ArticlePDF Available

Evaluating the Cost-Effectiveness of an Integrated Program to Reduce Maternal and Neonatal Mortality in Ghana

The cost effectiveness of a quality
improvement program to reduce maternal
and fetal mortality in a regional referral
hospital in Accra, Ghana
David M. Goodman
*, Rohit Ramaswamy
, Marc Jeuland
, Emmanuel K. Srofenyoh
, Cyril
M. Engmann
, Adeyemi J. Olufolabi
, Medge D. Owen
1Hubert-Yeargan Center for Global Health, Department of Obstetrics and Gynecology, Duke University
Medical Center, Durham, NC, United States of America, 2Gillings School of Public Health, University of
North Carolina, Chapel Hill, NC, United States of America, 3Sanford School of Public Policy & Duke Global
Health Institute, Duke University, Durham, NC, United States of America, 4Ridge Regional Hospital, Ghana
Health Service, Accra, Ghana, 5Department of Pediatrics, University of Washington & SeattleChildren's
Hospital, Seattle, WA, United States of America, 6Department of Anesthesiology, Duke University Medical
Center, Durham, NC, United States of America, 7Department of Anesthesiology, Wake Forest School of
Medicine, Winston-Salem, NC, United States of America
To evaluate the cost-effectiveness of a quality improvement intervention aimed at reducing
maternal and fetal mortality in Accra, Ghana.
Quasi-experimental, time-sequence intervention, retrospective cost-effectiveness analysis.
Data were collected on the cost and outcomes of a 5-year Kybele-Ghana Health Service
Quality Improvement (QI) intervention conducted at Ridge Regional Hospital, a tertiary
referral center in Accra, Ghana, focused on systems, personnel, and communication. Mater-
nal deaths prevented were estimated comparing observed rates with counterfactual projec-
tions of maternal mortality and case-fatality rates for hypertensive disorders of pregnancy
and obstetric hemorrhage. Stillbirths prevented were estimated based on counterfactual
estimates of stillbirth rates. Cost-effectiveness was then calculated using estimated disabil-
ity-adjusted life years averted and subjected to Monte Carlo and one-way sensitivity analy-
ses to test the importance of assumptions inherent in the calculations.
Main outcome measure
Incremental Cost-effectiveness ratio (ICER), which represents the cost per disability-
adjusted life-year (DALY) averted by the intervention compared to a model counterfactual.
PLOS ONE | July 14, 2017 1 / 19
Citation: Goodman DM, Ramaswamy R, Jeuland
M, Srofenyoh EK, Engmann CM, Olufolabi AJ, et al.
(2017) The cost effectiveness of a quality
improvement program to reduce maternal and fetal
mortality in a regional referral hospital in Accra,
Ghana. PLoS ONE 12(7): e0180929. https://doi.
Editor: Alash’le G. Abimiku, University of Maryland
School of Medicine, UNITED STATES
Received: June 20, 2016
Accepted: June 25, 2017
Published: July 14, 2017
Copyright: ©2017 Goodman et al. This is an open
access article distributed under the terms of the
Creative Commons Attribution License, which
permits unrestricted use, distribution, and
reproduction in any medium, provided the original
author and source are credited.
Data Availability Statement: Data are available
from Figshare (
Funding: Funding in support of the project was
provided by the International Association for the
Study of Pain, the Lacy Foundation, and the
Obstetric Anaesthetists’ Association.
Competing interests: A.J.O. and M.D.O. report
grants from the Making Every Baby Count Initiative
(September 2013). E.K.S and D.M.G. have no
From 2007±2011, 39,234 deliveries were affected by the QI intervention implemented at
Ridge Regional Hospital. The total budget for the program was $2,363,100. Based on pro-
gram estimates, 236 (±5) maternal deaths and 129 (±13) intrapartum stillbirths were averted
(14,876 DALYs), implying an ICER of $158 ($129-$195) USD. This value is well below the
highly cost-effective threshold of $1268 USD. Sensitivity analysis considered DALY calcula-
tion methods, and yearly prevalence of risk factors andcase fatality rates. In each of these
analyses, the program remained highly cost-effective withan ICER ranging from $97-$218
QI interventions to reduce maternal and fetal mortality in low resource settings can be highly
cost effective. Cost-effectiveness analysis is feasible and should regularly be conducted to
encourage fiscal responsibility in the pursuit of improved maternal and child health.
Despite recent and historic global decreases in maternal and child mortality during the Millen-
nium Development Goals (MDGs) era, this global reduction masks immense regional varia-
tion, and the prevention of maternal death continues to be a high-priority for national
governments and the public health community. In 2015, over 300,000 maternal deaths
occurred worldwide; 201,000 (66%) of these occurred in sub-Saharan Africa (SSA).[1] Within
SSA, maternal mortality ratios remain fifty times higher than in high-income countries.[1]
Meanwhile, stillbirths remain a silent and devastating problem. Each year, 2.6 million still-
births occur; 1.2 million of these lives are lost intrapartum. In SSA, stillbirth rates are 8 times
those of high-income settings; given current trends, it will take 160 years for stillbirth rates in
SSA to fall to the levels found in high-income countries.[2]
Providing maternal and fetal health (MFH) in low- and middle-income countries (LMICs)
requires a coordinated effort of antenatal care, delivery of skilled services, and emergency
obstetric care, and depends on overcoming demand-side and supply-side barriers to quality
care. In the context of MFH, demand-side barriers include common obstacles to accessing
care such as finanicial limitations, transportation challenges, and insufficient patient educa-
tion.[3] A thorough discussion on efforts to improve demand-side financing is available from
the WHO.[4] Recent successful examples include the expansive incentive-based Janani Surak-
sha Yojana program in India[5], a voucher-scheme in Uganda[6], and maternal insurance in
Nigeria[7]. These programs report incremental cost-effectiveness ratios (ICER) of $46-$302
per disability-adjusted life-year (DALY) averted and have been deemed highly cost effective.
As governments and non-governmental organizations (NGOs) work to address demand-
side barriers, more deliveries will be shifted towards institutions that will need to prepare for
the increased number of deliveries. Supply-side interventions will need to address the care
delivered by the health system and include the actions of providers and the environment in
which care is given. For example, it is known that 15% of pregnancies require advanced emer-
gency treatment for complications that are often difficult to predict or prevent.[8,9] Unfortu-
nately, many low-resource comprehensive emergency obstetric care CEmOC facilities cannot
deliver high quality emergency care and are not prepared for the shift towards greater institu-
tional delivery.[10±12] Barriers to its provision include a lack of leadership and patient
Cost-effectiveness of a quality improvement program in West Africa
PLOS ONE | July 14, 2017 2 / 19
conflicts of interest. Kybele, Inc. is the sub-
recipient of a grant awarded to PATH (Program for
Appropriate Technology in Health). M.D.O. is the
President and Founder of Kybele, Inc.
centeredness, absence of resources, and poor operational systems.[10±14] These deficiencies
in turn lead to inadequate management of complex obstetric cases and untimely death, and
underscore the need for a comprehensive approach that would strengthen the capacity of
high-risk referral centers.
Improving CEmOC requires careful analysis and trade-offs, especially in resource-limited
countries such as Ghana where maternal mortality and stillbirth rates remain high, and public
health resources limited. Yet there is a dearth of empirical evidence regarding the impact and
cost-effectiveness of CEmOC capacity-building efforts. Some argue that CEmOC interventions
are more costly than other measures to reduce maternal and neonatal death, and require
expertise that is mostly unavailable in low income countries.[15,16] Others report a lack of evi-
dence on the disease burden, cost, and effectiveness of intervention packages aimed at referral-
level facilities.[17] Responding to such arguments, the WHO Executive Board has recognized
the need to enhance international cooperation, institutional and operational capacity, and
infrastructure for public health.[18]
Indeed, there are few if any retrospective cost-effectiveness analyses of quality improvement
(QI) programs aimed at improving CEmOC in low-resource settings, although some studies pro-
vide insights about other CEmOC interventions. Broughton and co-authors describe a successful
partnership with the Ministry of Health in Niger that reduced mortality by focusing on active
management of the third stage of labour and immediate essential newborn care. They report an
ICER of $291 in 2015 USD.[19] In 2003, McCord et al. reported on an intervention using local
surgeons to train local general physicians to perform life-saving surgeries including cesarean sec-
tions in a small hospital in Bangladesh at a cost of $57 per DALY [20]. Also in 2003, a summary
of the Save the Mothers Initiative, a Uganda-Canada collaboration, described the costs of an
intervention to increase the availability and utilization of CEmOC services, without commenting
on its effects.[21] In Mozambique, a partnership between the University of Michigan and the
Instituto Superior de Ciencias de Saude showed that training surgical technicians was cost-effec-
tive at $39 per major surgery compared with $144 for surgeons and obstetricians.[22]
It is more common in the MFH literature to use prospective modelling to estimate the cost-
effectiveness of potential interventions. Prospective studies have been important for policy
advocacy related to MFH interventions, as they routinely show that such investment is benefi-
cial for society. Alkire et al. estimated that a program to reduce obstetric fistula by providing
access to cesarean delivery would have an ICER of $277 in Ghana.[23] Adam et al. analyzed
effectiveness data from several sources and calculated an ICER of $270 (2015 USD) for inter-
vention packages that encompassed CEmOC in SSA.[17] Erim et al. modeled the cost-effec-
tiveness of MFH interventions in Nigeria. They emphasized the need to couple cost-effective
family planning access and contraception if maternal mortality targets are to be reached, given
that ªthere is a threshold above which further reductions in mortality. . .are not possibleº with-
out high-quality CEmOC. [24] Helping tertiary referral obstetric hospitals in low-resource
countries to provide high-quality care may be feasible if QI methodologies used more fre-
quently in high-resource hospitals are also introduced using coaching and consistent monitor-
ing. Unfortunately, to the best of our knowledge, there are no published analyses that assess
the question of whether such QI interventions are cost effective.
