Biases in the measurement of health-related quality of life in the EQ-5D-5L among those with serious physical traumatic injuries: the ADVANCE cohort study.

PURPOSE: Physical traumatic injuries are likely to impact on a person’s Health-Related Quality of Life (HRQOL). We present data on HRQOL, from which EQ-5D-5L index scores are derived, and overall perceived health in UK Armed Forces personnel who sustained serious physical combat injuries and compare them to demographically similar personnel without such injuries (uninjured group). To ensure that perception of HRQOL is similar between groups, we also assess for measurement bias through Differential Item Functioning (DIF). METHODS: 577 personnel in an injured group (including subgroups of limb amputation injuries (n = 160) and non-amputation injuries (n = 417)) and 564 personnel in the uninjured group who sustained no such injuries were assessed in the ADVANCE cohort study. Health index values were derived from the EQ-5D-5L, overall perceived health from the EQ-visual analogue scale, and mobility from a Six-Minute Walk Test (6MWT). DIF was identified using multiple indicator multiple causes modelling. RESULTS: Compared to the uninjured group, personnel injured with associated limb loss had significantly higher likelihood of reporting low health index scores (Relative Risk Ratio (RRR) 7.18 (95% Confidence Interval (CI) 4.25, 12.29; relative to high health index scores)), however no difference in the probability of reporting low or moderate overall perceived health, relative to high, was observed. Those injured without limb loss had a significantly higher probability of reporting low health index values (RRR 3.58 (95%CI 2.57, 5.03)) and low overall perceived health (RRR 1.91 (95%CI 1.34, 2.66)). Measurement bias was observed in items regarding mobility and anxiety/depression for those injured with associated limb loss and self-care for those with injured without. Differences in 6MWT only partially explained the bias observed in the mobility item. CONCLUSION: Personnel who experience serious traumatic injuries perceive aspects of HRQOL differently to personnel who experience no such injuries, with heterogenous biases expressed depending on presence of limb amputation. Researchers evaluating HRQOL in health interventions may need to account for this bias if comparing individuals with different traumatic injuries/conditions.