Health and demographic surveillance systems: a step towards full civil registration and vital statistics system in sub-Sahara Africa?

The main purpose of a civil registration and vital statistics system (CRVS) is to record the occurrence of key (or vital) events in the lives of the people in a given geographic area. In 2003, the United Nations defined these as comprising a live birth, death, fetal death, marriage, divorce, adoption, legitimisation of birth, recognition of parenthood, annulment of marriage, or legal separation [1]. The greatest advantage of a CRVS is that it is continuously collecting data and therefore, it is possible to study the changes in a population’s vital events over a period of time. The earliest CRVS was in the form of church or parish registries in Europe (as early as the 14th century) and Japan, which monitored births, deaths and marriages in separate registers. These evolved into civil registries or CRVS in the 19th and 20th centuries [1].

The best CRVS is mandatory, permanent and continuous. In addition, it should be complete, in that it registers all vital events and it must provide high quality data in a timely manner. Completeness can be measured by comparing the percentage of recorded vital events with an estimate that has been established independently for that population. Some of the best CRVSs can be found in developed countries; one example of such a CRVS is the national registry of England and Wales, established by a Parliamentary Act in 1837. In Australia, each state and each territory maintains its own registry. In the United States, the vital registry is the purview of each state and in some cases, of a city. Several countries in Central and South America have a complete CRVS. In Asia, Sri Lanka stands as the one of the few nations with comprehensive CRVS (more than 90% of the events covered) that is compulsory [1]. While some CRVSs are not complete, they have been established by law to maintain specific aspects of a registry. An example is the Maldives, where in July 1993, a law was enacted requiring all births and deaths to be certified and registered by the Ministry of Health [1‐4].

The information contained within the CRVS is a legal record that documents each vital event and defines a person’s entitlement to certain rights within a population. The statistics formed by aggregating these records reveal trends in the population. They can be used to evaluate the needs of a population so as to set up programmes that will meet these needs [1, 5, 6].

Vital statistics are used to monitor levels and trends in the fields of population, public health and administrative planning. For public health decision-makers and policy promulgators to be effective, they require access to timely and complete data on births, deaths, causes of deaths and illness in their jurisdictions and CRVS has proven to be a reliable source of these data. The data collected by CRVSs can also be used directly to monitor and evaluate the impact of any public health initiative that has been undertaken by governing bodies to improve the quality of health in their communities [1, 5‐9].

CRVS is either non-existent or incomplete in most of sub-Saharan Africa (SSA). In many instances, although vital registration is conducted, there is inefficiency in the compilation of vital statistics, resulting in the paucity or poor quality of the data generated. In countries like South Africa and Zimbabwe, studies of death were carried out using civil registration data of deaths in selected cities in each country [10]. In most cases, the deaths registered are not medically certified and in fact, deaths due to HIV/AIDS are often attributed to other causes [10]. As such, statistics regarding adult mortality in SSA are calculated largely using prediction models and estimation procedures. It should be noted that South Africa is making strides to generate fairly reliable vital statistics from its CRVS and Statistics South Africa (StatsSA) plays a leadership role in promoting the development of CRVS in SSA. The World Health Organization (WHO) states that there are only five African countries in which there are vital registries that collect data on more than 25% of the population. This means that most people in SSA are born, live, and die without any record of any of these or other events in their lives [6, 10‐12]. According to the United Nations (UN), few countries in SSA use civil registration as the source of the vital events data reported [13, 14]. According to UN Statistics, among the 14 countries which cite civil registration as the source of vital data, only four have more than 90% coverage of the events and these countries are mostly islands (Table 1).

The lack of an effective civil registration system (CRS) and CRVS presents immense challenges for evidence-based health and population policies. Political will, instability (war), lack of adequate resources, a culture of non-data-use, lack of an effective evidence based-decision making process are some of the issues that explain the lack of viable CRS and VRS in SSA [1, 7, 12, 15].

In most countries in SSA, CRVS is not the first priority in the political agendas and ministries do not approach registration systems in/when seeking data that they can use in their decision-making processes. The International Institute for Vital Registration and Statistics (IIVRS) found that the biggest disincentive to establishing and running CRVSs in the late 1970s was a lack of commitment on the part of the political class [7]. The critique on political will is caused by an over-reliance on the Ministries of Health to contribute to data collection and analysis. These ministries usually focus on service provision rather than on making investments on population-based data collection and analysis. Consequently, there is need for a national dialogue involving key stakeholders who use vital statistics in their various businesses.