This study provides evidence that helps to address this knowledge gap, focusing on a five-
year intervention that aimed to reduce maternal and fetal mortality in Accra, Ghana. Kybele,
an international NGO that promotes safe childbirth through innovative partnerships, worked
with the Ghana Health Service (GHS) to improve CEmOC at a large urban referral hospital in
Accra. The intervention comprised education, leadership development, systems strengthening,
and QI.[25,26] As discussed elsewhere, the program proved highly effective at reducing mater-
nal and fetal mortality rates.[14,27] Still, it is important to recognize that not every change is
Cost-effectiveness of a quality improvement program in West Africa
PLOS ONE | July 14, 2017 3 / 19
an improvement[28]), and not every improvement is cost-effective. In order to be accountable
to the GHS and other donors, we therefore present a CEA of the intervention which should
help direct future efforts to build on this program, as well as the work of other similar
This study considers the cost effectiveness of a five-year collaboration between Kybele, Inc.
[29] and the GHS at the Ridge Regional Hospital (RRH) in Accra, Ghana (January 2007-
December 2011). RRH is the highest-volume obstetric referral center within the GHS with a
90-bed maternity unit providing CEmOC. Approximately 70% of deliveries at RRH are refer-
rals from other hospitals and polyclinics located in and around Accra. In 2004, RRH experi-
enced 2,000 deliveries per year that rose to 4,793 deliveries in 2006 (institutional reports, not
published). By the end of 201 when the intervention ended, RRH managed 9,357 deliveries.
This 95% increase in patient volume over the course of the collaboration was a result of de-
mand-side incentives in Ghana, changes in the referral patterns in Accra, and recognition of
improving quality by the community.[25,27]
Based on previous work of Kybele, the GHS invited the organization to lead a QI program
and build up provider capacity for leading other similar interventions in the future. The overall
aim of the partnership was to reduce the maternal mortality ratio (MMR, defined as maternal
deaths per 100,000 live births) and stillbirth rates (SBR, defined as intrapartum stillbirths per
1,000 live births) by 50% and to establish an ªObstetric Center of Excellenceº at RRH[25].
Resources were not available to build a new facility or to significantly increase staff in the
midst of the changes in demand for services at RRH. Instead, volunteers from high-volume
obstetric departments in the United States and United Kingdom made triannual visits to
Ghana to coach and mentor providers and administrators at RRH to optimize the protocols
and processes through systematic QI. The RRH-Kybele team jointly identified deficiencies in
delivery of care, and developed solutions guided by a strategic plan. The plan specified 97
improvement activities categorized into three bundles based on personnel, systems manage-
ment, and quality communication.[27] The hypothesis was that long-term investment would
lead to a strengthened system that would display stepwise improvement leading to a reduction
in maternal and fetal mortality (Fig 1). Institutional review board approval was granted by
Wake Forest University and the GHS for the conduct of the work.
A complete analysis of the changes over time has been published elsewhere along with
more descriptive supplemental material.[27] The Kybele-GHS partnership is unique in that it
did not focus primarily on providing materials commonly associated with healthcare costs,
such as medications and supplies. The partnership instead worked primarily to identify gaps
and create processes that enabled the staff to use their resources more effectively. Of the 97
processes that were designed, the mean implementation was rate 68%, and 26 processes were
found to be positively correlated with mortality reduction.[15] Activities included developing
leadership capabilities, facilitating communication, and improving timeliness of care. For
example, morning meetings between doctors, house officers, and midwives were organized to
allow daily discussion of challenges and successes.[14] Redesigning the workflow to create a
triage process similarly did not require hiring new staff, but rather facilitated more efficient
labor allocation, and improved safety for women.
Description of program costs
Program costs were prospectively collected and tracked the multiple sources of financing of
the intervention. All dollar values presented in this study were adjusted for inflation and
Cost-effectiveness of a quality improvement program in West Africa
PLOS ONE | July 14, 2017 4 / 19
standardized to 2015 USD. Kybele as an organization covered its own administration, logistics,
and educational materials. Members of Kybele±practicing professionals from North America
and England with backgrounds leading high-quality teams in obstetrics, midwifery, anesthesiol-
ogy, neonatology, and public health±provided voluntary (uncompensated) services. The dura-
tion of each trip was five to 12 days. Individual participants provided their own funding for
airfare, visas, and other expenses associated with travel to Ghana; these costs were included in
our cost analysis. The main intervention trips were triannual, although occasional shorter trips
were made for special meetings with the GHS. The GHS provided in-country accommodation
including housing, meals, and transportation for volunteers and arranged medical privileges.
A detailed project budget was kept throughout the course of the intervention and is shown
in Table 1, organized by organization. Eighty-six professionals traveled to Ghana 176 times; 34
Fig 1. Analytical model for Kybele-GHS institutional framework of change model. GHS, Ghana Health Service; HDoP,
hypertensive disorders of pregnancy; OH, obstetric hemorrhage; CFR, case fatality rate; QI, quality improvement; Dollar values in
$2015 USD.
Cost-effectiveness of a quality improvement program in West Africa
PLOS ONE | July 14, 2017 5 / 19
of whom (40%) returned multiple times. The value of volunteered professional time was
included although it was not money actually spent. This represents the opportunity cost of the
professionals offering their services within their home institutions. Values were determined
based on the U.S. Department of Labor, National Occupational Employment and Wage Esti-
mates, Role and Occupation Code.[30] Team members' time contributions were assigned as
follows based on 10-hour work days: physicians-$1100/day, nurse anesthetist $750/day, mid-
wife/engineer/nurse practitioner/consultants $500/day, nurse/resident/biostatistician $330/
day. Assigning value to volunteer time is challenging, and therefore was subjected to 25% vari-
ation in sensitivity analyses. Twelve Ghanaians were also sponsored to visit the US or England
for short-term hospital observations with costs primarily incurred by Kybele. Triannual mor-
tality conferences were organized during the program that improved communication between
GHS, RRH, and referring hospitals. The value of Ghanaian attendees' time at meetings and for
other meeting costs totaled $91,670.
Capital was invested by the GHS including equipment purchases and minor renovations
amounting to $120,910. Half of this investment was for beds that had depreciated by the end
of the intervention. The other investments included painting of walls, hanging of privacy cur-
tains, and repairing of floors in the labour ward. In 2008, 2009, and 2011, Kybele provided
refurbished equipment to the hospital, including two ultrasound machines, fetal monitors,
blood pressure monitors, anesthesia and theatre equipment, along with a variety of smaller
medical devices and supplies amounting to $452,320. These donations were either depreciated
at the time of donation or consumables that were not subject to annualization. Midwives were
posted to RRH during this time, but the hires were proportional to the delivery volume and
did not make up for the chronically low staff ratio of 6.6±7.1 midwives per 1,000 deliveries.
As seen in Table 1, the value of professional time accounted for 51% of the total budget.
Individual participants contributed 14% of the budget, Kybele provided 22%, and the GHS
covered 12%. The total cost of the program was $2,363,100. This analysis does not consider the
costs of delivering care such as changes in medication usage and inpatient length of stay, but
rather focuses on the cost of delivering the program. There was no budget for increasing deliv-
ery of services, so the QI project was structured to improve care in a cost-neutral way, mainly
by reducing delays and improving communication. Given that most of the funding came from
international sources, purchasing-power parity adjustments were not made in order to be able
to assess the cost-effectiveness relative to standard threshholds discussed in the literature.
Estimating disability-adjusted life years (DALY) for maternal and fetal
death in Ghana
The DALY is the most commonly used metric for quantifying the burden of disease in a given
population in low- and middle-income countries.[31±34] The DALY indicates the number of
years of healthy life lost due to death or disability. Disability-adjusted life years are the sum of
the present value of future years of life lost through premature mortality (YLL) and disability
(YLD), which accounts for the relative severity of mental or physical impairments that stem
from the disease using a disability weight.
The Global Burden of Disease (GBD) project provides important guidance on the appropriate
methods for computing DALYs, but several issues related to cost effectiveness analysis using
DALYs remain controversial in the literature. In particular, the most recent editions of the
GBD have not discounted future DALYs [20,22], thereby removing assumptions that individ-
uals and societies value years of life lost in the future at a reduced rate compared to current
Cost-effectiveness of a quality improvement program in West Africa
PLOS ONE | July 14, 2017 6 / 19
years of life lost. They also have not applied age weighting, which considers the value of years
of life lost to vary with age. Methodologies used by the WHO have historically used both of
these practices, however, on the basis of arguments made by many health economists.[34,35]
This analysis takes that controversy into account and presents the results of both methods. The
methods used for discounting DALYs have been well-discussed by other authors.[32] In this
study we used standard values for age weighting and discounting as seen in Table 2.
