Tanzania: CRVS Strengthening with SAVVY Implementation

In Tanzania, SAVVY started in October 2009 as a five-year project whose implementation began in early 2011. The main objective has been to strengthen the capacity of government to collect and use mortality surveillance data to assist in managing its national HIV/AIDS programs by expanding community-based identification and reporting of AIDS deaths.

The Tanzania team at the SAVVY workshop in February. Photo by Kathy Doherty, MEASURE Evaluation.

By contrast, Civil Registration of Vital Statistics (CRVS) systems in Tanzania have been in place since colonial days. The procedures and policies for CRVS are regulated by various public policies under several legal frameworks: The Child Development Policy 1996, the National Health Policy 2007, the National Population Policy 2006, the Births and Deaths (Registration) Act. Cap. 108, the Law of Marriage Act. Cap. 29. and the Statistics Act, 2015. Therefore, there is no single CRVS law, which leads to a resulting need for coordination and linkages among key actors and a means to better transmit and share registration data. In addition, the CRVS system in Tanzania is not yet comprehensive. SAVVY can potentially provide an important and cost-effective learning platform for a more robust and unified CRVS by which registration activities can be implemented at a smaller scale, and improved and refined before a roll-out or scale-up nationally.

For example, SAVVY results show that deaths from HIV/AIDS have gone down from 9.7 percent in 2013 to 8.9 percent in 2014. But, cardiovascular disease, cancers, acute febrile illness including malaria, and accidental deaths—especially road transport accidents—are killing more people in Tanzania than some of the known communicable diseases. Yet, the health system is not yet aligned with that reality, nor are health services and resource allocations. SAVVY provides an important evidence-informed means of assessing the current threats to health, by zone and nationally. Complementary SAVVY goals include monitoring trends in other mortality indicators relevant to the Millennium Development Goals.

It’s important to note that SAVVY data are nationally representative, with a good mix of social demographics in terms of urban and rural mix, geographic mix (highlands, lowlands, and the lake zone), education levels, and they include death reports from communities that often are not using health facilities.  

In Tanzania, the Ifakara Health Institute (IHI) has led the implementation of SAVVY, in partnership with the Ministry of Health and Social Welfare (MoHSW), the National Bureau of Statistics (NBS), and the National Institute for Medical Research (NIMR), funded by the Centers for Disease Control and Prevention (CDC). SAVVY was implemented in 23 districts of mainland Tanzania from 2011 to 2014.

The next challenge for Tanzania is to secure more funding for SAVVY to continue. It has contributed detailed data on current national vital statistics that the country’s decision makers can use to understand threats that endanger citizens’ health and that contribute to mortality rates.  Absent a fully scaled-up CRVS system, SAVVY has provided much of the heavy lifting to provide evidence for health program decisions.

But the funding question is fundamental. Stakeholders for these efforts include the government and private sector partners and external donors that are committed to SAVVY as a precursor toward scaling up to a full CRVS system. Additional stakeholders may be found in communities that see how their participation in the SAVVY data collection provide essential information to government that can improve their health and their lives.

The SAVVY team also plans to accelerate report production—to get the results to decision makers and stakeholders alike. Strategies include replacing traditional paper handling with digital and mobile solutions, complementing physician coding by automated computer coding disseminating study results at local council authorities or lower-level health facilities, and, participating in the implementation of the country’s CRVS strategy, including reviewing of business processes for registration and providing technical and logistical support on training in verbal autopsy and ICD-10 coding. In addition, the CRVS strategy proposes using SAVVY tools for evaluating the performance of the CRVS system as it is being demonstrated in selected wards.

A regional workshop held in Lilongwe, Malawi, February 23 – 25, 2016

The majority of countries in sub-Saharan Africa lack a fully functioning civil registration and vital statistics (CRVS) system. These countries rely primarily on population-based surveys and censuses to generate fertility and mortality data, which are essential element in determining health issues and services needs in any country.

In recent years, a number of countries in the region have initiated efforts to strengthen their nascent CRVS systems. In particular, Malawi, Tanzania, and Zambia have implemented Sample Vital Registration with Verbal Autopsy (SAVVY) as part of a sustained, incremental effort to gather this essential data. SAVVY was developed by MEASURE Evaluation and the U.S. Census Bureau and is a family of methods that allows the direct measurement of vital events and the determination of causes of death in a nationally representative sample of small areas, or in selected ‘sentinel’ locations. The components of SAVVY include demographic surveillance that registers resident population, mortality surveillance to report and register deaths in a resident population, and verbal autopsy (VA), to determine likely cause of death through interviews with next of kin and caretakers.

The workshop[1] was designed to improve knowledge among the three countries about processes, institutions and structures required to implement SAVVY, and to share details on how each have designed and implemented SAVVY within their countries. Specifically, the three countries shared how they work with civil registration agencies locally, their steps for linking SAVVY findings with facility-based health information systems and newly-developed CRVS systems so that the causes of death SAVVY finds can be captured in the CRVS system, and to educate and inform stakeholders and donors of SAVVY accomplishments and potential, both to generate ongoing donor support and to build a constituency for SAVVY and new CRVS systems to complement each other.

Attendees included officials from ministries of health (MOH), partners implementing SAVVY activities, national statistics offices and registration officials, and donors such as the Bill and Melinda Gates Foundation, the Centers for Disease Control and Prevention (CDC), Data for Health Initiative (D4H), Karonga Prevention Study/Malawi Epidemiology and Intervention Research Unit (KPS/MEIRU), the U.S. Census Bureau, The World Health Organization (WHO), and MEASURE Evaluation.

