Repeated measures in rheumatoid arthritis reduced the required sample size in a two-armed clinical trial

J Clin Epidemiol. 2008 Sep;61(9):940-4. doi: 10.1016/j.jclinepi.2007.12.004. Epub 2008 Jun 6.


Objectives: To investigate whether repeated measures in patients with rheumatoid arthritis will reduce the between subject variation and if so, to determine the optimal number of measures to effectively reduce the number of participants required in controlled clinical trials.

Study design and setting: A prospective observational study. Thirty-eight rheumatoid arthritis patients with a stable disease reported level of joint pain, fatigue and patient global assessment of disease activity on VAS scales as well as the Rheumatoid Arthritis Disease Activity Index (RADAI) daily during a total of 42 days.

Results: In all measures, the variation within each individual was substantial over a 42-day period. By increasing the number of measurements from one to five, the standard deviation (SD) decreased from 5.5% (RADAI score) to 11% (Pain VAS) resulting in a reduction in the number of patients needed in a clinical trial from 11% to 22%, respectively. When we used from 6 to 42 individual measurements, the decrease continued but the reduction was of a smaller magnitude.

Conclusion: The use of up to five repeated measurements per patient will decrease the number of patients required in a two armed clinical trial by as much as 22%.

MeSH terms

  • Adult
  • Aged
  • Arthritis, Rheumatoid / epidemiology
  • Arthritis, Rheumatoid / therapy*
  • Clinical Trials as Topic
  • Disease Progression
  • Female
  • Humans
  • Male
  • Middle Aged
  • Observer Variation
  • Pain Measurement / methods*
  • Prospective Studies
  • Reproducibility of Results
  • Sample Size
  • Sensitivity and Specificity
  • Severity of Illness Index
  • Treatment Outcome