How Healthcare Machine Learning Is Improving Care Management: Ruth’s Story
Liz, a nurse care manager in the local community health system’s care management program, loves her work. Each morning, she arrives at work eager to be involved in developing, implementing, evaluating, and managing patient care plans for approximately 200 chronic disease patients in her panel. On any given day, she could play the role of a nurse, social worker, teacher, or advocate for her patients, in addition to serving as a key liaison between patients, their families, other members of the care team, and community support agencies involved in each patient’s care. Being a skilled clinician and very relationship-centered, Liz was particularly well suited for the critically important clinical, communication, and coordination role of a care manager.
One morning, upon gaining access to the care management system, Liz’s attention was immediately drawn to a system alert regarding Ruth, an 80-year-old retired marketing executive who was part of her panel. Ruth lived in a nearby high-rise retirement community that was specifically designed to allow this fiercely independent patient to remain at home for as long as possible—a key retirement goal. The retirement center is wired with ubiquitous sensors—in the walls, appliances, and every room. There were even sensors in Ruth’s cane to monitor her movements, speed, and location.
Liz was concerned because there was evidence that Ruth was clinically deteriorating. She was spending more time in bed, eating less, and not taking her medications as rigorously as usual. Furthermore, Ruth’s blood pressure was down and her weight was rising. Fortunately, this was only a two-day trend and Liz recognized that she needed to act before things got worse. Liz called Ruth and learned that she had developed a cough, fever, and chills two days previously. Liz suspected that Ruth had acquired a viral infection that was spreading throughout the community, resulting in Ruth being unable to eat or take her medications appropriately. The weight gain could indicate a worsening of Ruth’s heart failure. Liz arranged for a home visit later that day. Examination confirmed the diagnosis of a viral illness and early-stage heart failure.
Liz moved Ruth’s digital medication dispenser to a bedside stand so she could resume taking her medications regularly, including her heart medication. She also instructed Ruth on how to stay hydrated, and arranged for the retirement community staff to deliver Ruth’s meals to her bedside over the next few days, temporarily adjusting her usual diet to accommodate her early-stage heart failure. Subsequent phone calls and one additional visit confirmed that Ruth responded to treatment. Her viral symptoms subsided and her appetite improved, along with her vital signs. Within four days, Ruth was back to herself, enjoying life again.
Clearly, this is a nice example of the value of an effective care management program. It also highlights the value of home monitoring and the use of care management systems—and, eventually, healthcare machine learning, predictive analytics, and artificial intelligence (AI)—to allow care providers to be the best they can be.
Healthcare Machine Learning Will Address Massive Growth in Healthcare Data
Like Liz, a growing number of care managers have patients that are supported by digital devices at home. This number is expected to grow dramatically over the next several years—easily into the millions for large health systems. As the number of patients using these devices grows, it will result in massive healthcare data streams being delivered to the associated health system. It would be impossible and too costly to hire enough people to monitor these extremely large data streams daily. However, they need to be regularly monitored to detect problems early and maintain optimal population health.
This massive growth in data ultimately will result in the need for intelligent data systems to identify or even predict problems to support care managers and health systems as they strive to optimize the health of populations under their care. Enter the need for healthcare machine learning, predictive analytics, and AI.
Healthcare Machine Learning Has an Increasingly Important Role in Care Management
Predictive analytics has been defined as the practice of extracting information from existing data sets to determine patterns and predict future outcomes and trends. Predictions made solely for the sake of making a prediction are a waste of time and money. In healthcare, prediction is most useful when that knowledge can be transferred into action. The willingness to intervene is the key to harnessing the power of historical and real-time data. Importantly, to best gauge efficacy and value, both the predictor and the intervention must be integrated within the same system and workflow routinely used by clinicians as they deliver care.
Machine learning and AI have become commonplace and indispensable for solving complex problems in most scientific fields and many industries. They are also rapidly becoming some of the most-discussed, perhaps most-hyped topics in healthcare analytics. Healthcare can learn valuable lessons from previous success in other disciplines to jumpstart the utility of predictive analytics for improving patient care, chronic disease management, health system administration, supply chain efficiencies, financial management, and cost control. The opportunity that currently exists for healthcare systems is to define what predictive analytics means to them and how it can be used most effectively to make improvements.
Ultimately, predictive analytics will play an increasingly important role in care management and population health programs. It will undoubtedly help identify patients like Ruth who need special attention. This is a relatively simple problem for such powerful predictive tools to solve.
Effective, Comprehensive Care Management Requires Healthcare Machine Learning
Healthcare machine learning, predictive analytics, and AI will allow health systems and care management teams to make care more efficient and appropriate as we manage ever-growing populations of patients in the face of always finite resources. That is, it will allow them to move from simply identifying a patient with a straightforward acute need (e.g., this patient has the flu and heart failure, and needs special assistance) to a more comprehensive, thoughtful, and appropriate approach to care management (e.g., this patient has the flu, heart failure, diabetes, asthma, and depression and, based on all of his or her characteristics, risks, beliefs, desires, and genetic profile, will have the best chance of the desired outcome if this specific bundle of interventions and treatments are applied, in this manner, from these specific resources, at this specific time that optimizes the health of both the individual and population at the lowest possible cost). Healthcare machine learning will drive the industry toward truly comprehensive, effective care management.