Compliance with Emergency Obstetric

Care Referrals Among Pregnant Women

in an Urban Informal Settlement of Accra,
Ghana

Adanna Uloaku Nwameme, James

F. Phillips & Philip Baba Adongo

Maternal and Child Health Journal

ISSN 1092-7875

Matern Child Health J

DOI 10.1007/s10995-013-1380-0

1 23
Your article is protected by copyright and all
rights are held exclusively by Springer Science
+Business Media New York. This e-offprint is
for personal use only and shall not be self-
archived in electronic repositories. If you wish
to self-archive your article, please use the
accepted manuscript version for posting on
your own website. You may further deposit
the accepted manuscript version in any
repository, provided it is only made publicly
available 12 months after official publication
or later and provided acknowledgement is
given to the original source of publication
and a link is inserted to the published article
on Springer's website. The link must be
accompanied by the following text: "The final
publication is available at link.springer.com”.

1 23
 Author's personal copy
Matern Child Health J
DOI 10.1007/s10995-013-1380-0
Compliance with Emergency Obstetric Care Referrals Among
Pregnant Women in an Urban Informal Settlement of Accra,
Ghana
Adanna Uloaku Nwameme • James F. Phillips •
Philip Baba Adongo
 Springer Science+Business Media New York 2013
Abstract This study examines the situation that women
face when they require emergency obstetric care in urban
Accra. The analysis clarifies the referral options available
to women needing emergency obstetric care, assesses the
constraints they face in accessing the obstetric referral
system and identifies the drawbacks associated with the
obstetric referral system for women living in informal
settlements of a rapidly growing district in Ghana. This
research was a descriptive cross-sectional study using
structured questionnaires administered to antenatal care
clinic attendees and in-depth interviews of principal
healthcare personnel. Sixty-five women had referrals in
their previous pregnancies of which 62 went to the referral
centre at varying time intervals. Three respondents did not
go due to lack of financial resources and preference for
traditional services. With regard to adherence to referral
advice, lack of finances was the major constraint (46.2 %)
followed by client complaints about the attitudes of nurses
at the referral centres (10.8 %), fear of surgery (7.7 %) and
concerns about the distance to referral centres (4.6 %).
Moreover, analysis identified a significant positive associ-
ation between parity and time elapsed between service
A. U. Nwameme (&)
P. B. Adongo
Department of Social and Behavioural Sciences, School
of Public Health, University of Ghana, P.O. Box LG 13, Legon,
Accra, Ghana
e-mail: adanwameme@yahoo.com
P. B. Adongo
e-mail: adongophilip@yahoo.com
J. F. Phillips
Heilbrunn Department of Population and Family Health,
Columbia University Mailman School of Public Health,
60 Haven Avenue, B-2, New York, NY 10032, USA
e-mail: jfp2113@columbia.edu
encounter and compliance with referral (p = 0.001). Major
constraints are faced by women when they attempt to uti-
lize referral healthcare services such as financial problems,
perceived remoteness of care points, fear of surgical pro-
cedures and concerns about the discourteous attitude of
nurses at the referral centres. Healthcare providers how-
ever, emphasized other elements of the climate of care such
as their perceived need for increasing staff strength.
Keywords Referral System
Emergency Obstetric
Care (EmOC)
Maternal Mortality
Urban informal
settlements
Ghana
Introduction
Global commitment to reducing maternal mortality is
increasingly focused on concerns about the lack of sub-
stantial progress in Sub-Saharan Africa. Over 800 women
die daily from complications associated with pregnancy
and delivery, of which over half occur in Sub-Saharan
Africa [1]. Of these deaths, nearly three-quarters are
attributable to direct causes that could be prevented with
effective referral and skilled obstetric care, such as hem-
orrhage, obstructed labour, sepsis, unsafe abortion and
eclampsia [2]. Death and disability amongst women of
childbearing age in developing countries is largely due to
complications during pregnancy and childbirth with an
annual estimate of 284,000 deaths [1]. Although the official
Millennium Development Goal aim for 2015 is 54 deaths/
100,000 live births [3], maternal mortality ratio in Ghana is
presently estimated to be 350 deaths per 100,000 live births
[1]. Clearly, high maternal mortality represents a sustained
development challenge for Ghana.
123
 Author's personal copy
Any effective maternity referral system aims to provide
timely treatment of obstetric complications. Among health
deprived impoverished women in developing countries,
estimates predict that access to essential obstetric care
could avert about half of the maternal deaths [4]. Essential
Obstetric Care (EOC) is a term used to describe the ele-
ments of obstetric care needed for the management of
normal and complicated pregnancy, delivery and the
postpartum period [5]. At the health center level, basic
EOC requires parenteral antibiotics, parenteral oxytocic
drugs, parenteral sedatives for eclampsia, skills essential to
