Sustainability and Scale of the Saving Mothers, Giving Life Approach in Uganda and Zambia

Community Normative Change

In several languages in Uganda and Zambia, the greeting of a mother who has given birth includes an element of surprise and relief that she has survived the perils of childbirth: in Bemba, Mwapusukeni (“You have survived”); and in Luganda, Kulika omwana (“Thank God you have survived with this baby”). Overcoming fatalism and encouraging confidence in a health care system that can respond to the complications related to birth is an important step in increasing demand for facility-based deliveries. Also essential are skilled and competent birth attendants and communities that are engaged to champion this change. The most significant evidence of sustainability in this domain is the formalization and institutionalization of robust community volunteer groups to champion maternal and newborn health; continuation of activities and leadership for maternal health from change champions; and proliferation of maternity waiting homes by diverse funding actors in collaboration with local communities in Zambia. Birth plans may be sustainable, but it is too early to tell if host country governments will continue to print and distribute them, despite interest in and current commitments to doing so. In Uganda, vouchers were not directly sustainable but the findings from the voucher pilot will inform larger social protection schemes in the country.

SMGL worked with ministries of health and other partners to leverage and strengthen predominantly effective existing interventions and promoted change champions, community volunteer groups for education and referrals (SMAGs in Zambia and village health teams [VHTs] in Uganda), and improved access to delivery services through a system of vouchers in Uganda and maternity waiting homes in Zambia.

• Download figure
• Open in new tab
• Download powerpoint

Mothers wait with Safe Motherhood Action Group member at Lundazi Urban Health Centre, Zambia. © Amy Fowler/USAID

At the national level, both Uganda and Zambia expanded the formal role of community change agents for maternal survival by formalizing guidelines and training packages for SMAGs and VHTs, which were implemented in both SMGL and non-SMGL districts. SMGL also ensured that tested materials for birth planning and outreach—radio spots, local language documentaries, and counselling aids for community leaders—were made increasingly available.

The most significant findings, however, were at the district level and below, stimulated by increased information on safe motherhood at the local level and the transfer of ownership of maternal health to traditional leaders, communities, and women and their families (Table 3).

In both countries, change champions, particularly traditional leaders and chiefs in Zambia and district local councils in Uganda, were active and, according to multiple informants, directly increased male involvement in birth and delivery preparedness. This support continued regardless of availability of financial support from SMGL and was cited in focus group discussions as an important factor for sustainability. The district health office in Kalomo, Zambia, confirmed, “Now every time Chief Chikankata has a meeting with Headmen, he asks us to come. Previously these were only ‘women's issues’ but now everyone is involved. The male engagement has meant more women are delivering at facilities and has even reduced child marriage.” Local leadership said this also influenced demand for family planning from women and support from their husbands. In some cases, local ownership resulted in unintended consequences, such as chiefs developing by-laws requiring antenatal care and fining male members of the household who allowed women to deliver at home without a justification (Kalomo and Nyimba). While this was incorrectly attributed directly to the SMGL intervention, it was a sincere reflection of local ownership and embedding of SMGL activities into community structures, which undoubtedly increased antenatal care and facility deliveries.²

Printed birth plans under SMGL aimed to increase birth preparedness and awareness of danger signs, which contributed to the significant demand for maternity waiting homes by both women and district leadership.^22–25 Birth plans provide a facilitated opportunity for health care providers to discuss pregnancy, delivery, and postnatal care and to plan for social support, logistics, nutrition, and care with pregnant women and their partners. The results of a birth plan audit conducted by Communications Support for Health estimated that 139,200 people in Zambia were exposed to the birth plan. Results from a random sample stated that over 70% of respondents recalled having seen the birth plan: 51% of women said they used the birth plans to learn about danger signs, with 48% saying they used it to prepare logistics and 43% saying they used it to remind them to save money. Birth plans included smiley and frowny faces that allowed women and families to assess their experience of care as a feedback tool for facility staff.

