Choi B1,2, Oh D3, Lim W1,4, Bang S5, Shim J6
1Woosong University, Physical Therapy, Daejeon, Korea (South), 2Woosong University, Advanced Institute of Convergence Sports Rehabilitation, Daejeon, Korea (South), 3Woosong University, Graduate School, Dept of Rehabilitation, Daejeon, Korea (South), 4Advanced Institute of Convergence Sports Rehabilitation, Woosong University, Daejeon, Korea (South), 5Hangil Korean Medicine Hospital, Busan, Korea (South), 6Daejeon Wellness Hospital, Department of Rehabilitation, Daejeon, Korea (South)
Background: Palliative rehabilitation programs for cancer survivors are now being practised in many clinical settings across the world and are becoming increasingly important in allowing survivors gain positive outcomes in regards to quality of life (QOL). The routinely measured self-reported QOL measurements are often challenging to maintain construct validity when measurements are translated into other languages.
Purpose: To examine the construct validity of the health-related quality of life (HRQOL) measurements for cancer survivors with Korean versions of EUROQOL-5 Dimension (EQ-5D) and Quality of Life (QOL) scale developed through the World Health Organization (WHOQOL-BREF).
Methods: A total of 52 cancer survivors who are currently registered for ongoing conservative physical therapy treatments was recruited from a local rehabilitation hospital and an oriental medicine hospital in Korea. The Korean version of the EQ-5D and the WHOQOL-BREF were administered to the survivors. The construct validity of the two outcome measures was confirmed using the Rasch analysis (1-parameter Item Response Theory model) and was followed by item-person map comparisons.
Results: As a means of determining the unidimensionality of two measures, Rasch fit statistics of the EQ-5D were acceptable. However 3 items (i.e., pain/discomfort, activities of daily living, negative feeling) of the WHOQOL-BREF were misfit. The most challenging item was “pain” (item 4), while the least challenging item was “self-care” (item 2) in the EQ-5D. The most challenging item was “negative feeling” (item 26), while the least challenging item was “concentrate” (item 7) in the WHOQOL-BREF. By inspecting the item-person map representing how well the items of the measures capture the HRQOL expressed by the survivors, 5 items of the EQ-5D showed a large ceiling effect which was unable to measure the survivors with mid/high HRQOL levels and also demonstrated some serious gaps.
Conclusion(s): Despite 3 misfit items, twenty-six items of the WHOQOL-BREF measurement well targeted the survivors throughout the entire ranges of HRQOL level. Since the EQ-5D failed to measure the cancer survivors with mid/high HRQOL levels, few items covering the gaps should be added to efficiently measure the HRQOL with fewer items compared to the WHOQOL-BREF measurement.
Implications: Construct validity of the quality of life measurements can be confirmed by applying Rasch model with item-person map and fit statistic analysis.
Keywords: construct validity, measurement, quality of life
Funding acknowledgements: This research is based on the support of 2018 Woosong University Academic Research Funding.
Purpose: To examine the construct validity of the health-related quality of life (HRQOL) measurements for cancer survivors with Korean versions of EUROQOL-5 Dimension (EQ-5D) and Quality of Life (QOL) scale developed through the World Health Organization (WHOQOL-BREF).
Methods: A total of 52 cancer survivors who are currently registered for ongoing conservative physical therapy treatments was recruited from a local rehabilitation hospital and an oriental medicine hospital in Korea. The Korean version of the EQ-5D and the WHOQOL-BREF were administered to the survivors. The construct validity of the two outcome measures was confirmed using the Rasch analysis (1-parameter Item Response Theory model) and was followed by item-person map comparisons.
Results: As a means of determining the unidimensionality of two measures, Rasch fit statistics of the EQ-5D were acceptable. However 3 items (i.e., pain/discomfort, activities of daily living, negative feeling) of the WHOQOL-BREF were misfit. The most challenging item was “pain” (item 4), while the least challenging item was “self-care” (item 2) in the EQ-5D. The most challenging item was “negative feeling” (item 26), while the least challenging item was “concentrate” (item 7) in the WHOQOL-BREF. By inspecting the item-person map representing how well the items of the measures capture the HRQOL expressed by the survivors, 5 items of the EQ-5D showed a large ceiling effect which was unable to measure the survivors with mid/high HRQOL levels and also demonstrated some serious gaps.
Conclusion(s): Despite 3 misfit items, twenty-six items of the WHOQOL-BREF measurement well targeted the survivors throughout the entire ranges of HRQOL level. Since the EQ-5D failed to measure the cancer survivors with mid/high HRQOL levels, few items covering the gaps should be added to efficiently measure the HRQOL with fewer items compared to the WHOQOL-BREF measurement.
Implications: Construct validity of the quality of life measurements can be confirmed by applying Rasch model with item-person map and fit statistic analysis.
Keywords: construct validity, measurement, quality of life
Funding acknowledgements: This research is based on the support of 2018 Woosong University Academic Research Funding.
Topic: Outcome measurement; Oncology, HIV & palliative care
Ethics approval required: Yes
Institution: Woosong University
Ethics committee: Institutional Review Board
Ethics number: 1041549-170711-SB-52
All authors, affiliations and abstracts have been published as submitted.