For the past decades, many countries in SSA have experienced civil wars. Countries like Angola, Cote d’Ivoire, Ethiopia, Liberia, Mozambique, Namibia, Rwanda, Sierra Leone, South Africa (Apartheid), and Uganda have been engaged in civil wars. Others like Chad, DRC, Somalia, and the Southern Sudan and Sudan continue to be ravaged by civil wars. Wars in SSA have destroyed the infrastructure and depleted human resources through disabilities, death, and displacement. These conditions make it virtually impossible to run any kind of effective registration system particularly in the face of massive internal isplacement of people fleeing the wars [15]. However, irrespective of the civil wars in some of the countries, efforts are ongoing in many others to generate vital statistics.

CRVSs are expensive to establish and maintain. For example, in 2000, Chile projected the annual cost of its CRVS to be USD 45 million of which 80% was generated by the system itself. This was done by attaching fees to services like issuing certificates, drivers’ licenses, and passports. Taxes raised by Chile’s CRVS provided the other 20%. Allocation of financial resources is another factor to consider. In Tanzania, in the 1990s, USD 30 million was used to gather health data using various systems but very little of it (USD 700,000) was spent in maintaining its CRVS; no population estimates were generated from the data it collected [7]. While Chile has shown that it is possible to finance a CRVS, the Tanzanian example shows competing interests might derail the process. Financing and allocating resources for a CRVS hence need to be carefully thought out and Ministries of Finance and Planning should be effectively be brought on board at the outset.

There has been little evidence-based decision making in most of SSA for many reasons, one of which is that the available data may not be aligned with the political views or interests of potential users. If there was high demand for data for decision-making, there would be greater incentive to put in place a system such as CRVS to generate quality and reliable information [7]. This is however not the entire story. The lack of the requisite local capacity to collect, curate, analyse and publish data contributes significantly to desire not to collect and make no use of the data. It is therefore extremely important to introduce strategies that would cater not only for data collection but also for their preparation, analysis and use.

Given the above thorny issues that must be dealt with, it is therefore clear that it will be some decades before full registration systems are established in SSA [1, 12]. The need for a viable alternative that provides a reliable source of vital data is critical and the emerging health and demographic surveillance systems (HDSSs) in LMICs could go a long way in addressing the chronic data gaps in SSA.

An HDSS consists of monitoring demographic and health characteristics of a population living in a well-defined geographic area [16‐18]. The process starts with a baseline census followed by regular update of key demographic events (birth, death and migration) and heath events. Figure 1 presents this HDSS concept. A typical HDSS may also include registration of marriages, divorces, changes in status and household relationships and fertility estimates. Such additional information is critical for a better understanding of the health and demographic dynamics of the population under observation. With such a dynamic platform that can monitor new health threats, can track population changes through fertility and migration rates, and can measure the effect of interventions on communities, the HDSSs can be used to generate evidence to guide health policy making.

HDSSs provide complementary and in some cases intermediate data to the other well-known methods, such as national censuses, and demographic and health surveys (DHSs). The main difference is that whereas HDSSs collect data prospectively and longitudinally on a district, for instance, censuses and DHSs collect data on a cross-sectional basis at generally long intervals (10 years for the census and 3–5 years for DHS) on either the whole population or on large nationally representative samples. The complementarity of HDSS data lies in the fact that the platform can be used to evaluate the population impact of health interventions at a community level and produce health and population indicators for the periods between two censuses or DHSs. Furthermore, the HDSS population can be used as a sampling frame for other studies and surveys. For example, morbidity, livelihood, and household panel surveys and cohort studies can use HDSSs as sampling frames allowing the surveys to be carried out within the rounds of the HDSS.

Data on causes of death provide critical statistics for policy makers at every level of governance –from local to national – to prioritise and address the health needs of their people. However, this information is not readily available in most low-income countries. In high-income countries, information on cause specific mortality data is usually made available through CRVSs linked to the health care system [3]. Unfortunately, in low-income countries, most deaths occur outside the health care system, making it difficult for a CRVS to obtain medically certified cause of death [4]. To address this challenge, the Verbal Autopsy (VA) [19] method was included in HDSSs to provide valuable information on patterns of causes of death. Despite some concerns (low sensitivity and specificity for some diseases) regarding this approach, there is general agreement that VAs are the best alternative method available for low-income countries to understand cause of death patterns in the absence of national vital registration systems. A very good development in VA use is a shift from using physician coding to arrive at cause of death to an automated procedure like Inter VA [20‐22] which is cheaper, faster and more consistent. The INDEPTH Network is working together with WHO and other organisations to make faster progress in this area.

