Associations between actigraphy-derived circadian rest-activity rhythms and cancer-related fatigue: Insights from the Harvest for Health Clinical Trial

About This Grant

PROJECT SUMMARY Cancer-related fatigue affects nearly all people undergoing cancer treatment and sometimes persists for months or years post-treatment. The etiology and pathophysiology of cancer-related fatigue are not well understood but may be related to disruption in circadian rhythms. Circadian rhythms describe the daily oscillations of metabolic, endocrine, and neuronal systems, all of which regulate energy levels. Behavioral interventions such as exercise and diet can regulate (or dysregulate) circadian rhythms, though circadian mechanisms are rarely leveraged in behavioral interventions. Ambulatory actigraphy data has become easy and inexpensive to collect, and there is a plethora of methods to analyze rest-activity rhythms from actigraphy data. However, timing is rarely incorporated into behavioral programs to address fatigue because the field has not yet defined unequivocal rest- activity rhythm parameters from actigraphy datasets that are associated with fatigue. Thus, the objective of this project is to quantify the associations between rest-activity rhythm parameters and persistent cancer-related fatigue using a battery of analysis methods. This study leverages the Harvest for Health study, which was a randomized controlled trial among older cancer survivors testing the effects of a 1-year gardening intervention vs. waitlist control. The dataset has actigraphy and fatigue data for 279 participants over two years. Participants wore an actigraph for 7 days and reported fatigue at baseline, 1 year, and 2 years. We will use traditional methods to quantify rest-activity rhythms (i.e., cosinar analysis, non-parametric measures) as well as novel hidden Markov modelling, based on a probabilistic framework. The hidden Markov models go beyond traditional methods in that they provide visual Day Profiles of activity and can quantify where in the day circadian disruptions are occurring. Aim 1 will assess the association between the rest-activity rhythm index parameters and fatigue from baseline to 1 and 2 years. We predict that higher hidden Markov-derived rhythm indices and other parameters that reflect stronger rest-activity rhythms will be associated with less fatigue. Aim 2 will assess the changes in rest-activity rhythm parameters over time. We predict that rest-activity rhythms will get stronger over time, as people get further from cancer treatment. Aim 3 will test the effects of the gardening intervention vs. control on Hidden Markov model-derived rhythm indices. We hypothesize that the gardening intervention will result in higher average rhythm index (stronger rest-activity rhythm) than the control at year 1. These findings will immediately be useful to researchers in the optimization of behavioral interventions (e.g., exercise, nutrition, sleep hygiene) to strengthen biological rhythms, address fatigue, and improve quality of life in cancer survivorship. The results and the resultant code to easily run hidden Markov models will inform future research and clinical applications. For example, these models will inform personalized recommendations for people struggling with fatigue based on a person’s rest-activity rhythm—should we focus efforts to improve sleep, encourage morning activity, minimize nutrient intake in the evening, or something else?—thereby accelerating people’s recovery from cancer.