the manual removal of placenta and of retained products.
Emergency Obstetric Care (EmOC) on the other hand,
expands upon this basic regimen to encompass services
such as surgery and blood transfusion for the treatment of
complications that arise during pregnancy and childbirth.
Maternal mortality will be significantly reduced by
improving the availability, accessibility, quality and use of
these services and early identification of these problems
will also reduce the complications of childbirth, including
birth asphyxia, which contribute up to a third of all neo-
natal deaths in some developing countries [6]. Throughout
Ghana, district hospitals typically comprise the first referral
level and therefore should offer comprehensive essential
obstetric care which includes the services offered by the
basic EOC as well as the EmOC. Such facilities are
regarded as ‘‘functional’’ if aforementioned services have
been continuously operational for 6 months prior to data
collection.
Reviews of the maternal referral system in Ghana have
consistently shown that maternal risk is elevated by service
capability lapses that concatenate with traditions that sup-
port home birth, lack of affordable and appropriate means
of transportation, long distances between primary and
secondary health institutions, lack of funding, unreliable
means of communication and the absence of a coherently
organized referral network [3]. As studies elsewhere in
West Africa have shown, addressing cultural and social
constraints to facility based delivery requires high quality
services, effective outreach, and supportive systems of
referral and emergency [7].
The Prevention of Maternal Mortality (PMM) Network,
a multi-site study on maternal mortality carried out in
Ghana, Nigeria and Sierra Leone between 1988 and 1996,
identified what is known today as the ‘‘three delays’’- delay
in the decision to take appropriate action; delay in arriving
at a health facility; and the delay within the health facility
while a fourth delay in the recognition of a problem
ab initio has also been proposed [8]. The PMM Network
determined that delay in actually receiving care immedi-
ately after arriving at the facility, is the most critical
determinant of effective referral and care [9]. Thus in
addition to developing maternal healthcare services,
123
Matern Child Health J
effective efforts to improve efficient delivery of care for
emergency obstetric complications require intermittent
assessments of existing services to gauge lapses in quality
of care and monitor any progress made so that recom-
mendations for interventions can be made based on clinical
evidence.
The study area, Ga East District, is the most rapidly
expanding urban area in Ghana, and yet has no referral
hospital. Only one ambulance is available for a population
exceeding 320,000 with no public district hospital to cater
for cases referred from the poly clinics and health centres.
For this reason, all referrals are made to a tertiary care
facility in the city called the Ridge Hospital, Accra, an
average of 20 kilometres from households of Ga East.
Considering the lack of adequate transportation, long dis-
tance to referral centres and low socioeconomic status of
the women, the referral system in the impoverished urban
centres is facing major challenges. Since a woman in need
of emergency obstetric care requires a short window of
referral time, the remoteness of Ridge Hospital exposes her
to higher health risks, maternal morbidity and mortality.
To foster official understanding of prevalence of
maternal mortality, regular assessments of referral systems
are needed to generate essential evidence that supports
policymakers’ efforts to accelerate progress with MDG5.
This study is focused on the component of this problem
that concerns constraints to emergency obstetric care
referral adherences among pregnant women in urban
Accra.
Methods
This study is a descriptive cross-sectional appraisal of the
views of service providers and their obstetric care clientele
at ante-natal care (ANC) clinics in three healthcare facili-
ties of Ga East District, which is located in the northern
part of Greater Accra Region of Ghana. It has a population
of 320,853 people and covers an area of 166 square kilo-
metres. Currently, of the 39 health facilities in the
municipality, majority are in the private sector. The public
health facilities are Madina Polyclinic, Kekele; Madina
Polyclinic, Rawlings Circle, the health centres at Abokobi
and Danfa; and the CHPS compound located at Taifa.
Pentecost Medical Centre is a mission facility at Madina
which currently serves as the municipal hospital.
In 2010, the Ga East Municipality recorded ANC cov-
erage of 8625 women (67 % of pregnant women) [10]. To
estimate statistical power for an assessment, we consulted a
study in Gutu, Zimbabwe with an antenatal and delivery
referral rate of 36 % [11]. Assuming a desired precision of
5 and 95 % confidence interval, a sample size of 355
women was calculated assuming simple random sampling.
 Author's personal copy
Matern Child Health J

Considering a 10 % rate of non-response, the desired Table 1 Socio-demographic characteristics of respondents

sample size was estimated to be 390 ANC attendees. Background characteristics Frequency (N = 390) Percent

