be attributed to the fact that, in most develop- the provision of EmOC signal functions in our study is supported
ing countries, the prevalence of blood transfusion for women in by previous studies in developing countries (Mony et al., 2013) in-
need is very low (Li et al., 2017). cluding Ethiopia (Admasu et al., 2011).
In the current study, postpartum sepsis was the least registered Nearly 13% (10,147/77,714) of the overall MNM cases were ac-
cause of MNM. This finding is consistent with findings of several counted to by the indirect causes of maternal deaths. HIV/AIDS is
studies conducted in multiple African countries where sepsis was the leading cause of maternal morbidity among the indirect causes
reported as the lowest cause of obstetric complications (van den of maternal deaths. Anaemia is the leading cause of the indirect
Akker et al. 2011; Adeoye et al., 2013; Nelissen et al., 2013). Reports maternal death and is the second most frequently observed cause
of a study conducted in Ethiopia also revealed that maternal infec- of the indirect MNM. Hepatitis is the least frequent cause of the
tion is the lowest among the top four causes of maternal mortality indirect MNM. Our findings are supported by other previous stud-
(Berhan and Berhan 2014). This lower rate of postpartum sepsis as ies where a high proportion of MNM due to the indirect causes of
a cause of MNM could be explained by the wider coverage of an- maternal morbidity was caused by HIV/AIDS and hepatitis was the
tibiotic administration for the treatment of sepsis (Windsma et al., least frequent observed cases of MNM (Abdella 2010). Reports of
2017). several studies showed that anaemia is the most frequent cause of
The administration of parenteral antibiotics was the most fre- maternal complications (Liyew et al., 2017; Tura et al., 2018). in our
quently practiced EmOC signal functions followed by administra- study, malaria cases were reduced when compared to previous re-
tion of parenteral anticonvulsant. Blood transfusion and caesarean ports (UNDP 2012), which might be explained by the effectiveness
section were the least frequently practiced CEmOC signal func- of the preventive health care approach in Ethiopia.
tions. Several existing study findings shows that administration
of parenteral antibiotic and anticonvulsant were the most fre- Maternal near miss indicators
quently practiced EmOC signal functions (Maswanya et al., 2018;
Bintabara et al., 2019). The infrequent performance of the CEmOC The overall national magnitude of MNMIR of the direct causes
might be attributed to policy restrictions, lack of trained workforce of maternal deaths among Ethiopian hospitals was 20.8%. Previ-
or lack of medical supplies (Maswanya et al., 2018). This might ous studies conducted in selected cities of Ethiopia (Berhane et al.,
also be attributed to the shortage of physician in rural and district 2012; Liyew et al., 2017; Tura et al., 2018) and other African coun-
hospitals. In Ethiopia, staff retention mechanisms was too poor re- tries (Nelissen et al., 2013; Tuncalp et al., 2013) revealed a lower
sulted in high staff turnover especially in rural and district facilities rate of MNMIR than the current findings. The observed variation
(Assefa et al., 2016). might be explained by disparities in the case definitions and the

Downloaded for Anonymous User (n/a) at National Library of Indonesia from ClinicalKey.com/nursing by Elsevier on April 14, 2020.
For personal use only. No other uses without permission. Copyright ©2020. Elsevier Inc. All rights reserved.
 A. Geleto, C. Chojenta and T. Taddele et al. / Midwifery 82 (2020) 102597 7

study design (Liyew et al., 2017). The current study included all of mothers who experience obstetric complications seek treatment
cases of mothers who experienced obstetric complications while from hospitals. This might increase the denominator while the nu-
all of the previous studies employed some forms of sampling merator remained constant. This might have caused us to overesti-
methods, which might have introduced some sorts of sampling mate the magnitude of MNMIR. Selection bias might be introduced
errors. Additionally, deliveries can be conducted in lower level as we included only hospitals’ data by excluding lower level health
health facilities including health centers and clinics that are more facilities where several births are attended. Therefore, the find-
proximal to the community while mothers who sustained obstet- ings of this study can only be generalized to hospitals in Ethiopia.
ric complications seek advanced management from hospitals. This Several deaths might not have been reported as many Ethiopian
might increase the MNM cases in hospitals, which might explain women gave birth at home, which might have affected the magni-
the higher MNMIR in our study, as our study was conducted only tude of MI.
in hospitals.
