The importance of geographic and demographic data from census for locating and mapping vulnerable populations

1. Geographical census data

Geography is critical to the census and data collection exercises, as it contextualizes and provides deeper meaning to statistical data. A geographically framed census will ideally feature georeferenced and geocoded data collection. As defined in the United Nations Principles and Recommendations for Population and Housing Censuses, Revision 3, georeferencing refers to the process of compiling collected census data and geographically referencing these data by locating the physical space on a map that each datapoint specifically coordinates to. This process then provides methods of analysing the geographic characteristics of those data. These data can also be mapped to effectively relate statistical data to the geographic area to which the census results refer. The process of geocoding is finding associated geographic coordinates, typically latitude and longitude attributes, from other geographic data for a datapoint or address.1

Therefore, a geographically framed census allows for more robust and rich statistical analyses and exercises, such as survey design, sampling frame selection, and most critically, enables the mapping of census data for full visualization. With accurate and complete census data, more effective surveys can be designed to obtain critical data not included in census enumeration. Furthermore, it is critical that surveys, which are frequently limited in the amount of people they collect data on, target ideal sampling frames, representing the ideal demographics for that survey. The selection of these sampling frames is best informed from timely, accurate, and georeferenced census data. However, the sampling frame selection process can become less accurate as time increases between the census and the sampling exercise. As time lapses, age structures, urbanization levels and household composition derived from the census might have become outdated, creating an implicit bias in economic and social indicators. The loss of sampling frame effectiveness clearly heightens the importance of an accurate and complete decennial census.2

The full visualization of census data via mapping exercises is arguably the most important outcome of a geographically framed census. The incorporation of geospatial data has an exponential impact on the maps and analyses that countries can produce from their collected census data. It is with geocoded census data that analysts can pinpoint and cluster particularly marginalized or vulnerable populations based on a given demographic feature. It is that information that allows the identification of populations aged 60 and over that may be more susceptible to a disease than younger populations, clusters of young people in need of educational services, or entire communities that are outside of emergency medical care service areas. Georeferenced and geocoded census data can be used to conduct comparative geospatial analyses of key demographic indicators, including comparative analyses such as the evaluation of healthcare accessibility across geographic areas for a given demographic. With these maps and analyses, governments and development actors are able to locate the most vulnerable populations and strategically target policies and programs where they are most pertinent.3

Combined, the crude numbers of people and their geographic distributions provide the essential denominators for planning. It is with these denominators that governments can effectively allocate resources such as infrastructure and social welfare programmes, election planning, and market analysis. The absence of these data in planning exercises can gravely impact communities, especially those with vulnerable and marginalized populations. Frequently, this is evidenced in the over-or under-allocation of healthcare resources, which can severely hinder overall healthcare delivery. Inaccurate denominators of the underlying population will lead public health officials to miscalculate the number of hospital beds or healthcare workers needed for a given geographical area. These miscalculations could be potentially devastating especially in times of crisis, or in geographic areas with vulnerable populations that may not have the resources or capabilities to seek healthcare services beyond their geographic area.

2. Census and the health sector

A nation’s census is a critical component of all national health planning and preparedness exercises. Public healthcare services, including hospitals, healthcare clinics, and healthcare programs, especially need-based insurance programmes for lower-income individuals or for older persons, are all reliant upon census data. It is with these data that governments can implement informed decennial service planning and allocation exercises. Often, a lack of accurate, timely census data can cause biases in the availability and accessibility of services for certain population groups and geographical areas.

Therefore, rigorous data collection at the individual level, will enable public health managers to perform detailed diagnostics, conduct targeted healthcare forecasting exercises, and when needed, develop informed responses to healthcare emergencies. Through these analyses and exercises, public health managers can understand, and visualize, where vulnerable populations are located, such as: where elders or persons with disabilities live and what specialized healthcare services they may require; how many are served by a given healthcare facility and how many people currently reside with access to that facility; what are the demographic characteristics of a given population and what healthcare services are they served by; how many hospital beds and what special equipment may be required to address the populations surrounding a healthcare facility; etc. Within the healthcare domain, the answers to these questions, and the underlying data that inform them, can determine adequate or ineffective planning, frequently with life and death consequences.

3. COVID-19 pandemic

In 1864, cholera was raging and John Snow first decided to explore the use of geographic information to map its impacts and spread. Snow’s mapping of cholera was one of the earliest known examples of using geographic data to address a health epidemic.

As the world battles a different pandemic, the importance of demographic and cartographic data to confront a public health crisis has once again been illuminated. COVID-19 continues to spread and infiltrate communities, necessitating national responses to mitigate its spread.4 Globally, national responses have included social distancing measures, massive procurement of personal protective equipment (PPE) for both the general public and healthcare workers, vaccine development campaigns, clinical trials of available medications and drugs, among many other efforts. Critically, almost all of these responses require some input of geographic and demographic data to optimize their effect.

At the healthcare level, these data are critical for informing the supply side of care, i.e. determining the number hospitals and health care workers needed for the potential number of patients within a given community or geographical area. The effective planning with census data will help to minimize misallocation of healthcare workers and medical supplies. Without census data, healthcare facilities could be gravely understaffed or undersupplied and unable to treat their patient load, while other healthcare facilities could be overstaffed and oversupplied, but with low patient volume.