Data Collection Procedures Age group

15–19 10 2.6
The data was collected using two different methods: 20–24 53 13.6
structured questionnaires were administered to all women 25–29 158 40.5
(except primiparous women) attending ANC clinics during 30–34 105 26.9
the study period and in-depth interviews were conducted 35–39 52 13.3
with the medical assistant-in-charge, medical officer-in- 40? 4 1
charge and the nursing administrator of the facilities No response 8 2.1
included in the study which are Abokobi Health Centre; Marital status
Madina Polyclinic, Kelele and Pentecost Medical Centre, Single 14 3.6
Madina. Married 359 92
In addition to the in-depth interview guide and struc- Cohabiting 17 4.4
tured questionnaires employed in the study, the Ghana Religion
Health Service Referral Guidelines, in combination with a Orthodox Christian 139 35.6
WHO EOC Facility Review checklist were also used. Unorthodox Christian 174 44.6
To ensure that high standards of data quality were sus- Moslem 72 18.5
tained, research assistants were duly trained and data col- No response 5 1.3
lected were doubly entered. Consent forms were issued to Level of education
all respondents prior to being interviewed and confidenti- Primary 51 13.1
ality was assured. Ethical approval for the study was Junior secondary school 203 52.1
obtained from the Ethical Review Board of the Ghana Senior secondary school 60 15.4
Health Service and permission sought and obtained from Tertiary 15 3.8
the Ga East District health authorities. Vocational 11 2.8
No education 50 12.8
Data Analysis
Occupation
Housewife 24 6.2
Bivariate tabulation was imposed to quanitative data with
Trader 297 76.2
results presented in tables and graphs. Chi square tests were
Civil seravant 24 6.2
also conducted on categorical variables with Fisher’s Exact
Student 2 0.5
Test used as deemed appropriate. Interviews were digitally
Seamstress 21 5.4
recorded, transcribed and subsequently analysed using the
Teacher 1 0.3
NVivo software.
Hairdresser 13 3.3
Healthcare provider 2 0.5
Study Limitations
Others 6 1.7
This study was carried out within a proposed time frame Parity
which led to the respondents being chosen from the clinic 1 170 43.6
ANC for questionnaire administration, thus introducing a 2 116 29.7
recall bias. Nevertheless, the questions were restricted to 3 64 16.4
present and previous pregnancies to counteract this effect. 4? 40 10.3

Results
Demographics Characteristics of ANC Attendees
ANC clients interviewed comprised an entire range of
women aged 17–46 years of which the modal 5 year age
class was represented by the 40.5 % who were in the
25–29 year age group. Most respondents were married
women (92.0 %), with the rest single (3.6 %) or cohabiting
(4.4 %), and were made up of orthodox Christians
(35.6 %), unorthodox Christians (44.6 %) and Moslems
(18.5 %) (Table 1). Three hundred and forty women
(87.2 %) had undergone some form of education ranging
from primary to tertiary level, while fifty women (12.8 %)
had no educational attainment whatsoever. The women
were engaged in different occupations with 76.2 % being

123
 Author's personal copy
Matern Child Health J

traders. Furthermore, the respondents all had children
ranging from one child (43.6 %) to more than four children
(10.3 %).
Factors Leading to Obstetric Referrals in Ga East
District
Whereas 89.7 % (349) of the respondents had been told
about the signs of pregnancy complications in the course
their antenatal visits, only 80.9 % (315) knew where to go
in the event of an actual emergency (Table 2). A statisti-
cally significant relationship did not exist between educa-
tional attainment and knowledge of where to go in the
event of an obstetric complication happening (p = 0.84),
with the least educated at greater odds of not knowing
where to go for emergency care.
Only 4.4 % (17) of respondents had been referred during
their present pregnancy (Fig. 1). Of these seventeen, six
were referred for pregnancy-induced hypertension, three
for antepartum haemorrhage, two for fetal malpresentation,
and one each for anaemia and malaria. In addition to these
cases, four referrals were made for non-pregnancy related
complications: difficulty in breathing, hepatitis B infection,
sickle cell anaemia and gastroenteritis.
Regarding previous pregnancies, 16.7 % (65) of the
respondents had been referred in their previous pregnancies
and the conditions for which the women were referred
included pregnancy-induced hypertension (7), obstructed
labour (7), antepartum haemorrhage (5), postpartum
haemorrhage (5), anaemia in pregnancy (4), primiparity
(4), malpresentation (4), ruptured uterus (3) and delayed
second stage (3) (Table 3). In addition to these, there were
two cases of neonatal complications and six of non-preg-
nancy related complications.
These obstetric conditions are in keeping with those
elicited from the key informant interviews. All three health

Table 2 Information
Variables Frequency (%) Chi square p value
concerning signs/symptoms of
obstetric complications Educated Not educated
(N = 389)
Information concerning obstetric complications
Provided at ANC 308 (79.2) 41 (10.5)
Not provided at ANC 33 (8.5) 7 (1.8)
Knowledge of where to go in the event of obstetric complications
Knows where to go 277 (71.2) 38 (9.7) 0.351 0.84
Does not know where to go 64 (16.5) 10 (2.6)

