Evaluation of Statistically Significant Findings in Orthopaedic Trauma Trials
Author Information
Author(s): Sung Jinsil, Siegel Judith, Tornetta Paul, Bhandari Mohit
Primary Institution: Boston University and McMaster University
Hypothesis
Statistically significant studies would not always report large treatment effects.
Conclusion
Statistically significant results in orthopaedic trials often report large risk reductions, but these are correlated with fewer total outcome events.
Supporting Evidence
- 76 randomized controlled trials with statistically significant findings were included in the analysis.
- The mean effect size for continuous outcomes was 1.7.
- 61% relative risk reduction was observed on average for dichotomous outcomes.
- Fewer outcome events were strongly correlated with larger treatment effects.
Takeaway
This study looked at trials for broken bones and found that when fewer people are studied, the results can show bigger improvements, but we should be careful about trusting those results.
Methodology
A comprehensive search for randomized controlled trials from 1995 to 2004 was conducted, focusing on orthopaedic trauma, with data abstracted by two reviewers.
Potential Biases
Small sample sizes in trials may lead to underpowered conclusions and inflated treatment effects.
Limitations
The study may have missed relevant trials due to indexing errors and focused only on orthopaedic trauma, limiting generalizability.
Participant Demographics
The trials included a total of 9757 patients, with sample sizes ranging from 10 to 424.
Statistical Information
P-Value
0.006
Confidence Interval
95% confidence interval: 1.43–1.97
Statistical Significance
p<0.01
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website