National responses have also had to factor in the populations with the highest risk of contracting and suffering severe medical complications from COVID-19, notably those over the age of 65 years with pre-existing medical conditions.5 The geographic distribution and mapping of these populations becomes critical for effective planning and preparedness. Without the geographic distribution of these populations, public health managers would be unable to effectively allocate resources to healthcare facilities that serve communities with higher vulnerable populations. Conversely, they could also over allocate resources to facilities that serve communities with lower risk profiles.

Beyond healthcare resource allocation, georeferenced census data, data that hold spatial information to identify the geographic location of phenomena across the Earth’s surface, have become critical to effective contact tracing programmes that many countries have implemented. These data provide a geospatial foundation for tracking movements and identifying populations that may be experiencing an acutely elevated risk and need to be isolated. Census data also provides national response officials with the ability to track the effects of the pandemic, such as the temporal effects in individual households and small businesses.

As global efforts are underway in developing a vaccine against COVID-19, the extent of the virus’ far-reaching impacts are far from known. The pandemic has caused significant disruption in almost every facet of life, including in social, economic, political dimensions, and even the 2020 census round. Notably, over 37% of censuses planned in the 2020 global round, which spans 2015–2024, are experiencing some form of delay or disruption. Further, over 80% of those scheduled to be conducted during the year 2020 are delayed or disrupted. As governments implemented lockdowns, social distancing measures, and other response efforts, most census authorities had to halt their planning, mapping, enumeration, and post enumeration activities.6

These delays and postponements are especially critical for those countries that conduct traditional censuses, such as those enumerating via face-to-face interviews and on a decennial life cycle. For those countries that conduct register-based or mixed methods censuses, the impacts of lockdowns or social distancing measures have been found to be less severe, as enumeration can largely proceed unimpeded. Furthermore, those countries that conduct their census every ten years, or have been unable to conduct or complete their census in recent rounds, will be reliant upon population projections to inform their national responses. As census data ages from the time of collection, these projections become less precise and could lead to the misallocation of critical healthcare resources.

However, even as the pandemic subsides enough for census activities to resume, any activity previously completed may be at risk of invalidity and incompleteness, particularly for those countries that were in cartography or enumeration phases. The pandemic may also negatively affect both the initial reporting, self-reporting online or by mail, and response to face-to-face visits. In ideal census environments, undercoverage or undercounting is a considerable concern for national authorities. Census undercoverage typically stems from the inability to reach vulnerable or hard-to-count populations such as indigenous, afro-descendants and migrants. Frequently, national authorities will conduct targeted awareness raising campaigns or other measures to ameliorate potential undercounts of traditionally hard-to-count populations. These targeted measures can include providing information on what a census is, why it could have specific importance for the targeted population, how to respond to questionnaires, translations of questionnaires or information on the census into local dialects, face-to-face interviews to collect data on non-respondents, etc. The pandemic is now increasing the risk of undercoverage in these populations, as well as in other populations such as the elderly. As people have become wary of opening their door for a stranger due to the risk of transmission, the completeness of data collected through enumerator’s home visits will likely suffer.7

Beyond incompleteness or invalidity, many countries that rely on traditional data collection methods for their census are now considering the possibility of changing their methods to adapt to the pandemic.8 This becomes particularly challenging in low and middle income countries (LMICs) where there may not be sufficient resources or technical capacity to build new collection infrastructures and procure requisite ICT equipment. However, according to surveys conducted by the World Bank and the United Nations Department of Economic and Social Affairs to monitor the impacts of the pandemic on statistical operations, the use of phone surveys, administrative data, and web surveys for data collection has grown substantially.9 Further, these surveys revealed that between May and July some countries had been able to resume some face-to-face interviews, with a decrease from 69% of NSOs reporting full office closures in May to only 31% reporting full office closures in July. This resumption of activity, along with the implementation of necessary safety measures, will be critical for those countries that cannot implement methodological changes.

The potential incomplete data collection caused by the pandemic could put entire communities at risk of a future public health crisis for the next decade. An undercount of the population will reduce funding – and therefore capacity – for vital healthcare resources. The financing and supplying of a hospital, with the accurate number of hospital beds and supplies to support the community it serves during a public health crisis, is predominantly reliant on accurate data. Further, the demographic composition of each community is critical for

determining the number of laboratory tests and vaccines that are needed, especially for those most vulnerable populations. If a vaccine is developed against COVID-19, a community with an undercounted population will not be allocated sufficient quantities.

As COVID-19 disrupts the 2020 census round, it is critical to acknowledge that the census is essential for national planning and preparedness, including for public health crises. The collection of georeferenced census data and conduct of spatiotemporal analyses and modeling are fundamental for nations to adequately prepare and address gaps in their healthcare system. However, the 2020 census round and subsequent national response and resilience initiatives are at risk, as COVID-19 jeopardizes the success of every nation’s census project and the rights of every individual to be counted.