Send to

Choose Destination
Arthritis Rheum. 2001 Jul;44(7):1599-607.

Stratification of flare intensity identifies placebo responders in a treatment efficacy trial of patients with osteoarthritis.

Author information

Piedmont Research Institute, Chapel Hill, North Carolina 27514, USA.



Studies evaluating osteoarthritis treatment often use increased arthritis activity ("flare") as a selection criterion, although no standardized assessments are available to quantify flare intensity and little is known about how this criterion affects treatment comparisons. This study evaluated the reliability of a flare assessment and how pretreatment flare intensity impacts conclusions on treatment efficacy.


Using data from a double-blind, randomized, controlled trial (n = 182), we compared 3 osteoarthritis treatments with placebo in patients who met 3 of 4 flare criteria. The Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) questionnaire was used to document levels of pain, stiffness, and physical functioning at baseline and at the final visit. Following factor analytic evaluation, the flare items were standardized and summed to create a flare intensity index, which was used to identify patient subgroups. Analysis of covariance was applied to compare change in WOMAC scale scores from baseline to final visit for assessment of treatment differences among the flare intensity subgroups.


The flare indicators appeared unidimensional. Analyses were stratified by tertiles of flare intensity. Mean WOMAC scores improved in the patients receiving active treatment who were categorized into the 2 lowest flare intensity subgroups, but mean WOMAC scores improved in patients in all 4 treatment groups (active and placebo) in the most intense flare subgroup.


Patients with higher intensity flares may be more likely to report substantial improvement in functional status regardless of treatment. Failure to account for flare intensity in analyses of data from pain trials with flare-based designs may inflate the risk of Type I and Type II errors in the interpretation of study results.

[Indexed for MEDLINE]
Free full text

Supplemental Content

Full text links

Icon for Wiley
Loading ...
Support Center