Declining maternal mortality ratio in Uganda: Priority interventions to achieve the Millennium Development Goal

1. Introduction 

Uganda has registered a decline in maternal mortality ratio (MMR: maternal deaths per 100,000 live births) over the past 10 years. MMR was 557 live births in 1995, this declined to 505 in 2001 and to 435 in 2006 [1]. Of concern to policy makers and program mangers is whether Uganda will achieve the Millennium Development Goal (MDG) target of 162 in 2015.

Previously we have reported that using the sector-wide approach (SWAp), Uganda has prioritized emergency obstetric care (EmOC), revitalization of family planning programs, and increasing skilled attendance at delivery [2], [3]. A high unmet need for EmOC has already been documented [2], [3]. Lack of emergency care was highly associated with maternal deaths in health facilities. Specifically, removal of retained products, assisted vaginal delivery and lack of blood transfusion were the key EmOC signal functions related to maternal deaths [3]. It is therefore important to assess other factors related to maternal mortality and provide data for effective advocacy on maternal health especially prioritization and resource allocation. This article focuses on the effect of health system factors on maternal deaths especially the state of health infrastructure, basic supplies, and staffing levels, since these are neglected issues in the advocacy strategy for maternal health in Uganda.

The importance of reducing maternal and child mortality in low income countries has been well recognized [4]. Maternal and infant mortality rates in a particular country reveal more about the state of the health system than any other indicators. It has been concluded that achieving low maternal and infant mortality rates requires a well-functioning health care delivery system that reaches communities with education and counseling, helps people avoid unwanted pregnancies, promotes good nutrition, screen risks, and responds to obstetric emergencies [4]. A health system has been defined as any set of activities whose primary intent is to improve or maintain health. The health system includes resources, actors and institutions related to financing, regulation and provision of health [11], [12].

Efforts to achieve the MDGs, especially on child and maternal mortality in Uganda, have resulted in the development of a road map to reduce maternal mortality and neonatal health. This strategy focuses on improving health systems like staffing levels, infrastructure, quality of care and community mobilization to create demand for services. The key interventions in this strategy are increasing access to EmOC, revitalization of family planning, and increasing antenatal care and skilled attendance at delivery [5].

How well a health system responds to obstetric emergencies is determined by many factors, as previous studies have shown [6], [7], [8], [9], [10]. Emergency obstetric care refers to care given to complications that threaten the lives of the mother and newborn [9]. In Uganda poor access to EmOC has already been documented [2], [3]. This was attributed to poor quality of care at health units but could be a reflection of inadequate health systems including community factors. To get baseline data for monitoring the scale-up of EmOC services in Uganda, a needs assessment on signal functions of EmOC was carried out. Data on maternal deaths and other health system factors were collected. This article highlights health system factors and presents a case for focusing on human resources and availability of EmOC if Uganda is to meet the MDG targets.

3. Results 

Availability of basic facilities and supplies was assessed in all 553 health units and is shown in Table 1.

Table 1. Availability of basic facilities, equipment and supplies in health units in Uganda, 2003

	Facility, equipment or supply	Percentage of health units with facility, equipment or supply
		HCIII	HCIV	Hospital	All
		n=359	n=107	n=87	n=553
	Running water	1.4	19.6	46.0	11.9
	Electricity	2.2	29.0	60.9	16.8
	Theater	Not applicable	15.9	35.6	24.7
	Laboratory with reagents	49.9	84.1	100.0	64.4
	Sulfadoxine–pyrimethamine⁎	87.5	91.6	96.6	89.7
	Chloroquine (injectable)⁎	79.9	81.3	95.4	82.6
	Approved staff positions filled (midwives)	98.9	73.8	93.1	93.1

Abbreviations: HCIII, level III health center; HCIV, level IV health center. ⁎To treat malaria.

3.1. Health centers III 

Most HCIIIs had no running water and electricity. On the other hand, almost half of them had a functional laboratory as well as sulfadoxine–pyrimethamine, chloroquine. And most had filled approved staff positions (midwives).

Only 11 of 359 (3.1%) HCIIIs were offering basic EmOC. The signal functions most often missing were assisted vaginal delivery (95.0%), removal of retained products (62.7%) and manual removal of retained placenta (65.2%) (Table 2).

