Content uploaded by Tegene Legese Dadi
Author content
All content in this area was uploaded by Tegene Legese Dadi on Aug 15, 2016
Content may be subject to copyright.
American Journal of Public Health Research, 2016, Vol. 4, No. 4, 120-127
Available online at http://pubs.sciepub.com/ajphr/4/4/1
© Science and Education Publishing
DOI:10.12691/ajphr-4-4-1
Risk Factors of Maternal Death in Jimma University
Specialized Hospital: A Matched Case Control Study
Tegene Legese1,*, Misra Abdulahi2, Anteneh Dirar2
1Collage of Health Science, Mizan-Tepi University
2Collage of Health Science, Jimma University
*Corresponding author: tege2004@gmail.com, emaye2008@gmail.com
Abstract Background: Maternal death has devastating effects on the family she leaves behind and country level.
Most of the literatures in our country are reviews of maternal death which are unable to determine the predictors of
maternal death and do not consider change of time since there is variation in care given and did not identify timing
of death. Objective: To assess risk factors of maternal death in Jimma University specialized hospital, Southwest
Ethiopia from January 2010 to December 2014. Methods: A time matched case control study was conducted on 600
charts, 120 cases and 480controls. Data was collected using checklist adapted from maternal death surveillance
review of Ethiopia guide line. Data were entered into epi data 3.1 and exported to Stata 13 for analysis. Conditional
logistic regression was done to identify the independent predictors of maternal death. The adjusted matched odds
ratio with the 95% confidence interval was reported and statistical significance was declared at p =<0.05. To ensure
confidentiality only code was written on the check list. Result: More than two third (68%) of death occurred during
post-partum period. Predictors of maternal death include: age group of 20 – 34 (AMOR= 0.299, 95% CI (0.113,
0.792)), being from rural area (AMOR = 2.594, 95%CI(1.001,6.726)), prolonged labour (AMOR=37.141,
95%CI(13.296, 103.750)), comorbidities (AMOR=9.631,95%CI(3.135, 29.588), referred cases from health center
(AMOR=4.011, 95% CI (1.113, 14.464) and other health institution (AMOR=6.029, 95%CI(1.565, 24.626)).
Conclusion: Duration of labour, age, comorbidities, residence and referral were the major factors that affect
maternal death.
Keywords: maternal death, risk factors of maternal death, maternal health, Jimma University
Cite This Article: Tegene Legese, Misra Abdulahi, and Anteneh Dirar, “Risk Factors of Maternal Death in
Jimma University Specialized Hospital: A Matched Case Control Study.” American Journal of Public Health
Research, vol. 4, no. 4 (2016): 120-127. doi: 10.12691/ajphr-4-4-1.
1. Introduction
Maternal death is the death of a woman while pregnant
or within 42 days of termination of pregnancy,
irrespective of the duration and the site of the pregnancy,
from any cause related to or aggravated by the pregnancy
or its management, but not from accidental or incidental
causes. Loss of a woman from family because of death
during pregnancy, delivery and within 6 weeks of
postpartum can threaten the survival of the entire family.
Women who die due to pregnancy-related causes are in
the prime of new born baby and young children lives and
are responsible for the health and well-being of their
families. Many women shoulder a double burden of
helping to support the family by working outside the home
and taking full responsibility for household duties and
child care. Yet, despite this vital role played by women in
society, the health needs of a women is neglected which is
evidenced by high level of maternal mortality in many
poor countries [1,2].
Women’s and children’s health is directly affected by
the social factors, institutional factors cultural factors and
economic status. Issues such as their low socio economic
status, harmful traditional practices, especially female
genital mutilation, early marriage, and low female literacy,
all have a direct negative impact on women’s health [3,4].
Poverty or low socio economic status results in poor child
and maternal health care services, lower skilled care and
poor access to health care services are some of the
consequences that have direct impact on health of mother.
These things results in high maternal mortality rates. As a
result, the use of antenatal care (ANC), skilled delivery
attendants and postnatal care (PNC) are recognized as key
maternal health services to improve health outcomes for
women and children [3,4].
Maternal death has devastating effects on the family she
leaves behind and country level. Women in developing
countries lose more disability-adjusted life years (28
million) to maternal causes than to any other. Different
literatures show that children less than 9 year are more
likely to be chronically malnourished and to die than
children who lost their father. Studies in developing
countries indicate that the risk of death for children less
than five years doubles or triples if their mother dies.
Other studies estimate that children whose mothers have
died are 3-10 times more likely to die within two years
than those whose parents are both alive. The other
consequence is economic burden on family, total lost
121 American Journal of Public Health Research
economic productivity and out of pocket expenditure of
households with maternal death is 10 times higher than
that in without maternal death [5,6,7,8]. Maternal mortality
of a single person was found to reduce per capita gross
domestic product by US$ 0.36 per year in WHO African
region. The study has demonstrated that maternal mortality
has a statistically significant negative effect on GDP [9].