The establishment of HDSSs started in the 1940s in SSA, with the Pholela Health Centre in South Africa being the first site created in 1940. This is a Community Oriented Primary Health Care programme established by the South African Ministry of Health to prevent and treat diseases prevalent in rural Natal (such as tuberculosis, syphilis, smallpox, typhoid, and measles), and to inform health policy [23‐25]. Niakhar in rural Senegal is the second site and was created in 1962 under the Niakhar Project [26]. Unlike Pholela, it was independent from the national Ministry of Health. Among other contributions, Niakhar has provided insight into the effectiveness of several public health interventions [27‐32].

Pholela and Niakhar have paved the way for numerous health research stations across the continent. Among these, the most acclaimed worldwide are Danfa in Ghana [33], Kasongo in the Democratic Republic of Congo [34], Kilombero in Tanzania [35], Agincourt in South Africa [36], Nouna in Burkina Faso [37], and Navrongo in Ghana [38]. With the increasing need to tackle new pandemics such as HIV/AIDS, to address the African urban health crisis, while devising strategies for health planning at the district level, a new generation of HDSS-based field stations have recently emerged, including Hlabisa (now Africa Centre) in South Africa, Nairobi in Kenya, and Rufiji in Tanzania.

More HDSS sites have been established within the past 40 years in areas that do not have a functional VRS. They may have been instituted differently but they all track population and demographic changes within their research sites or surveillance areas. Today, there is an international network of independent research centres that run HDSSs mainly in SSA, Asia and Oceania. The INDEPTH Network (http://​www.​indepth-network.​org) was established in 1998 and currently has a membership of 36 member centres which run 44 HDSSs with 32 of the HDSS in SSA [6, 12, 17, 39]. In a recent editorial (These HDSS sites are presented as a series of cohort profiles which follows a similar structure to forthcoming profiles [40].

The INDEPTH Network provides an environment in which researchers from the HDSS sites can jointly collect data and compare findings mostly through working groups formed around specific research agendas. The network offers a unique platform covering a total population of over 2.4 million individuals in Sub –Sahara Africa. Details on population coverage per country and HDSS site are presented in Table 2.

Many SSA countries had centralised health management and information systems (HMISs) which focused on morbidity and mortality reporting. Data flew from individual health units to the district and national levels. In the early 1990s, the Ministries of Health tried to implement HMISs which focus on using information at the point of collection [41]. However, since HMISs provide data that are based on information from users of health services, they show part of the picture because of the low coverage and use of health care in SSA (Figure 2).

To complement HMIS, data generated from HDSSs can be used to provide population-based data that can inform health systems decisions and performance. This is because, as mentioned before, HDSSs monitor vital events and other socio-economic data at the community level.

They provide accurate information that can be used to promulgate evidence-based health policies and then to monitor the impact of these policies on populations [6]. For instance, a study that compared mortality estimates of deaths by passive government surveillance and active HDSS in Western Kenya found that although there was under-reporting of peri-natal deaths, the distribution of child deaths was similar across the two systems for children aged 1 – 59 months. The authors also found a significant difference in cause of death between the two systems with the civil registry under-reporting malaria and pneumonia and over-reporting measles [42].

Another example of the benefits of HDSS in determining health service needs is the Tanzania Essential Health Interventions Project (TEHIP) model of Rufiji and Morogoro in Tanzania. In 1996, both the national statistics and the district statistics for Morogoro indicated that malaria was the number one reason for visits to health facilities. This came as no surprise to national and district level health service planners. In that year’s district health plan, when health programmes were being implemented, malaria was not singled out as a priority. A sentinel HDSS in the rural district found that 30% of total mortality burden was due to malaria, and 45% of mortality in children younger than five was due to malaria. They also found that 46% of deaths occurred without prior visits to a health facility. Therefore, the district had underestimated both the number of people who visited health facilities for malaria and the disease burden of malaria. Because of the HDSS data, the district reallocated its resources in its health plans so that there was a five-fold increase in resources being allocated to malaria control and a 20-fold increase in resources to malaria control in children younger than five.

