Exploring Variability in EQ VAS Scores: A Systematic Review and Meta-Regression of Population Health Studies
Author(s)
Ling Jie Cheng, PhD1, You-Shan Feng, PhD2, Jing Ying Cheng, BSN (Hons)3, Le Ann Chen, MPH4, Nan Luo, PhD4.
1Postdoctoral Fellow, University of Oxford, Oxford, United Kingdom, 2Medical University of Tübingen, Germany, Germany, 3Khoo Teck Puat Hospital, Singapore, Singapore, 4National University of Singapore, Singapore, Singapore.
1Postdoctoral Fellow, University of Oxford, Oxford, United Kingdom, 2Medical University of Tübingen, Germany, Germany, 3Khoo Teck Puat Hospital, Singapore, Singapore, 4National University of Singapore, Singapore, Singapore.
OBJECTIVES: The EQ Visual Analogue Scale (VAS), included in all EQ-5D instruments, is widely used to measure overall health status in population surveys. While its psychometric properties are well established in general populations, qualitative studies suggest substantial interpretative variability. This systematic review aimed to identify factors contributing to global variation in mean EQ VAS scores.
METHODS: We searched eight databases for English-language observational studies reporting EQ VAS scores in general population surveys, from inception to 24 January 2025. From each study, we extracted ten pre-specified variables: (1) comorbidity prevalence, (2) mean age, (3) gender ratio, (4) publication year, (5) EQ-5D version (3L/5L), (6) region (Europe, Americas, Asia, Oceania, Middle East), (7) sampling method, (8) administration mode, (9) survey venue, and (10) language. Regions followed UN Statistics Division classifications. We conducted random-effects meta-analysis, subgroup analysis, and meta-regression using the metafor package in R.
RESULTS: Of 16,525 records screened, 55 estimates from 51 studies were included, covering 24 language versions and 496,531 adults. Most data came from Europe (45.5%), South-East/East Asia (23.6%), and the Americas (18.2%). The average age was 46.3 years (SD 6.4). Surveys commonly used probabilistic sampling (72.7%) and interviewer administration (65.5%), especially household-based (87.3%). Univariate analyses identified comorbidity prevalence, age, region, administration mode, and language as significant contributors. In adjusted models, only comorbidity prevalence (β = -10.0, 95% CI -18.1 to -1.9) and the South-East/East/South Asia region (β = 7.6, 95% CI 1.4 to 13.9) remained significant, explaining 51.6% of variability.
CONCLUSIONS: Region and comorbidity prevalence explained most of the variation in EQ VAS scores. The difference between Asian and European populations exceeded the 5-point minimally important difference, suggesting possible ceiling effects. These findings highlight the EQ VAS’s potential for cross-regional comparisons and the need for more data from underrepresented regions to improve generalisability.
METHODS: We searched eight databases for English-language observational studies reporting EQ VAS scores in general population surveys, from inception to 24 January 2025. From each study, we extracted ten pre-specified variables: (1) comorbidity prevalence, (2) mean age, (3) gender ratio, (4) publication year, (5) EQ-5D version (3L/5L), (6) region (Europe, Americas, Asia, Oceania, Middle East), (7) sampling method, (8) administration mode, (9) survey venue, and (10) language. Regions followed UN Statistics Division classifications. We conducted random-effects meta-analysis, subgroup analysis, and meta-regression using the metafor package in R.
RESULTS: Of 16,525 records screened, 55 estimates from 51 studies were included, covering 24 language versions and 496,531 adults. Most data came from Europe (45.5%), South-East/East Asia (23.6%), and the Americas (18.2%). The average age was 46.3 years (SD 6.4). Surveys commonly used probabilistic sampling (72.7%) and interviewer administration (65.5%), especially household-based (87.3%). Univariate analyses identified comorbidity prevalence, age, region, administration mode, and language as significant contributors. In adjusted models, only comorbidity prevalence (β = -10.0, 95% CI -18.1 to -1.9) and the South-East/East/South Asia region (β = 7.6, 95% CI 1.4 to 13.9) remained significant, explaining 51.6% of variability.
CONCLUSIONS: Region and comorbidity prevalence explained most of the variation in EQ VAS scores. The difference between Asian and European populations exceeded the 5-point minimally important difference, suggesting possible ceiling effects. These findings highlight the EQ VAS’s potential for cross-regional comparisons and the need for more data from underrepresented regions to improve generalisability.
Conference/Value in Health Info
2025-11, ISPOR Europe 2025, Glasgow, Scotland
Value in Health, Volume 28, Issue S2
Code
P9
Topic
Epidemiology & Public Health
Topic Subcategory
Public Health
Disease
No Additional Disease & Conditions/Specialized Treatment Areas