Home > Journals > European Journal of Physical and Rehabilitation Medicine > Past Issues > Articles online first > European Journal of Physical and Rehabilitation Medicine 2020 Sep 16

CURRENT ISSUE
 

JOURNAL TOOLS

eTOC
To subscribe
Submit an article
Recommend to your librarian
 

ARTICLE TOOLS

Publication history
Reprints
Permissions
Cite this article as

 

 

European Journal of Physical and Rehabilitation Medicine 2020 Sep 16

DOI: 10.23736/S1973-9087.20.06451-5

Copyright © 2020 EDIZIONI MINERVA MEDICA

language: English

Randomized controlled trials in non-pharmacological rehabilitation research: a scoping review of the reporting of sample size calculation, randomization procedure, and statistical analyses

Mohammadreza AMIRI 1, Dinesh KUMBHARE 2

1 Toronto Rehabilitation Institute, University Health Network, Toronto, Canada; 2 Department of Medicine and Institute of Biomaterials and Biomedical Engineering, University of Toronto, Toronto, Canada


PDF


BACKGROUND: The RCTs are often considered as the gold standard for clinical trials and researchers argue that the quality of RCT reports should be of the highest standards due to their clinical significance.
OBJECTIVES: To review the quality of reporting of the sample size calculation methods, and randomization procedures, and assess whether the statistical analyses correlate as reported in the trials’ methods and results sections in non-pharmacological, physiological rehabilitation randomized controlled trials (RCT) interventions.
METHODS: A systematic electronic search was conducted in Cochrane Central from 1 January 2019 to 16 December 2019. Titles and abstracts were analyzed for inclusion independently by two authors, and disagreement(s) were resolved by a third reader. Studies were included if they met the following criteria: i) Assessed and reported a type of non-pharmacological rehabilitation RCT (e.g. physiotherapy); ii) randomized intervention to patients with a disease comparing to healthy or patients without intervention as the comparison group; iii) published in an indexed journal; and, iv) original research, available full text, human study, published in 2019, and written in English. The following information was extracted from the included articles: journal impact factor (JIF), sample size calculation methods (SS), randomization procedure (RND), and statistical analyses (STAT) reported. Analyzing the full text, whether
SS and RND were reported or not and whether the STAT correlateds with the Methods and Results sections. The prevalence of each statistical method was derived from the Methods section of the report and compared if it was reported in the Results section. The continuous variable of JIF was tested for normality and used for independent t-test for equality of means between categories. In addition, using Downs and Black checklist the methodological quality of the articles was assessed and categorized to be poor, fair, good, and excellent based on the checklist’s score. Finally, the association between the assessed quality (categorical variable) of the articles and the reporting variables (categorical variables) was analyzed utilizing the Person χ2.
RESULTS: One hundred and nighty-four articles were retrieved from the systematic search out of which 99 (51%) were included for data extraction and further analyses. About one in five (20.2%) and two in five (37.4%) did not properly and adequately report the SS and RND while one in five (19.2%) there was at least one mismatch in STAT. The JIF was not significantly associated to the quality of reporting of SS (t = 1.974, p = 0.051), RND (t = 0.309, p = 0.758), and STAT (t = -0.275, p = 0.784). This finding could indicate that the quality of the journal did not assure the quality of the reporting these methods. However, there was a significant association between the assessed quality of the article measured with the Down’s and Black checklist and the reporting of SS (X2 = 29.149, df = 2, p <0.0001), RND (X2 = 55.079, df = 2,p <0.0001), and STAT (x2 = 25.778, df = 2,p < 0.0001).
CONCLUSIONS: Recent reporting quality of non-pharmacological rehabilitation RCTs was investigated. We found that the quality of the article but not the quality of the journal in which it is published in may be associated to the quality of reporting in sample size methods, randomization processes, and statistical analyses reporting. The quality of study reporting may be enhanced utilizing a guideline that addresses the required information in sample size calculation, randomization of individuals, and proper statistical analyses used and reported.


KEY WORDS: Randomized control trials; Non-pharmacological rehabilitation; Sample size; Randomization

top of page