Handling Missing Data in Cancer Prognostic Studies
Author Information
Author(s): Burton A, Altman D G
Primary Institution: Cancer Research UK Clinical Trials Unit, University of Birmingham
Hypothesis
How is missing covariate data reported and handled in cancer prognostic studies?
Conclusion
Missing covariate data is common in cancer studies, and many articles do not adequately report or handle this issue.
Supporting Evidence
- Missing covariate data was an issue in almost all reviewed prognostic studies.
- Only 10 out of 100 articles explored differences between complete and incomplete data cases.
- Complete case analysis and available case analysis were the most commonly reported methods for handling missing data.
Takeaway
Sometimes, researchers don't have all the information they need about patients, which can make their results less reliable. This study looked at how often this happens in cancer research.
Methodology
A review of 100 articles published in 2002 from high-impact cancer journals was conducted to assess the reporting and handling of missing covariate data.
Potential Biases
Bias may arise from the handling of missing data, potentially affecting the validity of the study conclusions.
Limitations
The review only included articles from specific journals and may not represent all cancer studies.
Participant Demographics
The studies reviewed included a range of cancer types, with breast cancer and digestive organ cancers being the most common.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website