The current study revealed existence of regional variation in
Conclusions
MNMIR that ranged from 9.1% in Gambella region to 38.8% in Ben-
shangul Gumuze. Likewise, previous studies conducted in various
The current study revealed a higher MNMIR as compared to
parts of Ethiopia have reported different MNMIR, 101/10 0 0 live
most previous studies in African countries including Ethiopia. Ob-
births in Tigray (Berhane et al., 2012), 80/10 0 0 live births in Ad-
structed labour was found to be the leading cause of MNM while
dis Ababa (Liyew et al., 2017) and 8.01/10 0 0 live births in Harari
pregnancy-induced hypertension was the leading cause of mater-
(Tura et al., 2018). The regional variations in MNMIR might be at-
nal deaths. It was demonstrated that the provision of the signal
tributed to variation in the geographical distribution of health in-
functions of BEmOC was more frequent than the CEmOC. Regional
frastructure, magnitude of the catchment population and the urban
variations exist in the magnitude of MNMIR and MI that might be
rural distribution of the hospitals.
attributed to the significant difference in the provision of the sig-
Finally, the current study revealed an acceptable national MI of
nal functions of EmOC among regions. However, MI due to both
the direct causes of maternal mortality although excess deaths oc-
the direct and indirect causes of maternal deaths at the national
curred in some regions as MI in these regions exceeds one. Our
level appears to be within the acceptable level although higher
findings were consistent with findings of previous studies. For ex-
level of MI were noted in some regions. Therefore, evidence-based
ample, Tura et al. (2018) found a MI of nearly 0.17 (Tura et al.,
maternal health interventions should be designed for management
2018) while a study in Tanzania reported a mortality index of 0.13
of obstetric complications.
(Nelissen et al., 2013). It should be noted that in most developing
The high incidence of MNM in Ethiopian hospitals might have
countries, the majority of women who sustained obstetric compli-
occurred because of the lower coverage of the signal functions of
cations arrive at the hospitals after their conditions are advanced,
EmOC; especially of the CEmOC. Admitting the possibility for se-
carrying minimal chance to recover (Liyew et al., 2017).
lection bias, higher MI than the acceptable level in some regions
Similarly, the national MI for the indirect causes of mater-
can be considered as an indicator of poor quality of obstetric care
nal mortality was within the acceptable threshold although excess
at hospitals found in those regions (Say et al., 2009). Therefore,
mortalities were observed in Afar, Gambella, Somali and Harari re-
health managers should strengthen evidence-based practice to im-
gions. The MI in the first three regions might be high because
prove quality of obstetric care at hospitals in regions where MI was
these regions are pastoral areas where health service coverage is
higher. While developing strategies for the reduction of MNM, pol-
minimal (UNDP 2012). However, MI in Harari region might be high
icy makers need to consider equitable distribution of resources in
because it is the smallest region and serves as a referral center for
order to close the observed gap of regional variations in the pro-
several remote regions hence hosting several complicated referral
vision of the EmOC signal functions. Further studies with stronger
cases.
design, which include home deliveries, should be conducted in or-
der to obtain accurate magnitude of MNMIR and MI. Future studies
Strengths and limitations of the study
should also include the data of lower level health facilities where
the majority of deliveries were conducted.
This study has several strengths. The findings of this study are
highly representative and can apply to all regions of Ethiopia as we
used national representative dataset. These findings might also ap- Ethical approval
ply to other developing countries with similar socioeconomic char-
acteristics. Experts from national and international partners were The survey was granted ethical approval from the Scientific and
involved in the data collection and management processes from Ethical Review Office of the EPHI (approval number: EPHI-6-13-
the commencement to the finalization of the survey. Hence, the 728) on 6 June 2016. Ethical approval for this study was also ob-
analysed dataset were of high quality. This study presented all tained from the Human Research Ethics Committee (HREC) of The
MNM cases and the EmOC interventions provided for women in University of Newcastle, Australia (approval number: H-2018-0245)
order to estimate the quality of obstetric care. Finally, this study on 15 August 2018. The protocol developed to conduct this anal-
utilized data of both public and private hospitals. Hence, the re- ysis had received ethical approval from the Scientific and Ethical
ported MNM indicators represent all types of hospitals irrespective Review Office of Ethiopian Public Health Institute before the data
of the managing authority. were accessed (Protocol number: EPHI-IRB-048-2018) on 25 July
However, this study suffered from certain limitations. The WHO 2018.
recommends a follow-up period of up to 42 days postpartum to
define MNM. However, the EPHI conducted a survey by which the
Funding sources
maternal health indicator at hospitals were collected during a sin-
gle hospital visit. This might have resulted in the high incidence of
No funding sources to declare.
