How an Innovative AI-Based Smartphone Application Is Addressing Patients’ Palliative Care Needs

A Conversation With Mihir Kamdar, MD

Get Permission


Jamie H. Von Roenn, MD, FASCO

Jamie H. Von Roenn, MD, FASCO

Addressing the evolving needs of cancer survivors at various stages of their illness and care, Palliative Care in Oncology is guest edited by Jamie H. Von Roenn, MD, FASCO. Dr. Von Roenn is ASCO’s Vice President of Education, Science, and Professional Development.


During the 2019 ASCO Annual Meeting, Mihir Kamdar, MD, presented findings from his randomized clinical trial showing that a smartphone application was successful in decreasing patients’ pain by nearly 25% and pain-related hospitalizations by 69%.1 The study’s findings are especially significant because the trial participants were all coping with pain related to stage IV solid-organ cancers, including gastrointestinal, breast, and lung cancers. According to the American Cancer Society, 30% of patients with newly diagnosed cancer, up to 50% of patients undergoing treatment, and between 70% and 90% of patients with advanced cancer experience some degree of pain.2

Dr. Kamdar, Associate Director of the Division of Palliative Care and Director of the Cancer Pain Clinic at Massachusetts General Hospital, and his colleagues at Partners Pivot Labs, led by Dr. Kamal Jethwani, developed the smartphone app, called ePal, which employs artificial intelligence (AI) technology and clinical algorithms to monitor and measure levels of patients’ pain to optimize cancer pain management. In this study, 112 patients undergoing treatment in a palliative care clinic were randomly assigned to either a control group that received standard oncology care, including in-person palliative care, or an intervention group that received ePal-based care plus standard oncology care for 8 weeks. The researchers also conducted a patient chart review to identify pain-related hospital admissions and emergency department visits and compared the risk between the two study cohorts.

Our study shows that technology actually brings patients and clinicians closer in a way that makes patients feel more connected and does not overwhelm clinicians.
— Mihir Kamdar, MD

Tweet this quote

Integrated into ePal were a series of functions, including mechanisms for tracking pain and the ability to determine the source of poorly controlled pain; educational coaching messages and a library of mind/body meditation approaches; videos of patients’ own medical team members answering questions on how best to manage pain; and a tool to request nonurgent prescription refills.

In a wide-ranging interview with The ASCO Post, Dr. Kamdar discussed the results from the study. He also addressed how AI technology could be used to improve patient care without increasing the outpatient clinical burden in both academic and community settings and how improvements in the design of the next iteration of ePal could reduce myriad side effects of cancer and its treatment, including constipation, dyspnea, and fatigue.

Technology to Improve Patients’ Quality of Life

Why did you create ePal, and how did the patients enrolled in your study use the smartphone app to track their pain levels and report them to their oncology team?

Cancer pain is a huge problem for patients, and it affects their quality of life. As clinicians, we see patients every few weeks to monitor treatment effectiveness and any cancer-related health problems, such as pain. But if we knew in advance of those in-clinic visits what pain issues patients were having, we could address them as soon as they developed, sparing patients admissions to the hospital or trips to the emergency room.

With ePal, patients could quickly report their pain level using a sliding pain scale of 0 to 10 and an animation of a patient in various degrees of pain to assess their own level of discomfort, so we had both a numeric and a visual way for patients to gauge their pain. What is unique about this app is it uses AI to identify the patient’s specific barrier to achieving pain control and then it can decide whether the problem warrants immediate attention from the clinician or if it is something that could be alleviated through ePal’s education and empowerment tools.

In the study, if the patient was experiencing severe, new, or escalating pain beyond the level that was preset in the algorithm, notification would be sent to our on-call nurse, who would typically respond to the patient within 15 to 20 minutes. After office hours, the app would guide the patient on how to effectively reach the on-call palliative care clinician. If a patient’s pain was well controlled, the app would check back with the patient every 2 to 3 days.

