Human resources for health (HRH) are essential to health system functioning and to achieving universal health coverage. The global distribution of HRH is highly unequal: documented disparities include much higher physician density in Europe compared with sub-Saharan Africa and large gaps between high-income and low-income countries. Countries with fewer health workers are likely less resilient to shocks from infectious disease outbreaks. This study examined whether epidemic disasters are associated with short-term declines in HRH density and whether such declines are distributed unequally across regions and income groups.
The analysis combined two primary public datasets covering 194 countries and territories for the period 1990–2019. Epidemic-event counts were extracted from the Emergency Events Database (EM-DAT). Cadre-specific HRH density estimates (number per 10,000 population) were sourced from the Global Burden of Disease study. Additional country-level covariates—urban population proportion, population density, gross domestic product per capita, and the Socio-demographic Index (SDI)—were included from publicly available sources. The study used country-year records to derive associations between the number of epidemic events and HRH density in the subsequent year.
To estimate the association between epidemic occurrence and HRH density, the authors applied a Generalized Estimating Equation (GEE) model. The model adjusted for geographic and socioeconomic characteristics, including income level, SDI, urbanization, and population density. Using the country- and year-specific epidemic counts, observed HRH density, and the epidemic–HRH density association from the GEE model, the researchers calculated the annual average absolute number of HRH losses and the average relative number of HRH losses per 10,000 HRHs attributable to epidemic events at global, regional, and national scales.
Over the 1990–2019 period, EM-DAT recorded 1,185 epidemic events across the 194 countries and territories studied. The GEE model found that an increase of one epidemic event in the preceding year was associated with an average decrease of 57.5 × 10−3 HRHs per 10,000 population in the following year (95% confidence interval 18.5 to 96.4; p = 0.004). The authors report that, under the assumption that the observed associations were causal, this corresponded to substantive annual losses in the global health workforce attributable to epidemics.
Assuming causality, the study estimated a mean global annual loss of 17,549 HRHs (95% CI 5,661 to 29,437) attributable to epidemics, equivalent to 2.57 lost per 10,000 HRHs (95% CI 0.83 to 4.31). The burden was not evenly distributed:
These estimates summarize average losses attributable to epidemics over the study period and highlight substantial heterogeneity across settings.
The study used cadre-specific HRH density inputs from GBD (covering 16 types of health worker), and the main outcome measure reported was change in overall HRH density per 10,000 population. The author summary notes particular concern for nurses, midwives, and physicians when discussing workforce protection and retention, but the primary quantitative estimates presented in the abstract and main results focus on aggregate HRH losses attributable to epidemic events at country, regional, and income levels.
The authors acknowledge multiple limitations inherent to an ecological, country-level analysis. Potential sources of bias include uncontrolled confounding and reverse causation (for example, low HRH density could increase epidemic occurrence and be associated with measured epidemic counts). The authors also note potential underestimation of epidemic-attributable HRH losses in countries with weaker disaster-surveillance systems, and that country-level data cannot fully account for individual-level mechanisms such as health worker migration, endemic disease burden, or direct mortality among health workers. These limitations mean the reported attributable estimates should be interpreted cautiously, particularly regarding causal inference.
The study’s findings indicate that epidemics are associated with measurable declines in HRH density and that these declines disproportionately affect low- and middle-income countries and certain regions such as South Asia and sub-Saharan Africa. The authors suggest that epidemic preparedness and response planning should include targeted measures to protect and retain health workers—especially nurses, midwives, and physicians—in lower-resource settings. Protecting HRH may reduce the exacerbation of pre-existing workforce shortages following infectious disease outbreaks.
All datasets used (EM-DAT, GBD HRH estimates, World Bank indicators, and SDI) and the analysis code are reported to be publicly available via the cited repositories. The study was supported by grants from the National Natural Science Foundation of China and the Beijing Nova Program; the funders did not influence study design, analysis, or publication decisions. The authors declared no competing interests.