How to get from the multidimensional health assessment questionnaire to Stanford health assessment questionnaire disability index scores in patients with rheumatoid arthritis, psoriatic arthritis and axial spondyloarthritis: Development and validation of a conversion algorithm

E. Svensson, K. Løngaard, L. Midtbøll Ørnbjerg, Rikke Hodal Meincke, J. K. Pedersen, L. Dreyer, N. Steen Krogh, D. V. Jensen, M. L. Hetland

Research output: Contribution to journalConference abstract in journalResearchpeer-review


Background: In the DANBIO quality registry in Denmark, patients with rheumatoid arthritis (RA) psoriatic arthritis (PsA) and axial spondyloarthritis (axSpA) have reported Patient Reported Outcomes (PROs) including the Stanford Health Assessment Questionnaire Disability Index (HAQ-DI) for nearly twenty years as part of routine care. Patients’ feedback have stressed a need for a shorter registration of disability (1). While the shorter Multidimensional Health Assessment Questionnaire (MDHAQ) is preferred by patients, the original HAQ-DI is the preferred tool in observational studies. Thus, a conversion algorithm between the MDHAQ and HAQ-DI scores is warranted.Objectives: To develop and validate a simple conversion algorithm between MDHAQ and HAQ-DI scores in RA, PsA and axSpA patients.Methods: Patients registered in DANBIO with a diagnosis of RA, PsA or axSpA who had completed both HAQ-DI and MDHAQ simultaneously at a visit +/- 30 days from start of conventional synthetic (cs)DMARD or biological (b)DMARD were eligible for the analysis, and randomly divided into development and validation cohorts stratified by diagnosis. The conversion algorithm was developed in the RA development cohort using linear regression with HAQ-DI as the dependent variable and MDHAQ as the independent variable. The predicted HAQ (pHAQ) scores were then calculated by applying the conversion algorithm to the MDHAQ scores in the RA, PsA and axSpA validation cohorts. The pHAQ was validated against the HAQ-DI in the validation cohorts regarding criterion, correlational and construct validity.Results: We included 8983/4410/1760 patients with RA/PsA/axSpA, respectively. The conversion algorithm pHAQ=0.15+MDHAQ*1.08 had the best fit (R2=0.83) in the RA development cohort.Criterion validity: The correlation coefficients between HAQ-DI/pHAQ and patient global score at baseline were 0.66/0.65. In groups of patients with high and low disability (defined as patient global score ≥50), standardized mean difference was -1.4 for HAQ-DI, and -1.4 for pHAQ.Correlational validity: Correlation coefficients between HAQ-DI/pHAQ and ΔHAQ-DI/ΔpHAQ between baseline and first follow-up visit were r=0.91 and r=0.87, respectively. Correlation coefficients between HAQ-DI/pHAQ and pain score, DAS28CRP and physician global score were 0.63/0.64, 0.55/0.55 and 0.34/0.34, respectively. A Bland-Altman plot showed good agreement of HAQ-DI and pHAQ across all functional states.Construct validity: HAQ-DI/pHAQ at the first follow-up visit after baseline was comparable between Patient Acceptable Symptom State groups (PASS=No: mean 1.17 vs 1.18/PASS=Yes: 0.55 vs 0.60). Similar results were seen for the external anchor (Figure 1).In PsA and axSpA validation cohorts, similar results were found.Conclusion: A conversion algorithm from MDHAQ to HAQ-DI was developed in ≈ 4500 RA patients. In separate large validation cohorts of RA, PsA and axSpA patients, the predicted HAQ calculated from the MDHAQ scores showed good criterion, correlational and construct validity comparable to the original HAQ-DI. The results suggest that for research purposes the MDHAQ can be converted to HAQ-DI if a full HAQ-DI has not been performed.References: [1] Primdahl J. et al. Arthritis Care Res 2019 (in press).Acknowledgments: The authors thank all Danish patients and Departments of Rheumatology, who conscientiously report to the DANBIO registry.Disclosure of Interests: Elisabeth Svensson: None declared, Katja Løngaard: None declared, Lykke Midtbøll Ørnbjerg Grant/research support from: Novartis, Rikke Meincke: None declared, Jens Kristian Pedersen: None declared, Lene Dreyer: None declared, Niels Steen Krogh: None declared, Dorte Vendelbo Jensen: None declared, Merete L. Hetland Grant/research support from: BMS, MSD, AbbVie, Roche, Novartis, Biogen and Pfizer, Consultant of: Eli Lilly, Speakers bureau: Orion Pharma, Biogen, Pfizer, CellTrion, Merck and Samsung Bioepis
Original languageEnglish
JournalAnnals of the Rheumatic Diseases
Issue numberSuppl. 1
Pages (from-to)897-898
Number of pages2
Publication statusPublished - 1. Jun 2020
EventEULAR 2020 -
Duration: 3. Jun 20203. Jun 2020


ConferenceEULAR 2020

Cite this