Table 2. Availability of EmOC signal functions in health units in Uganda, 2003

	EmOC signal function	Percentage of health units providing the signal function
		HCIII	HCIV	Hospital	Total
		n=359	n=107	n=87	n=553
	Parenteral antibiotics	80.2	87.9	98.9	84.6
	Parenteral oxytocis	74.4	82.2	96.6	79.4
	Parenteral sedatives	37.9	42.1	87.4	46.5
	Manual removal of retain placenta	34.8	49.5	86.2	45.8
	Removal of retained products	37.3	59.8	95.4	50.8
	Assisted vaginal delivery	5.0	15.9	62.1	16.1
	All of the above	3.1	7.5	52.9	11.8
	Blood transfusions	NA	13.1	86.2	45.9
	Cesarean delivery	NA	9.3	86.1	43.8

Abbreviations: EmOC, emergency obstetric care; HCIII, level III health center; HCIV, level IV health center.

3.3. Maternal deaths 

Causes (direct and indirect) of maternal deaths were recorded in all 553 health units (Table 3).

Table 3. Obstetric complications, deaths and cause-specific case fatality rates recorded in health units in Uganda, 2003

		Obstetric complications (a)	Maternal deaths (b)	Case fatality rate (maternal deaths as a proportion of obstetric complications) (b)/(a)
	Direct causes
	Ruptured uterus	590	83	14.1
	Hemorrhage	5766	356	6.2
	Postpartum sepsis	1628	60	3.7
	Ectopic pregnancy	584	14	2.4
	Pre-eclampsia/eclampsia	2603	49	1.9
	Obstructed/prolonged labor	14,824	187	1.3
	Complications of abortion	16,394	93	0.6
	Total direct causes	38,677	842	2.2
	Indirect causes
	HIV/AIDS	287	59	20.6
	Sickle cell disease	61	6	9.8
	Anemia	2794	89	2.1
	Malaria	23,657	306	1.3
	Total indirect causes	26,799	460	1.7

3.3.1. Case fatality rates 

Table 3 shows maternal deaths as a proportion of obstetric complications — the case fatality rate (CFR). HIV/AIDS had the highest case fatality rate (20.6%), followed by ruptured uterus (14.1%), and sickle cell disease (9.8%). Note that all CFRs but one (abortion complications) exceeded the standard set in the UN Process Indicators of 1%.

3.3.2. Proportional mortality (data not shown) 

Hemorrhage was the most common direct cause (42.3%) of maternal deaths, followed by obstructed labor (22.2%). Complications of abortion were responsible for 11.1% and ruptured uterus accounted for 9.9% of the deaths. Among the indirect causes of maternal deaths, malaria accounted for 65.5% of the cases, followed by anemia (19.3%) and HIV/AIDS (12.8%).

Nearly half (48.5%) the deaths took place in hospitals; 22.5% occurred in HCIVs and 29.0% in HCIIIs (Table 4). However, 58.9% of complications were treated in hospitals compared with 23.8% in HCIIIs and 17.2% in HCIVs.

Table 4. Number of deliveries, complications and deaths recorded in three levels of health units in Uganda, 2003

	Maternal health indicator	HCIII	HCIV	Hospital	Total
		n=359	n=107	n=87	n=553
	Deliveries
	Number	43,202	33,066	117,761	194,029
	Percent	22.3	17.0	60.7	100.0
	Obstetric complications
	Number	15,576	11,318	38,582	65,476
	Percent	23.8	17.2	58.9	100.0
	Direct obstetric
	Number	4392	4514	29,771	38,677
	Percent	11.4	11.7	76.9	100.0
	Indirect obstetric
	Number	11,184	6804	8811	26,799
	Percent	41.7	25.4	32.9	100.0
	Maternal deaths (number)	378	293	631	1302
	Percent	29.0	22.5	48.5	100.0
	Direct obstetric	111	284	447	842
	Percent	13.2	33.7	53.1	100.0
	Indirect obstetric	267	9	184	460
	Percent	58.0	2.0	40.0	100.0
	Case fatality rate
	All maternal causes	2.4	2.6	1.6	2.0
	Direct obstetric causes	2.5	6.3	1.5	2.2
	Indirect obstetric causes	2.4	0.1	2.1	1.7

3.4. Institutional risk factors 

Factors related to maternal deaths in health facilities were assessed.

3.4.1. Geographical factors 

Using the Western region as the reference, health facilities in the eastern region had an elevated risk of maternal death (OR 3.7, P<0.0001), compared with those in the central and northern regions.

3.4.2. Level of health care 

Women at HCIVs had a higher risk of maternal deaths (OR 3.9, P<0.01) compared with those at HCIIIs (OR 2.4, P<0.002) and hospitals (reference) (Table 5).

