Network analysis of demographics, dietary habits and health-related quality of life among northern Chinese population.

Introduction: Previous studies have linked demographics and dietary habits to health-related quality of life (HRQoL), but the interrelationships among these factors have not been extensively explored using network analysis. We aimed to describe network patterns of demographics, dietary intake and HRQoL in northern Chinese population. Methods: We conducted a population-based cross-sectional study using baseline survey data from the Shaanxi cohort of the Regional Ethnic Cohort Study in Northwest China, collected from June 2018 to May 2019 (n=32 110). HRQoL was assessed using the Short-Form Health Survey 12, dietary intake was evaluated via a validated semiquantitative food frequency questionnaire, and demographic information was collected. Mixed graphical models were used for network analysis. We derived centrality indices and evaluated network model's stability and accuracy. Results: Dietary foods such as beef, mutton, beans, aquatic products, potatoes, poultry and beans intake were the most central and bridge food groups in the dietary network. Significant complex interaction was found between HRQoL, key demographics and dietary intake. Physical Component Score (PCS) positively correlated with chronic diseases (0.45), education (0.26) and income (0.15), but negatively correlated with age (-0.25), occupation (-0.18), residence (-0.06), and sex (-0.06). Mental Component Score (MCS) had positive correlations with residence (0.20), age (0.11) and chronic diseases (0.10), but negative correlations with marital status (-0.17), education (-0.09) and income (-0.04). PCS had positive correlations with wheat (0.14), fresh fruits (0.13), beef (0.10) and mutton (0.09). MCS had positive correlation with oil (0.12), wheat (0.11) and tea (0.09), but negative correlations with rice (-0.08), carbonated beverage (-0.08), poultry (-0.07) and animal innards (-0.06). Among various factors, education, income, chronic disease and physical activity, along with multiple food items, exhibited the highest centrality indices. The network was stable (stability coefficient of 0.75) for all centrality measures. Conclusions: Identifying demographics and dietary factors with high centrality indices in multidimensional network may provide opportunities for enhancing HRQoL, suggesting potential avenues for health interventions.