Skip to main content

Trending From ACCC

Integrating Artificial Intelligence and ePROs Into Clinical Workflows

Authored by

Association of Community Cancer Centers


J Clin Pathways. 2021;7(3):34.

ACCCThe cost of health care in America has grown at a startling rate in the past several decades. According to a 2019 survey of highly developed countries by the Organisation for Economic Co-operation and Development (OECD), the United States spends 16.9% of its gross domestic product on health care—almost double the OECD average of 8.8%—and spends nearly 20% more on drugs.

Statistics show that the quality of care has not necessarily kept pace with those increasing costs. According to 2015 statistics from the OECD, mortality for 6 of the 7 leading causes of death are higher in the United States than in comparable countries (cancer-related death being the only exception). These two trends run in opposition to the goals of value-based care—increased quality at reduced cost—and can threaten a cancer center’s participation in value-based programs like the Oncology Care Model.

In a session at the Alternative Payment Model Coalition Workshop at the ACCC 47th Annual Meeting & Cancer Center Business Summit, Jeff Hunnicutt, CEO of Highlands Oncology Group, shared the story behind two of his program’s innovative initiatives that have driven down the cost of care while improving care quality for their patients.

Electronic Patient-Reported Outcomes (ePROs)

No matter the field of medicine, understanding how a patient is responding to treatment is critical to ensuring positive patient outcomes. But in oncology, there are often gaps between office visits where cancer care teams have traditionally been unable to monitor their patients. In that time, patients may experience hardships and symptoms that lead to unnecessary emergency room (ER) visits or discontinuation of their treatment. “So many things can happen inside of a patient’s journey of care when they’re not inside of our own walls,” said Mr Hunnicutt. Finding a way to be available to patients between visits—in a manner where patients feel comfortable sharing how they feel—became a priority for Highlands Oncology Group (Highlands), a large independent practice in northwestern Arkansas.

In June of 2020, Highlands began to utilize Expain, a mobile-based app that allowed patients to submit ePROs from their phone. Highlands targeted patients on active treatment and educated them about the app and their chemotherapy treatment simultaneously, integrating its installation and use into the clinical workflow. The cancer center could then send notifications to patients on their phones, prompting them to respond in the app and indicate if they were experiencing any side effects from treatment. This information is sent to a centralized triage staff that could reference easy-to-follow triage pathways and contact providers and patients if any action needed to be taken.

As of March 5, more than 2000 patients had enrolled in Expain at a 99% opt-in rate (83% for established patients). Most patients (78%) used the app more than once a week based on notifications that are customized to their treatment, diagnosis, and risk level. With the acceleration in adoption of new technology in response to the COVID-19 pandemic, older patients have been more receptive to the use of the app than they were for previous app-based solutions the practice had tried, Mr Hunnicutt said.

Mr Hunnicutt shared a story of the app in clinical practice. A patient used the app to report a headache on a Friday afternoon. Triage staff flagged her report, and her care team asked her to come in for an MRI; the woman felt silly bothering her care providers with a simple headache, but she agreed. Without the ability to
report her current state, her oncologist may never have seen that her disease had progressed: she had developed brain metastases.

Artificial Intelligence

With all the possible tools and interventions available to cancer teams, identifying the right tool to use at the right time for each patient can become a problem in itself. “We live in a world that’s completely data-driven, where we have what seems like infinite data points confronting us every day,” said Mr Hunnicutt. “It is way too much work for us to be manually sorting through all that data.”

To meet this time-intensive task, Highlands partnered with Jvion and CardinalHealth to develop an artificial intelligence (AI) program that could both improve care and increase the accuracy of their risk assessments. By analyzing a patient’s health record and treatment, the AI could determine a patient’s risk for ER visit, inpatient admission or readmission, depression, mortality, functional decline, and need for pain management or hospice care.

Fully integrated in the practice workflow in less than 6 months, the final software was interoperable with Highlands’ electronic health record, practice management system, and health information exchange. Highlands and the vendor rolled out the new program in such a way as to address concerns as they arose. Mr Hunnicutt said that this integration has reduced staff workload and improved the use of tailored interventions, increasing the quality of the patient experience.

Due to transformational changes like these two innovations, Mr Hunnicutt said that the practice’s participation in value-based care models has been largely successful. In the Oncology Care Model, for example, the average cost of care per Medicare
beneficiary dropped by more than $250, and ER visits and hospital
admissions/readmissions had declined.

To access an on-demand replay of the Alternative Payment Models Coalition Workshop—as well as every session held at the ACCC 47th Annual Meeting & Cancer Center Business Summit—visit

Back to Top