Penn Medicine uses AI chatbot 'Penny' to improve cancer care – Healthcare IT News
Penn Medicine's Smilow Research Center
Photo: University of Pennsylvania Health System
Over the past 25 years, the field of oncology has witnessed a dramatic shift in cancer care delivery with the development of new therapeutics, a desire to shift care from inpatient and clinic settings to the home and widespread implementation of electronic health records.
THE PROBLEM
One area of innovation has been the development of oral anti-cancer drugs, including cytotoxic chemotherapies that patients can self-administer at home. It is estimated that approximately 25% of new anti-cancer agents in development are intended for oral administration, according to the NCCN Task Force Report: Oral Chemotherapy.
The use of these drugs allows patients to become more autonomous in their care and enables them to spend more time at home, said Dr. Parul Agarwal, an assistant professor of clinical medicine at the University of Pennsylvania Health System who is on the organization’s “Penny” team, particularly taking leadership in its role in gastrointestinal cancers.
“However, there are several issues to consider when administering oral chemotherapies at home,” she noted. “Medication schedules can be complex, and there is a concern for compromised efficacy if patients take less than the prescribed doses, and an increased risk of potentially life-threatening toxicity if more than the prescribed dose is administered.
“Recognizing these challenges, several mitigation strategies have been evaluated, including intensive, provider- and pharmacist-driven outpatient monitoring programs,” she continued. “In addition, measuring medication adherence and appropriate identification and management of toxicities for patients on oral chemotherapies is challenging.”
Capecitabine is a particularly complex oral chemotherapy that serves as the treatment backbone for many gastrointestinal malignancies. It is manufactured in two pill sizes and prescribed twice daily, dosed by body surface area, and treatment regimens are cyclical, leading to on- and off-treatment days.
Dr. Parul Agarwal, University of Pennsylvania Health System
“In addition, capecitabine often is dosed with other intravenous chemotherapies, other oral chemotherapy regimens, or concurrently with radiation that can lead to further complications with administration,” Agarwal explained. “Typically, monitoring of these agents to ensure safe management is provider- and resource-intensive, and literature documents frequent occurrences of errors and resultant avoidable complications.
“Our Penny pilot using an augmented intelligence chatbot is a novel strategy to improve medication adherence and manage toxicities related to oral chemotherapy agents with a primary goal of improving patient outcomes and secondary goal of improving healthcare utilization metrics,” she added.
PROPOSAL
In order to improve medication adherence and monitor and manage toxicities with the ultimate goal of delivering high-quality patient care, the algorithmically driven, augmented intelligence chatbot Penny was proposed.
Penny uses text-based, bi-directional, conversational interactions to guide patients through potentially complex regimens and reduce potential mistakes they might make as they go through their courses of chemotherapy.
MEETING THE CHALLENGE
Patients with gastrointestinal cancer on the oral chemotherapy capecitabine alone, or capecitabine in combination with other anti-cancer drugs or radiation, were able to participate. The Penny chatbot was bidirectional and messages could be initiated by the chatbot or the patient.
To address issues related to medication adherence and dosing, the chatbot sent twice-daily medication reminders to patients about the dose of capecitabine and when to take it. In the first phase of the pilot, participants received a message asking them to respond “TAKEN” after each dose of capecitabine. This was later modified to a weekly survey to assess medication adherence.
To address the toxicities associated with oral anti-cancer agents, the chatbot also initiated a weekly survey to elicit patient-reported symptoms. These symptoms were triaged using a combination of algorithmic surveys and natural language processing, a form of AI.
Participants also could initiate a message with the chatbot at any time throughout the process to report a symptom. All interactions between participants and the chatbot were monitored in real time by the study team.
RESULTS
“The primary outcome of this study was to evaluate feasibility of using the augmented intelligence chatbot from a patient safety perspective,” Agarwal explained. “Nearly 4,000 medication related-text messages were exchanged and approximately 93% were accurately interpreted by Penny.
“A medication adherence rate of approximately 70% was measured in participants who responded ‘TAKEN,’ though the actual medication adherence rate was likely much higher,” she continued. “Over 500 symptom-related text messages were exchanged and approximately 98% were correctly interpreted by Penny.”
Through an iterative process, patient engagement in symptom assessment increased from 25% to approximately 70%, and the majority of messages that were incorrectly interpreted were acted upon by the research team in real time, she added.
More than 50% of participants completed qualitative exit interviews. These interviews demonstrated an overwhelming level of patient engagement with clear feedback that Penny provided an additional layer of support, increasing their confidence in taking the medication and interacting with their care team, she said.
ADVICE FOR OTHERS
“The iterative manner in which this intervention was implemented led to several improvements in Penny’s natural language processing function and resulted in higher levels of patient engagement as the pilot progressed, which is crucial to healthcare innovations in this realm,” said Agarwal.
“This process required a high level of engagement from all parties involved – the healthcare team, research team and the Memora Team, who we partnered with for a technological platform,” she continued. “Penny identified care delivery inconsistencies with schedule, dosing and monitoring of capecitabine among gastrointestinal cancer providers, leading to initiatives to decrease unnecessary variations in care.”
A patient journey dashboard was created to clearly outline schedule and dosing with the aim of facilitating high-quality care delivery.
“Moving forward, this dashboard will be integrated into the EHR to become a standard tool for clinical teams to use to monitor their patients and adjust dosing as needed,” said Agarwal.
Follow Bill’s HIT coverage on LinkedIn: Bill Siwicki
Email him: bsiwicki@himss.org
Healthcare IT News is a HIMSS Media publication.
More Whitepapers
More Webinars
© 2023 Healthcare IT News is a publication of HIMSS Media