MNM as all women who experienced obstetric complications, irre-
spective of WHO recommendations were included in the current
study. The analysis was done on the hospitals’ data by excluding Clinical trial registry and registration number
data of the lower level health facilities. In Ethiopia, while there
were deliveries conducted at lower health facilities, the majority Not applicable.

Downloaded for Anonymous User (n/a) at National Library of Indonesia from ClinicalKey.com/nursing by Elsevier on April 14, 2020.
For personal use only. No other uses without permission. Copyright ©2020. Elsevier Inc. All rights reserved.
8 A. Geleto, C. Chojenta and T. Taddele et al. / Midwifery 82 (2020) 102597

Declaration of Competing Interest IAEG, U. (2016). Final list of proposed sustainable development goal indicators. re-
port of the inter-agency and expert group on sustainable development goal in-
dicators (E/CN. 3/2016/2/Rev. 1).
None declared Jackson, R., Tesfay, F.H., Gebrehiwot, T.G., Godefay, H., 2017. "Factors that hinder or
enable maternal health strategies to reduce delays in rural and pastoralist areas
in Ethiopia. Trop. Med. Int. Health 22 (2), 148–160.
Acknowledgments Kalhan, M., Singh, S., Punia, A., Prakash, J., 2017. Maternal near-miss audit: lessons
to be learnt. Int. J. Appl. Basic Med. Res. 7 (2), 85.
Kasahun, A.W., Wako, W.G., 2018. Predictors of maternal near miss among women
We would like to acknowledge the University of Newcastle,
admitted in Gurage zone hospitals, South Ethiopia, 2017: a case control study.
Australia for providing a full PhD scholarship for the principal au- BMC Pregnancy Childbirth 18 (1), 260.
thor of this paper. Our gratitude extends to the EPHI and the re- Li, S., Chen, M., He, G., He, L., Wei, Q., Li, T., Liu, X., 2017. Risk factors for blood trans-
fusion in women with postpartum hemorrhage: a case-control study. Sichuan
searchers who conducted the survey for their permission to access
da xue xue bao. Yi xue ban= J. Sichuan Univ. Med. Sci. Ed. 48 (6), 937–940.
the data. We also extend our appreciations to the participant hos- Liyew, E.F., Yalew, A.W., Afework, M.F., Essen, B., 2017. Incidence and causes of ma-
pitals, data collectors, supervisors and the partners who have coop- ternal near-miss in selected hospitals of Addis Ababa, Ethiopia. PLoS ONE 12 (6),
erated during the data collection and those who provided financial e0179013.
Maswanya, E., Muganyizi, P., Kilima, S., Mogella, D., Massaga, J., 2018. Practice of
support for EPHI for the survey. Finally yet importantly, we would emergency obstetric care signal functions and reasons for non-provision among
like to thank Natalia Soeters for providing language revision and health centers and hospitals in Lake and Western zones of Tanzania. BMC
proof reading of the manuscript. Health Serv. Res. 18 (1), 944.
Mony, P.K., Krishnamurthy, J., Thomas, A., Sankar, K., Ramesh, B.M., Moses, S., Blan-
chard, J., Avery, L., 2013. Availability and distribution of emergency obstetric
References care services in Karnataka state, South India: access and equity considerations.
PLoS One 8 (5), e64126.
Nelissen, E.J., Mduma, E., Ersdal, H.L., Evjen-Olsen, B., v. Roosmalen, J.J., Stekelen-
Abate, T., Dile, M., 2015. Proportion of maternal near misses and associated factors
burg, J., 2013. Maternal near miss and mortality in a rural referral hospital in
in referral hospitals of Amhara regional state, Northwest Ethiopia: Institution
northern Tanzania: a cross-sectional study. BMC Pregnancy Childbirth 13, 141.