Fig. 1 Distribution of
conditions that called for
referrals in the current
pregnancy

123
 Author's personal copy
Matern Child Health J

Table 3 Distribution of conditions that called for referrals in the Table 4 The referral process
previous pregnancies
Variables Referrals in the Referrals in the
Obstetric complications Frequency Percent present previous
pregnancy pregnancy
Pregnancy-induced hypertension 7 11.3 [frequency (%)] [frequency (%)]
Antepartum haemorrhage 5 8.1
Time interval between referrals and reporting at the referral centre
Postpartum haemorrhage 5 8.1
Within 24 h 14 (82.3) 53 (85.5)
Obstructed labour 7 11.3
24–48 h 2 (11.8) 3 (4.8)
Sepsis 1 1.6
2–5 days 0 4 (6.5)
Ruptured uterus 3 4.8
More than 10 days 1 (5.9) 2 (3.2)
Fetal distress 1 1.6
Issuance of an accompanying referral letter
Neonatal complications 2 3.2
Referral letter given 15 (88.2) 57 (91.9)
Anaemia in pregnancy 4 6.5
Referral letter not given 2 (11.8) 5 (8.1)
Primiparity 4 6.5
Presence of hospital staff in transportation vehicle
Poor maternal effort in labour 1 1.6
Accompanied by 1 (5.9) 18 (29.0)
Fetal malpresentation 4 6.5
hospital staff
Non-pregnancy related complications 6 9.7
Not accompanied by 16 (94.1) 44 (71.0)
Twin pregnancy 2 3.2 hospital staff
Macrosomia 2 3.2 Mode of transportation
Intrauterine death 1 1.6 Ambulance 0 14 (22.6)
Delayed second stage 3 4.8 Public transport 10 (58.8) 19 (30.6)
Placenta praevia 1 1.6 Private transport 7 (41.2) 29 (46.8)
Retained placenta 1 1.6
Amniotic fluid embolism 1 1.6
Postmaturity 1 1.6 who went to the referral centre, 53 (85.5 %) arrived within
24 h, 3 (4.8 %) within 48 h, 4 (6.5 %) within 2–5 days and 2
(3.2 %) in more than 10 days. The three respondents who did
facilities listed obstructed labour, anaemia in pregnancy, not go to the referral centres cited lack of money and more trust
postpartum haemorrahage, pregnancy-induced hyperten- in the locally available services as their reasons.
sion, delayed first stage of labour, malaria in pregnancy, According to the key informants, the referrals that did
ectopic pregnancy, abnormal fetal presentation, cord pro- not have any accompanying nurse or midwife were not
lapse and fetal distress as the major emergency obstetric emergency referrals but those identified as risky and refer-
referrals they had to deal with. red to a higher level of healthcare for further management.
This situation often translated to some patients arriving late
Accessibility to Emergency Obstetric Care in the Ga at the referral centres as one of the interviewees pointed out.
East District
‘‘Some patients see referrals as a punishment but
when we explain to them, they go. Another problem
Of the 17 women referred to referral centres in their present
is that patients don’t report early sometimes.’’ (In-
pregnancies, ten went by public transport and seven made
depth Interview, Nursing Administrator).
their own private arrangement for transportation. Fifteen of
the women were given a referral letter but only one was A relationship was evident between parity and time
accompanied to the referral centre by a hospital staff as lapse (in previous pregnancies) from being referred to
required in the Referral Policies and Guidelines of the reporting at the referral centre (p = 0.001) showing that
Ghana Health Service [12]. Of the 17 women, fourteen the more children the women had, the less likely they were
reported to the referral centre within 24 h, two within 48 h to heed to referral advice. Conversely, no relationship was
and one woman in more than 10 days as shown in Table 4. found between marital status and delays in reporting at the
Sixty-five women had referrals in their previous pregnancies referral centres (p = 0.23) (Table 5).
and 62 went to the referral centre out of which only 14 (22.6 %) All the respondents were asked what constraints women
were transported by hospital ambulance. Nineteen (30.6 %) in their community faced with regard to adhering to referral
went by public transport and 29 (46.8 %) made private advice. Financial problems were the major cause of con-
arrangements for their transportation. Though 57 (91.9 %) of cern (46.2 %) followed by the attitude of nurses at the
the women were given a referral letter, only 18 (29 %) were referral centres (10.8 %), fear of surgery (7.7 %) and the
accompanied by a hospital staff. Regarding the respondents distance to referral centres (4.6 %) (Table 6).

