An index of access to essential infrastructure to identify where physical distancing is impossible

Limited infrastructure poses a significant challenge for many low- and middle-income countries (LMICs), particularly in Africa. Inadequate infrastructure, such as electricity and roads, hinders sustained economic growth at the macro-level. At the micro-level, millions lack access to basic private infrastructure like clean water, adequate living space, and improved sanitation. This lack of infrastructure also compromises a nation's ability to prevent and control infectious disease outbreaks, as exemplified by the COVID-19 pandemic. By the end of 2021, the pandemic had resulted in over 288 million confirmed cases and more than 5.4 million deaths globally. While Africa reported fewer cases and deaths than other regions, this underreporting is partly due to limited testing facilities and potentially high infection rates obscured by a lack of testing. Low vaccination rates further exacerbate the issue. Containing the spread of the virus and preventing future variants requires both vaccination campaigns and investments in infrastructure that enable households to mitigate infection risks, preventing the need for devastating long-term measures like prolonged school closures. In response to the pandemic, many African countries implemented public health policies, including physical distancing regulations. However, these regulations proved less effective due to limited infrastructure, which hindered the ability of populations to maintain sufficient physical distance, despite significant economic costs associated with lockdowns and restrictions. This study aims to understand the geographic distribution of critical infrastructure patterns in Africa to better respond to current and future epidemics and pandemics. It seeks to measure a country's preparedness to prevent, detect, and manage infectious disease outbreaks, focusing on how limited private infrastructure restricts the effectiveness of physical distancing measures.

Existing indices attempt to measure a country's capacity to respond to infectious disease outbreaks, but mostly focus on health and governance systems rather than household capabilities for physical distancing. The WHO's monitoring of International Health Regulations (IHR) assesses a country's health system capacity across 13 dimensions; however, African countries lag behind. The Infectious Disease Vulnerability Index (RAND Corporation) considers seven dimensions influencing a country's vulnerability, revealing that 22 of the 25 most vulnerable countries are in Africa. Both indices provide only country-level aggregates, neglecting within-country variations. Brown et al. attempted to address this by creating a Home Environment for Protection (HEP) index, but it differs from the proposed study in several key aspects: indicators, level of aggregation (national versus subnational), weighting scheme (equal weighting versus PCA), and consideration of population density. The proposed Physical Distancing Index (PDI) in this study complements existing indices by addressing these limitations and focusing specifically on household-level infrastructure limitations related to physical distancing.

This study proposes a Physical Distancing Index (PDI) to measure the capacity of households to maintain physical distance, accounting for both infrastructure limitations and population density. The PDI is built using principal component analysis (PCA) on five indicators derived from Demographic and Health Surveys (DHS): lack of private toilets, lack of private drinking water, lack of ICT infrastructure (access to mobile phones), lack of private transportation, and lack of space (people per sleeping room). Data for 34 African countries and 519 first-level subnational regions were used. Population density data from the Socioeconomic Data and Applications Center (SEDAC) were integrated to weight the PDI, reflecting the increased risk of infection in densely populated areas with limited private infrastructure. Bayesian distributional regression was used to generate pixel-level (5 x 5 km grid) PDI estimates for selected countries, providing high-resolution risk maps. The PDI is normalized between 0 and 1, with higher values indicating a lower capability for physical distancing and a higher risk of disease transmission. The study compares the PDI with existing indices (IHR, Infectious Disease Vulnerability Index, and HEP) and GDP per capita, analyzing correlations and identifying discrepancies. The relationship between the PDI and reported COVID-19 cases at the subnational level in nine countries was also examined.

The spatial analysis of the PDI reveals considerable heterogeneity across Africa, with high-risk areas concentrated in Western Africa (Ghana, Gambia, Togo, Sierra Leone, Benin, Liberia, Senegal, and Côte d'Ivoire). These regions combine high population densities with limited infrastructure for physical distancing. Even within countries, considerable spatial heterogeneity exists, with high-risk hotspots identified in various regions (e.g., Western Kenya, Southern/Central Côte d'Ivoire, North-Western Tanzania, and North-Eastern South Africa). Pixel-level PDI maps for Ghana, Ethiopia, Kenya, and South Africa further highlight subnational heterogeneity. A strong positive correlation (0.4 to 0.9) was observed between the PDI and the number of reported COVID-19 cases at the subnational level in nine countries, indicating the PDI's ability to identify potential disease hotspots. Analysis of individual indicators shows various infrastructure challenges across countries: sanitation (Ghana, Liberia), water (Rwanda, Burundi), overcrowding (Gambia, Senegal), communication/transportation (Madagascar, Democratic Republic of Congo). A weak negative correlation was found between the PDI and GDP per capita, indicating that economically advanced countries have a lower risk of disease transmission, but the relationship is non-linear with high heterogeneity among poorer countries. The PDI demonstrates limited correlation with other indices focusing on health systems or general vulnerability, highlighting its unique contribution. Despite South Africa's better infrastructure compared to other African nations, its high COVID-19 case numbers underscore the importance of population behavior and other factors beyond infrastructure.

This study's findings highlight the importance of access to essential private infrastructure for effective physical distancing and disease control. Many African households lack such infrastructure, limiting the effectiveness of governmental regulations and contributing to high disease transmission risks despite costly national measures. The high correlation between the PDI and COVID-19 cases underscores the PDI's value in identifying hotspots for targeted interventions, such as vaccination campaigns and infrastructure investments. The weak correlation with existing indices emphasizes the PDI's unique contribution in assessing household-level capacity for physical distancing. The findings suggest the need for short-term measures like providing basic safety materials and prioritizing vaccination for vulnerable populations in high-risk areas. Long-term solutions involve addressing the lack of essential infrastructure to improve public health and quality of life.

This research demonstrates the critical role of access to essential private infrastructure in preventing infectious disease transmission. The developed PDI provides a valuable tool for identifying high-risk areas within countries, enabling targeted interventions. Future research should investigate the role of population behavior, integrate additional infrastructure indicators, and explore the PDI's applicability to other infectious diseases. Addressing the infrastructure crisis in Africa is crucial not only for pandemic preparedness but also for improving overall quality of life.

The study has several limitations: limited indicators beyond water, sanitation, space, communication, and transportation; exclusion of factors like virus importation probability and the effect of specific policies; absence of data on population behavior; inability to capture small-scale heterogeneity due to data limitations; and missing data for some countries. While the PDI effectively identifies subnational hotspots, it cannot predict national-level outbreaks independently. The study's results are limited to the available data and may not capture all nuances of the complex interplay of factors influencing disease transmission.