Navicular Drop Measurement in People With Rheumatoid Arthritis: Interrater and Intrarater Reliability

Abstract

Navicular drop (ND) measurement may be a valuable examination technique for patients with rheumatoid arthritis (RA). However, no data exist on reliability for this technique in patients with RA. The purposes of this study were: (1) to determine interrater and intrarater reliability of ND measurements in people with RA, (2) to compare ND values of people with RA with published normative data, and (3) to investigate ND measurement error associated with the use of skin markings. Ten women (20 feet) with RA consented to participate. Patients completed demographic and function questionnaires. Navicular height (NH) measurements were taken by 2 physical therapists and 1 physical therapist student, following four 1-hour training sessions, using standardized methods and a digital height gauge. Four different NH measurements were taken 3 times on each foot by each of the 3 examiners during a morning session and then repeated during an afternoon session on the same day. Navicular drop values were calculated, including ND1 (as reported in the literature), ND2 (compensating for skin error), and ND3 (single-limb stance). Intraclass correlation coefficients (ICCs) and standard errors of measurement (SEMs) were used to establish reliability. Means (+/-SD) for each ND measure for sessions 1 and 2, respectively, were as follows: ND1=8.36+/-5.29 mm and 8.29+/-5.24 mm, ND2=9.95+/-5.44 mm and 9.57+/-5.37 mm. The ICCs (2,1 and 2,k, respectively) for all interrater measurements ranged from .67 to .92 (SEM=2.0-3.3 mm) and from .85 to .97 (SEM=1.1-2.0 mm). The ICCs (2,1 and 2,k, respectively) for intrarater measurements ranged from .73 to .95 (SEM=1.3-2.8 mm) and from .90 to .98 (SEM=0.7-1.6 mm). Paired t tests showed the means of ND1 and ND2 for each examiner and for both sessions were significantly different. The results suggest that ND measurements for people with RA can be taken reliably by clinicians with varied experience. The ND values for our subjects were slightly greater than reported normal values of 6 to 8 mm. Error associated with skin markings was statistically significant for all sessions and examiners.