Hospices Leverage Machine Learning to Improve Care – Hospice News

Hospice providers are using new machine learning tools to identify patients in need of their services earlier in the course of their illnesses and to ensure that patients receive appropriate levels of home visits in their final days.

While hospice utilization is rising, lengths of stay for many patients remains too short for them to receive the full benefit of hospice care. Hospice utilization among Medicare decedents exceeded 50% for the first time in 2018, according to the National Hospice & Palliative Care Organization (NHPCO). More than 27% of patients in 2017, however, were in hospice for seven days or less, with another 12.7% in hospice for less than 14 days, NHPCO reported.

While not a panacea, machine learning systems have the ability to help hospices engage patients earlier in their illness trajectory. Machine learning, a form of artificial intelligence, uses algorithms and statistical models to detect patterns in data and make predictions based on those patterns.

With machine learning you actually begin with the outcome for example, the people who were readmitted to the hospital and those who were not. Then the software itself is able to learn what the rules are, all the differences between the people who did or did not get admitted, said Leonard DAvolio, founder and CEO of the health care performance improvement firm Cyft, and assistant professor at Harvard Medical School. The advantage of the special software being able to learn these patterns, instead of the human telling it the patterns, is that software can consider so many more factors or variables than a human could, and it can do it in microseconds. Its basically checking all of the patterns that have come before and predicting the next step forward.

Machine learning systems have the ability to analyze data from claims, electronic medical records or other sources of information to predict when a patient maybe in need of hospice or palliative care, as well as which patients are at the highest risk of hospitalization, among others.

Minnesota-based St. Croix Hospice a portfolio company of the Chicago-based private equity firm Vistria Group uses a system from the recently launched technology firm Muse Healthcare, which applies a predictive model to hospice clinical data to determine which patients are likely to pass away within the forthcoming seven to 12 days.

Patients and families tend to need more intense levels of service as the patient nears their final moments. For this reason, regulators mandate that hospices collect data on the number of visits their patients receive during the last seven and the last three days of life as a part of quality reporting programs.

The U.S. Centers for Medicare & Medicaid Services (CMS) requires hospice providers to submit data for two measures pertaining to the number of hospice visits a patient receives when death is imminent. The three-day measure assesses the percentage of patients receiving at least one visit from a registered nurse, physician, nurse practitioner, or physician assistant in the last three days of life.

A second seven-day measure assesses the percentage of patients receiving at least two visits from a social worker, chaplain or spiritual counselor, licensed practical nurse, or hospice aide in the last seven days of life.

CMS currently publicly reports hospices performance on the three-day measure on Hospice Compare, but to date has not published results on the seven day measure, citing a need for further testing.

Since adopting machine learning, St. Croix has achieved a rate of 100% compliance with the visits during the final three days of life requirement, according to the companys Chief Medical Officer Andrew Mayo, M.D.

I really view it as a sixth vital sign. It provides our clinical team with additional information that helps them make decisions about care. It doesnt replace the need for human contact and evaluation, Mayo told Hospice News. Quite the contrary, it can trigger increased involvement at a time where patients, their families and caregivers may need increased hospice involvement and guidance.

Research indicates that automatic screening and notification systems makes identification of patient needs more efficient, saving hospice and palliative care teams the time consuming task of reviewing charts, allowing them to reach out to patients before receiving a physician referral or order.

Early identification of patient needs can allow hospices to more frequently apply the Medicare service intensity add-on (SIA), which increases payment to hospices for nursing visits close to the end of life.

CMS introduced SIA in 2016 to allow hospices to bill an additional payment on an hourly basis for registered nurse and social worker visits during the last seven days of a patients life in addition to their standard per diem reimbursement.

There is a top line revenue opportunity there. Along with that, were talking about additional visits in the last seven days. Theres value in terms of the outcomes that are being tracked, and those outcomes affect quality scores for the providers, Bryan Mosher, data scientist for Muse Healthcare, told Hospice News. Theres a longer term value there for them there.

Link:
Hospices Leverage Machine Learning to Improve Care - Hospice News