While not unique to SMGL, the scale of distribution of birth plans, mentoring of health workers to use the birth plans for counseling, and systematic inclusion in the package of support for safe motherhood made the tool a routine part of facility outreach and counseling and, as demand for services increased, was appreciated and requested by expecting mothers and facility staff. District MOH leadership in Zambia, recognizing the importance of the plans, committed to continued printing and distribution beyond SMGL.²⁴

In Uganda, the transport voucher system pilot implemented during SMGL successfully enabled women to deliver their babies at facilities and facilitated adherence to the national plan of 4 antenatal visits, while addressing the first and second delays. Because it is not inherently sustainable without being embedded in another program, the voucher system will be scaled up under the World Bank's Uganda Reproductive Health Voucher Project and is being considered as a component of the national health insurance scheme. In Zambia, maternity waiting homes were the MOH's central tool for increasing access to health facility delivery and were cited as in high demand by all districts. Waiting shelters were associated with 49% of facilities by 2016, which was an increase of 69% from the baseline. District health officials cited these shelters as contributing to sustained access to care but highlighted concern that demand would soon exceed supply if the government was not able to continue new shelter construction. Communities and districts are now requesting new maternity waiting homes through other donor support and continuing to contribute materials and labor through communities.²⁵

For sustainable changes in community norms, several informants highlighted the need to stay vigilant in reaching local leaders and influencers with locally relevant information. In Uganda, this included active engagement of traditional birth attendants as village health workers (VHWs). In Zambia, sustainable changes included maintaining incentives and support structures for SMAGs and traditional leaders, particularly as the approach extends to new districts. Routine costs of VHWs and SMAGs have been largely absorbed by the host governments and other donor partners but will need continued oversight to ensure sustainability, especially for larger purchases such as bicycles. In Zambia, districts involved in both SMGL and the CoC program began budgeting for SMAG training, monthly meetings, and SMAG incentives from their own government grants and from donor funding. Engaging traditional leaders by, for example, paying honoraria as was done under SMGL in Zambia, was considered costly and controversial but was still determined to be a critical investment and a powerful and sustainable tool for shifting norms. Several district and provincial health officials in Zambia stressed the need for long-term sustainability plans to tackle cross-sectoral challenges that were beyond the remit of SMGL—such as roads, literacy, and poverty—in order to continue making significant progress in maternal and newborn health. Host countries and funding partners need to better coordinate investments to address the thorny underlying roots of poor maternal and newborn health to catalyze a more rapid decline in maternal and perinatal morbidity and mortality.

Governance, Leadership, and Accountability

Multiple respondents echoed the sentiments on increased ownership and leadership of maternal health issues that the external evaluation of SMGL captured in 2013: “[Before SMGL] there were many mothers dying in silence. At least now when mothers die, people notice, and they try to learn from it. It's a big issue. Now when a mother dies, we know before lunch.”²⁶ Ministry of Health district leadership noted that SMGL created champions and leaders—well beyond just medical professionals—for maternal health across the district. Increased visibility and leadership and greater accountability were the most notable findings within this domain. This was exemplified in the maternal death review process in both Uganda and Zambia, which started by bringing in key stakeholders under SMGL, with health sector officials and communities identifying roles each can play in reducing maternal mortality and, in Zambia, with the district commissioners chairing this committee. This is now a routine process for maternal death reviews, an important factor for districts sustaining ownership and action on findings.

At the national level, SMGL incorporated and codified evidence-based interventions through new and updated policies, guidelines, and training materials (Box 2). These core documents are used routinely for planning and training and will continue as guiding documents for the health sectors in both countries. In Uganda, an MOH official noted that “the Ministry of Health sees SMGL as a learning opportunity for the rest of the system.”²⁶ This was realized over the course of SMGL in both countries through testing improvements to change policies, tools, and systems.

The ability of SMGL to mobilize and sustain public–private partnerships did not prove to be as sustainable and scalable as anticipated. It was difficult to attract new global partners after the decision was made, at the end of Phase 1, to select only 1 additional sub-Saharan African country, Nigeria. In Zambia, while 2 local partners, LaFarge Cement and Stanbic Bank, joined SMGL and made important contributions, they have not yet committed to longer-term agreements beyond their initial corporate social responsibility investments. However, there was increased awareness of maternal health issues among these large companies that demonstrated the pull of the success of SMGL to bring new players into the big push for reducing maternal mortality.

Health System and Service Delivery

At the national level, the revised policies and guidelines in both countries benefited districts beyond SMGL and will be cornerstones as the approach continues to national scale. Significant outcomes included introducing uterine balloon tamponade training and other skills sessions in the EmONC curriculum in Zambia to better control postpartum hemorrhage, which is the leading cause of maternal mortality. A significant shift from offsite training to onsite mentorship programs was made in both countries, which became the host governments' core approach. In Zambia, the high-frequency/low-dose method of mentorship was supported because it was positive and encouraged health workers while previous mentorship approaches had been viewed as negative and punitive. Tackle boxes for postpartum hemorrhage and eclampsia were innovations provided to health workers, based on the mentorship experience. In Uganda, most notably, the health professional associations took the lead in providing mentorship to health workers. Routine mentoring visits have been reported as an effective tool to increase knowledge and improve practical skills among health professionals.²⁷ Mentorship is less costly than retraining staff and helps develop a greater culture of shared responsibility between levels of the health system. As a formal tool used by the ministries of health, mentorship also led to local innovation. A former provincial health director in Zambia, Dr. Mathew Ngambi, described how doctors and midwives in Luapula Province formed WhatsApp groups for continued mentorship and provision of advice with a focus on maternal health. While conducting drills on treating eclampsia and postpartum hemorrhage, a mentor in Southern Province noted the time it took as the midwife rushed to various places in the health center to gather supplies, medications, and fluids. In response, the mentor designed tackle boxes containing all the essential supplies for these 2 conditions to be placed in arm's reach of the delivery table.

