In older adults, physical activity (PA) is promoted for public health preventive effort. It is also a major target in the rehabilitation process. Existing assessment tools in the clinical routine do not include any aspects of PA or participation and are often prone to floor or ceiling effects. The aim of this study was to document the process of rehabilitation by activity monitoring without floor or ceiling effects. Ninety-two stroke patients of a geriatric rehabilitation unit (mean age 82 years, +/- 6.21 years, 61% women) were included in an observational study to assess physical capacity (balance, chair rise, gait speed) and PA at admission and two weeks thereafter. PA was measured through an ambulatory activity monitor based on accelerometers and gyroscopes and showed no floor or ceiling effects. Floor effects were however documented for measures of physical capacity (admission 5-11%; follow-up 2-9%) and ceiling effects were registered for the balance test (admission 17.4%; follow-up 22.8%). Improvements were documented for measures of physical capacity as well as for PA (all p < 0.001). We conclude that the assessment of PA by activity monitoring is a valuable measure to document objectively the process of rehabilitation without floor or ceiling effects.