Modeling geostatistical incomplete spatially correlated survival data with applications to COVID-19 mortality in Ghana

Spatial Survival Model for COVID-19 in México

“…According to the DIC and WAIC, models incorporating a spatial term have a better predictive fit than Cox models without a frailty term. This is because classical survival models cannot account for the spatial correlation of the data and, therefore, should not be considered as the standard model in research when geographic information is available [12,16]. These results are in agreement with those reported by Thamrin et al [20], who compared a Bayesian spatial survival model with a nonspatial one to analyze the factors influencing the survival of dengue patients and found that the spatial model was more appropriate.…”

“…As of 24 November 2021, 176,268 confirmed cases of COVID-19 were registered in the State of México. The municipalities with more than 10,000 cases were located in the east and center of the state: Ecatepec de Morelos (22,613 cases), Nezahualcóyotl (16,992 cases), Toluca (13,647 cases), Naucalpan de Juárez (12,450), and Tlalneplantla de Baz (10,605 cases). The municipalities with less than 20 cases were located in the southeast of the state, including Zacazonapan (18 cases), Ixtapan del Oro (14 cases), and Otzoloapan (12 cases) (Figure 3).…”

“…Healthcare 2024, 12, x FOR PEER REVIEW 8 of 21 cases), Toluca (13,647 cases), Naucalpan de Juárez (12,450), and Tlalneplantla de Baz (10,605 cases). The municipalities with less than 20 cases were located in the southeast of the state, including Zacazonapan (18 cases), Ixtapan del Oro (14 cases), and Otzoloapan (12 cases) (Figure 3).…”

“…Statistical models that take geographic location into account are increasingly used in survival analysis [12]. This is because they (i) incorporate spatial variation in survival times [13]; (ii) implement the Bayesian approach for parameter estimation and the new tools of Geographic Information Systems [13][14][15]; (iii) surpass the classical models by reducing the bias in the estimates and incorporating geographic information of the data [16,17]; (iv) introduce information about the spatial locations of the data, which plays an important role in predicting survival as it serves as a proxy for unmeasured regional characteristics such as access to medical care [13,18]; and (v) provide relevant information for public health decision making, such as the extent and direction of disease spread or the locations of disease hotspots, allowing control measures to be more effective and facilitating the adequate allocation of health resources [19].…”