To further bolster training and mentoring and ensure knowledge retention, availability of updated protocols and guidelines is also crucial; HFAs reported availability increased at almost all facilities in Uganda during implementation of the SMGL initiative. For topics such as eclampsia and magnesium sulfate, postabortion care, and postpartum hemorrhage, the increase in availability of protocols and guidelines from baseline to the end of Phase 2 was significant, increasing from 9% to 74%, 8% to 50%, and 15% to 86%, respectively. We posit this will bolster long-term improvements in service delivery.

In both Uganda and Zambia, HFA results demonstrated that commodity security improved despite ongoing challenges. In Uganda, essential medicines, such as oxytocin and magnesium sulfate, became increasingly available at public health facilities during the 12 months preceding the HFA (2013 and 2017), although some essential antibiotics, such as gentamicin, and other routine antenatal care medications were less available during the same period. Zambia also saw improvements across several key drug availability indicators, although the overall picture was mixed. Hospitals had higher drug and HIV test kit stock rates, but one-quarter of the hospitals faced stock-outs for oxytocin. For medicines such as gentamicin and magnesium sulfate, availability improved at the end of Phase 1 but was reduced in Phase 2. Decreasing trends between end of Phase 1 and endline for several service delivery indicators might indicate that the focused and well-funded programming during Phase 1 was not yet institutionalized across districts and facility types, as funding decreased during Phase 2. These mixed results show the strategic value of leveraging existing supply chain systems and the impact of increased accountability for maternal mortality on the larger health system—since SMGL did not directly support the supply chain—in both countries. It also illuminates an area where additional focus may be required in order to truly sustain achievements (Table 4).

SMGL's direct financial support for capital investments in health facility infrastructure—such as increasing access to running water, electricity, and communications—and skill-building was an important factor for improving confidence in the health system and providing critical support to district and provincial health directors. Such investments are likely to have a long-term impact on service quality and availability, as this type of system strengthening can endure well beyond SMGL and translate into progress in other health areas. In focus group discussions, district and provincial leadership said that their staff and budgets were sufficient to maintain the current infrastructure investments to date; however, this deserves continued attention moving forward. Although further capital investments are planned under donor-supported scale-up activities in both Uganda and Zambia, it is unlikely that the host governments, without support, would continue those investments in the near-term. The number of facilities providing the 7 basic EmONC (BEmONC) signal functions increased from 3 at baseline to 8 at endline in the 4 district-regions in Zambia and from 3 to 9 in the 4 districts in Uganda. Using the World Health Organization minimum recommended level of 5 EmONC facilities per 500,000 population, 3 of the 4 original SMGL districts in both Uganda and Zambia achieved the recommended minimum. Similarly, facilities providing 9 CEmONC signal functions increased from 4 to 5 facilities in Zambia and 7 to 17 facilities by endline in Uganda across SMGL districts.

Capacity to provide a blood transfusion is essential for surgical delivery and resuscitation for severe obstetric hemorrhage and is a CEmONC signal function. To expand the number of facilities with CEmONC capacity in Uganda, SMGL provided blood bank refrigerators to level IV health centers and transported blood from the regional blood banks to these facilities. SMGL provided training for lab technologists on blood grouping and cross matching and for doctors and nurses in prescribing and delivering blood transfusions. In Zambia, where over one-third of blood transfused is for pregnant women, an increased supply of blood was provided to SMGL-supported districts and a pilot initiative began to provide fresh frozen plasma—which has a 1-year, compared to 1-month, shelf life and does not require cross matching—to selected health centers. The 2016 HFA in Zambia, showed that blood transfusion capabilities were maintained in all districts except for Kalomo, which lost this capacity when Zimba district was split off and the remaining district lacked a functioning operating theater. Since 2016, there have been important improvements in the district including increased training, new full-time district leadership, and completion of the only surgical theater in the district. The Zambian government has increased funding to the National Blood Transfusion Service, as external donor funds declined, and plans to develop blood banking hubs nationally in 2018–2019, which may address some of the constraints.²⁸ Inadequate blood supply, however, continues to be a challenge that needs additional funding in both Uganda and Zambia to reduce maternal deaths.

