Validation of an algofunctional index for osteoarthritis of the hand

Although hand osteoarthritis is common, it has been the focus of few therapeutic trials. In addition to the problems raised by clinical trials in osteoarthritis in general and to the difficulties due to the unforeseeable course of osteoarthritis of the trapezometacarpal and finger joints, the lack of a clinical tool for assessing pain and function over time is an additional obstacle. We propose an algofunctional index designed for evaluation and symptomatic follow-up of patients with digital osteoarthritis. The index is based on a physician-administered questionnaire on 10 daily activities involving the hands. The patient is asked to answer each item using a 4-point verbal scale, from "possible without difficulty" (0) to "impossible" (3 points); thus, total scores range from 0 to 30. This index has been used in a few clinical placebo-controlled trials and was found sensitive to change. The aim of this study was to assess the metrological qualities of this index, including consistency (internal and external), sensitivity and specificity (by scoring the index in different groups of subjects), intra-observer reproducibility, and ease of use. Three hundred patients were recruited by 25 rheumatologists: 100 had a painful attack of digital and/or trapezometacarpal osteoarthritis (mean age: 64.9 years) with a score of more than 40 mm on a visual analog scale for overall pain severity (mean: 57.3 +/- 14 mm), 100 had "inactive" hand osteoarthritis (mean age 67.0 years), and 100 had no diseases of the upper limbs. Specificity/sensitivity: the mean index score was 12.41 +/- 5.41 in patients with painful OA, 4.28 +/- 3.87 in "inactive" cases, and 0.59 +/- 1.23 in controls. External consistency: the overall mean score was well correlated with pain severity: r = 0.49 (p < 0.001). Internal consistency: principal component analysis identified a primary axis responsible for 44.2% of the variance and two secondary axes each responsible for slightly more than 9% of the variance. None of the questions seemed redundant. Intra-observer reproducibility: two evaluations done one hour apart in symptomatic patients yielded the following scores: 12.32 +/- 5.41 and 12.5 +/- 5.51 (correlation: 0.95; mean difference: 0.17 +/- 1.64; coefficient of variation: 9.32%). Kappa values for both measurements of each item ranged from 0.68 to 0.87. Ease of use: mean time needed to determine the score 2.5 +/- 2 min. The scoring process was considered simple by 100% of investigators and easy/very easy by 98% of patients.(ABSTRACT TRUNCATED AT 400 WORDS)