The average age of maternal death at RRH was determined to calculate DALYs. Maternal
deaths were grouped in 5-year increments from <15±49 years of age, and the proportion of
deaths from each age group was calculated based on the age for each maternal death observed
during the intervention for 2007±2011. It was assumed that reductions in maternal deaths at
RRH were proportionate for each age group. The years of life lost due to premature death were
calculated using Ghana-specific life expectancy for each age group based on values that were
interpolated from 2000 and 2012 standard life tables for Ghana.[36]
YLD is based upon previously described estimates that for every maternal death, 5±20
women are disabled due to complications.[23,37±41] This ratio is represented as the disability
multiplier, ªMº. Maternal injury is a significant contributor to the global burden of obstetric dis-
ease that is not represented by the MMR.[42] It was assumed that the ratio of maternal deaths
to maternal injuries was the same before and after the intervention and distributed proportion-
ally across age groups. Because this study takes place at a referral center, it was determined that
many of the morbidities women develop during home deliveries should not be considered. We
hypothesized that proper management of obstetric hemorrhage (OH) and hypertensive disor-
ders of pregnancy (HDoP) would help prevent morbidities such as loss of fertility due to cesar-
ean hysterectomy and stroke due to uncontrolled hypertension and eclamptic seizures. The
GBD uses a disability weight (D) of 0.3,[43] for these conditions that are assumed to last for life.
Table 1. Kybele-GHS partnership budget.
Kybele Participant GHS Time Value Total Cost
Jan-07 $4,120 $29,310 $8,240 $135,910 $177,600
Mar-07 $0 $2,290 $0 $6,300 $8,600
Jun-07 $340 $18,430 $2,180 $28,970 $49,900
Jan-08 $880 $21,710 $7,930 $103,150 $133,700
Mar-08 $3,530 $8,380 $1,210 $7,270 $20,400
May-08 $330 $7,270 $2,200 $11,680 $21,500
Sep-08 $6,940 $16,750 $1,980 $53,120 $78,800
Jan-09 $3,100 $22,890 $2,770 $97,770 $126,500
May-09 $770 $14,600 $1,660 $36,500 $53,500
Sep-09 $6,300 $28,420 $4,090 $67,240 $106,100
Jan-10 $4,460 $34,920 $8,700 $138,390 $186,500
May-10 $3,370 $27,090 $2,830 $75,720 $109,000
Sep-10 $4,130 $19,800 $4,460 $79,420 $107,800
Jan-11 $2,430 $29,750 $6,960 $153,710 $192,900
May-11 $9,810 $26,060 $6,120 $65,300 $107,300
Aug-11 $3,060 $0 $0 $5,800 $8,900
Sep-11 $9,920 $31,230 $9,500 $143,160 $193,800
US/UK Visits $13,740 $1,650 $0 $0 $15,400
Conference/Meals $0 $0 $91,670 $0 $91,700
CAPITAL $452,320 $0 $120,900 $0 $573,200
TOTAL $527,600 $340,600 $283,400 $1,209,400 $2,363,100
Percentage 22% 14% 12% 51% 100%
Cost-effectiveness of a quality improvement program in West Africa
PLOS ONE | July 14, 2017 7 / 19
This was conservatively estimated to occur with an M of 2, which indicates that for every life
saved through CEmOC, improved care prevented two near-misses as well.
There is discussion about whether or not to include stillbirth in DALY calculations.[44±46]
This study does not intend to resolve ethical debate about assigning value to life lost in utero,
but recognizes the immense cost to families when a fetus dies during delivery. The WHO, in
the 2013 Global Health Estimates, recommends that stillbirths be considered as years of life
lost and bases the value on standard life tables for life expectancy at birth.[31] For obvious rea-
sons, there is no YLD component to stillbirth. The only stillbirths considered were fresh still-
births that occurred as a result of intrapartum complications. In this study, stillbirths were
thus included in the base analysis and the sensitivity analysis.
Estimating the number of deaths and DALYs averted through the
This study compares the number of maternal deaths avoided due to the partnership interven-
tion to a ªno-interventionº counterfactual that was not actually observed, but rather estimated
from the quasi-experimental pre- and post-intervention evaluation of the program. Typically,
NGOs would use mortality rates as the standard measure for improvement over time. As men-
tioned previously, RRH experienced a period of significant growth just before and during the
intervention when the volume and acuity of patients changed significantly. A full discussion of
the analysis of these changes has been published previously[27], but we will summarize the
salient points of the analysis here. To determine what likely would have occurred if the pro-
gram had not taken place, we considered two scenarios. The first is one in which the maternal
mortality ratio (MMR) was assumed to remain steady. Using this method, the baseline 2007
MMR was used to predict the number of maternal deaths that would have occurred had the
intervention not been present. Any reduction in the number of maternal deaths is assumed to
be an improvement over this counterfactual baseline, but attribution of causality is difficult,
given that there are demand- and supply-side factors that contribute to MMR.
Table 2. Sensitivity analysis parameters.
Parameter Low Mid High Distribution of variation
Average Age 26.9 29.9 32.9 Uniform
Remaining Life Expectancy 39.1 43.4 47.7 Uniform
Stillbirth Life Expectancy 55.1 61.2 67.3 Uniform
K (age weighting) Ð 1 Ð Used in discounting formula
C (constant) Ð 0.1658 Ð
r (discounting time) Ð 0.03 Ð
B (age weight function) Ð 0.04 Ð
Disability weight (discounting) 0.2 0.3 0.4 Uniform
Disability multiplier (M) 1 2 3 Uniform
Professional time $907,050 $1,209,400 $1,511,750 Uniform
2008±2011 HDoP Prevalence -1.96*SE*Observed yearly value +1.96*SE Normal
2008±2011 OH Prevalence -1.96*SE Observed yearly value +1.96*SE Normal
2008±2011 HDoP CFR -1.96*SE Observed yearly value +1.96*SE Normal
2008±2011 OH CFR -1.96*SE Observed yearly value +1.96*SE Normal
2008±2011 Maternal Deaths -1.96*SE Observed yearly value +1.96*SE Normal
2008-2011Stillbirths -1.96*SE Observed yearly value +1.96*SE Normal
SE: Standard error for sample
Cost-effectiveness of a quality improvement program in West Africa
PLOS ONE | July 14, 2017 8 / 19
The second approach is to treat the case-fatality rates of different maternal complications as
being in steady state, and to consider the difference between predicted deaths and observed
deaths as the measure of improvement. Cases and fatalities are directly observable and
improvements in the management of high-risk cases can be attributed to the performance of
the program. The program collected data on the prevalence (prev) and CFR for the two most
common causes of maternal death at RRH, OH and HDoP, which accounted for 59% of the
maternal deaths during the intervention. The hypothesis is that the QI intervention led to
improvements in these two causes of mortality in particular through improved preparation,
communication, and protocol adherence. Maternal deaths caused by other etiologies were
considered stable in this analysis. Applying the steady-state assumption for the CFR, we com-
pared observed outcomes to predicted fatalities based on CFRs for these two common condi-
tions in 2007, which can be considered the reference year.
condition ¼Number of deaths linked to condition in a year
Number of patients presenting with condition in that year
Deaths Averted
HDoP ðCFR2007
HDoPÞ  Deathsn
HDoP þPrevN
OH  ðCFR2007
OH ÞDeathsn
OH Deaths n
Following calculation of the deaths avoided, the DALYs avoided were calculated based on the
same set of assumptions described above, for determining the DALY burden of maternal
deaths and intrapartum stillbirth death.
Finally, we are unable to calculate a CFR for stillbirths, because we do not have data on the
prevalence of fetal distress or birth asphyxia. Therefore, for stillbirths, we first assumed that
the counterfactual stillbirth rate (SBR) would remain steady at 9/1,000 life births as seen in
2007. The avoided stillbirths were then obtained by subtracting the number of intrapartum
stillbirths observed from this counterfactual number.
Assessing the cost-effectiveness of treatment
The relative cost effectiveness of the QI program was determined using the ICER, which shows
the program cost effectiveness as measured in estimated attributable DALYs averted due to the
ICER ¼CostQI CostNull
The WHO-CHOICE project used the work of the Commission of Macroeconomics and
Health[47,48] to argue that interventions having an ICER less than the country-specific GDP
per capita can be considered highly cost effective. If the ICER is less than three times the GDP
per capita, the intervention is deemed cost effective. These thresholds, while somewhat arbi-
trary and atheoretic, have been used by other studies in maternal health, and more generally in
a range of assessments of the cost effectiveness of health interventions.[17,40,49] The GDP per
capita in Ghana averaged over 2007±2011 was $1268; this represents the benchmark for a
highly-cost effective intervention in this study. [50]
Sensitivity analysis
Assumptions made in the analysis were subjected to sensitivity analysis using Monte Carlo
simulations run in Crystal Ball (Oracle, Redwood Shores, CA), which is an add-in program to
Cost-effectiveness of a quality improvement program in West Africa
PLOS ONE | July 14, 2017 9 / 19
Microsoft Excel (Microsoft, Redmond, WA). The assumptions made for calculating DALYs
using discounting and standard formulas were varied as uniform distributions around high
and low estimates as shown in Table 2. The performance of the program and changes in the
acuity of disease were varied around a normal distribution for each parameter. Using 10,000
trial simulations, 95% confidence intervals for maternal and fetal deaths prevented, and cost
effectiveness outcomes, were obtained.
