Assessing Known Groups and Convergent Validity of the EQ-5D-5L in Assessing the Health-Related Quality of Life of Individuals With Obesity Who Have and Have Not Received Bariatric Surgery
Author(s)
Jia Jia Lee, BSc, MPH1, Asim Shabbir, MBBS (PAK.), MMed (S'pore), FRCS (Edin.)2, Nan Luo, PhD3.
1Saw Swee Hock School of Public Health, National University of Singapore, Singapore, Singapore, 2Department of Surgery, National University Hospital, Singapore, Singapore, 3National University of Singapore, Singapore, Singapore.
1Saw Swee Hock School of Public Health, National University of Singapore, Singapore, Singapore, 2Department of Surgery, National University Hospital, Singapore, Singapore, 3National University of Singapore, Singapore, Singapore.
OBJECTIVES: Asians face higher cardiovascular risk and qualify for bariatric surgery (BS) at lower BMI. Given the association between BMI and health-related quality of life (HRQoL), we examined the known-groups and convergent validity of EQ-5D-5L among Asians with and without BS in Singapore.
METHODS: We recruited obese adults from weight management clinics, including those eligible for BS and those who underwent BS at least 6 months ago. Participants completed EQ-5D-5L, selected BODY-Q subscales, PROMIS overall health and QOL questions, and demographic questions. BS participants also completed QOLOS excess skin scale. Known-groups validity was assessed using Cohen’s d by comparing EQ-5D-5L scores (EQ VAS and EQ-5D-5L index) across BS status, sleep quality, overall health, and QOL. Convergent validity was examined using Pearson’s correlations (r) between EQ-5D-5L scores and BODY-Q and/or QOLOS.
RESULTS: We analyzed 80 participants (non-BS: 43; BS: 37). VAS consistently distinguished between known groups, showing significant differences with large effect sizes by BS status (Cohen’s d=1.15), sleep quality (non-BS: d=1.00; BS: d=0.87), overall health (non-BS: d=1.37; BS: d=1.55), and overall QOL (non-BS: d=0.75; BS: d=0.90). Index showed smaller or insignificant differences in several comparisons among non-BS participants (non-BS: d=0.08 (sleep), 0.56 (health), 0.84 (QOL); BS: d=0.91 (sleep), 0.89 (health), 0.63 (QOL)). The correlations of VAS with BODY-Q/QOLOS ranged between weak and moderate among non-BS (r: 0.35-0.57) and BS participants (r: 0.24-0.55). The index had very weak to weak correlation among non-BS participants (r: 0.07-0.39) and ranged from weak to strong (r: 0.34-0.64) among BS participants.
CONCLUSIONS: VAS shows stronger known-groups and convergent validity than the index . Although index performed better among BS participants, its limited ability to distinguish between BS and non-BS groups suggests that condition-specific dimensions are needed when using EQ-5D-5L to assess HRQoL of obese Asians. Results will be updated with larger sample during ISPOR as data collection is ongoing.
METHODS: We recruited obese adults from weight management clinics, including those eligible for BS and those who underwent BS at least 6 months ago. Participants completed EQ-5D-5L, selected BODY-Q subscales, PROMIS overall health and QOL questions, and demographic questions. BS participants also completed QOLOS excess skin scale. Known-groups validity was assessed using Cohen’s d by comparing EQ-5D-5L scores (EQ VAS and EQ-5D-5L index) across BS status, sleep quality, overall health, and QOL. Convergent validity was examined using Pearson’s correlations (r) between EQ-5D-5L scores and BODY-Q and/or QOLOS.
RESULTS: We analyzed 80 participants (non-BS: 43; BS: 37). VAS consistently distinguished between known groups, showing significant differences with large effect sizes by BS status (Cohen’s d=1.15), sleep quality (non-BS: d=1.00; BS: d=0.87), overall health (non-BS: d=1.37; BS: d=1.55), and overall QOL (non-BS: d=0.75; BS: d=0.90). Index showed smaller or insignificant differences in several comparisons among non-BS participants (non-BS: d=0.08 (sleep), 0.56 (health), 0.84 (QOL); BS: d=0.91 (sleep), 0.89 (health), 0.63 (QOL)). The correlations of VAS with BODY-Q/QOLOS ranged between weak and moderate among non-BS (r: 0.35-0.57) and BS participants (r: 0.24-0.55). The index had very weak to weak correlation among non-BS participants (r: 0.07-0.39) and ranged from weak to strong (r: 0.34-0.64) among BS participants.
CONCLUSIONS: VAS shows stronger known-groups and convergent validity than the index . Although index performed better among BS participants, its limited ability to distinguish between BS and non-BS groups suggests that condition-specific dimensions are needed when using EQ-5D-5L to assess HRQoL of obese Asians. Results will be updated with larger sample during ISPOR as data collection is ongoing.
Conference/Value in Health Info
2025-11, ISPOR Europe 2025, Glasgow, Scotland
Value in Health, Volume 28, Issue S2
Code
MSR40
Topic
Methodological & Statistical Research
Topic Subcategory
PRO & Related Methods
Disease
Diabetes/Endocrine/Metabolic Disorders (including obesity), Surgery