Lessons Learned: The Difficulties of Incorporating Intensity Principles Into Inpatient Stroke Rehabilitation.

The trial was pseudo-randomized by alternating successive admissions according to clinic needs, so group assignment was not truly random and allocation was not concealed. The treating physical therapists also delivered the intervention, collected outcomes, and knew the design, which can bias treatment and assessment. The sample was small, with 16 intensity completers and 19 usual-care completers, and the authors noted the sample size as a limitation. Compliance was low: the intensity group received 6.54 sessions versus an expected 14.5, so the results may reflect a partially delivered program. Patients had to provide informed consent and follow motor commands, so those with severe dementia or more impaired cognition were excluded. Some baseline scores were set to zero or replaced with the lowest performer's time when patients could not complete measures, which can distort change. Only 49 individuals were consented out of approximately 700 stroke admissions at the sites, so the participants may not represent the full inpatient stroke population. Baseline walking speeds were 0.45 m/s for control and 0.40 m/s for intervention, and the authors noted this may indicate selection bias toward higher functioning patients. Speech therapy and other therapies shared a total 3 hours per day, limiting time for additional physical therapy sessions.

The supplied text does not include a funding or conflict-of-interest declaration. It states the investigation was not funded for additional treatment time.

Do not read the larger gains as proof that the intensity program worked as intended, because patients received far fewer sessions than planned and the trial was pseudo-randomized with therapists also doing assessments.