Estimating DALYs averted through the Kybele-GHS partnership
Baseline delivery and outcome data are presented in Table 3. During the intervention period,
MMR decreased by 22% from 496 to 385 maternal deaths per 100,000 live births. Based on a
steady-state assumption for the MMR, 43 maternal deaths were averted with an observed
annual rate of mortality reduction ranging from -6 to -24% over five years. Because significant
changes in volume, acuity, and CFR occurred over the course of the partnership, however, the
estimated number of deaths prevented using a steady-state MMR may be underestimated (Fig
2). Applying a steady-state CFR instead, we predict that 245 maternal deaths were averted.[15]
Ghana as a country experienced an annualized rate of change (ARC) of -3.6% over the course
of the intervention.[51] Accounting for this trend in improvements occurring in the country
would thus explain only 9 of these maternal deaths averted. The intrapartum stillbirth rate
decreased 52%, and an estimated 129 stillbirths were prevented (Fig 3).
Years of life lost (YLL) per maternal death were calculated as described above, with and
without discounting and age-weighting. The age-specific life expectancy for a woman dying in
pregnancy ranged from 54 to 29 depending on her age when she died. Using the proportions
of deaths occurring in each age group, the total YLLs for an average maternal death were
found to be 28.6 (43.4 undiscounted). Estimating the years of life spent disabled requires esti-
mation of an average disability weight due to disability, as well as the duration of that disability.
Given these assumptions, we obtained an additional 17.18 YLD (26.0 undiscounted) averted
Table 3. Baseline data for the Kybele-GHS partnership.
Parameter 2007 2008 2009 2010 2011
Total Deliveries 6049 7465 8230 8133 9357
Observed MMR (per 100,000 live births) 496 388 328 369 385
Number of Obstetric Hemorrhage cases (Prevalence, %) 54 (0.9) 97 (1.3) 321 (3.9) 342 (4.2) 487 (5.2)
Number of Hypertensive disorders of pregnancy cases (Prevalence, %) 321 (5.3) 582 (7.8) 996 (12.1) 1033 (12.7) 1357 (14.5)
OH CFR (%) 14.8 5.1 1.9 2.0 1.6
HDoP CFR (%) 3.1 1.3 1.1 1.1 1.1
Observed number of maternal deaths 30 29 27 30 36
Due to OH 8 5 6 7 8
Due to HDoP 10 8 11 11 15
Due to Other causes 12 16 10 12 13
Estimated number of maternal deaths based on steady-state MMR assumptions 30 37 41 40 46
Maternal deaths prevented due to steady-state MMR assumptions 0 8 14 10 10
Estimated number of maternal deaths based on steady-state CFR assumptions 30 47 93 97 130
Maternal deaths prevented due to steady-state CFR assumptions 0 18 66 67 94
Observed number of intrapartum stillbirths 55 40 48 45 40
Observed SBR (per 1,000 live births) 9.0 5.4 5.8 5.5 4.3
Estimated number of stillbirths based on steady-state SBR assumptions 55 68 75 74 85
Stillbirths prevented based on steady-state SBR assumptions 0 28 27 39 45
Cost-effectiveness of a quality improvement program in West Africa
PLOS ONE | July 14, 2017 10 / 19
for every maternal death prevented. For stillbirths, the calculated YLL was 31.4 (61.2 undis-
counted) years.
In total, therefore, the DALY per maternal death was 45.8 (69.4 undiscounted). Base calcu-
lations show that 236 maternal deaths were averted by the program leading to 10,838 DALYs
avoided. If the 129 prevented stillbirths are also included, an additional 4,038 DALYs were
Cost-effectiveness analysis
The cost effectiveness of the Kybele-GHS partnership can be compared to the GDP thresholds
for cost-effectiveness. Including all 14,876 DALYs averted, the ICER was found to be $158
($129-$195). This is 8 times lower than the 2007±2011 Ghanaian GDP per capita, and indicates
that the intervention was highly cost-effective. Without taking stillbirth into account, the pro-
gram was still very cost effective with an ICER of $218 ($179±282) for 10,838 maternal DALYs
averted. With the more conservative steady-state MMR assumption, and not considering still-
births, yields an ICER of $1212, which remains below the cost-effectiveness threshold.
Fig 2. Maternal deaths predicted to have occurred by assumption. MMR, maternal mortality ratio (maternal deaths per 100,000live births); CFR, case-
fatality rate.
Cost-effectiveness of a quality improvement program in West Africa
PLOS ONE | July 14, 2017 11 / 19
Sensitivity analysis: Considering the changing environment
This study is unique in that it reports on the results from a completed intervention, but several
assumptions needed to be made, notably regarding the unobserved counterfactual. Table 2
describes the model parameters and describes how uncertainty was handled for each of these.
We varied performance as a normal distribution centered about an observed mean to account
for uncertainty. We used Monte Carlo simulations to take random draws from the specified
parameter distributions in order to derive the credibility estimates presented in Table 4. The
Fig 3. Stillbirths predicted to have occurred by assumption. SBR, stillbirth rate (intrapartum stillbirth per 1000 live births).
Table 4. Sensitivity analysis*.
Parameter DALYs 95% CI CER 95% CI
Discounted maternal deaths and disability 10,838 8,679±13,561 $218 $170-$282
Discounted maternal death, disability and stillbirths 14,876 12,588±17,637 $158 $129-$195
Undiscounted calculations maternal deaths and disability 16,422 12,910±20,776 $144 $111-$189
Undiscounted maternal deaths, disability and stillbirths 24,300 20,477±28,844 $97 $79-$120
*Based on model estimates of 236 (±5) maternal deaths and 129 (±13) stillbirths prevented
Cost-effectiveness of a quality improvement program in West Africa
PLOS ONE | July 14, 2017 12 / 19
tornado chart in Fig 4 further displays the 5
and 95
percentile estimates for each assumption
and how that variation individually affects the estimate of the ICER. The most significant assump-
tion is the disability multiplier, i.e. the assumption that for every maternal death prevented, 1 to 3
women would also not suffer disabilities. This assumption accounts for 39% of the variation in
the ICER estimate. The next most significant assumption was the value of professional time that
was estimated as part of the budget. Centered around a base estimate of $1.21 million, changes by
± $0.27 million alter the ICER by $18/DALY avoided, which corresponds to 22% of the variation
observed. The disability weight assumption ranging from 0.2±0.4 contributes an additional 12%
of the variation, but is based on the values used in the Global Burden of Disease[43]. Finally, vari-
ability in the average age of maternal death contributes 3% of the variation.
Main findings
We present the cost-effectiveness of a systems-based QI intervention using prospectively col-
lected financial records and outcome data. This study serves to provide accountability to
Fig 4. Tornado Chart of ICER for discounted calculations of maternal deaths, disability, and stillbirths showing all variables contributing 1% of
the variation in the estimates. ICER, incremental cost-effectiveness ratio; Upside, this assumption makes the program more cost effective; downside, this
assumption makes the program less cost effective.
Cost-effectiveness of a quality improvement program in West Africa
PLOS ONE | July 14, 2017 13 / 19
Kybele for their actions, and to direct inform the planning of other similar interventions. Our
analysis shows that with $2.4 million invested, including $1.2 million worth of dontated pro-
fessional time, the Kybele-GHS partnership was able to prevent between 43 and 236 maternal
deaths and 129 intrapartum stillbirths. This amounts to 14,876 discounted DALYs (24,300
undiscounted DALYs), and leads us to conclude that the intervention was highly cost-effective,
with an ICER of $158 ($129-$195) 2015 USD. Sensitivity analysis strengthens this conclusion;
variation in key assumptions resulted in ICERs ranging from $79 to $282.
It is not typically feasible to systematically control interventions such as this, which makes
attribution of causality difficult. It is likely that the ªno interventionº counterfactual would
have involved increasing mortality rates as a consequence of the increasing prevalance in OH
and HDoP, and that the increase would have been on the order of 236 maternal deaths. Such a
counterfactual would have led to an MMR similar to those observed in other referral hospitals
in Ghana, which ranged from 913±1004 maternal deaths per 100,000 live births.[52,53] Our
study shows that RRH was able to achieve significantly lower CFRs than these other hospitals
over the intervention period.
The GHS and RRH continued to provide care through the national health insurance
scheme and patient fees. It is likely that these charges would have been present without the
partnership as the caseload at the hospital increased and in accordance with the treatments
administered. The QI intervention was intended to promote higher quality and more efficient
care, which could translate into lower costs. For example, the provision of more rapid treat-
ment for conditions such as OH or HDoP, may help mitigate expensive treatments such as
blood transfusions and intensive care admissions. We cannot, however, assess how the costs of
clinical care changed at RRH during the course of the intervention.
Cost-effectiveness is not routinely reported by organizations working to improve MFH,
so our analysis addresses an important knowledge gap. To the extent possible in this retro-
spective analysis, the CHEERS guidelines[54] for reporting cost-effectiveness were followed.
