Reliability of Fitness Tests Using Methods and Time Periods Common in Sport and Occupational Management

CONTEXT: Fitness testing is used frequently in many areas of physical activity, but the reliability of these measurements under real-world, practical conditions is unknown. OBJECTIVE: To evaluate the reliability of specific fitness tests using the methods and time periods used in the context of real-world sport and occupational management. DESIGN: Cohort study. SETTING: Eighteen different Cirque du Soleil shows. PATIENTS OR OTHER PARTICIPANTS: Cirque du Soleil physical performers who completed 4 consecutive tests (6-month intervals) and were free of injury or illness at each session (n = 238 of 701 physical performers). INTERVENTION(S): Performers completed 6 fitness tests on each assessment date: dynamic balance, Harvard step test, handgrip, vertical jump, pull-ups, and 60-second jump test. MAIN OUTCOME MEASURE(S): We calculated the intraclass coefficient (ICC) and limits of agreement between baseline and each time point and the ICC over all 4 time points combined. RESULTS: Reliability was acceptable (ICC > 0.6) over an 18-month time period for all pairwise comparisons and all time points together for the handgrip, vertical jump, and pull-up assessments. The Harvard step test and 60-second jump test had poor reliability (ICC < 0.6) between baseline and other time points. When we excluded the baseline data and calculated the ICC for 6-month, 12-month, and 18-month time points, both the Harvard step test and 60-second jump test demonstrated acceptable reliability. Dynamic balance was unreliable in all contexts. Limit-of-agreement analysis demonstrated considerable intraindividual variability for some tests and a learning effect by administrators on others. CONCLUSIONS: Five of the 6 tests in this battery had acceptable reliability over an 18-month time frame, but the values for certain individuals may vary considerably from time to time for some tests. Specific tests may require a learning period for administrators.