Human resources for health is an area that will continue to require significant support post-SMGL despite substantial strides made, particularly in Uganda. In Uganda, facilities were supported under SMGL to reduce vacancy rates and recruit health staff—at the government rate—who would be absorbed with the increase in salaries as stipulated in the Wage Bill⁹ when it comes into effect in 2019. Almost three-quarters (74%) of health workers hired directly by SMGL in Uganda were eventually absorbed into the health system and all those staff continue to be paid by the government at salary levels stipulated in the national policy. This was an important boost for the sustainability of activities and availability of service providers in the associated facilities. In Zambia, “retired but not tired” midwives were hired directly by SMGL for health centers that lacked a nurse or midwife; however, they could only be given 1-year renewable contracts due to retirement laws and regulations. Enrolled nurses were given additional midwifery training, in part due to lessons learned from SMGL. As Zambia works to expand the pipeline of skilled health workers, SMGL highlighted the opportunity for the government to consider further involvement in and formalization of hiring options for the skilled but retired or out-of-work midwife cadre to support health services.

• Download figure
• Open in new tab
• Download powerpoint

A service provider examines a client at the SMGL-supported Kakumiro Health Center in Uganda. © Amy Fowler/USAID

Overall analysis of the sustainability domain questions in the HFA reports demonstrated significant improvements and prospects for sustainability across indicators for human resources, health facility infrastructure, and access to EmONC services in both Uganda and Zambia.²⁸ There was some reduction in performance of some indicators following Phase 1, when there was a funding gap of 1 year due to a significant break in implementing partners in both Uganda and Zambia due to procurement timelines as well as an extended delay in funding reaching both countries. During Phase 2 there was a planned reduction in funding during the second year by 25% from the first year, a 50% reduction the following year, and a 75% reduction in the final year to naturally push for greater government ownership. These circumstances provided an opportunity to demonstrate the sustainability of interventions and the ability of local and national governments to assimilate activities from the “big push” into national maternal and newborn health programs. Through this lens, Uganda had greater success in sustainable improvements for human resources for health. However, in both countries, early investments in infrastructure are likely to continue to pay dividends in access to quality care and services.

Strategic Investments, Efficiency, and Sustainable Financing

While SMGL was not designed to specifically increase national maternal health budgets, we hoped that the programmatic results would drive governments to increase domestic resource allocation for continued quality maternal and newborn services.⁸ In the end, SMGL did not contribute to significant changes in either country's domestic resource mobilization or increased government funding. SMGL maximized allocative efficiency by identifying high-impact activities that, when clustered, addressed the 3 delays and provided health system benefits to population groups beyond mothers and newborns, such as emergency transportation, improved infection prevention, water supply and electricity at health facilities, and availability of surgical services and operating theaters—all critical components of a functioning health system. SMGL districts in Zambia cited a significant change in the district's annual plans for maternal health. Several districts cited the value of a ‘blueprint’ for health systems areas of improvement and, in particular, cited an important increase in planning and prioritization around the first and second delays (demand and access to services). In Zambia, for example, multiple district partners highlighted the increased investment in community structures and health promotion to increase demand. This approach was recently codified with requirements from the MOH that 10% of all budgets must directly benefit the community. In Zambia, SMGL funds were channeled directly to 2 of the 4 focus provinces that oversaw implementation in the SMGL districts—approximately US$300,000 per province per year—from the U.S. Government and Sida (for 1 year) to support the SMGL approach. This direct support further engaged the districts and resulted in a natural transfer of central activities from SMGL support to direct funding. As already noted, SMGL directly influenced greater investments under the CoC program in Zambia and the World Bank Reproductive, Maternal, and Child Health Services Improvement project in Uganda, leveraging 5-year investments of US$125 million in Zambia and US$140 million in Uganda. In Zambia, this translated into prioritizing interventions under SMGL for direct funding from donors and direct host-country resources and, in Uganda, the Sharpened Plan Investment Case⁷ for RMNCAH built heavily on the SMGL approach.

