Article Text


AB0232 Reliability and responsiveness of the radai, a self-assessed rheumatoid arthritis disease activity index
  1. J Fransen,
  2. A Forster,
  3. D Uebelhart,
  4. BA Michel
  1. Department of Rheumatology and Institute of Physical Medicine, University Hospital Zurich, Zurich, Switzerland


Background The Rheumatoid Arthritis Disease Activity Index (RADAI) is a self-administered questionnaire on signs and symptoms of rheumatoid arthritis (RA). The main goal of the RADAI is to evaluate the course of RA disease activity over time. The RADAI consists of 5 items and produces a single index from 0–10; higher scores are indicating higher levels of disease activity. The RADAI has shown to be cross-sectionally valid1 and responsive to clinical important worsening of disease activity.2

Objectives The study objectives are: 1) to determine the test-retest reliability of the RADAI, and 2) to determine if the RADAI is responsive to clinical important improvements in disease activity, using an intervention of known efficacy (Infliximab) as external standard.

Methods Until now, 12 RA patients (ACR criteria) have been included (1 male), having high or moderate disease activity (DAS28 >3.2);3 10 patients already had a 6-week follow-up. The intervention consisted of treatment with Infliximab and Methotrexate, which is likely to induce a relatively large improvement after 6 weeks. The RADAI was self-administered 1 week before the first infusion with Infliximab (T0). At the first infusion (T1), the patient filled in the RADAI again and the physician assessed the DAS28.

After 6 weeks (T2), the RADAI and the DAS28 were again assessed. For analysis of reliability (T0 and T1) the ICC3,1 and the Limits-of Agreement [4] were used. Responsiveness was studied by judging the change in RADAI at T2 in relation to the Limits-of-Agreement and the DAS28 response criteria.3

Results The RADAI scores at T0 and T1 were mean (sd) 5.0 (1.7) and 4.8 (1.9). The mean (sd) difference was -0.2 (0.8), (paired t-test, p = 0.41). The ICC3,1 was 0.89 (p < 0.05). The Upper Limit-of-Agreement was at 1.6 and the Lower at -2.0. According to the DAS28 response criteria, 3 patients were classified as having no response at T2, 3 had a moderate and 4 a good response. The change in RADAI did not exceed the Lower Limit-of-Agreement for 4 of the 7 responders. The responders had a median (range) change in RADAI of -2.1 (-0.9 to -8.0), signed rank test, p = 0.02).

Conclusion The data collection is ongoing till N = 20. The first results on reliability and responsiveness are promising. However, if the Limits-of-Agreement turn out to be much wider than -1 and +1, the reliability of the RADAI should be improved.


  1. Fransen J, et al. Rheumatology 2000;39(3):321–7

  2. Fransen J, et al. Arthritis Rheum., in press

  3. Van Gestel AM, et al. Arthritis Rheum. 1998;41(10):1845–50

  4. Bland J, Altman D. Lancet 1986;1(8476):307–10

Statistics from

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.