Domain Attrition in the IWQOL Series and the Challenges of Factor-Only Item Reduction
Author(s)
Natalia Zdorovtsova1, Saniya Deshpande, MSc2, Matthew Hankins, PhD3.
1Health Analytics Associate Consultant, Lane Clark and Peacock, London, United Kingdom, 2Health Analytics, Lane Clark & Peacock, London, United Kingdom, 3Health Analytics, LCP Health Analytics, London, United Kingdom.
1Health Analytics Associate Consultant, Lane Clark and Peacock, London, United Kingdom, 2Health Analytics, Lane Clark & Peacock, London, United Kingdom, 3Health Analytics, LCP Health Analytics, London, United Kingdom.
OBJECTIVES: Progressive shortening of patient-reported outcome measures (PROMs) is often driven by factor-analytic item reduction with limited patient input. Using the Impact of Weight on Quality of Life (IWQOL) family as a case study, we explored how successive refinements—from the 74-item IWQOL (1995) to the 31-item IWQOL-Lite (2001), the 22-item IWQOL-Lite-Clinical Trials version (IWQOL-Lite-CT, 2017), and finally the 20-item IWQOL-Lite-CT (2019)—alter domain coverage and the potential clinical meaning that is retained or lost.
METHODS: We conducted a review of development and validation papers and scale manuals. Items and domains eliminated between versions were catalogued and qualitatively examined for content relevance based on current obesity literature.
RESULTS: The original IWQOL captured eight domains and 74 items derived from interviews with ≈20 patients. IWQOL-Lite retained only five domains: Physical Function (11 items), Self-Esteem (7), Sexual Life (4), Public Distress (5) and Work (4), removing Health, Social/Interpersonal, Activities of Daily Living and Comfort with Food. 58% of items were discarded solely on statistical grounds. IWQOL-Lite-CT initially comprised 22 items, which were further reduced to 20 items (2019) spanning two broad domains: Physical (7) and Psychosocial (13). Notably, the final refinement eliminated two stigma-related items based on statistical redundancy. Although cognitive interviews informed wording, the earlier omitted domains were not restored; nuanced and salient concepts such as appetite control, eating enjoyment, day-to-day self-care, and weight-related stigma remain absent.
CONCLUSIONS: The IWQOL trajectory illustrates how psychometric Procrusteanism can lead to the removal of content that matters to patients. When PROMs are shortened without patient input, clinically-salient domains may disappear, potentially undermining sensitivity to modern treatment benefits. Future instrument refinement should combine statistical criteria with iterative patient engagement.
METHODS: We conducted a review of development and validation papers and scale manuals. Items and domains eliminated between versions were catalogued and qualitatively examined for content relevance based on current obesity literature.
RESULTS: The original IWQOL captured eight domains and 74 items derived from interviews with ≈20 patients. IWQOL-Lite retained only five domains: Physical Function (11 items), Self-Esteem (7), Sexual Life (4), Public Distress (5) and Work (4), removing Health, Social/Interpersonal, Activities of Daily Living and Comfort with Food. 58% of items were discarded solely on statistical grounds. IWQOL-Lite-CT initially comprised 22 items, which were further reduced to 20 items (2019) spanning two broad domains: Physical (7) and Psychosocial (13). Notably, the final refinement eliminated two stigma-related items based on statistical redundancy. Although cognitive interviews informed wording, the earlier omitted domains were not restored; nuanced and salient concepts such as appetite control, eating enjoyment, day-to-day self-care, and weight-related stigma remain absent.
CONCLUSIONS: The IWQOL trajectory illustrates how psychometric Procrusteanism can lead to the removal of content that matters to patients. When PROMs are shortened without patient input, clinically-salient domains may disappear, potentially undermining sensitivity to modern treatment benefits. Future instrument refinement should combine statistical criteria with iterative patient engagement.
Conference/Value in Health Info
2025-11, ISPOR Europe 2025, Glasgow, Scotland
Value in Health, Volume 28, Issue S2
Code
PCR66
Topic
Patient-Centered Research
Topic Subcategory
Instrument Development, Validation, & Translation, Patient-reported Outcomes & Quality of Life Outcomes
Disease
Diabetes/Endocrine/Metabolic Disorders (including obesity)