Initially, we worried that the app would generate an excessive number of notices to the clinic and increase the number of call-backs to patients, but that did not happen. We only saw, on average, one call triggered by the app for every 2 patients over 2 months, so the app did not add to the clinical burden. This is an important finding because it shows that these digital health devices may have an important role in helping to remedy our workforce shortage in palliative care.

We heard from patients that just knowing there was a clinician on the other end of the app made them feel more confident. Clinicians often worry that technology is going to replace them, but that is not true. Our study shows that technology actually brings patients and clinicians closer in a way that makes patients feel more connected and does not overwhelm clinicians.

More Anxiety

According to your study results, patients using ePal had about a 25% reduction in their pain level and improvement in their attitudes regarding cancer pain treatment, but they also experienced higher levels of anxiety than patients in the control arm. Do you know why patients using the app experienced greater anxiety?

This is a great point of discussion. When I talk with colleagues working in the fields of digital and mobile health, they say it is not uncommon to see increases in patient anxiety with the use of this type of technology. It is important to put into context the difference in the level of anxiety between the control and intervention groups.

With the Generalized Anxiety Disorder 7-item (GAD-7) scale that we used, although the increase in anxiety was statistically significant, it was not clinically significant. You have to have a change of 5 points or more to achieve a clinically significant finding, but the change we saw in the level of anxiety was under 3 points, so it was still in the mild range. Interestingly, patients who used the app more than twice a week did not have an increase in anxiety.

My sense is that patients who were not using the app as frequently became more anxious after receiving reminders asking about their pain. In the future, we want to develop the app so that it’s tailored to the patient’s needs—if the patient doesn’t want constant reminders about reporting pain levels, the app can adapt and schedule fewer reminders, but it will still be able to monitor changes in the amount of pain patients are experiencing.

It is good to know that using this type of technology may increase patient anxiety, because it gives us an opportunity to think about how we might use digital health devices to address the issue and actually reduce anxiety.

Using Technology to Optimize Patient Care

What have you learned from the results of your study that you can now take to the clinic?

There are a couple of important lessons from our study. One is that there is a role for digital health in patients with advanced cancers. This is important because much of the digital health world is focused on wellness and not on the seriously ill. In addition, we have shown that AI can be used not only to gather patient-reported outcomes, but to analyze them and optimize patient care as well.

This was a single-site study in a palliative care clinic. We now want to expand this study into the community setting. We are in the process of redesigning ePal with the goal of bringing this technology to the community and providing virtual palliative care in settings where traditional palliative care is not available.

We have shown that AI can be used not only to gather patient-reported outcomes, but to analyze them and optimize patient care as well.
— Mihir Kamdar, MD

Tweet this quote

Also, we know that patients benefit from palliative care intervention at the time of diagnosis, regardless of disease stage, and not just in the advanced-care setting. We ought to be thinking about using digital health devices earlier in the course of illness. We don’t want to wait until patients are having serious issues to provide palliative care.

In the next iteration of ePal, we will still focus on pain, but we hope to add modules for treatment side effects such as nausea, constipation, and fatigue, as well as additional features, including the ability to interface with patients’ electronic health records. We are also adding video elements so we can actually get a visual of how sick the patient is, to more quickly and accurately determine whether the patient needs to come into the clinic or if the problem can be solved over the phone. And we are thinking of incorporating voice technology into the app, for patients who may have trouble typing their information on a keypad.

The more these mobile devices are integrated into patient care, the more comfortable patients will be with using them, and they will become a seamless part of oncology care. That is where this technology and health care are going. 


DISCLOSURE: Dr. Kamdar has stock or other ownership interests in Amorsa Therapeutics and has served in a consulting or advisory role for Amorsa Therapeutics and Vivtex.


1. Kamdar M, Centi AJ, Agboola S, et al: A randomized controlled trial of a novel artificial intelligence-based smartphone application to optimize the management of cancer-related pain. 2019 ASCO Annual Meeting. Abstract 11514. Presented June 1, 2019.

2. American Cancer Society: Addressing state policy barriers to cancer pain management. Available at Accessed July 9, 2019.