Table 5. Factors associated with maternal deaths in health facilities in Uganda

	Covariates	Odds ratios (95% CI)	Level of significance	Percentage of maternal deaths averted 100 (1-risk ratio)
	Central region	1.1 (0.9–1.3)	P>0.465	–
	Northern region	0.9 (0.7–1.2)	P>0.52	–
	Eastern region	3.7 (3.2–4.3)	P<0.0001	–
	Western (reference)	–
	Availability of staff (midwives)	0.20 (0.17–0.23)	P<0.0001	80%
	Availability of electricity	0.39 (0.34–0.45)	P<0.0001	61%
	Availability of a theater	0.56 (0.48–0.67)	P<0.0001	44%
	Availability of a laboratory	0.71 (0.32–0.79)	P<0.0001	29%
	Availability of sulfadoxine–pyrimethamine	0.68 (0.58–0.81)	P<0.002	32%
	HCIIIs	2.4 (1.5–5.3)	P<0.002	–
	HCIVs	3.9 (1.3–8.5)	P<0.001	–
	Hospitals (reference)	–

Abbreviations: HCIII, level III health center; HCIV, level IV health center.

3.4.3. Available resources 

Availability of midwives (OR 0.20, P<0.0001), theater (OR 0.56, P<0.0001), electricity (OR 0.39, P<0.0001), laboratory (OR 0.71, P<0.0001) and sulfadoxine–pyrimethamine (OR 0.68, P<0.002) were protective for maternal deaths at health facilities (Table 5).

4. Discussion 

Our results show that most health facilities – at all levels – in Uganda lack basic equipment and infrastructure necessary to provide quality of care. Availability of running water, electricity, and functional operating theaters in facilities expected to have them was lacking. The implication of these findings is that unless resources are allocated to the development of health infrastructure and improving human resources, Uganda is unlikely to meet the MDG on maternal health.

Most (97.2%) health facilities expected to offer basic EmOC services were not able to do so at the time of the study. Most HCIVs were not functional and therefore not able to provide comprehensive EmOC. Indeed, barely half of hospitals were able to provide all the elements of basic EmOC (assisted vaginal delivery was not practiced in many). To address the issue of high maternal mortality, these centers need to be operational as they are more likely than hospitals to be within the reach of most pregnant women. In the recent past, the Ministry of Health in Uganda has constructed theaters at HCIVs but these still need to be equipped and staffed by appropriate personnel recruited and motivated to provide services [5].

In many countries of sub-Saharan Africa, high levels of maternal mortality now coexist with high levels of HIV prevalence. HIV/AIDS may increase pregnancy related mortality directly (puerperal sepsis) or indirectly (anemia or tuberculosis) [13]. Table 3 shows that HIV/AIDS had the highest CFR among all causes of maternal deaths. This finding has implications for women accessing information and services for family planning, voluntary counseling and testing (VCT) for HIV and antiretroviral drugs. Uganda has had a success in reducing HIV prevalence among pregnant women [18], [19]; but uptake of VCT has been notably low [18] There is need to focus on integrating VCT into antenatal care to reduce high AIDS-related maternal mortality in Uganda.

Analysis of local data at facility level is important in two ways; one is to provide data for decision making, prioritization and resource allocation, and second to provide evidence for monitoring the implementation of activities such as the integration of services and quality of care. All these have policy implications and create opportunities for policy reviews. Continuing analysis of the magnitude of maternal mortality and the related causes is necessary to monitor national and international commitments to improve maternal health to acceptable levels.

This research found a facility-based maternal death rate of 671/100,000 live births. Although this figure is not representative of all women in Uganda, it highlights a problem in the quality of care women receive once they reach the health units. Women in eastern Uganda seem to have a high risk of death compared with the central and western regions. The protracted armed conflict in this region has led to collapse of infrastructure and difficulty in accessing care, pregnant women (and others) are therefore more likely to present with more severe complications. Health seeking behavior in conflict areas and how emergency care can be delivered under such circumstances needs further study and action. Surprisingly women in the northern region were less likely to die at health units although this is a conflict area. This has been attributed to the close proximity of internally displaced camps to HCIIIs; and this region has previously recorded a high proportion of facility deliveries [20].

The importance of the health system in delivery of quality of care cannot be overemphasized. Previous studies have highlighted the impact of this on maternal and child health [15], [16], [17]. This study highlights the role of human resources especially the effect of staffing levels (midwives) on maternal mortality. The policy implication of these findings is that the Ministry of Health, the Ministry of Finance and other development partners need to prioritize and allocate more resources to the development of infrastructure and human resources and to improve quality of care especially the availability of basic supplies and equipment.