Maternal mortality rate (MMR) globally has fallen by
45% from 523,000 (380/100,000) in 1990 to 289 000
(210/100,000) in 2013 yielding an average annual decline
of 2.6%. It is one of the statistics showing the largest
degree of disparity between developed and developing
countries. In 2013, Developing countries account for 99%
(286 000) of the global maternal deaths with sub-Saharan
Africa (SSA) alone accounting for 62% (179 000) followed
by Southern Asia (69 000). The MMR in developing
regions was 14 times higher than in developed regions. In
general in SSA there is 49% reduction from 990 to 510
per 100,000 but eastern Africa reduces MMR by 57%
from 1000 to 440 per 100,000 which are highest in SSA.
Ethiopia decreased maternal death from 1400 to 420 per
100,000 MMR between 1990 and 2013 which is making
progress to achieve MDG 5a by 2015. Ethiopia is grouped
under high MMR with 420/100,000 in 2013 [2,3,10].
The 2011 Ethiopia demographic and health survey
(EDHS) (676/100,000) shows that there is bit increment of
MMR from 2005 EDHS (673/100,000), even though the
95 % CI is coinciding. 2005 EDHS report (673/100,000)
and joint estimation by WHO, UNICEF, UNFPD and
world bank in same year (740/100,000) have wide
variation [10,11,12,13,14].
Each maternal death has a story to tell and can provide
indications on practical ways to address problems. There
was four times risk of maternal death in the age group 35–
49 years compared to the group of 25–34 years. Being
primiparas and multipara increases the risk of maternal
death having no attendance of ANC, distance from
hospital, having comorbid complications on admission 9
times, having elevated pulse 10.7 times, and referral [15,16].
Most of the literatures are reviews of maternal death
which are unable to determine the predictors of maternal
death and previous literatures did not identify timing of
death except single study done in Bonke wereda
(community based cross sectional study). Besides high
number of maternal death, there is scarcity of literatures in
Ethiopia and particularly in the study set up, regarding
duration and factors associated with maternal death.
1.1. Significance of Study
Maternal death is widely regarded as one of life’s most
tragic outcomes. There is a big pain in the death of a
woman who is engaged in the act of creating life, and her death
is an incomparable loss for any children who are left behind.
Even though, there are limited numbers of literatures in
Ethiopia in general and study setup in particular, the
already available literatures does not consider change of
time, unable to determine the predictors of maternal death,
old literatures and did not identify timing of death.
Therefore, the aim of this study was to assess duration
and factors affecting maternal death in case of in Jimma
university specialized hospital (JUSH). Information generated
from this study can be used by health professionals, health
care planners, managers and policy makers to save
women’s lives by improving the quality of care provided.
It is intended that this information will help or contribute
to change policies and practices that will lead to
improvements in maternal health. It is also intended that
this study can be used as input for JUSH, researchers and
academicians.
2. Objectives
2.1. General Objective
• To assess risk factors of maternal mortality in
JUSH, Ethiopia from January 2010 to December
2014.
2.2. Specific Objectives
• To describe duration or timing of death of
mothers in JUSH from January 2010 to
December 2014
• To determine risk factors of maternal mortality in
JUSH from January 2010 to December 2014.
3. Methods
3.1. Location of Study
The study was conducted at Jimma university specialized
hospital (JUSH). JUSH is located in Jimma town which is
around 355 Km from Addis Abeba in south west direction,
Jimma zone, Oromia regional state, Ethiopia. JUSH is one
of the oldest public hospitals found in the south western
part of the country that runs under Jimma University. It is
currently the only teaching and specialized hospital in south
west part of Ethiopia. The hospital serves as referral site
and provides specialized care for south-western Ethiopia
with a catchment population of about 15 million. The
specialized care provided for the population are in internal
medicine, surgery, obstetrics and gynecology, pediatrics,
ophthalmology, psychiatry and dermatology [17].
3.2. Study Period
The study was conducted from March 14 – April 14,
2015 among women who visited maternal health services
from January 2010 to December 2014.
3.3. Study Design
Institution based matched case control study was
conducted.
3.4. Population
The source Population of this study was
All charts of mothers who visited JUSH for maternal
health service utilization.
Whereas the Study Population was
Cases: all charts of mothers who died during pregnancy,
delivery and 42 days after delivery in JUSH from January
2010 to December 2014
Controls: all charts of mothers who visited JUSH for
maternal health service utilization from January 2010 to
December 2014
American Journal of Public Health Research 122
Sample population
The sample population was randomly selected cards of
cases and controls during the study period.
