Data including demographics, type of HCT (e.g., allogeneic or autologous), preexisting physical conditions (e.g., chronic joint injury), CRF, steroid use data, ECOG and KPS scores are collected from patients' medical charts at time of enrollment. Patients are prescribed participation in a primarily self-directed physical activity program which encourages them to spend 6 hours out of bed daily and to perform 30 minutes of light-to-moderate daily aerobic activity. Patients who are able to maintain independent mobility undergo physical therapist assessment 2 times a week until hospital discharge. Patients wear a physical activity monitoring device and daily activity and sleep data are collected continuously during hospital LOS.

Review of medical chart

Will focus on trends in scores between patients with different durations of activity and sleep. Results, from both descriptive and inferential analyses, will be presented in table and/or graphical format.

Will focus on trends in scores between patients with different durations of activity and sleep. Results, from both descriptive and inferential analyses, will be presented in table and/or graphical format.

Will use interval-censored regression to estimate distributions of activity from one day to the next. This repeated-measures analysis with a polynomial model will generate average curves for activity, each spanning average LOS (i.e., activity duration plotted on day one, day two, day three, etc.). Will also run diagnostics to verify the assumptions of a linear model (e.g., statistical independence of observations and lack of undue influence of outliers on model fit).

Will use interval-censored regression to estimate distributions of sleep from one day to the next. This repeated-measures analysis with a polynomial model will generate average curves for sleep, each spanning average LOS (i.e., sleep duration plotted on day one, day two, day three, etc.). Will use multiple linear regressions to control for potential confounders in predicting sleep duration. Will also run diagnostics to verify the assumptions of a linear model (e.g., statistical independence of observations and lack of undue influence of outliers on model fit).