This is a single study-based estimate that therefore serves as a valuable source for understand-
ing the potential cost of long-term coaching, leadership development, and quality improve-
ment measures that others might attempt in similar facilities. Nonetheless, the incremental
effectiveness of such interventions may vary considerably as a function of the heterogeneity of
health systems and demand-side influences, and should be the subject of additional future
One important issue that this analysis ignores is that Ridge Hospital serves as a referral hos-
pital, and since the Kybele±GHS partnership was initiated, outreach to referral centers, educa-
tional modules and strengthening of protocols has also occurred. The benefits of this outreach,
which may be significant, were not included here. Further analysis at the regional levels is
therefore urgently needed to deepen our understanding of the value of those aspects of the
Strengths and limitations
This study has several advantages. It retrospectively analyses a real-world intervention
rather than relying on hypothetical constructs developed prospectively. This project was
established as a shared initiative between Kybele and the GHS. By achieving local buy-in from
the GHS, Kybele was able to create a sense of partnership that allowed prioritization and access
to decision makers. The project provided the development of leadership and QI skills to sus-
tain healthcare improvements at Ridge and at other sites within the GHS. Cost-sharing by
the government indicated commitment to the program, and increases the likelihood of its
Cost-effectiveness of a quality improvement program in West Africa
PLOS ONE | July 14, 2017 14 / 19
There are also weaknesses present that are important to acknowledge. The project was con-
ducted using volunteer time donated by foreign experts. This makes the pecuniary cost of the
program modest, and required us to estimate the value of the donated professional time. Relat-
edly, most hospitals in low-income settings will not have ready access to skilled personnel able
to support similar programs. The QI intervention was extremely time intensive and may not
be reproducible in other settings due to constraints on manpower and funding. Third, the
study was implemented in a single hospital. The estimates of impact on maternal and fetal
mortality are based on pre- and post-intervention comparisons; as such, the precise number of
deaths averted cannot be known, especially in light of the significant changes in the number of
births and disease acuity observed at the hospital over time. These changes were not antici-
pated at the project outset and may have resulted from changes in referral patterns in Accra,
insurance incentives promoting institutional delivery, and community recognition of improv-
ing care at RRH. Considering the facilities around the world in which women deliver, RRH
is a moderately well-staffed and equipped hospital. This limits external validity as the cost
effectiveness of this type of intervention would vary across hospital and clinic settings. Fi-
nally, it is difficult to estimate the effectiveness of interventions addressing maternal health
without also considering neonatal health as well, which were not carefully followed and
Assessing the long-term effectiveness of a program also requires careful consideration of
whether the changes induced can be sustained over time. It is our hope that the coaching and
leadership development model helped to change the culture at RRH in such a way that
improvements can now be led by local providers with gradually decreasing intervention from
Kybele. During the intervention, leaders and staff who displayed a high level of ability and
interest in the QI process were identified as ªclinical champions.º Since 2011, much of the
work at RRH has focused on enabling these champions to lead and sustain improvement. Fre-
quently, they have been asked by the GHS to provide trainings at other facilities. Assessing the
longer-term effects of the intervention will be the subject of future research.
As governments continue to emphasize and incentivize institutional delivery as part of their
MFH improvement packages, the need for high-quality CEmOC will continue to rise. We
report that it may be highly cost-effective to support CEmOC with QI methodologies led by
visiting consultants. Over time, the goal is to develop enough internal capacity to lead these
efforts in LMICs. We believe that NGOs can play an important role in building that capacity
by partnering with ministries of health.
The Kybele-Ghana Health Service Partnership was able to reduce maternal and fetal mortality
over the course of five years using quality improvement methodologies. Applying an approach
that predicted the impact of this intervention based on an assumption of steady-state CFRs, it
was estimated that the program prevented 236 (±5) maternal deaths and 129 (±13) stillbirths,
which amounted to 14,876 DALYs averted. The best estimate of the ICER was therefore found
to be $158, a value well below the very cost effective thresholds based on per capita GDP in
Ghana ($1,268 USD). This study supports the hypothesis that collaborative partnerships using
quality improvement methodologies can produce cost-effective outcomes. Additional similar
experiences need to be studied in other settings, but it seems likely that quality improvement
interventions that address systems issues in other resource-constrained settings and at larger
scale will often be similarly cost-effective.
Cost-effectiveness of a quality improvement program in West Africa
PLOS ONE | July 14, 2017 15 / 19
Author Contributions
Conceptualization: David M. Goodman, Emmanuel K. Srofenyoh, Cyril M. Engmann,
Adeyemi J. Olufolabi, Medge D. Owen.
Data curation: David M. Goodman, Emmanuel K. Srofenyoh, Adeyemi J. Olufolabi, Medge
D. Owen.
Formal analysis: David M. Goodman, Rohit Ramaswamy, Marc Jeuland.
Funding acquisition: Emmanuel K. Srofenyoh, Adeyemi J. Olufolabi, Medge D. Owen.
Investigation: Adeyemi J. Olufolabi, Medge D. Owen.
Methodology: David M. Goodman, Rohit Ramaswamy, Marc Jeuland, Medge D. Owen.
Project administration: David M. Goodman, Medge D. Owen.
Resources: Marc Jeuland, Medge D. Owen.
Software: David M. Goodman, Rohit Ramaswamy, Marc Jeuland.
Validation: Marc Jeuland, Cyril M. Engmann.
Visualization: David M. Goodman, Marc Jeuland.
Writing ± original draft: David M. Goodman, Rohit Ramaswamy, Marc Jeuland, Emmanuel
K. Srofenyoh, Cyril M. Engmann, Adeyemi J. Olufolabi, Medge D. Owen.
Writing ± review & editing: David M. Goodman, Rohit Ramaswamy, Marc Jeuland, Emman-
uel K. Srofenyoh, Cyril M. Engmann, Adeyemi J. Olufolabi, Medge D. Owen.
1. Alkema L, Chou D, Hogan D, Zhang S, MollerA-B, Gemmill A, et al. Global, regional, andnational levels
and trends in maternal mortality between 1990 and 2015, with scenario-based projections to 2030: a
systematic analysis by the UN Maternal Mortality Estimation Inter-Agency Group. Lancet. 2015 Nov 13;
2. Lawn JE, Blencowe H, Waiswa P, Amouzou A, Mathers C, HoganD, et al. Stillbirths: rates, risk factors,
and acceleration towards 2030. Lancet. 2016 Jan 18;
3. Lee ACC, Lawn JE, Cousens S, Kumar V, Osrin D, Bhutta ZA, et al. Linking families and facilities for
care at birth: what works to avert intrapartum-related deaths? Int J Gynaecol Obstet. 2009 Oct; 107
Suppl 1:S65±85, S86±8.
4. Gupta Indrani, Joe William, Rudra Shalini. Demand Side Financing in Health: How far can it address the
issue of low utilization in developing countries? [Internet]. World Health Organization; 2010 [cited 2016
Oct 25]. Available from:
5. Gopalan SS, Durairaj V. Addressing maternal healthcare through demand side financial incentives:
experience of Janani Suraksha Yojana programin India. BMC Health Serv Res. 2012 Sep 15; 12:319. PMID: 22978630
6. Alfonso YN, Bishai D, Bua J, Mutebi A, Mayora C, Ekirapa-Kiracho E. Cost-effectiveness analysis of a
voucher scheme combined with obstetrical quality improvements: quasi experimental results from Uganda.
Health Policy Plan. 2015 Feb; 30(1):88±99. PMID: 24371219
7. Gomez GB, Foster N, Brals D, Nelissen HE, Bolarinwa OA, HendriksME, et al. Improving Maternal
Care through a State-Wide Health Insurance Program: A Cost and Cost-Effectiveness Study in Rural
Nigeria. PLoS One. 2015; 10(9):e0139048. PMID:
8. WHO | Monitoring emergency obstetric care[Internet]. WHO. [cited 2015 Nov 9]. Available from: http://
9. Jahn A, De Brouwere V. Referral in Pregnancy Concepts and Strategies [Internet]. John Snow Interna-
tional|Europe; Available from:
10. Adegoke AA, Campbell M, OgundejiMO, Lawoyin T, Thomson AM. Place of birth or placeof death: an
evaluation of 1139 maternal deaths in Nigeria. Midwifery. 2013 Nov; 29(11):e115±21.
1016/j.midw.2012.11.018 PMID: 23415349
Cost-effectiveness of a quality improvement program in West Africa
PLOS ONE | July 14, 2017 16 / 19
11. Vesel L, Manu A, Lohela TJ, Gabrysch S, Okyere E, AsbroekAHA Ten, et al. Quality of newborn care: a
health facility assessment in rural Ghana using survey, vignette and surveillance data. BMJ Open.
2013; 3(5).
12. Nesbitt RC, Lohela TJ, Manu A, Vesel L,Okyere E, Edmond K, et al. Quality alongthe continuum: a
health facility assessment of intrapartum and postnatal care in Ghana. PLoS One. 2013; 8(11):e81089. PMID: 24312265
13. Bhutta ZA, Das JK, Bahl R, Lawn JE, Salam RA, Paul VK, et al. Can available interventions end pre-
ventable deaths in mothers, newborn babies, and stillbirths, and at what cost? Lancet. 2014 Jul 26; 384
(9940):347±70. PMID: 24853604
14. Ramaswamy R, Iracane S, Srofenyoh E, Bryce F, Floyd L, KallamB, et al. Transforming Maternal and
Neonatal Outcomes in Tertiary Hospitals in Ghana: An Integrated Approach for Systems Change. J
Obstet Gynaecol Can. 2015 Oct; 37(10):905±14. PMID: 26606708
15. Campbell OMR, Graham WJ, Lancet Maternal Survival Series steering group. Strategies for reducing
maternal mortality: getting on with what works. Lancet. 2006 Oct 7; 368(9543):1284±99.
