Reflections on cases from Tanzania and India
By Genevie Fernandes
Last month, I got this fantastic opportunity to join a group of statisticians and social scientists working in Sierra Leone, Ethiopia, Madagascar, India and Canada, for a workshop on mortality estimation in low and middle-income countries (LMICs). After a week of lectures, hands-on training in statistics, and discussions based on country experiences, I left the programme convinced about why we need to strengthen national health information systems for improving public health.
A robust health information system is one of the building blocks of any health system. A national health information system will ideally collect data on a periodic basis from civil registration statistics including birth, fertility, and death; epidemiological surveillance; health service records; household-level health surveys; health financing trends, plans and accounts; need, access, and consumption for medicines and diagnostics; and health workforce and infrastructure. Many LMICs have weak and fragmented health information systems, which arise not only because of financial constraints but also due to inadequate technical capacity and political will.
In this post, I reflect on the need to strengthen existing national health information systems in LMICs to collect data on two vital events i.e., birth and death, or in Mark Twain’s words, ‘the two most important days of our life’.
First, let’s look at why registering births is so important.
The births of around one fourth of the global population of children under-five have never been registered. Without a formal birth certificate, a child can be denied health care and education, raising difficulties in later life for employment in the formal sector, to vote, gain a passport and so many other basic rights as a citizen of the country. Lack of reliable data on births also makes it challenging for governments to track child health outcomes, plan interventions and allocate appropriate resources in targeted areas where it is needed the most.
With less than 13% of under-five children having birth certificates, Tanzania faces a major challenge of having one of the lowest rates of birth registration and certification in Africa. However, in 2013, the government of Tanzania with the support of UNICEF and the Canadian government, launched the decentralized birth registration initiative, which is proving to be an innovative and sustainable model for improving birth registration and certification at the facility level in low-resource settings. Prior to this initiative, barriers such as long distance to registration centres, multiple visits to the facility for registration and certification, direct ($1.6 registration fee) and indirect (travel and loss of wage) costs, and low awareness deterred Tanzanian parents from registering their child’s birth.
This new initiative addresses these barriers, by setting up decentralized facilities thereby drastically reducing travel distance, time and cost, waiving off registration fees, making birth registration and certification a one-step, one-visit process, and using mixed methods of communication including local community and religious leaders to generate awareness and demand for this service. Birth registration data is captured using mobile phones and simple excel dashboards to depict progress at the national, regional, district, ward and health facility-level in real-time. This initiative has led to a birth registration and certification coverage of 30% in a span of six years, and is currently expanding to cover the entire country.
Now, moving on to why counting the dead is vital for improving the health of the living.
In the absence of a well-functioning national health information system, the number and causes of deaths that occur outside of a health facility, often do not get recorded accurately. Take the case of rural India where most people die at home and get cremated or buried within a day or two; deaths often don’t get registered. When they do, registration may merely capture the name of the deceased, but the actual cause of death rarely gets included.
Mortality information is vital for several reasons. First, families need to know the cause of death of their loved one, as it gives them closure and also prepares them for any preventive or protective measures if required. Second, without accurate information on the number of deaths and what kills people, governments are unable to allocate financial and technical resources for appropriate programs and interventions; it limits the monitoring of health goals and accountability over health expenditures; and as seen recently in the case of Ebola, the lack of timely mortality data misses identifying new or re-emerging disease outbreaks. Without reliable mortality data, governments in LMICs and external donors track progress on health goals by relying on estimates from the United Nations (UN) agencies and most recently the Global Burden of Disease (GBD) Study undertaken by the Institute of Health Metrics and Evaluation (IHME).
How do the UN agencies and IHME estimate the mortality and morbidity burden in LMICs where existing health information is scarcely available?
UN agencies have access to the available published and unpublished sources of data on vital statistics including birth, fertility and mortality in member states. This national data is compiled and modelled by UN experts using softwares such as MortPak, after which member states are consulted, and estimates are made publically available. Similarly, for the GBD estimates, the IHME sources data from surveys, sample and vital registration systems from countries; however, these data correspond to only 30% of all global deaths. While the GBD estimates have highlighted changing global trends over time and provided benchmarks to monitor progress in the era of the Millennium Development Goals (MDGs) and Sustainable Development Goals (SDGs), we need to bear in mind that these estimates are only estimates and not actual measurements. The GBD estimates rely on heavy modelling where reliable information is scarce, and do not offer disaggregated data for regions and groups within the country, which makes it difficult to set national priorities and allocate resources for targeted programs and interventions. Furthermore, understanding the complex modelling process and interpreting the range of these findings can often be challenging for national governments and their health ministries.
While these global morbidity and mortality estimates are an extremely valuable starting point to check country progress over time, they still have limited use at the national level for programme planning, funding, and monitoring, which is why we need to strengthen existing health information systems at the country-level and improve vital data collection.
The Million Death Study (MDS) in India is one such innovative example where existing resources have been built upon to collect cause-specific mortality data and also inform national priority setting, policy-making and funding decisions. Since 1971, the government of India has been implementing the Sample Registration System (SRS) where births and deaths are followed in thousands of randomly selected areas every ten years. The MDS trained government staff who collect data for the SRS, to conduct verbal autopsies by interviewing household members where a death has taken place and gathering information on the cause of death by filling out a simple half page narrative in the local language. Each narrative is converted to an electronic record after which two physicians from a selected online cohort, independently review and code the cause of death using the World Health Organization’s (WHO) International Statistical Classification of Diseases and Related Health Problems, Tenth Revision. In the case of disagreement between the physicians, the record is reviewed and resolved by a senior physician.
Between 1998 and 2014, the MDS will have monitored nearly 14 million people in 2.4 million households, providing nationally reliable and disaggregated data by cause, age and region. Cause-specific mortality data from the MDS have informed the government’s policies and financial allocations on health areas including tobacco control and malaria; it has also raised attention to the issue of sex-selective abortions in the country. The running costs of the MDS are less than US $1 million per year, which includes all field costs and physician coding costs (the per death cost of field work and physician coding is about US $1), which makes this a reliable, reproducible, low-cost and sustainable model for mortality measurement in low-resource settings.
Both these cases illustrate the feasibility of improving existing health information systems in low-resource settings with the help of political priority by governments, innovative public and private sector financing, and most importantly, by emphasising a bottom-up approach of connecting with population data at its source. We must strive to develop and strengthen national health information systems that go beyond the measurement of global health goals such as the SDGs, but are also capable of supporting day-to-day management, long-term planning, and policy development for the entire national health system.
Genevie attended the Global Health and Mortality Summer Institute organised by the Dalla Lana School of Public Health and the Centre for Global Health Research at the University of Toronto, thanks to the generous funding provided by the Global Health Governance Programme at the University of Edinburgh. The case studies from India and Tanzania were kindly shared by Professor Prabhat Jha (Founding Director - Centre for Global Health Research) and Mr. Bhaskar Mishra (Child Protection Specialist – UNICEF Tanzania) at the Global Health and Mortality Summer Institute.