3.5. Inclusion and Exclusion Criteria
Inclusion criteria: cases that fulfill the standard
definitions of maternal mortality given by international
classification of disease – 10 and controls which came for
maternal health service utilization.
Exclusion criteria: cases that are registered on the log
book but whose charts’ were missed. Charts that didn’t
include the assessment of admission and status of mother
(dead or alive) during discharge were excluded from the
study. Family planning users, ANC users are excluded.
3.6. Sample Size Determination and Sampling
Procedures
Sample size determination
Sample size was determined by two population
proportion using Epi info version 7 by taking the
following assumptions: 95% CI, and 80% power, odds
ratio of 0.06 (odds of nurses attending the delivery over
odds of delivery attended by other) from study done in
Tigray region, case to control ratio of 1:4 and the
prevalence of exposure among controls were 6.5%
(proportion of controls attended by nurse) taken from
unmatched case control study done in Tigray regional
state [12]. A total of 600 charts, 120 cases and 480
controls were included in the study.
Ascertainment of cases = the occurrence of death was
ascertained based on the information on chart. If death is
reported on the chart, it was taken as a case.
Sampling procedures
Charts of both cases and controls that fulfill the
inclusion and exclusion criteria were selected from
maternity ward, delivery ward, operation theatre, PNC
ward and gynecologic ward.
First cases were identified from log book or registration
book from respective wards and then sampling frame was
prepared. Hundred twenty cases were selected randomly
from sampling frame by using random number of cases in
SPSS version 21.
After identifying cases included in the sample the time
of procedures was identified for each selected cases. For
each case four controls that were the nearby survived
women after delivery within 48 hour were selected. But if
there are more than four controls, who are candidates,
lottery method is used in selection. Selected 120 maternal
death and 480 controls charts were reviewed.
Charts that had missing values of more than 30% were
replaced by random selection from available charts. When
cases are replaced the control group is also replaced with
controls nearer to cases.
3.7. Variables
Dependent variable
Maternal death
Matching variable
Time of event or death
Independent variable
Socio-demographic factors: Age, residence, distance
of home from Jimma bus station
Delivery history: Length of labor, birth attendant,
mode of delivery and place of delivery
Obstetric history: ANC, parity and gravidity
Antenatal/ intra natal risks: Placenta previa, previous
ceaserean section, obstructed labour, multiple gestation
and malpresentation
Comorbidities: Chronic hypertension, anaemia, HIV
positive and malaria
3.8. Data Collection Instrument
Data collection instrument was adapted from Maternal
Death Surveillance and Response Technical Guideline of
Ethiopia [18]. The checklist consists of socio-demographic
data, obstetric and delivery history, presence of
comorbidities, cause of death, antenatal and intranatal
risks and presence of complications.
3.9. Data Collectors
Three midwives and three health officers were recruited
and trained for two days. The training covered: the
contents of the tool, ethical considerations and way of
extraction of necessary information from chart. Two
supervisors, one midwife and one public health officer,
monitored and followed data collection while the principal
investigator supervised the overall data collection process.
3.10. Data Processing and Analysis
Data were entered into Epi Data version 3.1 and then
exported for analysis to STATA version 13.0. Data were
cleaned for inconsistencies and missing values. Variables
that were missed in more than 40% of total sample were
excluded from analysis.
Simple frequencies were done to see the overall
distribution of the study subject with the variables under
study and to see any missing data. Multicollinearity test
was performed to see collinearity of variables.
Bivariate conditional logistic regression analysis was
used to determine the association between different factors
and the outcome variable and to select candidate variables
for multivariate conditional logistic regression. Variables
that had p value less than or equal to 0.20 was entered into
multivariate conditional logistic regression to identify
independent predictors of maternal death.
Significant variables from multivariate conditional
logistic regression model were interpreted and discussed.
Confidence interval of 95% was used to see the precision
of the study and the level of significance was taken at p
value =<0.05. The adjusted matched odds ratio with the
95%CI was reported.
3.11. Data Quality Assurance
To assure the quality of data, data collectors were
trained and did pretesting of the checklist tool until they
became well conversant with the instrument. Every day
filled checklist was reviewed and checked for
completeness and relevance by supervisors and principal
investigator.
After data collection, each filled checklist was given a
unique code by the principal investigator. Data was
entered using Epi Data version 3.1 then exported to
STATA version 13.0 for analysis. Accuracy was improved
through double entry, 10 % of entered data was reentered
123 American Journal of Public Health Research
with the actual filled tool and compared the dataset using
STATA. Frequencies were used to check for entry errors,
missed values and outliers. Any errors identified were
corrected after revision of the original data using the code
numbers.