10.1016/S0140-6736(06)69381-1 PMID: 17027735
16. Adam T, Lim SS, Mehta S, Bhutta ZA, Fogstad H, Mathai M, et al. Cost effectiveness analysis of strate-
gies for maternal and neonatal health in developing countries. BMJ. 2005 Nov 12; 331(7525):1107. PMID: 16282407
17. Graham WJ, Cairns J, Bhattacharya S, Bullough CHW, Quayyum Z, Rogo K. Maternal and Perinatal
Conditions. In: Jamison DT, Breman JG, Measham AR, Alleyne G, Claeson M, Evans DB, et al., editors.
Disease Control Priorities in Developing Countries [Internet]. 2nd ed. Washington (DC): World Bank;
2006 [cited 2015 Nov 9]. Available from:
18. WHO Executive Board. Strengthening essential public health functions in support of the achievement of
universal health coverage [Internet]. World Health Organization; [cited 2016 Oct 25]. Available from:
19. Broughton E, Saley Z, Boucar M, Alagane D, HillK, Marafa A, et al. Cost-effectiveness of a quality
improvement collaborative for obstetric and newborn care in Niger. Int J Health Care Qual Assur. 2013;
26(3):250±61. PMID: 23729128
20. McCord C, Chowdhury Q. A cost effective small hospital in Bangladesh: what it can mean for emer-
gency obstetric care. Int J Gynaecol Obstet. 2003 Apr; 81(1):83±92. PMID: 12676406
21. Lalonde AB, Okong P, Mugasa A, Perron L. The FIGO Save the Mothers Initiative: the Uganda-Canada
collaboration. Int J Gynaecol Obstet. 2003 Feb; 80(2):204±12. PMID: 12566200
22. Kruk ME, Pereira C, Vaz F, Bergstro
m S, Galea S. Economic evaluationof surgically trained assistant
medical officers in performing major obstetric surgery in Mozambique.BJOG Int J Obstet Gynaecol.
2007 Oct; 114(10):1253±60.
23. Alkire BC, Vincent JR, Burns CT, Metzler IS, Farmer PE, Meara JG. Obstructed labor and caesarean
delivery: the cost and benefit of surgical intervention. PLoS One. 2012; 7(4):e34595.
1371/journal.pone.0034595 PMID: 22558089
24. Erim DO, Resch SC, Goldie SJ. Assessing health and economicoutcomes of interventions to reduce
pregnancy-related mortality in Nigeria. BMC Public Health. 2012 Sep 14; 12:786.
1186/1471-2458-12-786 PMID: 22978519
25. Srofenyoh E, Ivester T, Engmann C, OlufolabiA, Bookman L, Owen M. Advancing obstetricand neona-
tal care in a regional hospital in Ghana via continuous quality improvement. Int J Gynaecol Obstet. 2012
Jan; 116(1):17±21. PMID: 22040863
26. Thaddeus S, Maine D. Too far to walk: maternal mortality in context. Soc Sci Med 1982.1994 Apr; 38
27. Srofenyoh EK, Kassebaum NJ, Goodman DM, Olufolabi AJ, Owen MD. Measuring the impact of a qual-
ity improvement collaboration to decrease maternal mortality in a Ghanaian regional hospital. Int J
Gynaecol Obstet. 2016 Aug; 134(2):181±5. PMID: 27177512
28. Eliezer Yudkowsky. Harry Potter and the Methods of Rationality [Internet]. Self-published; [cited 2016
Nov 18]. Available from:
29. Kybele, Inc. [Internet]. Kybele, Inc. [cited 2016 Apr 30]. Available from:
30. United States Department of Labor, Bureau of Labor Statistics. OccupationalEmployment Statistics
[Internet]. 2014. Available from:
31. Fox-Rushby JA, Hanson K. Calculating and presenting disability adjusted life years (DALYs) in cost-
effectiveness analysis. Health Policy Plan. 2001 Sep; 16(3):326±31. PMID: 11527874
32. GBD 2013 Risk Factors Collaborators, Forouzanfar MH, Alexander L, Anderson HR, Bachman VF, Bir-
yukov S, et al. Global, regional, and national comparative risk assessment of 79 behavioural,environ-
mental and occupational, and metabolic risks or clusters of risks in 188 countries, 1990±2013: a
Cost-effectiveness of a quality improvement program in West Africa
PLOS ONE | July 14, 2017 17 / 19
systematic analysis for the Global Burdenof Disease Study 2013. Lancet. 2015 Dec5; 386
(10010):2287±323. PMID: 26364544
33. WHO | Metrics: Disability-Adjusted Life Year (DALY) [Internet]. WHO. [cited 2015 Oct 23]. Available
34. Murray CJL, Ezzati M, Flaxman AD, Lim S, Lozano R, Michaud C, et al. GBD 2010: design, definitions,
and metrics. Lancet. 2012 Dec 15; 380(9859):2063±6.
PMID: 23245602
35. Susan Horton, Carol Levin. Reproductive, Maternal, Newborn, and Child Health. In: Disease Control
Priorities [Internet]. Third. [cited 2016 Oct 25]. p. 338. Available from:
36. Global Health Observatory Data Repository [Internet]. WHO. [cited 2015 Oct 23]. Available from: http://
37. Storeng KT, Baggaley RF, Ganaba R, Ouattara F, Akoum MS, Filippi V. Paying the price: the cost and
consequences of emergency obstetric care in Burkina Faso. Soc Sci Med 1982. 2008 Feb; 66(3):545±
38. Donnay F. Maternal survival in developing countries: what has been done, what can be achieved in the
next decade. Int J Gynaecol Obstet. 2000 Jul; 70(1):89±97. PMID: 10884537
39. Requejo J, Bryce J. Countdown to 2015:Decade Report (2000±2010) Takingstock of maternal, new-
born and child survival [Internet]. Geneva, Switzerland: World Health Organization; 2010 [cited 2015
Aug 2]. Available from:
40. Foster S, Bryant M. A Framework for Estimating Benefits of Investing in Maternal,Newborn, and Child
Health [Internet]. Geneva, Switzerland: Partnership for Maternal, Newborn, and Child Health; 2013
Feb [cited 2015 Aug 2]. Available from:
41. Filippi V, Ronsmans C, Campbell OMR, Graham WJ, Mills A, Borghi J, et al. Maternal health in poor
countries: the broader context and a call for action. Lancet. 2006Oct 28; 368(9546):1535±41. https:// PMID: 17071287
42. Koblinsky M, Chowdhury ME, Moran A, Ronsmans C. Maternal morbidity and disability and their conse-
quences: neglected agenda in maternal health. J Health Popul Nutr. 2012 Jun; 30(2):124±30. PMID:
43. Salomon JA, Haagsma JA, Davis A, de Noordhout CM, Polinder S, Havelaar AH, et al. Disability
weights for the Global Burden of Disease 2013 study. Lancet Glob Health. 2015 Nov; 3(11):e712±23. PMID: 26475018
44. WHO | Global Health Estimates [Internet]. WHO. [cited 2015 Nov 5]. Available from:
45. Phillips J, Millum J. Valuing Stillbirths. Bioethics. 2015 Jul; 29(6):413±23.
12120 PMID: 25395144
46. Kuznik A, Habib AG, Manabe YC, Lamorde M. Estimating the Public Health Burden Associated With
Adverse Pregnancy Outcomes Resulting From Syphilis Infection Across 43 Countries in Sub-Saharan
Africa. Sex Transm Dis. 2015 Jul; 42(7):369±75.
PMID: 26222749
47. WHO | WHO-CHOICE [Internet]. WHO. [cited 2016 May 1]. Available from:
48. Evans DB, Edejer TT-T, Adam T, Lim SS. Methods to assess the costs and health effects of interven-
tions for improving health in developing countries. BMJ. 2005 Nov 12; 331(7525):1137±40. https://doi.
org/10.1136/bmj.331.7525.1137 PMID: 16282411
49. Hu D, Bertozzi SM, Gakidou E, Sweet S, Goldie SJ. The costs, benefits, and cost-effectiveness of inter-
ventions to reduce maternal morbidity and mortality in Mexico. PLoS One. 2007; 2(8):e750. https://doi.
org/10.1371/journal.pone.0000750 PMID: 17710149
50. Gross Domestic Product per capita [Internet]. Available from:
51. Kassebaum NJ, Bertozzi-Villa A, Coggeshall MS, Shackelford KA, Steiner C, Heuton KR, et al. Global,
regional, and national levels and causes of maternal mortality during 1990±2013: a systematic analysis
for the Global Burden of Disease Study 2013. Lancet. 2014 Sep 13; 384(9947):980±1004. https://doi.
org/10.1016/S0140-6736(14)60696-6 PMID: 24797575
52. Adu-Bonsaffoh K, Oppong SA, Samuel OA, BinlinlaG, Obed SA, Samuel OA. Maternal deaths attribut-
able to hypertensive disorders in a tertiary hospital in Ghana. Int J Gynaecol Obstet. 2013Nov; 123
(2):110±3. PMID: 23969337
Cost-effectiveness of a quality improvement program in West Africa
PLOS ONE | July 14, 2017 18 / 19
53. Lee QY, Odoi AT, Opare-Addo H, Dassah ET. Maternal mortality in Ghana: a hospital-based review.
Acta Obstet Gynecol Scand. 2012 Jan; 91(1):87±92.
