BACKGROUND AND AIMS: The aim was to derive health state utility scores in ulcerative colitis (UC) by establishing the relationship between the physician-rated ulcerative colitis disease activity index (UCDAI) and a patient reported EQ-5D by statistically mapping the two instruments. METHODS: In a randomised controlled trial comparing oral plus enema mesalazine treatment with oral mesalazine treatment alone (PINCE), UCDAI and EQ-5D scores were collected in parallel from patients with active UC. From these data, multinomial logistic regression was used to estimate response probabilities to each of the five domains of the EQ-5D index from assessment of UC disease severity using original and abbreviated (no endoscopy) versions of the UCDAI. Predicted EQ-5D responses were converted by Monte Carlo simulation to the EQ-5D index for predicting health-related quality of life (HRQoL). The reliability of the algorithm was tested using UCDAI scores from a second mesalazine RCT (PODIUM). RESULTS: The abbreviated-UCDAI showed comparable explanatory performance to the full UCDAI. For patients in remission, mean utility was 0.939, 0.944, and 0.940U for PINCE(estimated), PINCE(observed), and PODIUM, respectively. Mild/moderate and relapsing cases showed mean utilities of 0.801, 0.811, and 0.775, respectively; whilst for those in severe relapse, the mean utilities were 0.630, 0.700 and 0.660 units, respectively. The mean squared error between actual and predicted utilities from observations in PINCE was 0.019. CONCLUSION: Response mapping of UC activity to EQ-5D domains produced reliable estimates of patient-rated health state utility consistent with UCDAI rated severity. Comparing abbreviated-UCDAI and full UCDAI suggests that inclusion of endoscopy scores has limited predictive value in estimating patient HRQoL.
BACKGROUND AND AIMS: The aim was to derive health state utility scores in ulcerative colitis (UC) by establishing the relationship between the physician-rated ulcerative colitis disease activity index (UCDAI) and a patient reported EQ-5D by statistically mapping the two instruments. METHODS: In a randomised controlled trial comparing oral plus enemamesalazine treatment with oral mesalazine treatment alone (PINCE), UCDAI and EQ-5D scores were collected in parallel from patients with active UC. From these data, multinomial logistic regression was used to estimate response probabilities to each of the five domains of the EQ-5D index from assessment of UC disease severity using original and abbreviated (no endoscopy) versions of the UCDAI. Predicted EQ-5D responses were converted by Monte Carlo simulation to the EQ-5D index for predicting health-related quality of life (HRQoL). The reliability of the algorithm was tested using UCDAI scores from a second mesalazine RCT (PODIUM). RESULTS: The abbreviated-UCDAI showed comparable explanatory performance to the full UCDAI. For patients in remission, mean utility was 0.939, 0.944, and 0.940U for PINCE(estimated), PINCE(observed), and PODIUM, respectively. Mild/moderate and relapsing cases showed mean utilities of 0.801, 0.811, and 0.775, respectively; whilst for those in severe relapse, the mean utilities were 0.630, 0.700 and 0.660 units, respectively. The mean squared error between actual and predicted utilities from observations in PINCE was 0.019. CONCLUSION: Response mapping of UC activity to EQ-5D domains produced reliable estimates of patient-rated health state utility consistent with UCDAI rated severity. Comparing abbreviated-UCDAI and full UCDAI suggests that inclusion of endoscopy scores has limited predictive value in estimating patient HRQoL.