To calculate the sample size for this study we used data from the Functional Mobility Test (FMT). The FMT was designed to evaluate an astronaut’s ability to complete challenging locomotor maneuvers similar to those encountered during an egress from a space vehicle following long-duration space flight [20 (link)]. To perform the FMT subjects walked at a self-selected pace through an obstacle course set up on a base of medium density foam. The foam provided an unstable surface that increased the challenge of the test. The 6.0 m × 4.0 m course consisted of several pylons and obstacles made of foam. Subjects were instructed to walk through the course as fast as possible without touching any of the objects on the course. FMT data were used to calculate the sample size as it is the only pre/post flight data currently available with the longest recovery times (~15 days). If the other tests that are used in this study are similar to the FMT in sensitivity to spaceflight, we would fully expect to reject H0 for all tests. For example, from FMT results on 18 long-duration international space station (ISS) subjects, we found the mean change in log transit times to be 1.68 log sec with a standard deviation of 0.60 log sec [20 (link)]. With such a large signal-to-noise ratio and normally distributed differences the power of the t-test against H0 is virtually 1.0, even with as few as 10 subjects. Even if an outcome has only half the sensitivity of the FMT to spaceflight, the power with 10 subjects would still be 0.975. However, it is also important to have enough subjects to accurately estimate the mean change. If the sensitivity of a test to spaceflight were similar to that of the FMT, it would take about 13 subjects to produce a coefficient of variation of 10% for the estimated mean change post flight with respect to preflight performance. Therefore, under the assumption that sensitivities are comparable, we will require 13 long-duration astronaut subjects. We plan to target 15 subjects so we have a reserve of 2 subjects to account for subject attrition.
By including at least as much bed rest participants and control group participants as astronauts, we will ensure enough power to detect potential changes over time for these populations too.
Free full text: Click here