Structural Equation Modeling (SEM) of Cysticercosis in School-Aged Children in Tibetan Rural Farming Areas of Western China: Implications for Intervention Planning

Abstract

Neurocysticercosis (NCC) significantly contributes to morbidity in developing countries. We recently published a study of prevalence and risk factors in school-aged children in three mountainous areas in Sichuan province of western China. Using structural equation modeling (SEM) on data from that study to guide intervention planning, here we examine risk factors grouped into three broad interventional categories: sociodemographics, human behavior, and sources of pork and pig husbandry. Because neuroimaging is not easily available, using SEM allows for the use of multiple observed variables (serological tests and symptoms) to represent probable NCC cases. Data collected from 2608 students was included in this analysis. Within this group, seroprevalence of cysticercosis IgG antibodies was 5.4%. SEM results showed that sociodemographic factors (β = 0.33, p < 0.05), sources of pork and pig husbandry (β = 0.26, p < 0.001), and behavioral factors (β = 0.33, p < 0.05) were all directly related to probable NCC in school-aged children. Sociodemographic factors affected probable NCC indirectly via sources of pork and pig husbandry factors (β = 0.07, p < 0.001) and behavioral variables (β = 0.07, p < 0.001). Both sociodemographic factors (β = 0.07, p < 0.05) and sources of pork and pig husbandry factors (β = 0.10, p < 0.01) affected probable NCC indirectly via behavioral variables. Because behavioral variables not only had a large direct effect but also served as a critical bridge to strengthen the effect of sociodemographics and sources of pork and pig husbandry on probable NCC, our findings suggest that interventions targeting behavioral factors may be the most effective in reducing disease.