Abstract
A secondary data analysis of the 2022 Philippine National Demographic and Health Survey (PNDHS) was conducted to explore the underlying structure of knowledge regarding communicable and noncommunicable diseases using multilevel confirmatory factor analysis (CFA). The PNDHS data consist of two levels: level-1 represents within-household data (household questionnaire), and level-2 represents between-household data (primary sampling unit (PSU)). Therefore, a two-level CFA and two-level variance CFA were performed. Furthermore, a multigroup analysis assessed the structural differences between males/females and urban/rural groups. In the PNDHS survey, 30,372 household interviews were completed. Knowledge levels for cancer, heart disease, diabetes, dengue fever, TB, and COVID-19 were 96.7%, 94.9%, 97.8%, 98.4%, 96.7%, and 92.8%, respectively. The two-level CFA indicated that the coefficient loadings of each item for both levels were statistically significant (Z-test, P < 0.001). Regarding two-level variance CFA, the variance at level-1 was higher than that at level-2 (13 and 6.7, respectively). The multigroup analysis revealed that the model was non-invariant (not equal) across gender and residence (likelihood ratio test; P < 0.001, P < 0.001, respectively). In conclusion, level-1 has greater effect than does level-2 because the variance in level-1 is greater than that in level-2, the knowledge of COVID-19 has the lowest loading compared to other items, and rural/urban areas and females/males exhibit different levels of health knowledge.