Factors that influence the presence of diabetes in patients at the Víctor Ramos Guardia hospital – 2021

Abstract

The study was conducted with patients with diabetes, an indicator of health in developing nations based on biological, social and cultural factors, in order to understand the determinants of the presence of diabetes. Factors significantly associated with diabetes include biological aspects, such as age (p = 0.006 < 0.01) and weight (p = 0.025 < 0.05); cultural factors, such as educational level (p = 0.017 < 0.05) and dependency (p < 0.05); and socioeconomic factors, highlighting marital status (p = 0.033 < 0.05) and income (p = 0.011 < 0.05). The results also indicate significant associations between diabetes and variables such as age (p = 0.006), weight (p = 0.025), educational level (p = 0.017), marital status (p = 0.033), income (p = 0.011) and dependency (p < 0.05). The probability of the patient with diabetes using the binary logistic regression model was: z = 2.618 + 1.00X1 + 1.327X2 + 2.572X3 + 1.639X4 1.034X5 + 2.800X6. The predictive model obtained was validated using the Hosmer-Lemeshow and Desviance tests, demonstrating its adequacy by not presenting significant discrepancies between the observed data and the predictions. With an optimal cut-off point of 0.3125, a correct classification rate of 77.7% was achieved, with a specificity of 78.4% and a sensitivity of 83.5%. The surface under the ROC curve (0.8524; 95% CI: 0.8181-0.9067) reflects an excellent discriminative capacity. These findings highlight the importance of addressing the identified risk factors to improve diabetes prevention and control strategies in vulnerable populations.