PMID: 21793813
54. Husereau D, Drummond M, Petrou S, Carswell C, MoherD, Greenberg D, et al. Consolidated Health
Economic Evaluation Reporting Standards (CHEERS) statement. BJOG Int J Obstet Gynaecol. 2013
May; 120(6):765±70.
Cost-effectiveness of a quality improvement program in West Africa
PLOS ONE | July 14, 2017 19 / 19
... Of these challenges infection, anemia, preterm delivery and low birth weight cannot be ruled out among pregnant women despite increase ANC uptake and skilled delivery services among the FMHCP bene ciaries (10,45,46). The consequences are that the increase in utilization do not necessarily translate to desired outcomes of reduced stillbirth rate and hence poses a constraint to the attainment of the 12 per 1000 lives target as set by WHO (8,47). ...
Full-text available
Background: The inception of the 'Free’ Maternal Health Care Policy (FMHCP) in Ghana since 2008 has seen drastic improvement in maternal healthcare utilization, particularly antenatal care uptake and skilled delivery utilization. However, its impact against late pregnancy outcomes has yet to be examined contextually. Purpose: This study aimed to describe the implications of the ‘free’ policy funding constraints on maternal healthcare utilization and explain how factors play to contribute to stillbirth despite the increase in maternal healthcare utilization. Methods: The study adopted an intrinsic-case study method using qualitative techniques to collect primary data through one-on-one interviews with service providers, and focus group discussions for pregnant women. The study then interviewed an expert in healthcare practice and policy implementation as a key informer to triangulate the data for analysis. Deductive thematic analysis guided the approach to writing the report, where sub-themes also emerged inductively and are reported with verbatim quotes to aid explanatory power. Results: The study found that within increased skilled attendance, comes top-up payments for medicines and laboratory services among pregnant women despite the policy being described as ‘free’. The study also found that routine medicines such as folic acid, ferrous sulphate and multivitamin tablets are often in short supply in the ‘free’ policy credentialed facilities and affecting service provision and quality of care. Conclusion: We conclude that the increase in maternal healthcare utilization do not necessarily translate to desired late pregnancy outcomes in its current form, particularly within the context of acute shortages of medicine commodities. The Ministry of Health and its agencies should perhaps set agendas to ensure regular supply of drug commodities in hospitals and health centers to boost quality of care.
Full-text available
The Global Burden of Disease (GBD) study assesses health losses from diseases, injuries, and risk factors using disability-adjusted life-years, which need a set of disability weights to quantify health levels associated with non-fatal outcomes. The objective of this study was to estimate disability weights for the GBD 2013 study. METHODS: We analysed data from new web-based surveys of participants aged 18-65 years, completed in four European countries (Hungary, Italy, the Netherlands, and Sweden) between Sept 23, 2013, and Nov 11, 2013, combined with data previously collected in the GBD 2010 disability weights measurement study. Surveys used paired comparison questions for which respondents considered two hypothetical individuals with different health states and specified which person they deemed healthier than the other. These surveys covered 183 health states pertinent to GBD 2013; of these states, 30 were presented with descriptions revised from previous versions and 18 were new to GBD 2013. We analysed paired comparison data using probit regression analysis and rescaled results to disability weight units between 0 (no loss of health) and 1 (loss equivalent to death). We compared results with previous estimates, and an additional analysis examined sensitivity of paired comparison responses to duration of hypothetical health states. FINDINGS: The total analysis sample consisted of 30 230 respondents from the GBD 2010 surveys and 30 660 from the new European surveys. For health states common to GBD 2010 and GBD 2013, results were highly correlated overall (Pearson's r 0·992 [95% uncertainty interval 0·989-0·994]). For health state descriptions that were revised for this study, resulting disability weights were substantially different for a subset of these weights, including those related to hearing loss (eg, complete hearing loss: GBD 2010 0·033 [0·020-0·052]; GBD 2013 0·215 [0·144-0·307]) and treated spinal cord lesions (below the neck: GBD 2010 0·047 [0·028-0·072]; GBD 2013 0·296 [0·198-0·414]; neck level: GBD 2010 0·369 [0·243-0·513]; GBD 2013 0·589 [0·415-0·748]). Survey responses to paired comparison questions were insensitive to whether the comparisons were framed in terms of temporary or chronic outcomes (Pearson's r 0·981 [0·973-0·987]). INTERPRETATION: This study substantially expands the empirical basis for assessment of non-fatal outcomes in the GBD study. Findings from this study substantiate the notion that disability weights are sensitive to particular details in descriptions of health states, but robust to duration of outcomes.
Full-text available
Background While the Nigerian government has made progress towards the Millennium Development Goals, further investments are needed to achieve the targets of post-2015 Sustainable Development Goals, including Universal Health Coverage. Economic evaluations of innovative interventions can help inform investment decisions in resource-constrained settings. We aim to assess the cost and cost-effectiveness of maternal care provided within the new Kwara State Health Insurance program (KSHI) in rural Nigeria. Methods and Findings We used a decision analytic model to simulate a cohort of pregnant women. The primary outcome is the incremental cost effectiveness ratio (ICER) of the KSHI scenario compared to the current standard of care. Intervention cost from a healthcare provider perspective included service delivery costs and above-service level costs; these were evaluated in a participating hospital and using financial records from the managing organisations, respectively. Standard of care costs from a provider perspective were derived from the literature using an ingredient approach. We generated 95% credibility intervals around the primary outcome through probabilistic sensitivity analysis (PSA) based on a Monte Carlo simulation. We conducted one-way sensitivity analyses across key model parameters and assessed the sensitivity of our results to the performance of the base case separately through a scenario analysis. Finally, we assessed the sustainability and feasibility of this program’s scale up within the State’s healthcare financing structure through a budget impact analysis. The KSHI scenario results in a health benefit to patients at a higher cost compared to the base case. The mean ICER (US$46.4/disability-adjusted life year averted) is considered very cost-effective compared to a willingness-to-pay threshold of one gross domestic product per capita (Nigeria, US$ 2012, 2,730). Our conclusion was robust to uncertainty in parameters estimates (PSA: median US$49.1, 95% credible interval 21.9–152.3), during one-way sensitivity analyses, and when cost, quality, cost and utilization parameters of the base case scenario were changed. The sustainability of this program’s scale up by the State is dependent on further investments in healthcare. Conclusions This study provides evidence that the investment made by the KSHI program in rural Nigeria is likely to have been cost-effective; however, further healthcare investments are needed for this program to be successfully expanded within Kwara State. Policy makers should consider supporting financial initiatives to reduce maternal mortality tackling both supply and demand issues in the access to care.
Full-text available
Background: The Global Burden of Disease, Injuries, and Risk Factor study 2013 (GBD 2013) is the first of a series of annual updates of the GBD. Risk factor quantification, particularly of modifiable risk factors, can help to identify emerging threats to population health and opportunities for prevention. The GBD 2013 provides a timely opportunity to update the comparative risk assessment with new data for exposure, relative risks, and evidence on the appropriate counterfactual risk distribution. Methods: Attributable deaths, years of life lost, years lived with disability, and disability-adjusted life-years (DALYs) have been estimated for 79 risks or clusters of risks using the GBD 2010 methods. Risk–outcome pairs meeting explicit evidence criteria were assessed for 188 countries for the period 1990–2013 by age and sex using three inputs: risk exposure, relative risks, and the theoretical minimum risk exposure level (TMREL). Risks are organised into a hierarchy with blocks of behavioural, environmental and occupational, and metabolic risks at the first level of the hierarchy. The next level in the hierarchy includes nine clusters of related risks and two individual risks, with more detail provided at levels 3 and 4 of the hierarchy. Compared with GBD 2010, six new risk factors have been added: handwashing practices, occupational exposure to trichloroethylene, childhood wasting, childhood stunting, unsafe sex, and low glomerular filtration rate. For most risks, data for exposure were synthesised with a Bayesian meta-regression method, DisMod-MR 2.0, or spatial-temporal Gaussian process regression. Relative risks were based on meta-regressions of published cohort and intervention studies. Attributable burden for clusters of risks and all risks combined took into account evidence on the mediation of some risks such as high body-mass index (BMI) through other risks such as high systolic blood pressure and high cholesterol. Findings: All risks combined account for 57·2% (95% uncertainty interval [UI] 55·8–58·5) of deaths and 41·6% (40·1–43·0) of DALYs. Risks quantified account for 87·9% (86·5–89·3) of cardiovascular disease DALYs, ranging to a low of 0% for neonatal disorders and neglected tropical diseases and malaria. In terms of global DALYs in 2013, six risks or clusters of risks each caused more than 5% of DALYs: dietary risks accounting for 11·3 million deaths and 241·4 million DALYs, high systolic blood pressure for 10·4 million deaths and 208·1 million DALYs, child and maternal malnutrition for 1·7 million deaths and 176·9 million DALYs, tobacco smoke for 6·1 million deaths and 143·5 million DALYs, air pollution for 5·5 million deaths and 141·5 million DALYs, and high BMI for 4·4 million deaths and 134·0 million DALYs. Risk factor patterns vary across regions and countries and with time. In sub-Saharan Africa, the leading risk factors are child and maternal malnutrition, unsafe sex, and unsafe water, sanitation, and handwashing. In women, in nearly all countries in the Americas, north Africa, and the Middle East, and in many other high-income countries, high BMI is the leading risk factor, with high systolic blood pressure as the leading risk in most of Central and Eastern Europe and south and east Asia. For men, high systolic blood pressure or tobacco use are the leading risks in nearly all high-income countries, in north Africa and the Middle East, Europe, and Asia. For men and women, unsafe sex is the leading risk in a corridor from Kenya to South Africa. Interpretation: Behavioural, environmental and occupational, and metabolic risks can explain half of global mortality and more than one-third of global DALYs providing many opportunities for prevention. Of the larger risks, the attributable burden of high BMI has increased in the past 23 years. In view of the prominence of behavioural risk factors, behavioural and social science research on interventions for these risks should be strengthened. Many prevention and primary care policy options are available now to act on key risks.
