A network for students interested in evidence-based health care

The effect of intensity of therapy upon length of stay in a neurological rehabilitation setting: a critical review

Posted on 4th January 2017 by

Evidence Reviews
A randomised control trial to determine the effect of therapy upon length of stay in a neurological rehabilitation setting. 


Neurological events such as stroke or traumatic brain injury can have significant consequences for the individual and their families. Currently, levels of optimal therapy for inpatient rehabilitation in the UK are uncertain. However, outside the UK, facilities are often required to provide at least 3 hours of therapy per day in order to be accredited as a ‘rehabilitation facility’. Research into this area can be extremely difficult (although not impossible) due to the varying symptoms and consequences of disease. Previous research has looked at the intensity of therapy and length of stay, but hasn’t taken into account the effects of impairments and disability mix among participants.

The study: What they did and how they did it

The researchers recruited 161 participants with various neurological diagnoses. Participants were subsequently randomized into control and intervention groups. The aim of this study was to determine the effect of intensity of therapy (physiotherapy and occupational therapy) upon length of stay in a neurological rehab setting and also to test the hypothesis that patients were discharged at a similar level of functioning. Due to the method of timetabling in this particular setting, randomisation into groups occurred the week before admission. As a result, 20 participants were not included in the trial due to consent refusal. The participants were randomised into a control and intervention group by block randomisation prior to admission to allow for timetabling.

The intervention group received 67% more therapy in any given period compared to the control group. The medical staff making decisions on discharge were blinded to the participants’ group allocations. The primary outcome was length of stay in the rehab unit. Extent of dependency was also recorded using the modified Barthel index (a scale to measure performance in activities of daily living) and cognitive impairments were measured using the mayo scale. These were tested upon admission and prior to discharge.

The results

This study found that, after adjusting for confounders, higher levels of therapy led to a non-significant reduction of 5 days length of stay was found in the experimental group. However when taking impairment mix into account, this difference increased to a significant 14 day reduction. Although the results varied greatly depending on which method of data analysis was used. No differences were found in Barthel scores between groups and mayo scale scores are not reported on, or accounted for, in the results.

Strengths and weaknesses of the study

To evaluate the study systematically, the CASP RCT tool will be used as a guideline.

  • The trial clearly identifies and addresses the population, intervention used, control group and outcome measures for the study.
  • Participants were randomized using block randomisation which prevents unequal group sizes, minimizes bias and increases the validity of the trial. Randomisation however occurred prior to consent, to allow for timetabling. Following consent refusal of 20 participants, this left group sizes of 75 in the experimental group and 66 in the control group. Nonetheless, the groups remained similar in characteristics such as gender, age, days since event, outcome measure scores and neurological condition.
  • Therapists treated a mix of control and experimental group participants, which could lead to some potential bias.
  • Due to the nature of the trial, blinding of the therapists was not possible. However those making decisions about discharge were blinded to which groups the participants were allocated.
  • The length of stay in a neurological rehabilitation setting was the primary outcome of this trial and is stated clearly. The Barthel index was used as an outcome measure to determine the level of dependency of the participants. This outcome measure has excellent intra and inter rater reliability and excellent validity (Hseuh et al. 2002). The Mayo scale was also used as an outcome measure. However the results were not reported on, suggesting the researchers are selectively reporting data.
  • The sample size of the study (141 participants) was relatively large. However 161 participants were initially recruited as the study justifies their sample size by stating 160 patients (80 in each group) were needed to give the study power of 80%. Intention to treat analysis was used on all participants who consented to the trial. This is used to assess the clinical effectiveness of the treatment and takes into account that not everyone will adhere to treatment (NICE 2016).
  • The trial was designed to take into account confounding factors and impairment mix. However results varied significantly depending on which regression model was used to analyse the data. This highlights the vulnerability of the trial design when applied in the real clinical setting.
  • Participants for this trial were recruited from 1995 – 1997 and the study was published in 2002, considering this trial is approximately twenty years old and the varied results I wouldn’t consider this trial to be clinically significant or valid.


The clinical relevance of this study is unclear. Whether or not the difference between the high intensity therapy group and the control group was statistically significant depended on the type of statistical analysis the researchers used. Further research is needed in this area, with more complex trial designs and analysis methods to determine the true effect.


CRITICAL APPRAISAL SKILLS PROGRAMME (CASP) – MAKING SENSE OF EVIDENCE., 2013. CASP Randomised Controlled Trial Checklist [online]. [Viewed 30 November 2016]. Available from: http://media.wix.com/ugd/dded87_40b9ff0bf53840478331915a8ed8b2fb.pdf

HSUEH, I.-P., LIN, J.-H., JENG, J.-S. and HSIEH, C.-L., 2002. Comparison of the psychometric characteristics of the functional independence measure, 5 item Barthel index, and 10 item Barthel index in patients with stroke. Journal of Neurology, Neurosurgery & Psychiatry [online]. vol. 73, no. 2, pp. 188–190 [viewed 1 December 2016]. Available from: http://jnnp.bmj.com/content/73/2/188.full

NICE., 2016. Glossary [online]. NICE [viewed 1 December 2016]. Available from: https://www.nice.org.uk/glossary?letter=i

SLADE, A., TENNANT, A. and CHAMBERLAIN, M. A., 2002. A randomised control trial to determine the effect of therapy upon length of stay in a neurological rehabilitation setting. Journal of rehabiliation medicine [online]. June, pp. 260–266 [viewed 30 November 2016]. Available from: https://www.researchgate.net/profile/Anita_Slade/publication/11027252_A_randomised_controlled_trial_to_determine_the_effect_of_intensity_of_therapy_upon_length_of_stay_in_neurological_rehabilitation_setting/links/54e759610cf2cd2e02938d32.pdf84


Jane Arthur

A physiotherapy student from Queen Margaret University, Edinburgh. View more posts from Jane

Leave a Reply

Your email address will not be published. Required fields are marked *

Subscribe to our newsletter

You will receive our monthly newsletter and free access to Trip Premium.