3.12. Operational and Definitions of Terms
Maternal health service: Health services including
abortion, visit for pregnancy and pregnancy related
complications during antenatal period, delivery and post-
natal care within 6 weeks.
Time: time at which the event occurred. Time of
control is the nearby time from time of case within one
day.
Cause of death = the causes written on the chart is
considered as cause of death.
Presence of comorbidity = the presence of at least one
medical disease was considered.
Prolonged labour = duration of labour more than 24
hours.
3.13. Ethical Consideration
Prior to data collection, ethical clearance was obtained
from Research and Ethics committee of the College of
health sciences of Jimma University. Written permission
letter was also received from JUSH CEO and other
concerned bodies in the study set up. In order to establish
anonymous linkage only the codes, not the names of the
participant from the chart, was registered on the
questionnaire. During the training of data collectors and
supervisor, ethical issues was addressed as important
component of the research.
4. Result
4.1. Descriptive Analysis of Variables
Socio - demographic characteristics
A total of 120 cases and 480 control charts were
included in the study. From the variables, for which there
was a plan to collect information for, the following are
incomplete in 99% of charts: ethnicity, educational status,
marital status, occupation, and income. As a result they
are excluded from analysis.
More than half (52.5%) of cases were in the age group
20-34 whereas nearly two third of controls were in the age
group 20-34. Although about three quarter (75.83%) of
cases were from rural area, more than half (54.58%) of
controls were from urban residence. Forty eight (40%) of
cases came from 11-50km distance whereas hundred fifty
four (32%) controls came from <= to 10 km distance.
Most of cases (70.8%) and more than half (51.9%) of
controls were referred from health centers. From not
referred women most of them are not died (Table 1).
Fifty (41.67%) cases were prim-gravid women while
228 (47.7%) controls were gravida 2-4. Similarly most of
cases (43.3%) were primi-paras whereas most of controls
(47.7%) were para 2-4. Only 3(2.5%) cases and 12 (2.5%)
controls had history of previous C/S. More than two third
(67.5%) of cases were women who did not have ANC
follow up whereas majority of controls (96.5%) had ANC
follow up. From females who are booked for ANC only
7.8% females are dead. In the contrary from females who
don’t have ANC follow up most of them are dead. (Table 1).
Although 40 (34%) cases had obstructed labor, only 59
(12.7%) controls had obstructed labor. More than one
quarter (26.7%) cases had malpresentation whereas only
13.75% of controls had malpresentation. More controls
(12.3%) had prolonged rupture of membrane than cases
(4.2%). Most of cases (83.3%) had prolonged labour while
most of controls (82.9%) length of labour less than 24 hrs.
Majority (92%) of cases and controls (81%) were attended
by doctors. Fifty five (46%) cases have undergone
ceaserean section or hysterectomy procedures followed by
assisted vaginal delivery (27%). However, more than half
(53%) of controls were delivered/gave birth by
spontaneous vaginal delivery (Table 1).
4.2. Pregnancy Risk and Complication
In 97.5% (586) of women there are no multiple
gestations. From 14 women who have multiple gestations
only three (21.4%) are dead. From total deaths in
occurring in women only three women (2.5%) have
multiple gestations. The rest 97.5% of women are alive.
Placenta previa only occur in 17 (2.83%) of women. From
which 6 (32.3%) are dead. Abruptio placenta occur only in
23(3.83%) of women. From which 14 (61%) are dead
which covers 12% from total death. only 9% (54) of
women have APH from which 38 (70%) of them are dead.
It covers 32% of total maternal death. Only 46 (8%) of
women have PIH from which 23(50%) are dead in which
it covers 19% from total deaths. 81% of women who dead
did not have PIH.
Only 45(7.5%) of women have intra-partum or post-
partum complications (chorioamnitis, wound infection and
post-partum psychosis) from which 24(53%) of them are
dead in which 20% of death from total death. Post term
pregnancy occurred in only 26(4.33%) of women from
which only 2 deaths occurred. 79 (13%) of women face
other types of comorbidity from which 40(51%) of them
dead which is 33% from total death.
Although 28 (23.3%) cases have experienced uterine
rupture whereas only 9 (<2%) of controls have
experienced uterine rupture. Similarly, though 37 (30.8%)
cases have developed PPH only 5 (1%) controls have
developed PPH. Seventeen (14%) cases have experienced
puerperal sepsis while 7 (1.5%) controls have developed it.
Even though three quarter (75%) of cases have
experienced hemorrhage only 30 (6%) have experienced it.
Thirty four (28%) cased had anemia whereas nearly 10%
of controls had anemia (Table 1).
