Classification of the Condition of Cancer Patients Receiving Home Health Care with Machine Learning Methods

Abstract

Determining the health status of cancer patients is of vital importance in the cancer treatment process. This process plays a critical role in assessing patients' quality of life and supporting the treatment process. We thought that the use of machine learning in the field of cancer treatment and patient care could contribute to better patient outcomes and increased quality of life. Evaluation results of cancer patients who received home health care from XXXXXX and Research Hospital between January 2013 and August 2017 were discussed and 1000 patient files in home health service patient records were prospectively examined. In this article, cancer types were classified with machine learning methods using the Visual Analog Scale (VAS), Karnofsky performance scale, ECOG, Katz and Bartel scores to determine the quality of life of cancer patients receiving home health care. This study includes the evaluation results of 132 patients, 69 women (mean age 60.31±9.61) and 63 men (mean age 62.36±9.58). The DT classifier was noted to exhibit 83.3% accuracy and had the highest sensitivity in the lung cancer type, with a sensitivity of 88.9%. SVM classifier reached the highest accuracy compared to other classifiers with 90.2% accuracy. SVM has the highest sensitivity in lung cancers, with a sensitivity of 97.8%. The ANN classifier achieved 88.6% accuracy for all cancer types.The use of machine learning algorithms may provide a more sensitive and objective way to evaluate patients' response to treatment. The machine learning model allows determining the type of cancer using the feature space based on VAS, Karnofsky performance scale, ECOG, Katz and Bartel scores. This situation can also be constructed as an indicator in early diagnosis or risk group determination, and thus can contribute to improving home health services and increasing the quality of life of cancer patients. The results of this study may contribute to studies aimed at developing more effective strategies for the care and treatment of cancer patients.