Problems with Using 'Optimal' Cutpoints in Cancer Prognosis
Author Information
Author(s): DG Altman
Primary Institution: ICRF Medical Statistics Group, Centre for Statistics in Medicine, Institute of Health Sciences, Oxford OX3 7LF, UK
Hypothesis
The use of the 'optimal' cutpoint method in cancer prognosis is misleading.
Conclusion
The 'optimal' cutpoint method can lead to high false-positive rates and misinterpretation of tumor marker significance.
Supporting Evidence
- The minimum P-value method has a false-positive rate of around 40%.
- The reported P-values can be misleadingly small.
- The cutpoint derived from this method lacks clinical meaning.
- Using this method can distort results in subsequent meta-analyses.
Takeaway
Some doctors use a method to find the best cut-off point for cancer tests, but it can give wrong results and make things look better than they are.
Methodology
The study critiques the minimum P-value method used in cancer prognosis analyses.
Potential Biases
The method transfers bias from univariate to multivariate analyses.
Limitations
The method can lead to misleading results and does not provide a true indication of the importance of tumor markers.
Statistical Information
P-Value
0.002
Statistical Significance
p<0.05
Want to read the original?
Access the complete publication on the publisher's website