A separate study in this supplement by Ben Johns et al.²⁹ was conducted to determine costs—incremental costs and incremental cost per death averted—of the SMGL district strengthening approach. They found that the incremental cost for maternal and newborn care per SMGL-supported district in 2016 was US$845,000 in Uganda and US$760,000 in Zambia. This translates into about US$38 per facility birth in Uganda and US$95 per facility birth in Zambia in 2016. In Zambia, the cost per death averted was US$12,514, or $206 per life-year saved. In Uganda, the cost per death averted was US$10,311, or $177 per life-year saved. The researchers concluded that the approach was cost-effective, with the cost of life-year gained as a percentage of gross domestic product (GDP) being 26% and 16% in Uganda and Zambia, respectively. The authors concluded that the SMGL program “could be paid for by increasing health spending from 7.3% to 7.5% of GDP in Uganda or 5.4% to 5.8% of GDP in Zambia.” This is far less than domestic investments for other health areas, such as HIV and AIDS.³⁰ The analysis did not take into consideration the costs associated with the considerable ripple effects a maternal death has on children, families, and communities.^31,32

The district strengthening approach, as demonstrated in SMGL, represents a substantial cost-effective health investment, one that low- and middle-income countries can afford.²⁹

Strategic Information

At the national, provincial, and district levels, improvements in strategic information were viewed not only as sustainable systems-level improvements in the way data are made available, used, and analyzed, but also as a catalyst for increasing leadership and prioritization of maternal and newborn health. The former provincial health director of Luapula Province noted, “Success of SMGL really grew out of improved information, audits and the ability to use data to hone in on problems and demonstrate that with planning and focused resources the problems could be addressed. We were able to see the results of our work and started asking other districts for similar information.”

In both Uganda and Zambia, maternal and perinatal death surveillance and review systems (MPDSR) were institutionalized and training programs and guidelines developed with the support of SMGL. Progress in MPDSR has been sustained in both countries in Phase 2 of SMGL. At the hospital level, for example, according to the HFA, 93% of hospitals in Uganda and 100% in Zambia conducted maternal death reviews in 2016. In Zambia, 75% of all deaths in SMGL districts were reported and reviewed in 2016. In focus group discussions, district leadership felt the process had become routine and could be sustained without external support. In both countries, non-healthsector leadership is now involved in death reviews, which continue to build political will and address maternal health appropriately within the broader community context, further adding to the likelihood of sustainability. In Zambia, district commissioners were appointed as heads of the district MDSR committee and required to submit their maternal death reports to the national government; this ensured a high level of political investment in reporting maternal deaths and in understanding the multiple causes of death. In Uganda, the birth weight and age-at-death boxes for an intervention and evaluation system (BABIES) matrix²⁸—a simple data tool to better track and understand newborn deaths and stillbirths—facilitated understanding of the timing and causes of perinatal deaths through closely tracking fresh and macerated stillbirths and early neonatal deaths. It is now being used comprehensively in 6 of 13 SMGL districts and has been proposed in the Uganda MPDSR guidelines.

SMGL strengthened Health Management Information System (HMIS) data in both countries through vigorous review of registers and health information aggregation reports by implementing partners with district health office staff. The data collection approach, implemented on a quarterly basis, included data collection from all possible data sources in a health facility across all existing departments—data that were later triangulated with the health information aggregation report. This worked as a validation process for HMIS and consequently improved site-level data collection and reporting. Ministry of Health district staff in Zambia commented that it became routine for SMGL team members to call and question them about their data quarterly. Consequently, district health workers took more interest and began reviewing data at the district headquarters before submitting, a process that continued after SMGL funding ended and is likely to be sustained. In Uganda, Pregnancy Outcome Monitoring Survey data were used to update the monthly maternity register summaries that were entered and reported into the national District Health Information System 2 database. This community data collection tool, which remains in place, was strengthened so that it could be used for pregnancy and MPDSR. SMGL supported procurement and supply of facility tools and data registers and facilitated orientation of health workers on the registers, if needed. One district health director in Zambia commented that the district would print registers if they were not available from the national level. This is not a universal position at the district level but, with continued external funding under the CoC program and heightened scrutiny at the national and provincial levels, these minor but critical tools are likely to remain in place for the foreseeable future.

In both Uganda and Zambia, the focus on routine data reviews beyond the facility resulted in routine, more meaningful, review meetings, where gaps were identified and addressed with decision makers in the room. These were simple but important problem-solving exercises. For example, a province in Zambia realized that a facility that had made only limited improvement in the number of antenatal care visits did not have a motorbike and reallocated one from an incoming shipment so the facility could conduct routine outreach to remote communities. This small data-based problem-solving process was repeated over time and helped address many health system issues that had previously seemed insurmountable. One district health director in Zambia noted that “as the SMGL metrics were absorbed into their performance standards, they [health care workers] became more accountable.” She felt this would be a key driver of sustainability.