Workshop Proceedings

Malawi’s Chief of Health Service, Dr. Charles Mwansambo, in his opening remarks, highlighted the need for evidence and quality data to inform policies and decision-making. Recognizing the long timeline for creating a robust civil registration system, Dr. Mwansambo focused on the need for interim measures, such as SAVVY, to generate vital events data. Malawi—among the three countries, the one in the earliest stages of the design and implementation of its CRVS system—is currently doing a phased roll out of birth registration at health facilities in a number of districts while death certification still relies on patient request.

Tanzania, on the other hand, has a long tradition of civil registration, an existing CRVS Strategic Plan, but lacks a coordinated legal and implementation environment. Renewed focus is on reaching goals of registering births with 90 days and deaths within 30 days, whether manually or electronically. Lastly, Zambia has registration offices in all districts in a decentralized system and CRVS is part of the national strategic action plan for 2015 through 2019, which includes the implementation of SAVVY as well as linking the CRVS system with other systems.

Across all countries, the goal was to be able to provide nationally-representative statistics. The first hurdle in that regard is how to construct a representative sample for the SAVVY exercise. Tanzania used a probability proportional to size (PPS) and systematic probability sample of districts, enumeration areas, and households. Malawi designed its sample to be representative at the national, urban/rural, regional and ministry of health “zone levels” in 37 SAVVY sites. Similar to the other countries, Zambia relied on census data to complete its sampling. A one-stage sampling of population segments within each province was completed, stratified by rural and urban areas and 76 segments were selected, allowing for both national and provincial-level estimates.

The process of conducting verbal autopsies differs in each country. Zambia utilizes employees of the MOH who use WHO questionnaires and follow the standard SAVVY process of conducting verbal autopsies for deaths reported by community key informants (CKIs). Unlike Zambia, Tanzania selected interviewers who expressly did not have a clinical background, in order to avoid bias. Tanzania also used the WHO questionnaires with slight modification for the local context. In Malawi, interviewers are health surveillance assistants who are non-clinical health workers in their districts.

For all countries, once interviews are conducted, medical personnel are engaged for death certification and coding the cause of death. In Malawi, a team of trained clinicians meets quarterly to review completed VA questionnaires. Two doctors review the data and, if the cause of death differs, the two doctors must come to consensus. In Tanzania, 50 physicians have been trained to code questionnaires. Each questionnaire is reviewed by two doctors, who must come to a consensus on cause of death. A similar process is completed in Zambia. Challenges identified are incomplete questionnaires or difficulty in reading open-ended responses.

Reporting SAVVY Results

Tanzania described a robust system for reporting data back to the district through district health profiles and district mortality profiles. Data is also fed back to the national level to inform planning. The results of SAVVY in Tanzania show that it is a good representation of the national population conforming to census data. A similar finding was reported by Zambia, which noted two challenges to the reliability of results: 1. Avoiding inclusion of reported deaths outside the defined SAVVY area or outside the defined time period for sampling; 2. Cultural taboos against reporting the deaths of neonates. Malawi has not yet completed follow-up of deaths identified during the SAVVY baseline census, so it did not have data to share, but described the types of indicators that will be calculated.

Coordinating the SAVVY results with other data in country poses some challenges. Tanzania noted problems linking the SAVVY system with the CRVS system due to separate databases that do not speak to each other. While partners are coordinating for the implementation of SAVVY, further linkages are needed to use the data more effectively. Zambia has a national steering committee to ensure that the national strategic action plan includes SAVVY in its scope. The committee has a CRVS technical working group that is active in coordination and ensuring project objectives are met by each of the responsible government entities. Malawi has clearly defined roles for each government entity, similar to Zambia. All agreed that formalized cooperation among government agencies and compatible data sets that all can use are key issues for continued work.

Funding Going Forward

A common need among all countries was the need for additional funds to be committed for SAVVY to continue in the future. While international donors are interested in funding they are bound by donor priorities and strategies. In order to attract further funding, it’s critical that CRVS and SAVVY are included in national strategies and plans, so that donors know it is a country priority.

Each of the development partners and donors provided an overview of their strategies and activities related to CRVS and SAVVY that will guide their commitment for the future.

• CDC Zambia: CDC Zambia initially invested in the country’s SmartCare system, which has the capability to register births and could be integrated with CRVS. This digitized health card system is, however, limited by the country’s unreliable power grid.
• CDC Malawi: With PEPFAR funding, CDC Malawi has implemented a comprehensive CRVS approach to gathering vital statistics, with a focus on both infrastructure and systems. 
• Gates Foundation: This donor is interested in rolling out the Countrywide Mortality Surveillance for Action (COMSA) system, and sees the potential of working with existing SAVVY and CRVS systems. It is currently investing in minimally invasive tissue sampling (MITS) to determine precise causes of death.
• WHO: The WHO recognizes the increasing number of censuses and surveys to fill the gap as CRVS systems are strengthened. Their strategy is focused on registration rather than VA, asking countries to register all deaths (both hospital deaths and community deaths) before focusing on improving cause of death reporting.
• Data for Health Initiative (D4H): Funded by Bloomberg Philanthropies, is a four-year project in 20 countries in Africa, Asia, and Latin America, focused on civil registration, data use, and non-communicable disease surveillance. Zambia is in the early stages of a work plan to train cause of death coders and in Malawi to document processes, support training and supervision, and advocate for certification and coding training in clinical curricula.