Some trials are ground-breaking and grab a great deal of attention. Others fail to excite. What makes them different? A recent descriptive study published in Physiotherapy Theory and Practice offers some insights.
The study aimed to identify common characteristics of landmark physiotherapy clinical trials. Data were extracted from the PEDro Top 5 Trials in 2014-2019, 91 physiotherapy trials published in top medical journals (New England Journal of Medicine, Lancet, JAMA and BMJ) in 2014-2019, and 99 trials published in 2014-2019 that were randomly selected from PEDro. The characteristics that were either downloaded from PEDro or extracted from the articles were:
- trial features (total PEDro score and individual PEDro scale items; sample size; number of trial sites; use of prospective registration; positive or negative trial results; PEDro codes for subdiscipline, topic, problem, therapy, and body part)
- trial reach (number of citations in general and in guidelines; Altmetric score)
- journal characteristics (impact factor)
- author characteristics (number of publications and citations of first and last authors).
One-way independent ANOVA and Chi-squared tests were used to evaluate the between-group differences.
In terms of key trial features, the PEDro Top 5 Trials in 2014-2019 had a higher mean total PEDro score (8.0 out of 10) compared to the trials in top medical journals (6.9) and the random sample of physiotherapy trials (5.4), with the PEDro Top 5 Trials having the highest proportion of trials with concealed allocation, assessor blinding, low loss to follow-up and intention to treat analysis. Trials published in the top medical journals recruited the largest samples (mean 1,454 participants), followed by the PEDro Top 5 Trials (710) then the random set of trials (162). The PEDro Top 5 Trials were more likely to be multi-site (100%) and prospectively registered (100%) than trials published in the top medical journals (78% and 76%, respectively) and the random sample of trials (19% and 16%, respectively). The random sample of trials had the highest percentage of positive results (84%), followed by the PEDro Top 5 Trials (60%) then trials published in the top medical journals (49%). The problem was the only PEDro code was distributed differently across the trial groups.
For trial reach, the PEDro Top 5 Trials and trials published in the top medical journals had more citations in general (mean 25 for both) and in guidelines (40% vs. 37%) and higher Altmetric scores (mean 220 vs. 347) than the random sample of physiotherapy trials (mean of 2 citations, 4% cited in guidelines, mean of 17 on Altmetric).
The journal impact factor was substantially higher for the PEDro Top 5 Trials (mean 51.237) and trials published in the top medical journals (47.856) than for the random sample of physiotherapy trials (2.553).
The first and last authors of trials published in the top medical journals have more citations (mean 21,348) and publications (339) than the PEDro Top 5 Trials (6,050 and 179, respectively) and the random sample of physiotherapy trials (3,879 and 144, respectively).
The PEDro Top 5 Trials and physiotherapy trials published in the top medical journals have characteristics that may inform the design, conduct, and reporting of future physiotherapy trials. Key features are high methodological rigor, large sample sizes, multiple study sites, prospective registration, many citations (including in guidelines), high Altmetric scores, and senior authors who are highly cited and have large numbers of publications. The PEDro Top 5 Trials and physiotherapy trials published in top medical journals reached prominence despite being more likely to be ‘negative’ trials.