The HDSS was also instrumental in evaluating the impact of two programmes that were being implemented by the district, i.e. the Integrated Management of Childhood Illness (IMCI), a strategy that was devised by WHO, and the use of insecticide-treated bed nets (ITNs) [43, 44].

Centers running HDSSs are funded in several ways. Some are funded by private institutions, foundations and international public funds. There is an increasing number of them funded by in-country institutions such as universities. However, there is need to connect research at some of these HDSSs with national interests and priorities in order to increase interest from governments in supporting the HDSS sites and in using research findings from the systems. The focus on publishing results more in international peer-reviewed journals should be juxtaposed with a strategy to engage in national policy dialogue. The INDEPTH Network is trying to address this challenge by working with the centers to synthesise research results into policy briefs and specially written reports which address national priorities.

HDSSs are geographically localised, mostly in rural areas, and the population covered is in most cases not representative of the entire geographic area or is often small, to minimise operational costs. It has therefore been argued that one cannot use measures obtained from HDSS data to make inferences about other regions or at the national level. However, a historical look at the designing and validating of epidemiological surveillance in uncounted populations in Sweden concluded that “Unsubstantiated claims that local sub-national population data are “unrepresentative” or “only local” should not therefore predominate over likely representativity” [45]. Further, a comparison of HDSS and DHS approaches to mortality surveillance in Ethiopia revealed comparable overall mortality patterns. However, HDSS data were more sensitive to local variations and less subject to recall bias [46].

It has also been suggested that a population under such regular and intensive surveillance such as the one under an HDSS may have better health indicators compared to a population from another setting because the repeated measurement is seen as a sort of passive intervention. In addition, populations from HDSS areas are often exposed to studies that may provide interventions (e.g. clinical trials, testing an intervention, among others). Thus, indicators from HDSS sites should be extrapolated to other settings with caution. However, there is need for a detailed investigation to assess how different the population in an HDSS is, from the population in a non-HDSS setting [46]. It is also important to explore possible statistical techniques using data from different sources to correct for potential differences between HDSS populations and other populations.

As HDSSs continue to be run over a long period, a number of operational challenges and constraints are to be expected. For instance, respondent fatigue has been a case in point especially for HDSSs that update the population data at least three times a year. But HDSS managers have found solutions to this challenge since many of the systems have existed over decades and will continue to do so. Investigators from the community are used and the individuals are not always the same study subjects in nested projects within the HDSS. In addition, some centers provide health intervention to the community. For example, the African Population and Health Research Centre (APHRC) running the Nairobi HDSS provide free health camps to the population under surveillance twice a year.

Another challenge is the cost of operating the core HDSS. For example, a survey conducted by INDEPTH in 2009 put an average of USD 250,000 per year for an HDSS of 60,000 people with three update rounds per year in rural settings. This gives a per capita cost of USD 4.17. When this cost is compared with the value of the data generated and the platform created for a host of other studies, an HDSS can be considered cost effective. Funders who are reluctant to fund the surveillance infrastructure must be educated better on the utility of the platform. A few funders have started to agree to provide some support for running the core HDSS in INDEPTH projects they fund. Governments should also be made to appreciate the value of such systems and to contribute to the running cost of HDSSs in their countries. It is worth mentioning that Tanzania, Ghana and Burkina Faso are examples of countries where the Ministries of Health contribute in small but significant ways to the running of HDSSs in their countries. The INDEPTH Network will need to do more advocacies and the HDSS centre leaders will have to engage more in projects that will make their systems more relevant to the national agendas through direct engagement with the Ministry of Health.

HDSSs have faced challenges in providing timely data such as cause of death so that they can be used in annual sector reviews. This is largely due to the complex relational and longitudinal data generated by the HDSSs which require advanced demographic/statistical analyses, and less to do with the uninformed calls to share data irresponsibly. More capacity needs to be developed and strengthened in data management and analysis. Thanks to a few funders who are convinced about the role of INDEPTH in this area, the Network is focusing its energy to increase data management and analysis capacities in the South. INDEPTH contributed data to the Global Burden of Diseases Estimates by WHO and several datasets are now made freely available on the Network’s website: http://​www.​indepth-network.​org and an online repository (iSHARE) has been created to share HDSS data: http://​www.​indepth-ishare.​org.