Ørnbjerg LM, Svensson E, Løngaard K, Meincke RH, Pedersen JK, Dreyer L, Krogh NS, Jensen DV, Hetland ML. OUP accepted manuscript.
Rheumatology (Oxford) 2022;
61:3919-3929. [PMID:
35108380 DOI:
10.1093/rheumatology/keac040]
[Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 10/06/2021] [Revised: 01/12/2022] [Indexed: 11/15/2022] Open
Abstract
OBJECTIVES
To develop and validate in real-world patients a conversion algorithm from the Multidimensionel Health Assessment Questionnaire physical function scale (MDHAQ) to the Stanford Health Assessment Questionnaire disability index physical function scale (HAQ) score.
METHODS
From the DANBIO registry, 13 391 patients with RA (n = 8983), PsA (n = 2649) and axial spondyloarthritis (axSpA, n = 1759) with longitudinal data on HAQ and MDHAQ were included, stratified by diagnosis, and randomized 1:1 into development and validation cohorts. Conversion algorithms were developed by linear regression and applied in validation cohorts. From MDHAQ, the HAQ was calculated (cHAQ) and validated against the observed HAQ for criterion, correlational and construct validity.
RESULTS
For RA, we developed the conversion algorithm cHAQ = 0.15+MDHAQ*1.08, and validated it in the RA validation cohort. Criterion validity: HAQ and cHAQ had comparable discriminative power to distinguish between high and low patient global scores (standardized mean difference: HAQ:-1.29, cHAQ:-1.35). Kappa value between HAQ and cHAQ functional states indicated good agreement (0.83). Correlational validity: baseline HAQ and cHAQ, respectively, correlated well with patient global scores (r = 0.65/0.67). Bland-Altman plots showed good agreement across all functional states. Construct validity: HAQ and cHAQ discriminated equally well between patients reporting symptom state as acceptable vs not, and across responses to an external anchor. Aiming for a common algorithm, the RA conversion algorithm was validated for PsA and axSpA with similar results.
CONCLUSION
This study suggests that in observational datasets with only the MDHAQ available, a simple algorithm allows valid conversion to HAQ on the group level in RA, PsA and axSpA.
Collapse