Given the wealth of experience in pain research and the paucity of trials of fatigue as a primary endpoint, the analytic approaches adopted for analgesic trials should be extrapolated for fatigue research. Change scores, which can be summarized by subtracting the baseline data from subsequent data points, are likely to be the most informative outcomes. Because these scores take the baseline level of fatigue into account, they allow meaningful comparisons of fatigue levels over time. An alternative approach that also allows interpretable group comparisons is to control for baseline fatigue by entering it into an appropriate covariate analysis.

Fatigue studies are likely to collect a large amount of data related to the description of fatigue, potential etiologies, and comorbidities. Exploratory post-hoc analyses should be pursued in an effort to clarify the relationships among fatigue characteristics and other variables. There is so much yet to be learned about fatigue that this type of analysis is justified, and could potentially identify associations that were obscure as well as suggest future lines of research.