Intratester and intertester reliability of clinical measures of lower extremity anatomic characteristics: implications for multicenter studies.

摘要

OBJECTIVE: To determine whether multiple examiners could be trained to measure lower extremity anatomic characteristics with acceptable reliability and precision, both within (intratester) and between (intertester) testers. We also determined whether testers trained 18 months apart could perform these measurements with good agreement. SETTING: University's Applied Neuromechanics Research Laboratory. PARTICIPANTS: Sixteen, healthy participants (7 men, 9 women). ASSESSMENT OF RISK FACTORS: Six investigators measured 12 anatomic characteristics on the right lower extremity in the Fall of 2004. Four testers underwent training immediately preceding the study, and measured subjects on 2 separate days to examine intratester reliability. Two testers trained 18 months before the study (Spring 2002) measured each subject on day 1 to examine the consistency of intertester reliability when testers are trained at different times. MAIN OUTCOME MEASUREMENTS: Knee laxity, genu recurvatum, quadriceps angle, tibial torsion, tibiofemoral angle, hamstring extensibility, pelvic angle, navicular drop, femur length, tibial length, and hip anteversion. RESULTS: With few exceptions, all testers consistently measured each variable between test days (intraclass correlation coefficient>or=0.80). Intraclass correlation coefficient values were lower for intertester reliability (0.48 to 0.97), and improved from day 1 to day 2. Intertester reliability was similar when comparing testers trained 18 months before those trained immediately before the study. Absolute measurement error varied considerably across individual testers. CONCLUSIONS: Multiple investigators can be trained at different times to measure anatomic characteristics with good to excellent intratester reliability. Intratester reliability did not always ensure acceptable intertester reliability or measurement precision, suggesting more training (or more experience) may be required to achieve acceptable measurement reliability and precision between multiple testers.