The intermetacarpal distance method for assessment of active thumb radial abduction has excellent test-retest agreement, reliability, and precision in persons with non-operative thumb carpometacarpal osteoarthritis.

Background: Limitations in thumb radial abduction (i.e., carpometacarpal extension) are commonly experienced by persons with thumb carpometacarpal osteoarthritis. Restoring this deficit is often a focus of surgical and rehabilitative interventions. Because of this, clinical measures of radial abduction are needed. The ''gold-standard'' assessment of thumb radial abduction is goniometry however it has modest reliability in persons with thumb carpometacarpal osteoarthritis. Conversely, the intermetacarpal distance (IMD) method of assessing radial abduction has promising reliability in healthy persons and excellent inter-rater reliability in those with thumb carpometacarpal osteoarthritis. However, to date, there has been no exploration of its test-retest reliability and precision in persons with thumb carpometacarpal osteoarthritis. Further, while multiple trials are often performed in various hand therapy assessments, it is common practice to take a single measurement of hand joint range of motion. Yet, we do not know if multiple trials might enhance the reliability of these range of motion measurements.

Purpose: The current study aimed to (1) assess the test-retest agreement, reliability, and precision of the IMD method when measuring thumb radial abduction and (2) compare these psychometric properties when reporting one trial, the mean of two trials, and the mean of three trials of the IMD method in people with thumb carpometacarpal (CMC1) osteoarthritis (OA).

Study Design: Prospective Cohort study of test-retest reliability.

Methods: Purposive sampling was used to recruit adults with radiographically confirmed CMC1 OA. Participants' radial abduction was assessed using the IMD method to measure radial abduction of the affected hand on two separate occasions approximately two weeks apart. Three trials of the assessment were administered at each visit. The Bland Altman method was uses to assess agreement, the intraclass correlation coefficient (ICC) was calculated to examine the reliability, and the standard error of the measurement (SEM), minimum detectable change (MDC) and MDC percentage were calculated to determine the precision of the IMD method for one trial, the mean of two trials, and the mean of three. Descriptive data on demographics and IMD values of the sample were presented.

Results: Forty persons with CMC1 OA participated. The mean difference between trials ranged from 0.21 to 0.30 mm, no significant fixed biases (p≥0.48) or proportional biases (p≥0.41) existed between mean test and retest scores, and 38/40 (95%) of test-retest differences fell within the 95% limits of agreement. The Intraclass Correlation Coefficient (ICC) values were 0.942 (0.893-0.968), 0.970 (0.943-0.984), and 0.970 (0.942-0.984) for 1 trial, the mean of two trials, and then mean of three trials respectively. The ICC value of the mean of two trials was significantly higher than that of 1 trial indicating superior reliability however reliability for each approach fell into the excellent range (i.e., >0.90). The precision of one trial (MDC%=13.0) fell into the acceptable range while the precision of the mean of two (MDC%=9.1) and three trials (MDC%=9.1) fell into the excellent range.

Conclusions: The IMD method for assessing CMC1 radial abduction has acceptable agreement, excellent test-retest reliability, and acceptable-to-excellent precision when performing one trial, the mean of two or the mean of three trials in persons with CMC1 osteoarthritis. The precision and reliability of the mean of two or three trials were non-superior however the mean of two trials yielded superior reliability and precision to that of a single trial. Changes in IMD measurements of 1.9 mm and 1.4 mm are needed to exceed measurement error when administering one or two trials respectively and change must surpass 5.3 mm and 3.8 mm to exceed clinically meaningful thresholds for one or two trials respectively. We recommend taking the mean of two trials over a single trial when using the IMD method for assessing CMC1 radial abduction in persons with CMC1 OA.