Machine learning (ML) can provide significant value in the field of palliative care. However, according to a thorough analysis published in HRB Open Research, researchers still have a lot of unexplored ground to cover before the technology reaches its full potential.
“ML is a subset of artificial intelligence that is rapidly advancing capacity to identify patterns and make predictions using large datasets,” wrote Virginia Storick, Trinity College Dublin in Ireland, and colleagues. “In contrast to traditional analytic methods, where the analyst specifies data inputs according to hypotheses and/or conceptual models, ML approaches leverage computing power to identify patterns in available data and can make inferences without explicit user instruction.”
To see how researchers were exploring ML technology, Storick et al. analyzed seven databases for peer-reviewed studies of how it can improve palliative care for adults. The specific outcomes they looked for were survival, quality of life (QoL), place of death, costs and receiving “high-intensity treatment” near the end of life.
Ultimately, the team explored three studies that examined ML and palliative care. The first study involved predicting six-month mortality, the second involved predicting 12-month mortality and the third involved cross-referencing 12-month mortality with healthcare costs.
These are three key concerns the authors expressed about the current state of machine learning-based palliative care research:
1. Explore a variety of patient outcomes
“No included study examined patient outcomes such as QoL,” the authors wrote. “Evaluations of QoL are not straightforward because the outcome of interest is an individual and subjective concept where mortality is an observable binary state. Nevertheless, the reality is that living and dying with serious illness, and caring for those populations, is messy and complex. Studies characterizing palliative care need beyond mortality, for example those at risk of pain or unmet need or death anxiety, and accurately predicting risk would have the capacity to improve clinical decision-making and treatment pathways.”
2. Include the perspective of caregivers
While the patient is always going to be the center of these studies, researchers should also consider the role of unpaid caregivers. The authors noted that evaluations of “dyad and family outcomes” are becoming more and more common in palliative care research—a trend that did not carry over into the ML-based literature they read.
“Identifying caregiver needs in advance would also have vast potential benefit,” they added.
3. Cover as much time as possible
Storick and colleagues noted that the studies all included a timeframe of six to 12 months. If more time was covered, the findings could potentially have a more lasting affect on patient care.
“Treatment choices from diagnosis have the greatest scope to impact outcomes and costs, and so studies that can inform these choices are the most useful,” they wrote.
Overall, the authors explained, ML has shown “the potential to support clinicians in improved decision-making” when it comes to palliative care for adult patients. But the research up to this point has been lacking in many areas.
“Applications of ML approaches to policy and practice remainsformative,” they concluded. “Derived results depend on available data and must be interpreted in this context. Future research must not only expand scope to consider other outcomes and longer timeframes, but also address individual needs and preferences in the context of prognosis, and engage with the profound ethical challenges of this emerging field.”