Based Cross Sectional Study. Gynecol. Obstet. 05 (08).
Pattinson, R., Hall, M., 2003. "Near misses: a useful adjunct to maternal death en-
Abdella, A., 2010. Maternal mortality trend in Ethiopia. Ethiop. J. Health Dev. 24
quiries. Br. Med. Bull. 67 (1), 231–243.
(Special Issue 1), 115–122.
Ridge, A.L., Bero, L.A., Hill, S.R., 2010. Identifying barriers to the availability and use
Adeoye, I.A., Onayade, A.A., Fatusi, A.O., 2013. Incidence, determinants and perinatal
of magnesium sulphate injection in resource poor countries: a case study in
outcomes of near miss maternal morbidity in Ile-Ife Nigeria: a prospective case
Zambia. BMC Health Serv. Res. 10, 340.
control study. BMC Pregnancy Childbirth 13, 93.
Rulisa, S., Umuziranenge, I., Small, M., van Roosmalen, J., 2015. Maternal near miss
Admasu, K., Haile-Mariam, A., Bailey, P., 2011. "Indicators for availability, utilization,
and mortality in a tertiary care hospital in Rwanda. BMC Pregnancy Childbirth
and quality of emergency obstetric care in Ethiopia, 2008. Int. J. Gynaecol. Ob-
15, 203.
stet. 115 (1), 101–105.
Say, L., Pattinson, R.C., Gülmezoglu, A.M., 2004. "WHO systematic review of mater-
Assefa, T., Mariam, D.H., Mekonnen, W., Derbew, M., Enbiale, W., 2016. "Physician
nal morbidity and mortality: the prevalence of severe acute maternal morbidity
distribution and attrition in the public health sector of Ethiopia. Risk Manag.
(near miss). Reprod. Health 1 (1), 3.
Healthc. Policy 9, 285.
Say, L., Souza, J.P., Pattinson, R.C., w. g. o. M. Mortality, W.H.O., Morbidity, c., 2009.
Banke-Thomas, A., Wright, K., Collins, L., 2019. "Assessing geographical distribution
Maternal near miss–towards a standard tool for monitoring quality of maternal
and accessibility of emergency obstetric care in sub-Saharan Africa: a systematic
health care. Best Pract. Res. Clin. Obstet. Gynaecol. 23 (3), 287–296.
review. J. Glob. Health 9 (1), 010414.
Tuncalp, O., Hindin, M.J., Adu-Bonsaffoh, K., Adanu, R.M., 2013. Assessment of ma-
Berhan, Y., Berhan, A., 2014. CAUSES of maternal mortality in ETHIOPIA: a significant
ternal near-miss and quality of care in a hospital-based study in Accra, Ghana.
decline in abortion related death. Ethiop. J. Health Sci. (Special Issue) 15–28.
Int. J. Gynaecol. Obstet. 123 (1), 58–63.
Berhane, G., Gessessew, A., Roosmalen, J.V., Akker, T.V.D., 2012. Obstetric near-miss
Tunçalp, Ö., Hindin, M.J., Souza, J., Chou, D., Say, L., 2012. The prevalence of maternal
and maternal death: the case of Ayder teaching hospital, Mekelle, Ethiopia.
near miss: a systematic review. BJOG Int. J. Obstet. Gynaecol. 119 (6), 653–661.
Ethiop. J. Reprod. Health 6:1, 56–63.
Tura, A.K., Trang, T.L., van den Akker, T., van Roosmalen, J., Scherjon, S., Zwart, J.,
Bintabara, D., Ernest, A., Mpondo, B., 2019. Health facility service availability and
Stekelenburg, J., 2019. Applicability of the who maternal near miss tool in sub-
readiness to provide basic emergency obstetric and newborn care in a low-re-
-Saharan Africa: a systematic review. BMC Pregnancy Childbirth 19 (1), 79.
source setting: evidence from a Tanzania National Survey. BMJ Open 9 (2),
Tura, A.K., Zwart, J., van Roosmalen, J., Stekelenburg, J., van den Akker, T., Scher-
e020608.
jon, S., 2018. Severe maternal outcomes in eastern Ethiopia: application of the
Chhabra, P., 2014. Maternal near miss: an indicator for maternal health and mater-
adapted maternal near miss tool. PLoS One 13 (11), e0207350.
nal care. Indian J. Commun. Med. Off. Publ. Indian Assoc. Prev. Soc. Med. 39 (3),
UNDP (2012). Development brief: analysing regional performance and desparities in
132.
health outcome in Ethiopia.