123
 Author's personal copy
Matern Child Health J

Table 5 Relationship between

Variable Time Lapse [Frequency (%)] Chi Square p value*
marital status and time lapse;
and parity and time lapse 24 h 24–48 h 2–5 days More than 10 days
(N = 62)
Marital status
Married 49 (79.1) 2 (3.2) 3 (4.8) 2 (3.2) 13.63 0.23
Single/cohabiting 4 (6.4) 1 (1.6) 1 (1.6) 0
Parity
One 19 (30.6) 1 (1.6) 0 0 29.23 0.001
Two 12 (19.4) 0 1 (1.6) 0
Three 12 (19.4) 2 (3.2) 1 (1.6) 1 (1.6)
* p value was calculated using
More than four 10 (16.1) 0 2 (3.2) 1 (1.6)
the Fisher’s Exact Test

Table 7 EOC facility review using WHO guidelines

Table 6 Respondents’ opinion of the constraints faced by women in
their community regarding referrals essential obstetric Name of facility
care service
Constraints Frequency Percent readiness Abokobi Madina Pentecost
N = 390 health Polyclinic, Medical Centre,
Centrea Kekelea Madinab
Fear 30 7.7
Financial problems 180 46.2 Parenteral YES YES YES
antibiotics
Poor performance of referral centres 16 4.1
Parenteral oxytocics YES YES YES
Attitude of staff 42 10.8
Parenteral sedatives/ YES YES YES
Distance 18 4.6
anticonvulsants
Ignorance 3 0.8
Manual removal of YES YES YES
Previous bad experience 3 0.8 placenta
Transportation difficulties 5 1.3 Removal of retained NO NO YES
Financial problems and attitude of staff 13 3.3 products of
Financial problems and distance 21 5.4 conception
Financial problems and fear 24 6.2 Assisted vaginal NO NO YES
delivery
Financial problems and poor performance 11 2.8
of referral centres Blood transfusion NO NO YES
Fear and distance 3 0.8 Caesarean section NO NO YES
a b
Financial problems, distance and religious 3 0.8 Verdicts: Not an EOC facility, A comprehensive EOC facility
reasons
Fear, distance and financial problems 5 1.3
of placenta and so fell short of being tagged essential
Don’t know 13 3.3
obstetric care facilities.
At Abokobi Health Centre and Pentecost Medical Centre,
Madina, more than 75 % of skilled staff was trained in the use
Availability of Emergency Obstetric Care in the Ga of the referral guidelines while in Madina Polyclinic, Kekele,
East District only about 25 % had received any training. All the health
facilities complained of inadequate staff strength to deal with
Of the three health facilities included in the study, only the large number of obstetric patients that visit their facilities.
Pentecost Medical centre is an essential obstetric care
‘‘We are overwhelmed. You can see the number of
facility as it is the only facility that had offered the six
patients here today and we do not have enough staff
services required of a basic EOC facility in the last 6 months
to cater for them so we cannot even effectively
as stipulated by WHO- parenteral antibiotics, oxytocics and
practice the focused antenatal care.’’ (In-depth
sedatives/anticonvulsants, manual removal of placenta,
Interview, Nursing Administrator).
removal of retained products of conception and assisted
vaginal delivery (Table 7). In addition to these, the health Abokobi Health centre has 1 medical officer, 20 midwives/
facility also offers emergency obstetric care services- blood nurses and 13 community health nurses; Madina Polyclinic,
transfusions and caesarean sections- making it a compre- Kekele has 3 medical officers, 5 physician assistants, 20
hensive essential obstetric care facility (CEOC). midwives/nurses and 29 community health nurses while
Abokobi Health Centre and Madina Polyclinic, Kekele Pentecost Medical Centre, Madina has 4 medical officers, 16
only offered the parenteral medications and manual removal midwives/nurses and 6 community health nurses.