4.3. Timing of Death
Most deaths 82 (68%) occurred during post-partum
period while 25% deaths occurred during intra-partum
period. Only 8 (6.7%) happened during ante partum period
(Figure 1). Out of 8 deaths that occurred during ante-
partum period, five of them occurred during 14 – 28
weeks of ante-partum. Two deaths occurred in less than 14
weeks. Twenty eight deaths, out of 30 deaths during intra-
partum, occurred during 29 – 37 weeks of intra-partum.
Out of 82 deaths during post-partum period, 38 (46.34%)
happened between 4 - 7 days whereas 29 (35.37%)
happened within 3 days. The rest occurred in between 8 –
42 days.
American Journal of Public Health Research 124
Table 1. factors associated with maternal death in JUSH, May, 2015
variables
category
controls
cases
Total
CMOR (95% CI)
AMOR (95% CI)
age
<20
62(76.54)
19(23.46)
81
0.840
(0.448, 1.518) 1.582
(0.434, 5.762)
12.92
15.83
13.5
20 - 34
315(83.33)
63(16.67)
378
0.558
(0.356, 0.874) 0.299
(0.113, 0.792)
65.63
52.5
63
>=35
103(73)
38(26.95)
141
1
1
21.46
31.67
23.5
residence
rural
218(90)
91(10)
309(51.5)
2.250
(1.210, 4.210) 2.594
(1.001, 6.726)
45.42
75.83
42.5
urban
262(70.55)
29(29.45)
291(48.5)
1 1
54.58
24.17
43.67
Estimated km of area from jimma
<=10
154(92.77)
12(7.23)
166
0.1801
(0.0875, 0.3705) 0.499
(0.134, 1.864)
32.08
10
27.67
11 - 50
142(74.74)
48(25.26)
190
0.8116
(0.4700, 1.4014) 1.226
(0.436, 3.450)
29.58
40
31.67
51 - 150
102(79.69)
26(20.31)
128
0.6374
(0.3518, 1.1546) 0.539
(0.171, 1.698)
21.25
21.67
21.33
>=151
82(68.42)
34(31.58)
76
1
1
17.03
28.33
19.33
Referral
health center
249(74.55)
85(25.45)
334
6.6667
(3.0214, 14.7102) 4.011
(1.113, 14.464)
51.88
70.83
55.67
Other HI
85(75.22)
28(24.87)
113
6.4467
(2.8696, 18.4521) 6.209
(1.565, 24.626)
17.71
23.33
18.83
not referred
146(95.42)
7(4.58)
153
1
1
30.42
5.83
25.5
comorbidity
No
441(84.64)
80(15.36)
521
1
1
91.88
66.67
86.83
Yes
39(49.37)
40(50.63)
79
15.3090
(2.7965, 83.8048) 9.631
(3.135, 29.588)
8.13
33.33
13.17
Parity
Primi
249(74.55)
85(25.45)
334
6.6667
(3.0214, 14.7102) 0.498
(0.031, 7.927)
51.88
70.83
55.67
2 - 4
85(75.22)
28(24.87)
113
6.4467
(2.8696, 18.4521) 0.329
(0.042, 2.593)
17.71
23.33
18.83
>=5
146(95.42)
7(4.58)
153
1
1
30.42
5.83
25.5
Gravidity
Primi
168(77.1)
50(22.9)
218
0.667
(.396, 1.124) 1.565
(0.108, 22.634)
35
41.7
36.3
2 - 4
228(87.4)
33(12.6)
261
0.333
(0.195, 0.571) 1.489
(0.207,10.687)
47.5
27.5
43.5
>=5
84(69.4)
37(30.6)
121
1 1
17.5
30.8
20.2
attendant
doctors
388(77.91)
110(22.09)
498
2.7196
(1.3473, 5.489) 1.028
(0.277, 3.821)
80.83
91.67
83
Others
92 (90.25)
10(9.85)
102
1 1
19.17
8.33
17
obstructed
labour
No
419(84.14)
79(15.86)
498
1
1
87.29
65.83
83
Yes
61(59.8)
41(40.2)
102
3.4686
(2.1787, 5.5220) 0.695
(0.259, 1.869)
12.71
34.17
17
Malpresentation
No
414(82.47)
88(17.53)
502
1
1
86.25
73.33
83.67
Yes
66(67.35)
32(32.65)
98
2.3736
(1.4459, 3.8966) 0.873
(0.301, 2.535)
13.75
26.67
16.33
Duration of labour
<24 hrs
398(95.22)
20(4.78)
410
1
1
82.92
16.67
69.67
>=24 hrs
82(45.05)
100(54.95)
182
22.2254
(12.1072, 40.7996) 37.141
(13.296, 103.750)
17.08
83.33
30.33
Anemia No
434(83.46)
86(16.54)
520
1 1
90.42
71.67
86.67
Yes
46(57.5)
34(42.5)
80
3.5058
(2.1426, 5.7363) 2.204
(0.855, 5.679)
9.58
28.33
13.33
PROM No
421(78.54)
115(21.46)
536
1 1
87.71
95.83
89.33
Yes
59(92.19)
5(7.81)
64
0.293
(0.134, 0.759) 0.220
(0.036, 1.329)
12.29
4.17
10.67
125 American Journal of Public Health Research
Figure 1. Timing of death of women in JUSH, May, 2015
4.4. Predictors of Maternal Death
Multiple gestation, post term pregnancy and previous
ceaserean section were excluded from multivariate
conditional logistic regression analysis since they have
p>0.2 in bivariate conditional logistic regression. ANC
and delivery method were excluded from multivariate
conditional logistic regression analysis since they affect
the precisions of significant variables.
