Machine Learning Model for GI Health in Cancer Survivors
Author Information
Author(s): Han Claire J., Ning Xia, Burd Christin E., Tounkara Fode, Kalady Matthew F., Noonan Anne M., Von Ah Diane
Primary Institution: The Ohio State University
Hypothesis
Can machine learning models classify gastrointestinal health status in cancer survivors using telomere length and social determinants of health?
Conclusion
The study developed machine learning models that effectively classify gastrointestinal health in cancer survivors, highlighting the importance of telomere length and social determinants of health.
Supporting Evidence
- Random forest model achieved an AUC of 0.98 in training data.
- Gradient boosting machine model showed an AUC of 0.80 in test data.
- Longer telomere length and higher income were associated with better GI health.
- Social determinants of health significantly influenced GI health outcomes.
- Machine learning models can handle complex relationships among multiple variables.
Takeaway
This study used computers to help figure out how healthy people's stomachs are after cancer treatment, looking at things like their age and money situation.
Methodology
The study analyzed data from 645 adult cancer survivors using machine learning models to classify gastrointestinal health based on telomere length and social determinants.
Potential Biases
Potential overfitting due to differences in model performance metrics between training and test datasets.
Limitations
The study's cross-sectional design limits predictability, and the sample size for the test dataset was small.
Participant Demographics
Participants were primarily non-Hispanic Whites, with a mean age of 66.3 years, and included both men and women.
Statistical Information
P-Value
p<0.05
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website