Beyond knowledge: The role of digital health literacy and food environment in Iron Deficiency Anemia management among Bangladeshi university communities- A Structural Equation Modeling approach

Technical Summary: Beyond Knowledge: The Role of Digital Health Literacy and Food Environment in Iron Deficiency Anemia Management among Bangladeshi University Communities

Problem Statement
Iron Deficiency Anemia (IDA) remains a critical public health challenge in Bangladesh, particularly among university students and young adults. While factors such as imbalanced diets, hall-based catering, and low awareness of iron-rich foods are known contributors, the interplay between cognitive factors, digital capabilities, and structural environmental constraints is not fully understood. Existing literature suggests that knowledge alone is insufficient for behavior change; however, the specific mechanisms by which digital health literacy (the ability to find, evaluate, and use online health information) and the university food environment (accessibility, affordability, and quality of food) interact with health beliefs to influence dietary practices and productivity in this demographic require integrated analysis.

Methodology
The study employed a cross-sectional online survey design involving 2,600 participants (students and academic/administrative staff) across Bangladeshi universities. Data collection occurred between January and April 2026.

• Instrumentation: A structured questionnaire measured:
  • IDA Knowledge: A 7-item score (0–7) covering definition, causes, and sources.
  • Health Belief Model (HBM) Constructs: Perceived susceptibility, severity, benefits, barriers, and self-efficacy using Likert-type items.
  • Digital Health Literacy: Measured via the validated eHealth Literacy Scale (eHEALS).
  • Food Environment: Assessed across four dimensions: physical access, economic access, institutional quality, and cultural/social environment.
  • Outcomes: Dietary practices, self-reported IDA symptoms, and academic/work productivity.
• Statistical Analysis: The study utilized Structural Equation Modeling (SEM) with Stata 18 to test a conceptual framework where IDA knowledge influences dietary practices both directly and indirectly through HBM constructs. Digital health literacy and food environment were modeled as moderators and direct predictors.
  • Validation: Reliability was assessed via Cronbach's alpha, and construct validity via Confirmatory Factor Analysis (CFA) using Average Variance Extracted (AVE) and Composite Reliability (CR).
  • Hypotheses: The model tested mediation (H1), moderation by digital literacy (H2) and food environment (H3), and the predictive chain of diet to symptoms to productivity (H4).
  • Robustness: Multi-group analysis compared students vs. staff, and sensitivity analyses excluded diagnosed IDA cases and supplement users.

Key Results

• Sample Characteristics: The mean age was 25.1 years. 22.4% reported a prior IDA diagnosis, and 27.7% reported chronic digestive issues.
• Measurement Validity: Most scales demonstrated good to excellent internal consistency (Cronbach's $\alpha$ range: 0.73–0.96), with the exception of the productivity scale ($\alpha$ = 0.64). CFA confirmed satisfactory convergent validity and reliability for all latent constructs.
• Structural Pathways (H1 & H4):
  • IDA knowledge was positively associated with all HBM constructs (perceived severity, benefits, self-efficacy) and inversely with perceived barriers.
  • Knowledge influenced dietary practices both directly and indirectly through HBM constructs. The indirect effect was significant ($B=0.11, p<0.001$).
  • Dietary practices significantly predicted IDA symptoms ($\beta=0.22$), which in turn predicted productivity ($\beta=0.22$). The indirect path from diet to productivity via symptoms was significant ($B=0.25, p<0.001$).
• Direct Predictors and Moderation (H2 & H3):
  • Digital Health Literacy: Directly promoted healthier dietary practices ($\beta=0.14, p<0.001$). However, it significantly weakened the direct relationship between IDA knowledge and dietary practices ($\beta=-0.09, p<0.001$), suggesting that high digital literacy may introduce diverse information that dilutes the direct impact of specific factual knowledge on behavior.
  • Food Environment: Directly promoted healthier dietary practices ($\beta=0.19, p<0.001$). The hypothesized moderation of self-efficacy by the food environment was not statistically significant ($p=0.063$).
• Group Differences: Multi-group analysis revealed that the path from dietary practices to IDA symptoms was significant for students but not for staff. Digital health literacy significantly predicted dietary practices for students but not staff, whereas the food environment had a stronger effect on staff than on students.
• Model Fit: The full model showed acceptable fit for core structural paths (CFI=0.99, RMSEA=0.11) prior to outcome variables, but the full model including outcomes showed lower fit indices (CFI=0.82, RMSEA=0.16), attributed to the use of composite scores with heterogeneous response scales.

Key Contributions

1. Integrated Framework: The study is among the first to simultaneously examine IDA knowledge, HBM constructs, digital health literacy, and the food environment within a single Structural Equation Model in a Bangladeshi university setting.
2. Digital Literacy Nuance: It identifies a complex role for digital health literacy: while it independently promotes healthy eating, it acts as a negative moderator on the knowledge-behavior link, suggesting that high digital literacy may complicate the translation of specific medical knowledge into action due to information overload or conflicting sources.
3. Productivity Link: The research empirically establishes a pathway linking dietary practices to reduced IDA symptoms and, subsequently, to improved academic and professional productivity.
4. Demographic Specificity: It highlights distinct mechanisms between students and staff, noting that staff are more influenced by the food environment, while students are more responsive to digital health literacy cues.

Significance and Claims
The paper claims that managing IDA in university communities requires moving beyond simple knowledge dissemination. The findings suggest that interventions must be multi-faceted:

• Self-Efficacy: Strengthening self-efficacy and perceived susceptibility is crucial, as these HBM constructs mediate the effect of knowledge on behavior.
• Environmental Support: Improving the university food environment (accessibility and affordability of iron-rich foods) is a direct driver of dietary improvement, acting independently of individual knowledge levels.
• Digital Strategy: While digital health literacy is beneficial, interventions must account for its potential to dilute the direct impact of specific knowledge, implying a need for curated, credible digital content rather than just access to information.

The authors conclude that while the cross-sectional design limits causal inference, the robustness checks and large sample size support the validity of the proposed structural relationships. They advocate for future multi-component interventions that simultaneously target digital competencies, health beliefs, and institutional food environments to reduce the IDA burden and improve functional productivity in this population.