CSA and ICF, 2016. Ethiopian Demographic and Health Survey, Key indicators. CSA,
United Nations Population Fund, 2009. Monitoring Emergency Obstetric care: a
Addis Ababa Ethiopia. ICF Rockville, Maryland, USA.
Handbook. World Health Organization, New York.
Donati, S., Senatore, S., Ronconi, A.Regional Maternal Mortality Working, 2012. Ob-
van den Akker, T., van Rhenen, J., Mwagomba, B., Lommerse, K., Vinkhumbo, S., van
stetric near-miss cases among women admitted to intensive care units in Italy.
Roosmalen, J., 2011. Reduction of severe acute maternal morbidity and maternal
Acta Obstet. Gynecol. Scand. 91 (4), 452–457.
mortality in Thyolo district, Malawi: the impact of obstetric audit. PLoS One 6
EDHS, E., 2016. Demographic and Health Survey 2016: Key Indicators Report. The
(6), e20776.
DHS Program ICF.
Vandecruysa, H.I.B., Pattinsona, R.C., Pattinson, R.C., Pattinson, R.C., Macdonalda, A.P.,
Federal Ministry of Health Ethiopia (2015). Ethiopian health system transformation-
Mantel, G.D., 2002. Severe acute maternal morbidity and mortality in the Pre-
plan: 2015/16 - 2019/20 (2008-2012 EFY).
toria academic complex: changing patterns over 4 years. Eur. J. Obstet. Gynecol.
Firoz, T., Chou, D., von Dadelszen, P., Agrawal, P., Vanderkruik, R., Tunçalp, O.,
Reprod. Biol. 102 (1), 6–10.
Magee, L.A., van Den Broek, N., Say, L., 2013. Measuring maternal health: focus
World Wealth Organization (2004). Beyond the numbers: reviewing maternal
on maternal morbidity. Bull. World Health Organ. 91, 794–796.
deaths and complications to make pregnancy safer.
Gaym, A., Bailey, P., Pearson, L., Admasu, K., Gebrehiwot, Y.O.N.C.A.T. Ethiopian
WHO, UNICEF, UNFPA, WORLD BANK GROUP and U. N. P. DIVISION, 2015. Trends in
National Em, 2011. Disease burden due to pre-eclampsia/eclampsia and the
Maternal Mortality: 1990 to 2015 - Estimates by WHO, UNICEF, UNFPA. World
Ethiopian health system’s response. Int. J. Gynaecol. Obstet. 115 (1), 112–116.
Bank Group and the United Nations Population Division.
Getaneh, W., Kumbi, S., 2010. Use of magnesium sulfate in pre-eclampsia and
Windsma, M., Vermeiden, T., Braat, F., Tsegaye, A.M., Gaym, A., van den Akker, T.,
eclampsia in teaching hospitals in Addis Ababa: a practice audit. Ethiop. Med. J.
Stekelenburg, J., 2017. Emergency obstetric care provision in Southern Ethiopia:
48 (2), 157–164.
a facility-based survey. BMJ Open 7 (11), e018459.
Ghazal-Aswad, S., Badrinath, P., Sidky, I., Safi, T.H., Gargash, H., Abdul-Razak, Y.,
Worldometers (2019). Ethiopian population 2019.
Mirghani, H., 2013. Severe acute maternal morbidity in a high-income devel-
oping multiethnic country. Matern. Child Health J. 17 (3), 399–404.
Hogan, M.C., Foreman, K.J., Naghavi, M., Ahn, S.Y., Wang, M., Makela, S.M.,
Lopez, A.D., Lozano, R., Murray, C.J., 2010. Maternal mortality for 181 countries,
1980–2008: a systematic analysis of progress towards millennium development
goal 5. Lancet 375 (9726), 1609–1623.

Downloaded for Anonymous User (n/a) at National Library of Indonesia from ClinicalKey.com/nursing by Elsevier on April 14, 2020.
For personal use only. No other uses without permission. Copyright ©2020. Elsevier Inc. All rights reserved.