Being in the age group of 20 – 34 is protective than
being in the age group of >=35. A one women increase in
the age group of 20 – 34, decreased the odds that a woman
would die by 70.1% than being in the age group of >=35
(AMOR= 0.299, 95% CI (0.113, 0.792). Women who
came from rural area are more likely would die 2.594
times than women who came from urban area (AMOR=
2.594, 95% CI (1.001, 6.726).
Women who were referred from health center were
4.011 times more likely to die than women who were not
(AMOR=4.011, 95% CI (1.113, 14.464). Similarly
women who referred from other health institutions
(hospital and private or NGO clinics) were 6.209 times
more likely they would die than women who didn’t
referred (AMOR=6.209, 95% CI (1.565, 24.626). Women
who had prolonged labour were 37.141 times more likely
the women would die than women who had less than 24
hour duration of labour (AMOR= 37.141, 95% CI (13.296,
103.750). Women who had comorbidities were 9.631
times more likely they would die than women who didn’t
have comorbidities (AMOR= 9.631, 95% CI (3.135,
29.588) (Table 1).
5. Discussion
5.1. Timing of Maternal Death
Most (68%) of maternal death occurred in post-partum
period which is higher than finding from south west
Ethiopia in Bonke wereda in which 51% of maternal death
occurred in post-partum period. This might be because of
difference in number of maternal death included in the
study, 49 deaths were included in Bonke wereda but in
this research 120 maternal death were included in the
study. The other reason might be because of recall bias
introduced in south west Ethiopia study since it is
retrospective cross sectional house hold survey. However
study from maternity hospital of Nigeria shows about 61.9%
of the maternal deaths occurred in post-partum period
which is lower than this finding. This might be because of
only 84 maternal deaths in two year study period were
included with this number and study period if it is data of
five year it might be higher than this research finding
[22,23]. Higher number of maternal death in post-partum
period implies that the management following delivery
and during delivery might be poor. Prolonged period of
labour due to OL or due to not having timely intervention
might contribute a lot for death occurring in post-partum
period.
The second period in which higher number of death
occurred is the intra-partum period (25%). It is smaller
than finding from Bonke wereda, south west Ethiopia
which was 37%. This difference might be because of
difference in site of delivery, 97% of delivery in JUSH is
HI delivery in contrary to study of Bonke wereda, south
west Ethiopia in which 92% of delivery occurred in home
so that in home delivery complications are not managed
well so high number of death may occur in intra-partum
period. But it is much higher than from findings of
maternity hospital of Nigeria which is 13.1%. This
difference might be because of high antepartum death
(10.7%) and death because of post-abortal complications
(14.3%). in maternity hospital of Nigeria in which 25% of
women died before reaching to intra-partum period but in
our set up antepartum death including abortion
complication is only 7%. This might be because of high
prevalence of death associated with abortion relative to
JUSH. The other reason might be absence of blood bank,
difference in size and level of hospital and difference in
the number of health professionals (low number) during
the study period in maternity hospital of Nigeria might
contribute to poor quality of care. The other reason might
American Journal of Public Health Research 126
be dramatic reduction of maternal death related with
abortion due to the introduction of misoprostol or safe
mother hood services in our country.
Ante partum death occurred in south west Ethiopia is
12% which is higher than this research finding 7%. There
might be difference in awareness, this might create poor
follow up of women during ante partum in south west
Ethiopia study [19,20. These imply that proper care during
antenatal period, reduce home delivery, intra-partum
period and after abortion save the lives of many women in
antenatal and intra partum period. In general the major
reason of inflated postpartum death and relative small
number of death in intra-partum and post-partum in JUSH
as compared to Bonke wereda study is set up difference in
which health institution death represents for 14% of
females who utilize institutional delivery but the study of
Bonke wereda represents 86% of females who did not
utilize institutional delivery.
