Structural equation modeling to identify the direct and indirect risk factors of diabetes in adults: Findings from a national survey

Abstract

BackgroundFew previous studies have investigated the multiple pathways that contribute to diabetes mellitus (DM) because of the complex, simultaneous interplay of attributing covariates. Structural equation modelling (SEM) is a robust multivariate approach that measures both direct and indirect effects of variables by simultaneously utilizing several regression equations. The current study applied SEM to test a hypothesized model of the covariates affecting DM among the adult population of the Sultanate of Oman. MethodsData from a large nationally representative 2017 WHO STEPwise approach to surveillance survey were analyzed. Stata 16 software was used to perform SEM and path analysis of the sociodemographic, behavioral, anthropometric, and metabolic variables affecting normoglycemia and DM. A priori factor structure was hypothesized with special emphasis on observing direct and indirect effects, and the correlations that defined them. ResultsEight paths that directly affected DM status were established based on eight sociodemographic, metabolic, and behavioral variables (age, sex, educational status, physical activity level, body mass index, waist-to-hip ratio, systolic blood pressure, and family history of DM). The remaining variables (marital status, employment status, smoking, high-density lipoprotein level, total blood cholesterol level, fruit and vegetable intake, and type of oil used for cooking) showed variable indirect effects. ConclusionsThe results of this study further reinforce the evidence that lifestyle changes are vital for the prevention and control of DM. Individuals with a family history of DM and a high waist-to-hip ratio comprise a high-risk group and should be targeted with screening and lifestyle-intervention programs.