Comparing Risks of Adverse Effects in RCTs and Observational Studies
Author Information
Author(s): Su Golder, Yoon K. Loke, Martin Bland
Primary Institution: Centre for Reviews and Dissemination, University of York
Hypothesis
Do estimates of the risk of harm outcomes differ between randomized trials and observational studies?
Conclusion
There is no average difference in the risk estimate of adverse effects between meta-analyses of RCTs and observational studies.
Supporting Evidence
- 19 studies were included, yielding 58 meta-analyses.
- 93% of the estimates of harm from different study designs had overlapping confidence intervals.
- 64% of the results agreed on whether there was a significant increase or decrease in adverse effects.
Takeaway
This study looked at different types of research to see if they agree on the risks of side effects from treatments. It found that they mostly agree.
Methodology
The study involved a meta-analysis of meta-analyses comparing pooled risk estimates from RCTs and observational studies.
Potential Biases
Observational studies may be subject to confounding factors that could affect the validity of their findings.
Limitations
The review may have missed relevant studies and relied on authors' categorizations of study designs.
Participant Demographics
The studies included a diverse range of participants, but specific demographics were not detailed.
Statistical Information
P-Value
1.03
Confidence Interval
95% CI 0.93–1.15
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website