Machine Learning-Based Palliative Performance Status (PPS) Prediction in Palliative Care Patients Who Have Undergone Hypodermoclysis

Abstract

Palliative care aims to provide a better quality of life, in addition to preventing and alleviating the multidimensional suffering of patients with life-threatening diseases and is of great relevance for cancer patients. In this context, the need to improve techniques that minimize negative impacts, such as pain and other discomforts, is highlighted, and it is essential to disseminate knowledge about the use and benefits of hypodermoclysis. In addition, the study explores an innovative dimension by addressing the application of Machine Learning in Public Health, demonstrating its viability as a predictive tool in the context of palliative care. This study aimed to analyze the use of hypodermoclysis in patients admitted to the Oncology Palliative Care Unit of the Hospital de Apoio de Brasília – HAB/SES-DF, relate sociodemographic and pathological data and, above all, evaluate the effectiveness of Machine Learning algorithms in predicting the Palliative Performance Scale (PPS) and in analyzing its degradation based on sociodemographic variables and clinical evolution recorded in medical records.

The study also reinforces the potential of Machine Learning in Public Health, especially by demonstrating the ability of these algorithms to predict the clinical evolution of cancer patients, as evidenced by the prediction of PPS and its degradation over time. The results highlight the applicability of machine learning techniques as a promising approach to improve palliative care management, enabling more accurate and efficient decision-making, in addition to contributing to the planning of personalized interventions that meet the specific needs of each patient.