Full-text available
Estimates of the burden of disease assess the mortality and morbidity that affect a population by producing summary measures of health such as quality-adjusted life years (QALYs) and disability-adjusted life years (DALYs). These measures typically do not include stillbirths (fetal deaths occurring during the later stages of pregnancy or during labor) among the negative health outcomes they count. Priority-setting decisions that rely on these measures are therefore likely to place little value on preventing the more than three million stillbirths that occur annually worldwide. In contrast, neonatal deaths, which occur in comparable numbers, have a substantial impact on burden of disease estimates and are commonly seen as a pressing health concern. In this article we argue in favor of incorporating unintended fetal deaths that occur late in pregnancy into estimates of the burden of disease. Our argument is based on the similarity between late-term fetuses and newborn infants and the assumption that protecting newborns is important. We respond to four objections to counting stillbirths: (1) that fetuses are not yet part of the population and so their deaths should not be included in measures of population health; (2) that valuing the prevention of stillbirths will undermine women's reproductive rights; (3) that including stillbirths implies that miscarriages (fetal deaths early in pregnancy) should also be included; and (4) that birth itself is in fact ethically significant. We conclude that our proposal is ethically preferable to current practice and, if adopted, is likely to lead to improved decisions about health spending.
( Lancet 2016;387: 462–674) World leaders at the landmark Millennium Summit in September 2000 agreed to adopt the Millennium Development Goals (MDGs) with the goal of improving the lives of the world’s poor. Countries and international agencies agreed to monitor progress on these goals that addressed development and health outcomes between 1990 and 2015. One of the goals, MDG 5, strived for a reduction of 75% in the world-wide maternal mortality ratio over these 25 years. The objective of this current study was to project the levels and trends in maternal mortality for 183 countries for the period 2016 to 2030 based on the data obtained for MDG 5.
Objective: To evaluate the performance of a continuous quality improvement collaboration at Ridge Regional Hospital, Accra, Ghana, that aimed to halve maternal and neonatal deaths. Methods: In a quasi-experimental, pre- and post-intervention analysis, system deficiencies were analyzed and 97 improvement activities were implemented from January 2007 to December 2011. Data were collected on outcomes and implementation rates of improvement activities. Severity-adjustment models were used to calculate counterfactual mortality ratios. Regression analysis was used to determine the association between improvement activities, staffing, and maternal mortality. Results: Maternal mortality decreased by 22.4% between 2007 and 2011, from 496 to 385 per 100000 deliveries, despite a 50% increase in deliveries and five- and three-fold increases in the proportion of pregnancies complicated by obstetric hemorrhage and hypertensive disorders of pregnancy, respectively. Case fatality rates for obstetric hemorrhage and hypertensive disorders of pregnancy decreased from 14.8% to 1.6% and 3.1% to 1.1%, respectively. The mean implementation score was 68% for the 97 improvement processes. Overall, 43 maternal deaths were prevented by the intervention; however, risk severity-adjustment models indicated that an even greater number of deaths was averted. Mortality reduction was correlated with 26 continuous quality improvement activities, and with the number of anesthesia nurses and labor midwives. Conclusion: The implementation of quality improvement activities was closely correlated with improved maternal mortality.
An estimated 2·6 million third trimester stillbirths occurred in 2015 (uncertainty range 2·4–3·0 million). The number of stillbirths has reduced more slowly than has maternal mortality or mortality in children younger than 5 years, which were explicitly targeted in the Millennium Development Goals. The Every Newborn Action Plan has the target of 12 or fewer stillbirths per 1000 births in every country by 2030. 94 mainly high-income countries and upper middle-income countries have already met this target, although with noticeable disparities. At least 56 countries, particularly in Africa and in areas affected by conflict, will have to more than double present progress to reach this target. Most (98%) stillbirths are in low-income and middle-income countries. Improved care at birth is essential to prevent 1·3 million (uncertainty range 1·2–1·6 million) intrapartum stillbirths, end preventable maternal and neonatal deaths, and improve child development. Estimates for stillbirth causation are impeded by various classification systems, but for 18 countries with reliable data, congenital abnormalities account for a median of only 7·4% of stillbirths. Many disorders associated with stillbirths are potentially modifiable and often coexist, such as maternal infections (population attributable fraction: malaria 8·0% and syphilis 7·7%), non-communicable diseases, nutrition and lifestyle factors (each about 10%), and maternal age older than 35 years (6·7%). Prolonged pregnancies contribute to 14·0% of stillbirths. Causal pathways for stillbirth frequently involve impaired placental function, either with fetal growth restriction or preterm labour, or both. Two-thirds of newborns have their births registered. However, less than 5% of neonatal deaths and even fewer stillbirths have death registration. Records and registrations of all births, stillbirths, neonatal, and maternal deaths in a health facility would substantially increase data availability. Improved data alone will not save lives but provide a way to target interventions to reach more than 7000 women every day worldwide who experience the reality of stillbirth.
Objective: In Ghana, regional referral facilities by design receive a disproportionate number of high-risk obstetric and neonatal cases and therefore have mortality rates higher than the national average. High volumes and case complexity result in these facilities experiencing unique clinical, operational, and leadership challenges. In order to improve outcomes in these settings, an integrated approach to strengthen the overall system is needed. Methods: Clinical skills strengthening, quality improvement training, and leadership skill building have all been used to improve maternal and neonatal outcomes with some degree of success. We present here a customized model tailored to the particular context of tertiary referral hospitals that develops these three skills simultaneously, so that the complex interaction between clinical conditions, resource constraints, and organizational issues that affect the lives of mothers and babies can be considered together. This model uses local data to identify the drivers of poor maternal and neonatal outcomes and creates an integrated training package to focus on approaches to addressing these drivers. Based on this training, quality improvement projects are introduced to change the appropriate clinical or operational processes, or to strengthen organizational leadership. Results: In testing in one of the largest referral hospitals in Ghana, the model has been well received and has improved performance in several cross-cutting areas affecting the quality of maternal and neonatal care, such as triage, patient flow, and NICU hand hygiene. Conclusion: An integrated approach to systems strengthening in referral hospitals holds much promise for improving outcomes for mothers with high-risk pregnancies and babies in Ghana and in other low-resource settings.
Untreated syphilis in pregnancy is associated with adverse clinical outcomes to the infant. The study aimed to estimate the public health burden resulting from adverse pregnancy outcomes due to syphilis infection among pregnant women not screened for syphilis in 43 countries in sub-Saharan Africa. Estimated country-specific incidence of syphilis was generated from annual number of live births, the proportion of women with at least 1 antenatal care (ANC) visit, the syphilis prevalence rate, and the proportion of women screened for syphilis during ANC.Adverse pregnancy outcome data (stillbirth, neonatal death, low birth weight, and congenital syphilis) were obtained from published sources. Disability-adjusted life-year (DALY) estimates were calculated using undiscounted local life expectancy, the neonatal standard loss function, and relevant disability weights. The model assessed the potential impact of raising ANC coverage to at least 95% and syphilis screening to at least 95% (World Health Organization targets). For all 43 sub-Saharan Africa countries, the estimated incidence of adverse pregnancy outcomes was 205,901 (95% confidence interval [CI], 113,256-383,051) per year, including stillbirth (88,376 [95% CI, 60,854-121,713]), neonatal death (34,959 [95% CI, 23,330-50,076]), low birth weight (22,483 [95% CI, 0-98,847]), and congenital syphilis (60,084 [95% CI, 29,073-112,414]), resulting in approximately 12.5 million DALYs. Countries with the greatest burden are (in DALYs, millions) Democratic Republic of the Congo (1.809), Nigeria (1.598), Ethiopia (1.466), and Tanzania (0.961). Attaining World Health Organization targets could reduce the burden by 8.5 million DALYs. Substantial infant mortality and morbidity results from maternal syphilis infection concentrated in countries with low access to ANC or low rates of syphilis screening.