Municipality-Level Predictors of COVID-19 Mortality in Mexico: A Cautionary Tale

Abstract

Objective Local characteristics of populations have been associated with COVID-19 outcomes. We analyze the Municipality-level factors associated with a high COVID-19 mortality rate of in Mexico. Methods We retrieved information from cumulative confirmed symptomatic cases and deaths of COVID-19 as of June 20th, 2020 and data from most recent census and surveys of Mexico. A negative binomial regression model was adjusted, dependent variable was the COVID-19 deaths and the independent variables were the quintiles of the distribution of sociodemographic and health characteristics among the 2,457 Municipalities of Mexico. Results Factors associated with high MR of COVID-19, relative to Quintile 1 were; diabetes and obesity prevalence, diabetes mortality rate, indigenous population, economically active population, density of economic units that operate essential activities and population density. Among factors inversely associated with lower MR of COVID-19 were; high hypertension prevalence and houses without drainage. We identified 1,351 municipalities without confirmed COVID-19 deaths, of which, 202 had high and 82 very high expected COVID-19 mortality (Means=8 and 13.8 deaths per 100,000 respectively). Conclusion This study identified Municipalities of Mexico that could lead to a high mortality scenario later in the epidemic and warns against premature easing of mobility restrictions and to reinforce strategies of prevention and control of outbreaks in communities vulnerable to COVID-19.

Fatores de risco e distribuição espacial dos óbitos por COVID-19: revisão integrativa

“…18,20,23,[29][30][31] As for the spatial analysis, ten studies were selected, which investigated which risk factors were associated with deaths in the studied areas and in certain populations. 1,19,24,29,30,[33][34][35][36][37] In these investigations, different techniques were used, among them: empirical Bayesian estimate (EBE); 37 geographically-weighted random forest (GW-RF); 1 geographically-weighted regression (GWR); 33 spatial Durbin model; 30 Least Absolute Shrinkage and Selection Operator (LASSO); 24 space-time scanning techniques (discrete Poisson model); [34][35][36] spatial correlation (Moran's Index); 29,30,33,37 and Pearson's correlation. 29,34,37 With the use of such tools, a heterogeneous distribution of deaths and/or mortality rates was evidenced, with socioeconomic and environmental conditions 1,19,29,30,33 and population density 24,29,30,35,37 being explanatory factors for the occurrence of events in these territories in space and space-time.…”

“…47 Other studies have also shown other health conditions associated with deaths, such as neurological, 20,32 respiratory, kidney diseases 12,27,38 and obesity. 20,21,35 In addition to biological and clinical factors, other factors that influenced mortality due to COVID-19 were socioeconomic, environmental and spatial distribution of the disease in the area of residence. Research conducted in 209 countries around the world, 29 in which spatial analysis techniques were used, such as global and local Moran's Index, in addition to multivariate logistic regression, identified the association of economic factors and population density related to mortality from the disease, establishing that mortality rates were associated with low economic level and higher population density in low-and middle-income countries.…”

“…33 It is also noteworthy that, through spatial scan statistic, significant clusters of high relative risk were detected statistically significant for the occurrence of deaths, and priority locations for health interventions were indicated by management and health services and systems, namely the city of Manaus, in the state of Amazonas (Brazil) 34 , and in the states of Oaxaca, Yucatán and Sonora (Mexico). 35 This technique has been routinely used in several ecological studies carried out around the world dealing with communicable diseases, as it allows the identification of clusters, whether of low or high relative risk, of a given event in space and space, calculating radii in which values maximize the likelihood function related to the total number of cases observed. 49 Other significant clusters for mortality were identified through spatial analysis of the global and local Moran's Indexes in border counties in the Phoenix area of Arizona (USA), 30 state of West Bengal (India).…”

Application of Data Science for Cluster Analysis of COVID-19 Mortality According to Sociodemographic Factors at Municipal Level in Mexico

“…Ríos et al found that patients lived in municipalities with the highest overcrowding had a higher risk of dying from COVID-19 as compared to those who lived in a municipality with low overcrowding [63]. Likewise, Contreras et al and Villa et al reported population density as a factor associated to higher mortality rates and adverse results for COVID-19, respectively [64,65].…”

Impacto de los determinantes sociales de la salud sobre la tasa de contagios por COVID-19

“…diversas políticas se han adoptado para contener su propagación, sin embargo, determinantes en salud como pobreza, desempleo, inseguridad laboral, condiciones laborales, soporte social, afectan la evolución de la enfermedad. Además, factores como pertenecer a población indígena, población económicamente activa, dedicarse a actividades económicas esenciales y densidad de población, se han asociado a altas tasas de mortalidad por COVID-19 así como la prevalencia de enfermedades crónicas (2).…”