Implications of these findings are
• Management after terminating pregnancy might
be poor or the women might not utilize
postpartum care service.
• The risk of death during intra-partum period is
high when the delivery is at home.
5.2. Factors Affecting Maternal Death
The commonest age group in which most maternal
death (52.5%) occurred is 20 – 34. Most controls (65.63%)
are also found in this age group. Likewise study done in
Tigray regional states’ hospitals in Ethiopia, most deaths
(55%) occur in age group of 20 – 34 but it is not
significant predictor. This might be because of
representation of maternal death by less than half from
this finding. When we compare this finding with study
done in Tanzania Muhimbili national hospitals in 2011
most deaths (74.5%) occurs in age group of 20 – 34. This
difference might be because of over representation of 20-
34 due to high sample (141 deaths) and it might be
because of poor quality of care given like inadequate or no
blood transfusion. Even though most death occurred in
this age group it is protective than being in the age group
of >=35. A one women increase in the age group of 20 –
34, decreased the odds that a woman would die by 70.1%
than being in the age group of >=35 with AMOR of 0.299
and 95% CI (0.113, 0.792) [12,16,21,22]. This finding
implies that even though high maternal death occurred in
this age group from total deaths, it is favorable age group
when we compare death ratio within the age groups. From
the total sample included in the age group of 20-34 only
16.7% are dead which is less than age group of <20
(23.5%) and >=35 (26.9%). Favorability of 20-34 is also
seen in Tigray study in which only 15% of women in this
age group are dead compared to 27.3% in <20 and 36.5%
in >= 35 age group.
Three fourth of women who came from rural area are
died as per the finding of this research. They would die
2.594 times more likely than women who came from
urban area with AMOR of 3.572 and 95% CI (1.001,
6.726). Even though it is not significant predictor, study
from Tigray region shows 75.6% of women who died
came from rural part of area. This might be because of
women who came from rural area are at higher risk than
who came from urban women which is evidenced by from
women who don’t have ANC follow up 72.4%, from
women who were grand multi parity 72.9%, from women
who had APH 70.4%, from women who had PPH 69%,
from women who had prolonged labour 62.6% and from
women who had uterine rupture 78.4% of them came from
rural area.
These imply that women who live in rural area are poor
at utilization of maternal health services might be due to
different reasons like awareness problem or inaccessibility
or unavailability of maternal health services.
Around 94% of maternal deaths are referred cases.
Referred cases from health center and other HI are 4.011
(95% CI (1.113, 14.464)) and 6.209 (95% CI (1.565,
24.626)) times more likely they would die as compared to
women who are self-referred respectively. Study done in
tertiary hospital of Kenya shows referred cases would die
two times than who don’t referred AOR of 2.1 and 95%
CI (1.0–4.3) [16]. This might be because of cases are
referred to a higher health institutions when it is
complicated. This implies that there may be unnecessary
delay from health institutions and from the family which
is evidenced by 90% of referred women have prolonged
labour.
When the duration of labour of women is 24 hour or
more 37.141 times more likely the women would die than
women who had less than 24 hour duration of labour with
95% CI (13.296, 103.750). OL is the underlying factors
for different kind’s risks during intra-partum period.
Majorly it might lengthens the length of labour. Most
(69%) of women who had OL had prolonged labour or
unnecessary delay from patient side or HI which is
evidenced by 90% of referred women have prolonged
labour. The other reason might be around 40% of women
came from more than 51km so it lengthens time without
intervention. Research finding from Nigeria maternity
hospital AMOR - 2.86 and 95% CI (1.39, 5.9) shows
prolonged labour is the major risk factor for maternal
death. Study from Tigray shows women whose length of
labour is < 24 hours are protected from death than women
whose length of labour is >= 24 hour with AOR of 0.27
and 95% CI (0.07-0.89) [15,22,35]. The major consequence
following prolonged labour is uterine rupture – 78%
women who had uterine rupture prior they are exposed for
prolonged labour. These findings imply that prolonged
labour is the major risk factor and exposes the women for
another risk factor like uterine rupture. The other
implication is OL and referral might be the great
contributors for prolonged labour.
Seventy nine women (13%) had any type of
comorbidities. They are 9.631 times more likely they
would die than women who don’t have comorbidities with
AMOR of 9.631 and 95% CI (3.135, 29.588). In Kenya
tertiary hospital also shows presences of comorbidities are
significant predictors with AOR of 3.0 and 95% CI (1.7–
5.3) [16]. It is known that pregnancy by itself is immune
suppressive conditions if there is an additional disease it is
dangerous. In this research half of women who had
comorbidities are died.
5.3. Strength of Study
• Free of recall bias
• Time matching (indirectly controls for quality of
care, expansion of services…)
127 American Journal of Public Health Research
• Study design – case control is preferable for rare
event like maternal death and for determining
associated factors.
