Addressing Population Health Literacy: Why It Matters, Where Inequalities Emerge, and How To Intervene

다국어 초록 (Multilingual Abstract)

Health literacy (HLIT) is increasingly recognized as a critical public health capability that mediates individuals’ capacity to access, comprehend, and act on health information. Not only is HLIT important on the individual level, but it also affects population health outcomes and is related to health system efficiencies. However, prevailing research lacks a systematic method to analyze HLIT using population-level survey data and has often treated HLIT as a fixed individual attribute, overlooking its embeddedness within structural and contextual conditions. This study addresses this gap by introducing the Population Health Literacy Assessment through Multilevel Estimation (PHLAME) framework—an integrative analytic strategy designed to model HLIT as a socially stratified and spatially patterned outcome. Drawing on a nationally representative HLIT survey (N = 11,027), the study employs a sequential analytic design combining latent profile analysis, small-area prediction, and multilevel modeling. Individual HLIT scores were downscaled from metropolitan to district and county level using covariate-informed estimation to generate high-resolution spatial predictions. Area-level socioeconomic contexts were classified via latent profile analysis (LPA) of structural indicators (e.g., age dependency, basic livelihood support, educational attainment), yielding four district typologies: Affluent, Moderately Deprived, Severely deprived, and Average SES. Multilevel linear models nested individuals within 255 administrative districts to estimate the independent and interactive effects of place-based deprivation. Intraclass correlation analysis confirmed that 6.6% of HLIT variance was attributable to contextual differences. Incorporating LPA-defined SES profiles significantly improved model fit (ΔAIC = 4477, p < .001) and adding individual-level predictors and cross-level interactions further enhanced explanatory power (ΔAIC = 20, p < .001). Education and employment emerged as strong positive predictors of HLIT, while older age, female sex, and disability were negatively associated. Importantly, the effect of individual education was more pronounced in deprived districts, suggesting a structurally contingent return on personal resources. Findings provide robust evidence that HLIT is co-produced by individual characteristics and the broader socioeconomic environments in which people live. The PHLAME framework offers a scalable and policy-relevant template for mapping HLIT equity, identifying contextual leverage points, and guiding targeted, place-sensitive interventions to reduce HLIT disparities at the population level.

목차 (Table of Contents)

• CHAPTER 1. INTRODUCTION 1
• 1.1 Health literacy matters 1
• 1.1.1 Defining health literacy 1
• 1.1.2 Evolution of the Concept 1
• 1.1.3 Health Literacy as a Social Determinant of Health 2
• 1.2 Why, Where, and How 3
• 1.2.1 The Need for a Population-Level Approach 3
• 1.2.2 Research Gap: From Individual Data to Regional Inequalities 3
• 1.2.3 Aims and Objectives of This Thesis 4
• CHAPTER 2. LITERATURE REVIEW: Why Health Literacy Matters 5
• 2.1 Conceptual Background 5
• 2.2 Measuring Health Literacy 6
• 2.2.1 Instruments (HLS-EU-Q, TOFHLA, REALM, etc.) 6
• 2.2.2 KHPS and HLS-EU-Q16 in the Korean Context 6
• 2.2.3 Population Health Literacy Survey 7
• 2.2.4 Prediction Models and the Need for Small-Area Estimation 7
• 2.3 Health Literacy at the Micro Level 8
• 2.3.1 Health Literacy and Health Outcomes 8
• 2.4.1 Health Literacy and Population Health 9
• 2.4.2 Health Literacy and Health Systems 10
• 2.5 Health Literacy as a Social Determinant of Health 11
• 2.5. Nutbeam's Framework 11
• 2.5.2 Health Literacy as a Mediating Factor 11
• CHAPTER 3. DATA & METHODS 13
• 3.1 Research Questions 13
• 3.1.1 Part 1. Descriptive Epidemiology of HLIT 13
• 3.1.2 Part 2. Latent Health Literacy Profile 13
• 3.1.3 Part 3. Spatial Inequality in HLIT 14
• 3.1.4 Part 4. Contextual Determinants and Multilevel Dynamics 14
• 3.1.5 Part 5. Framework Validation and Policy Translation 15
• 3.2 Data & Sample Population 15
• 3.2.1 Korean National Health Panel Survey 15
• 3.2.2 Community Health Survey 16
• 3.2.3 Ethical Considerations 18
• 3.3 Measures 18
• 3.3.1 Health Literacy 18
• 3.4 Research Schematic 19
• 3.4.1 Population Health Literacy Assessment through Multilevel Estimation_PHLAME Framework 19
• 3.5 Research Methods 22
• 3.5.1 Exploratory Analysis 22
• 3.5.2 Welch's ANOVA & Games-Howell Post-hoc test 22
• 3.5.3 Multiple Linear Regression and Prediction Model 22
• 3.5.4 Latent Profile and Latent Class Analysis 23
• 3.5.5 Probabilistic downscaling to county and district level 24
• 3.5.6 Multilevel Linear Regression 26
• CHAPTER 4. RESULTS: Where Inequalities Emerge 34
• 4.1 The Vulnerable Population 34
• 4.1.1 Exploratory Analysis 34
• 4.1.2 Factors Associated with Health Literacy 42
• 4.1.3 Prediction Model of Health Literacy 44
• 4.1.4 Latent Profile Analysis 50
• 4.2 The Vulnerable Areas 62
• 4.2.1 Exploratory Analysis 62
• 4.2.2 Contextual Factors Related to Health Literacy 65
• 4.2.3 Area Level Socioeconomic Context: A Latent Class Analysis 65
• CHAPTER 5. DISCUSSION 80
• 5.1 Health Literacy in the Population: Patterns, Vulnerabilities, and Dimensions 80
• 5.2 Latent Profiles of Health Literacy: Typologies of Vulnerability and Strength 83
• 5.3 Spatial Inequality in Health Literacy: Mapping the Geography of Structural Disadvantage 85
• 5.4 Contextual Determinants and Multilevel Dynamics of Health Literacy 88
• 5.5 PHLAME Framework Validation and Policy Translation 89
• CHAPTER 6. CONCLUSION 91
• 6.1 Practice & Policy implications 91
• 6.2 Limitations & Future research 92
• 6.3 Final Remarks 92
• REFERENCES 94