Warning: This is an author submission that is not peer-reviewed or edited. Preprints - unless they show as "accepted" - should not be relied on to guide clinical practice or health-related behavior and should not be reported in news media as established information.
Machine Learning in Palliative Care: A Scoping Review of Applications
ABSTRACT
Background:
Palliative care has been increasingly positioned as a key element in the health system in response to the aging population and the growing numbers of those dying from life-limiting illnesses. The advent of big data is changing every industry, stimulating interest in using advanced analytical methods such as machine learning (ML). Studies leveraging advances of ML in healthcare pay little attention to PALC. Clinical decision making in PALC is complex and the prognostic uncertainties present unique challenges to care providers. ML based prediction models can help identify PALC candidates and improve the accuracy of prognostication and improve planning for health system resourcing.
Objective:
This study undertakes a scoping review of research on applications of ML methods in palliative care.
Methods:
The scoping review was conducted using the methodological framework of Arksey and O’Malley, with reporting guided by Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews. A comprehensive search of Medical Literature Analysis and Retrieval System Online, EMBASE, PsycINFO, Web of Science, Scopus and Cumulative Index to Nursing and Allied Health was performed for studies published by April 15, 2021. Two authors assessed eligibility and extracted data.
Results:
Results This scoping review applied a structured search of 6 databases in April 2021, revealing 46 publications. The studies applied ML for predicting mortality outcomes predominantly (55%), followed by other outcomes such as hospitalizations and palliative care interventions (21%), improving / assessing end of life process / quality indicators (13%), phenotyping for patient profiling and screening and disease classification (11%).
Conclusions:
We identified areas where ML methods are being applied to improve palliative care. The review can serve as a guide to identify gaps in knowledge and new areas where these methods can be applied for the benefit of palliative care patients, providers, policy makers and system planners.
Citation
