Differential psychometric properties of EQ-5D-5L and SF-6D utility measures in patients with low back pain

Abstract

Introduction: LBP is a common musculoskeletal disorder requiring multiple diagnostic tests and treatment, these can incur high medical costs. It is thus desirable to have appropriate utility measure for cost-utility analysis of various LBP-related interventions and their cost-effectiveness. This prospective study aims to examine the acceptability, validity and reliability of the EuroQoL 5-dimension 5-level (EQ-5D-5 L) and Short-Form 6-Dimension (SF-6D) health utility measures in patients with low back pain (LBP). Materials and Methods: Health-related quality of life (HRQoL) questionnaires including the generic 12-item Short Form Health Survey (SF-12), EQ-5D-5 L and low back/back-related questionnaires were administered at a specialty outpatient clinic. Responses to SF-12 items were transformed to SF-6D utility scores using the Hong Kong population scoring algorithm derived by Standard Gamble whereas response to EQ-5D-5 L were mapped onto EQ-5D-3 L response via interim mapping algorithms and then converted to EQ-5D-5 L utility scores using the Chinese-specific value set. Construct validity was determined by evaluating Spearman correlation between SF-12 scores and EQ-5D-5 L utility scores. Correlation between back-specific questionnaires and HRQoL scores were also assessed. Results: A total of 100 patients were recruited. No significant (>15%) floor and ceiling effects were observed for utility scores of EQ-5D-5 L and SF-6D. Detailed proportion of respondents for each EQ-5D-5 L domain indicated that Pain/Discomfort domain was the most prevalent problem. The SF-6D utility score had a strong Spearman rank correlation with SF-12 domain and summary scores (0.600-0.855). Moderate-to-strong Spearman rank correlations were observed between EQ-5D-5 L score (0.455-0.700) and SF-12 domain and summary scores, and between EQ-5D-5 L and SF-6D scores (0.625). Adequate construct validity was evident as both utility scores conceptually measure the similar construct. Patients with no history of previous spine surgery or no disc degeneration had significantly higher EQ-5D-5 L scores (p = 0.047; p = 0.010). Significant correlations were observed between both EQ-5D-5 L and SF-6D scores and back-specific questionnaires. Conclusions: Both EQ-5D-5 L and SF-6D instruments appeared to be applicable and valid measures in assessing the HRQoL of LBP patients. This is the first study to examine the differential psychometric properties and validation of the use of EQ-5D-5 L and SF-6D in the Chinese LBP population. The impact of utility score selection on assessing the effectiveness of clinical interventions for LBP can now be explored.