5.4. Limitation of Study
• Selection bias due to intentional selection of
controls
• Does not observe effect of some of socio
demographic variables like economy,
education…due to 99% missingnes of socio
demographic variables.
• Wider confidence interval – due to rarity of
events and lower sample size
6. Conclusion
Most of the women died in post-partum period
particularly in between 4-7 days.
Women who reside in rural area, presence of
comorbidities, women who referred from health
institutions and women who had prolonged labour
increases the likelihood of maternal death. Whereas being
in the age group of 20 – 34 was the protective.
References
[1] Divya A. Patel, Nancy M. Burnett, BS Kathryn M. Curtis P.
Maternal health epidemiology. 2003.
[2] WHO. International statistical classification of diseases and
related health problems. 10th revision. Vol.2 Instruction Manual.
2010 Edition. Geneva: 2010;2.
[3] Baden S. Challenges to women ’ s reproductive health : maternal
mortality. 1996;44(38).
[4] The world bank. The millennium development goals for health to
rising the challenges. 2004.
[5] Ye F, Wang H, Huntington D, Zhou H, Li Y, You F, et al. The
Immediate Economic Impact of Maternal Deaths on Rural Chinese
Households. 2012;7(6).
[6] Reed HE, Koblinsky MA, Mosley WH. The Consequences of
Maternal Morbidity and Maternal Mortality. 2000.
[7] Bryce J, Victora C, Berman P, Lawn J, Mason E, Starrs A, et al.
Fulfilling the Health Agenda for Women and Children The 2014
Report ER. 2014.
[8] BS Garg SC and SMZ. Safe motherhood : social , economic , and
medical determinants of maternal mortality Women.
2006;(September). Available from:
http://www.the-
networktufh.org/publications_resources/trainingmodules.asp.
[9] Kirigia JM, Oluwole D, Mwabu GM, Gatwiri D, Kainyu LH.
Effects of maternal mortality on gross domestic product ( GDP ) in
the WHO African region. Afr J Heal Sci. 2006;13(1):86-95.
[10] WHO, world bank, UNFPA U and U. Trends in Maternal
Mortality : 1990 to 2013. 2013.
[11] Koblinsky M, Tain F, Tesfaye S. Reducing maternal mortality and
increasing use of skilled birth attendance: Ethiopia and MDG 5.
2010;4(1).
[12] Gidey G. Patterns of maternal mortality and associated factors; a
case-control study at public hospitals in Tigray region, Ethiopia.
Int J Pharm Sci Res. 2013;4(5):1918-29.
[13] Central Statistical Agency [Ethiopia] and ICF International.
Ethiopia Demographic and Health Survey 2011. 2011.
[14] Central Statistical Agency [Ethiopia] and ORC Macro. Ethiopia
Demographic and Health Survey Preliminary Report. 2005.
[15] Evjen-olsen B, Hinderaker SG, Lie RT, Bergsjø P, Gasheka P,
Kvåle G. Risk factors for maternal death in the highlands of rural
northern Tanzania : a case-control study. BMC Public Heal 2008,.
2008;9.
[16] Yego F, Este CD, Byles J, Williams JS, Nyongesa P. Risk factors
for maternal mortality in a Tertiary Hospital in Kenya : a case
control study. BMC Pregnancy Childbirth. 2014;14(38).
[17] Jimma univeristy specialized hospital. services given in the
hospital. 2015.
[18] Federal Democratic Republic of Ethiopia Ministry of Health.
Maternal Death Surveillance and Response (MDSR) Technical
Guideline. 2012.
[19] Yaya Y, Eide KT, Norheim OF, Lindtjørn B. Maternal and
Neonatal Mortality in South-West Ethiopia : Estimates and Socio-
Economic Inequality. PLoS ONE 9(4) e06294. 2014;9(4).
[20] Olopade, F.E. and Lawoyin T. Maternal Mortality in a Nigerian
Maternity Hospital. African J Biomed Res. 2008;11(June):267-73.
[21] Garomssa H and ADD. Maternal mortality in Ambo Hospital: a
five year retrospective review. Ethiop J Reprod Heal. 2008;2(1):
2-13.
[22] Pembe AB, Paulo C, Brenda SD, Roosmalen J Van. Maternal
mortality at muhimbili national hospital in Dar-es-Salaam,
Tanzania in the year 2011. BMC Pregnancy Childbirth. 2015;1-7.
[23] Adelaja LM, Taiwo OO. Maternal and Fetal Outcome of Obstetric
Emergencies in a Tertiary Health Institution in South-Western
Nigeria. Int Sch Res. 2011;2011.
