Analytics, machine learning aid hospice in helping patients

Infinity Hospice is using a population health management platform to improve services to Medicare patients.

The end-of-life care organization, which serves the Phoenix, Reno and Las Vegas metropolitan areas, is using guidance from the newly developed application, which uses artificial intelligence and machine learning to determine which patients are entering the final stages of life.

Hospices seek to incorporate family members and palliative and other types of care to patients during their final days of life. But it can be difficult to determine when a patient is near the end, especially when the hospice is serving 200 patients, says Darren Bertram, CEO of Infinity Hospice.

Also See: New CMS website will aid consumers in selecting hospice care

Finances also come into play. How well a hospice cares for patients in their last week in part determines how much Medicare will pay the hospice. That’s why the hospice became a beta site for new software from WellSky that uses artificial intelligence and machine learning technology to ensure patients are with family at home and getting the right care.

Machine learning, for instance, uses real-time clinical, symptomatic and psychosocial data to assess whether a patient is in the last week of life and reduces the risk of medical service utilization errors.



“WellSky allowed us to increase the skilled care hours we spend with patients in the last seven days of their life by 61 percent,” Bertram explains. “We’ve had several occasions where we were able to send a nurse out to a patient’s home in time for them to be with the patient and the family on the day of their death.”

However, initial use of the software to identify the sickest patients caused concern among some clinicians, Bertram recalls. “A clinician may be in charge of 60 patients and knows who they all are and knows it because they’ve been doing this for 10 years, and now we changed their routines. This is the biggest change they have seen, even bigger than going from paper processes to electronic charting. It took three or four weeks to get everyone fully on board.”

Despite the concerns, automating the process of identifying patients most in need has been well worth the effort, Bertram adds. “We believe that one of the most important quality metrics is the time spent with patients in the last seven days of life.”