Biases in Chronic Disease Trials
Author Information
Author(s): Vollenweider Daniela, Boyd Cynthia M, Puhan Milo A
Primary Institution: Johns Hopkins University
Hypothesis
Methodological deficiencies in trials for chronic diseases could bias effect estimates and clinical practice.
Conclusion
Many chronic disease trials fail to define a primary outcome and use inadequate methods for subgroup analyses and handling missing data, leading to potential biases.
Supporting Evidence
- Only 21% of trials had a single primary outcome.
- 68% of trials did not report statistical comparisons of baseline characteristics.
- 42% of trials reported an intention-to-treat analysis.
Takeaway
This study looked at trials for chronic diseases and found that many don't clearly state what they're measuring, which can lead to mistakes in understanding the results.
Methodology
Survey of 161 randomized trials on chronic diseases, assessing trial design and analysis aspects.
Potential Biases
High risk of confounding and selection bias due to poor trial design.
Limitations
Did not assess how the trial design aspects affected the effect estimates.
Participant Demographics
Trials included patients with chronic obstructive pulmonary disease, heart failure, stroke, and type 2 diabetes.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website