123
 Author's personal copy
Matern Child Health J

Referral System Support Mechanisms mobile phone before transporting the patients to determine
whether or not a hospital bed is available so that alternative
Copies of the Ghana Health Service Referral Policies and clinical coverage can be pursued if necessary.
Guidelines were available at all the health facilities and the
‘‘There are hindrances between the two hospitals, no
staff usually adhered to the stipulated directions. Standard
beds, no doctors…all these contribute to the delays.’’
referral forms were available in all the facilities and were
(In-depth Interview, Nursing Administrator).
issued to the referral patients for onward transmission to
the referral centres. Referral registers were also available in
the facilities but were not properly documented. Pentecost
Discussion
Hospital however had patients’ information computerized
making for easier access to patients’ records and data. The
Availability of Emergency Obstetric Care Services
referring health facilities receive little or no feedback from
for Urban Women
the referral centre and thus have no way of following up on
the patients’ progress. Some of the key informants found
Findings state that direct causes such as haemorrhage,
this particularly distressing.
obstruction, sepsis, unsafe abortion and eclampsia account
‘‘We don’t receive any feedback from the hospitals. for about 75 % of all maternal deaths [2]. In accordance
At least it would help us understand what we could with this, major indications for maternal referrals in the Ga
have done better.’’ (In-depth Interview, Medical East District include pregnancy-induced hypertension,
Officer). obstructed labour, antepartum haemorrhage, postpartum
haemorrhage, anaemia in pregnancy, primiparity and fetal
‘‘What we need is proper coordination and collabo-
malpresentation.
ration. Ridge helps us by bypassing some of the
The healthcare providers at the health facilities under
bureaucracies and organizing mortality meetings.’’
study complained that they were hugely understaffed as
(In-depth Interview, Medical Officer).
they have more patients than the facilities can comfortably
attend to. This observation corresponds with reports from
Transportation and Communication Arrangements
the district that show that the doctor-patient ratio is
Between the Referral Levels
1:61,412 while the nurse-patient ratio is 1:2020 [10]. In
addition to this, occasionally the referral centres do not
The health centre at Abokobi depends on a van offered by
have enough bed spaces and so are not able to admit the
the District Assembly to transfer obstetric emergency cases
patients referred to them. The health centres then have to
to referral centres. Occasionally, they rely on Pentecost
make alternative arrangements leading to delays in
Medical Centre or Ridge Hospital to send an ambulance for
accessing appropriate healthcare. This congestion at the
the patient. Madina Polyclinic, Kekele uses the hospital
referral centres can only lead to suboptimal healthcare
van to transport patients to nearby Pentecost Medical
delivery. Miller et al. [13] conducted a study in the
Centre while relying on the National Ambulance Service
Dominican Republic and established that overcrowding
for referrals to Ridge Hospital. Pentecost Medical Centre
and poor technical ability in hospitals lead to an increase in
makes referrals to Ridge Hospital, Korle-Bu Teaching
maternal mortality ratio.
Hospital and 37 Military Hospital- all tertiary care facili-
The referral system does not exist alone but depends on
ties- using their own ambulance. In all cases, patients are
an effective support service system that includes transport
accompanied by a nurse in the vehicle. These arrangements
and communication.
are not without challenges. According to one key
informant,
‘‘We do not know how long our arrangement with the Transportation Support
District Assembly will last. Our problem is finance—
at least, we should be able to buy our own ambu- The study found that all the healthcare facilities involved
lance.’’ (In-depth Interview, Medical Assistant). had a means of transporting their emergency obstetric cases
to the referral centre but considering the nature of these
‘‘We still have a problem with access to an ambu-
transport arrangements, it goes without saying that the time
lance especially at night.’’ (In-depth Interview,
lapse between a referral and the patient’s arrival at the
Medical Officer).
referral centre could be unduly prolonged. For a woman
In the event of an obstetric emergency, the staff of refer- who is haemorrhaging postpartum, the average interval
ring health facilities always contact the referral centres by from onset of bleeding to death is 2 h [14]. Timely and

