PreprintArticleVersion 1Preserved in Portico This version is not peer-reviewed
Machine Learning-Based Assessment Indicates Patient-Reported Outcomes to Be Less Reliable than Clinical Parameters for Predicting 1-Year Survival in Cancer Patients.
Version 1
: Received: 31 October 2023 / Approved: 1 November 2023 / Online: 1 November 2023 (08:51:18 CET)
How to cite:
Salvador Comino, M. R.; Youseff, P.; Heinzelmann, A.; Bernhardt, F.; Seifert, C.; Tewes, M. Machine Learning-Based Assessment Indicates Patient-Reported Outcomes to Be Less Reliable than Clinical Parameters for Predicting 1-Year Survival in Cancer Patients.. Preprints2023, 2023110042. https://doi.org/10.20944/preprints202311.0042.v1
Salvador Comino, M. R.; Youseff, P.; Heinzelmann, A.; Bernhardt, F.; Seifert, C.; Tewes, M. Machine Learning-Based Assessment Indicates Patient-Reported Outcomes to Be Less Reliable than Clinical Parameters for Predicting 1-Year Survival in Cancer Patients.. Preprints 2023, 2023110042. https://doi.org/10.20944/preprints202311.0042.v1
Salvador Comino, M. R.; Youseff, P.; Heinzelmann, A.; Bernhardt, F.; Seifert, C.; Tewes, M. Machine Learning-Based Assessment Indicates Patient-Reported Outcomes to Be Less Reliable than Clinical Parameters for Predicting 1-Year Survival in Cancer Patients.. Preprints2023, 2023110042. https://doi.org/10.20944/preprints202311.0042.v1
APA Style
Salvador Comino, M. R., Youseff, P., Heinzelmann, A., Bernhardt, F., Seifert, C., & Tewes, M. (2023). Machine Learning-Based Assessment Indicates Patient-Reported Outcomes to Be Less Reliable than Clinical Parameters for Predicting 1-Year Survival in Cancer Patients.. Preprints. https://doi.org/10.20944/preprints202311.0042.v1
Chicago/Turabian Style
Salvador Comino, M. R., Christin Seifert and Mitra Tewes. 2023 "Machine Learning-Based Assessment Indicates Patient-Reported Outcomes to Be Less Reliable than Clinical Parameters for Predicting 1-Year Survival in Cancer Patients." Preprints. https://doi.org/10.20944/preprints202311.0042.v1
Abstract
Machine learning (ML) techniques can help predict survival among cancer patients and might help with a timely integration in palliative care. We aim to explore the importance of subjective variables self-reported and collected via electronic patient reported outcome measure (ePROM) for survival prediction. A total of 256 advanced cancer patients met the eligible criteria. We analyzed objective variables collected from electronic health records, subjective variables collected via ePROM and all clinical variables combined. We used logistic regression (LR), decision trees, and random forests to predict 1-year mortality. Receiver operating characteristic (ROC) curve - area under the curve (AUC) and the ML models feature importance were analyzed. The performance of all variables for predictions (LR reaches 0.80 [ROC AUC] and 0.72 [F1 Score]) does not improve over the performance of only clinical non-patient reported outcome (non-PRO) variables (LR reaches 0.81 [ROC AUC] and 0.72 [F1 Score]). Our study indicates that patient-reported outcome (PRO) variables, which measure subjective burden, cannot be reliably used to predict survival. Further research in this area is needed to clarify the role of self-reported patient's burden and mortality prediction via ML.
Keywords
Patient Reported Outcome Measure; Artificial Intelligence; Machine Learning; Predictive Analytics; Cancer Patients; Palliative Care
Subject
Medicine and Pharmacology, Medicine and Pharmacology
Copyright:
This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.