Using HAQ-DI to estimate HUI-3 and EQ-5D utility values for patients with rheumatoid arthritis in Spain

Background/ObjectiveUtility values are not usually assessed in clinical trials and do not allow cost-utility analysis to be performed with the data collected. The aim of this study was to derive relation functions so that Health Assessment Questionnaire – Disability Index (HAQ-DI) scores could be used to estimate Health Utilities Index - 3 (HUI-3) and EQ-5D utility values for patients with rheumatoid arthritis (RA).MethodsAn observational, cross-sectional, naturalistic, multicentre study was conducted. A total of 244 patients aged 18 years or older, with RA according to American College of Rheumatology diagnostic criteria, were recruited. Sociodemographic and clinical variables were recorded and patients completed three generic HRQoL questionnaires: the HAQ-DI, the HUI-3, and the EQ-5D. Two linear regression models were used to predict HUI-3 and EQ-5D utility values as functions of HAQ-DI scores, age, and gender.ResultsPatient mean age was 57.8 years old (standard deviation [SD], 13.3 years); 75.8% of the patients were women and 95.9% were white. Mean disease duration was 10.8 years (SD, 9 years). Patient distribution according to HAQ-DI severity was as follows: HAQ-DI < 0.5, 29%; 0.5 ≤ HAQ-DI < 1.1, 28%; 1.1 ≤ HAQ-DI < 1.6, 16%,1.6 ≤ HAQ-DI < 2.1, 15%; and HAQ-DI ≥ 2.1, 12%. HAQ-DI and EQ-5D mean scores were 1.02 (SD, 0.78) and 63.1 (SD, 20.3), respectively. Mean utility values for HUI-3 and time trade-off (TTO) were 0.75 (SD, 0.21) and 0.65 (SD, 0.3), respectively. The equations converting HAQ-DI scores to utilities were HUI-3 = 0.9527 – (0.2018 × HAQ-DI) +ε (R2=0.56), and TTO = 0.9567 – (0.309 × HAQ-DI) + ε (R2=0.54). Error distribution was non-normal. Age and gender were found to have no bearing on the utility functions.ConclusionsHAQ-DI scores can be used to estimate HUI-3 and EQ-5D utility values for patients with RA in data obtained from studies where utility values have not been collected.