Hospices Deploy AI for Swifter Clinical Decisions, Improved Compliance

Hospices in recent years have enlisted artificial intelligence in their quest to match the right patient to the right care at the appropriate time. As these systems continue to develop, some are finding potential benefits when it comes to regulatory compliance as well.

Hospices that use AI systems often seek to predict when a patient may become eligible for the benefit, usually through analysis of electronic health records or Medicare claims data. The objective is to ensure that no individual patients fall through the cracks of the health care system and to bring them onto service earlier in the course of their illnesses.

”The advantage for us is that we will get patients that are hospice eligible. This software helps identify those people who have a very high risk of dying within the next six months,” Trustbridge Chief Medical Officer Faustino Gonzalez, M.D., told Hospice News. “Hopefully, we will get those patients earlier. One of the benefits is that we build as strong a case as possible [for eligibility].”

The matter of hospice eligibility is a regulatory concern as well as a clinical imperative.

The question of hospice eligibility is one of the most frequently targeted issues in regulatory enforcement. Hospice organizations are under increasing scrutiny related to medical necessity complaints under the False Claims Act.

The U.S. Department of Health & Human Services Office of the Inspector General (OIG) in 2023 will conduct a nationwide audit of hospice eligibility. The audit will focus on patients who did not have a hospitalization or emergency department visit prior to electing hospice.

To predict eligibility, Trustbridge and its referral partner Palm Beach Accountable Care Organization (ACO) recently invested in the ClosedLoop AI platform. The system ranks patients based on certain clinical indicators to predict when they will become hospice appropriate.

The tech analyzes information from the ACO’s patient database to determine the probability of death within six months. Through their preferred provider agreement, Trustbridge is then able to approach the patient to begin goals-of-care conversations.

“The literature shows that these patients who enrolled in hospice earlier in that six month trajectory have better outcomes in terms of symptom control, and their families feel that they feel better about the care their loved one received,” Gonzalez said. “That’s important for us and for the ACO.”

AI is a constantly evolving technology. A defining characteristic of these systems is that they learn as they collect data, identifying additional patterns and predictive indicators.

“It needs to be able to react and improve. There’s a prevalence to use more and more machine learning, which means that more models are being built. And that particular technology keeps getting better,” Jeremy Powell, CEO of the predictive analytics solutions firm Acclivity Health, told Hospice News. “These models keep getting sharper and sharper, so you measure things differently, with greater precision.”

Acclivity recently worked with with Oregon-based Willamette Valley Hospice & Palliative Care and the health information exchange Reliance eHealth Collaborative to improve care coordination and reduce hospitalizations.

The Acclivity system notifies the hospice when a patient presents at an emergency department, hospital or urgent care center. This allows them to intervene rapidly, often before the admissions process is complete. After receiving the notification, Willamette can confer with the hospital to help determine the best course of action for the patient.

“We’re working together to build a connected community around a health plan and their need to better understand and engage members who meet certain criteria, where Willamette can literally solve for those patient needs,” Powell said. ‘Those include the patient’s palliative needs. The tool derives specifically to support quick referral decisions.”

Ensuring that patients come into hospice early enough to experience the full benefit is a common goal for providers. A second priority is making sure that patients receive sufficient and appropriate care near the end.

Illinois-based Transitions Lifecare has used the Medalogix Muse predictive analytics system to identify patients who are nearing the last seven days of life.

Patients typically experience the highest levels of symptom burden during the last few days before death, necessitating close monitoring.

For this reason, as of 2022, the U.S. Centers for Medicare & Medicaid Services (CMS) includes “Hospice Visits in the Last Days of Life” as a required quality measure. This replaces a previous measure, “Hospice Visits When Death Is Imminent,” formerly a component of the Hospice Item Set. The principal difference between the two is the data source, the use of claims as opposed to provider reports.

Since implementing its AI system, Transitions has seen its average number of home visits during the final week of life rise to nine, up from six — a more than 50% increase.

The system alerts Transitions staff to changes in a patient’s condition.

“One of the regional directors of nursing said there was a young cancer patient who had a change from a previous visit the day before [requiring additional care],” said Melissa Popp, executive director of Transitions Care Illinois. “The director said that the patient was not even on her radar.”

Investors are also taking an interest in these technologies. Last May, Chicago-based investment firm The Vistria Group made recapitalization and growth investments in Medalogix and Muse Healthcare. The two tech companies then consolidated under the Medalogix brand.

Joining Vistria as minority investors were home health and hospice giants LHC Group (NASDAQ: LHCG), Amedisys (NASDAQ: AMED) and Encompass Health Corp. (NYSE: EHC).

Amedisys recently reported that the Medalogix Care application helped the company reduce its days to requests for anticipated payment by 40%.

“This process is more efficiently aligning the team to the patient needs,” said Amedisys COO Chris Gerard, in a statement. “Further, we create a plan of care that is tailored specifically to each patient, which is more informed by our clinical team’s interdisciplinary experience and now supported by data.”