123
 Author's personal copy
affordable emergency care depends on reliable transporta-
tion and time lapses between complication onset and access
to a higher level of care can only be bridged by improved
road infrastructure, readily available transport vehicles and
funds for fuel and maintenance costs [15].
Communication Support
Results show that all the health facilities are able to make
verbal contact with the referral centres via mobile phone,
prior to the transportation of any patient. As such, they can
request for assistance in transporting the patient, informa-
tion on availability of bed space or confirmation of the
presence of a competent medical practitioner at the referral
centre. An efficient system of communication works in
tandem with emergency transport, and is required at all
level of healthcare to arrange for transportation, receive
advice about immediate medical management, and arrange
prompt referral [2].
Accessibility of Emergency Obstetric Care to Urban
Women
Most of the respondents cited financial problems, distance
and fear of clinical procedures as the major constraints
facing women in accessing the referral system; yet despite
these hurdles faced, many of the women reported to the
referral centres. In concurrence with this, Thaddeus and
Maine [16] found in their study that while distance and cost
are major obstacles in the decision to seek care, people often
consider the quality of care as more important than cost.
However, monetary costs are not the only source of worry
for these women- there are opportunity costs to bear in mind
as well. Even though the healthcare delivery services are
free, expenditures on any accompanying relative, care of the
children left at home, return trip costs and lost productivity
are all to be considered. Underprivileged women are dis-
proportionately affected as they are likely to live far away
from adequate healthcare services and so have to travel
further distances and suffer greater impacts from both the
direct and indirect costs of reaching care [17].
Furthermore, the women also considered the unfriendly
attitude of nurses at the referral centres a deterrent to
accessing a higher level of healthcare. Many of the re-
spondetns preferred TBAs for this reason, not only because
clients had more confidence in their services, but also
because TBAs were less expensive sources of care. Fatusi
and Ijadunola [18] conducted a study that cut across twelve
states in Nigeria and established that many women pre-
ferred the services of the TBAs over that of healthcare
delivery facilities. This inclination was due to the smaller
financial burden and nurturing offered by the TBAs as
123
Matern Child Health J
opposed to the ‘‘hostile and unsympathetic nature’’ of the
healthcare workers.
Tabulation revealed that time lapse in receiving care
increased directly with the number of children ever and
agrees with findings by Majoko [19] in his study conducted
in Gutu, Zimbabwe which established that multiparous
women with uncomplicated previous pregnancies tended to
ignore referral advice, evaluating the clinical directives
based on their own perception of risks.
This study found other barriers to accessing the referral
system to include transportation difficulties, ignorance
about options for receiving care and prior experience with
malevolent care providers at referral centres. These obsta-
cles, as one of the key informants pointed out, lead to the
women showing up late at the referral centres. In a study
conducted in Lusaka, Zambia it was established that many
healthcare seekers bypassed the health centres and sought
services from the national referral hospitals largely due to
availability of drugs at the hospitals [20]. Contrary to this
view, even though Ga East District is made up of peri-urban
communities, the women sought help at the health centres
and were reluctant to go to the tertiary healthcare facilities
chiefly because of poverty and distance. Though the refer-
ring healthcare facilities bear the cost of transporting the
emergency patients to the referral centres, where the
obstetric complication is not an emergency, the women are
expected to make transportation arrangement for them-
selves, an obvious problem for women experiencing an
episode of illness requiring immediate clinical intervention.
In this circumstance, financial considerations and issues of
distance, appropriate transportation and road infrastructure
are major complications. The fact that the healthcare
facilities at the community level are not equipped to offer
emergency obstetric care services does not help matters as
unnecessary and avoidable referrals are made as in the case
of referrals due to primiparity. Nkyekyer [21] pointed this
out in his study at Korle-Bu Teaching Hospital stating that
some of the referrals were perhaps unnecessary or
avoidable.
The purpose of the referral policies and guidelines of the
Ghana Health Service is to provide guidance and standards
to improve referral practices and ensure quality care at all
levels of the health services, both public and private [12].
The staffs of healthcare facilities are expected to maintain
good public relations with clients and among themselves.
However, more than 10 % of the respondents in the study
cited the unfriendly attitude of the nursing staff as deter-
rents to their acting on referral advice. In a study in Tan-
zania, the complacent attitude among staff was identified as
a major barrier to accessibility of healthcare with most of
the staff opining that maternal deaths were due to extra-
neous circumstances beyond their control such as delayed
 Author's personal copy
Matern Child Health J
arrivals, cultural factors, and lack of drugs and equipment
leading to a passive attitude amongst them [22].
Some of the referrals made were for obstetric compli-
cations that were not emergencies such as demographic
risks (such as primiparity and small stature) and obstetric
historical risks (such as previous postpartum haemorrhage)
thus, there was no ambulance service offered to convey the
patient and no accompanying nurse or midwife. Referral
measures however were adhered to in obstetric emergency
cases and the cost was borne by the health facility. This
policy has mitigated the detrimental effect of resource
constraints on compliance with referral directives and
delay in seeking care. More often than not, referral letters
accompanied all the referred patients but unfortunately, the
referring healthcare facility rarely receives any feedback
report from the hospitals. This lack of communication
hinders further management of patients back at the com-
munity level along the appropriate lines while so many are
lost to follow up.
Conclusion
The maternal referral system in the study area was found to
be functioning inefficiently due to the absence of a func-
tional transport system, inadequate workforce, unavail-
ability of bed spaces in referral centres, and lack of
feedback from referral centres to care providers at the
periphery. This procedural lapse is a breach in the proto-
cols as outlined by the GHS Referral Policies and Guide-
lines. With the ratios between health personnel and the
population being so poor, the staff strength in the health
centres is weak, straining the quality of healthcare delivery
services.
Even among women who comply with referral instruc-
tions, concerns are expressed about financial worries,
remoteness of services, fear of caesarean surgery and the
harsh attitudes of nursing staff at tertiary facilities. The
frequency of non-compliance increases with parity, perhaps
because quality of care is poor and experience with care
providers is often unpleasant. In response to experience with
care, the odds that women will be non-compliant with
referrals increases and the chance grows that providers will
be poorly trained. Clearly, policy, training and management
attention to the climate of care is needed so that facility
based delivery becomes a more positive and nurturing
experience for both routine deliveries and referral clientele.
In an effort to combat maternal death due to obstetric
emergencies, existing referral policies need to be rein-
forced while antenatal clinics carry out extensive health
education not only to impress on women, the need for
prompt action in the face of an emergency, but also the
need for providers to be supportive, thoughtful and sensi-
tive representatives of the maternal care system.
Acknowledgments This paper was supported by grants to Colum-
bia University by the Doris Duke Charitable Foundation Africa
Health Initiative for the Ghana Essential Health Intervention Project
(GEHIP).
References
1. World Health Organization, UNICEF, UNFPA, The World Bank.
(2012). Trends in maternal mortality: 1990–2010. Geneva: The
World Health Organization.
2. Holmes, W., & Kennedy, E. (2010). Reaching emergency
obstetric care: overcoming the ‘second delay’. Burnet Institute on
behalf of Compass, the Women’s and Children’s Health
Knowledge Hub. Melbourne, Australia.
3. Awoonor-Williams, J. K. (2010). Transportation and referral for
maternal health within the CHPS system in Ghana. PowerPoint
presentation at the Wilson Center workshop, 20th May.
4. Murray, S. F., & Pearson, S. C. (2006). Maternity referral systems
in developing countries: Current knowledge and future research
needs. Social Science and Medicine, 62, 2205–2215.
5. WHO Fact Sheet No 245, June 2000. Available from: https://
apps.who.int/inf-fs/en/fact245.html. Accessed 28 Jan, 2011.
6. Murray, S. F., Davies, S., Phiri, R. K., & Ahmed, Y. (2001).
Tools for monitoring the effectiveness of district maternity
referral systems. Health Policy and Planning, 16(4), 353–361.
7. Harrison, K. A. (1985). Childbearing, health and social priorities:
A survey of 22,774 consecutive hospital births in Zaria, Northern
Nigeria. British Journal of Obstetric and Gynaecology, 92(5),
1–119.
8. Senah, K. (2003). Maternal mortality in Ghana: The other side.
Research Review NS, 19(1), 47–55.
9. Post, M. (1997). Preventing maternal mortality through Emer-
gency Obstetric Care. Support for Analysis and Research in
Africa (SARA) Issues Paper.
10. Ghana Health Service and Ga East Municipal Health Adminis-
tration (2010). Annual Report.
11. Majoko, F., Nystrom, L., Munjanja, S. P., & Lindmark, G.
(2005). Effectiveness of referral system for antenatal and intra-
partum problems in Gutu district, Zimbabwe. Journal of
Obstetrics and Gynaecology, 25, 656–661.
12. Ghana Health Service (2006). Referral Policies and Guidelines.
13. Miller, S., Tekur, U., & Murgueytio, P. (2002). Strategic
assessment of reproductive health in the Dominican Republic.
USAID, Secretaria de Estado de Salud Pública y Asistencia
Social (SESPAS), Population Council.
14. Maine, D. (1987). Prevention of maternal deaths in developing
countries: program options and practical considerations. Inter-
national Safe Motherhood Conference.
15. Giovine, A., & Ostrowski, C. (2010). Improving transportation
and referral for maternal health: knowledge gaps & recommen-
dations. A technical report for the Woodrow Wilson International
Center for Scholars.
16. Thaddeus, S., & Maine, D. (1994). Too far to walk: Maternal
mortality in context. Social Science and Medicine, 38(8),
1091–1110.
17. Ensor, T., & Cooper, S. (2004). Overcoming barriers to health
service access: Influencing the demand side. Health Policy and
Planning, 19(2), 69–79.
123
 Author's personal copy
18. Fatusi, A. O., & Ijadunola, K. T. (2003). National study on
essential obstetric care facilities in Nigeria. Nigeria and UNFPA:
A technical report for the Federal Ministry of Health.
19. Majoko, F. (2005). Assessing antenatal care in rural Zimbabwe.
Acta Universitatis Upsaliensis. Digital Comprehensive Summa-
ries of Uppsala Dissertations from the Faculty of Medicine 78:65.
20. Atkinson, S., Ngwengwe, A., Macwan’gi, M., Ngulube, T. J.,
Harpham, T., & O’Connell, A. (1999). The referral process and
123
Matern Child Health J
urban health care in Sub-Saharan Africa: The case of Lusaka,
Zambia. Social Science and Medicine, 49, 27–38.
21. Nkyekyer, K. (2000). Peripartum referrals to Korle-Bu Teaching
Hospital, Ghana: A descriptive study. Tropical Medicine and
International Health, 5(11), 811–817.
22. Mbaruku, G., & Bergstrom, S. (1995). Reducing maternal mor-
tality in Kigoma. Tanzania. Health Policy and Planning, 10(1),
71–78.