One way health systems are seeking to reduce hospitalizations and ER visits for chemotherapy patients is by deploying electronic patient-reported outcomes (ePROs) to monitor side effects in real-time. John Deeken, M.D., head of the cancer service line at Virginia-based Inova Health, recently spoke with Healthcare Innovation about a four-year project that will integrate ePROs from the patient portal into existing workflows across nine sites at Inova.

Inova is one of several sites nationwide that PCORI is funding for this ePROs work. Deeken is president of the Inova Schar Cancer and Medical Director for the Inova Schar Head and Neck Cancer Program. He said Inova’s cancer program is the largest in the Washington, D.C., region and in the state of Virginia with more than 50 medical oncologists and more than 8,000 new cancer patients per year.

Healthcare Innovation: Dr. Deeken, could you describe the scope of the PCORI-funded project at Inova and why you wanted to focus on ePROs? 

Deeken: Most cancer patients are treated on the outpatient side, and many get chemotherapy. Cytotoxic chemotherapy can have side effects, from having people have their hair fall out to feeling fatigue or nauseated or cause them constipation or diarrhea or having their blood counts drop, where they're at risk for not being able to fight off an infection. How to manage that is one of the key challenges facing us. In outpatient hematology/oncology, one of the biggest expenses and most variable expenses is how much cancer programs or health systems have their patients go to the ER or be admitted for those chemotherapy toxicities. 

This has been something that CMS has been working on for the last eight or nine years. And we have been trying to see what we can do to help patients who don't want to go to the ER, don't want to be in the hospital for these toxicities. How can we better treat them before it gets so bad that they're sent to the hospital? Oftentimes we're giving chemotherapy every three or four weeks, so if something is happening with a patient, you don't realize it's brewing. By the time they call us, things are in such dire straits that it's hard to keep managing them on the outpatient side. So they have to go the ER and either get resuscitated or get fluids or get admitted to the hospital. 

HCI: Before this PCORI grant, did Inova Health already have some experience with ePROs? 

Deeken: We've done a number of pilot studies over the last few years, trying to use ePROs, with patients answering questions to surveys about how they are feeling. Do they have a fever? Are they nauseated? Do they have pain? 

When COVID first struck in February and March of 2020, we set up a system for our cancer patients who tested positive for COVID but weren't sick enough to be in the hospital yet. We got them an iPad, a Bluetooth thermometer, blood pressure cuff and pulse ox. They answered questions about how they were feeling, and they submitted their vitals and their symptoms three times a day, and we monitored it during, business hours, and on weekends. We had 30 patients go through that process, and only two ended up having to go to the hospital. So we knew that it worked in that acute, awful setting of the first days of COVID. 

Then over the last couple of years, we've modified that by asking the same ePROs questions and adding an Apple Watch and monitoring their symptoms that way and other ways — but the core of the intervention and all these pilot studies was the ePROs questions. 

HCI: Has this been studied before? 

Deeken: Over the last several years, other people have looked at this question. An NCI-funded research group, which is called the Alliance, did a study where they used ePros. The results of that study found it didn't improve overall survival, but they found significant improvements in terms of less hospitalizations, decreased costs, improved quality of life, better patient satisfaction. 

PCORI set up about 40 health systems in the country that are Implementation Science Centers. If we know something works in one study environment, can we make it work across whole health systems? So we're one of 16 health systems that got these awards and the grant starts in July. 

The starting gun goes off July 1 for us to roll this out over four years across all of our centers. So it's not a small feasibility study where just some percent of our patients have this made available; we are applying it across all our cancer locations and all of our patients to see, at the end of the day, if it helps in those parameters — lower hospitalizations, lower ER visits, better costs, and better patient satisfaction.

HCI: Since you're rolling it out across Inova, does that require workflow changes for nursing staff and the docs, too?

Deeken: We're hoping to use our current infrastructure with Epic MyChart. But instead of waiting for patients to call us when they're feeling bad at home, we're going to look at all of our chemotherapy regimens we give to all cancer patients — whether they have melanoma or lung cancer or acute leukemia — and have a gradation system where they are higher, middle or lower risk of these side effects, these 10 to 12, key symptoms that if we know about, we intervene early. We’re going to ask patients based on that risk profile of the chemotherapy regimen they're on with different frequencies. We're either going to ask them three times a week, or we're going to ask them once a week, or ask them every couple couple weeks, are they feeling these symptoms? 

Also, every once in a while, we're going to ask things such as: how distressed are they? Are they financially insecure? But we're going to be testing the hypothesis that we can do this at the right frequency for the regimen they're on and then that workflow will be audited. So instead of waiting for them to call us, we're going to be pushing questions to them through that portal. If they respond back to us with abnormally high results —  they're in pain, they're throwing up, whatever the question is, then that's going to automatically generate a prompt to the triage pool. We'll be calling the patient, figuring out what's going on, and seeing how we can intervene in real time before they're so sick they have to go to the hospital.

HCI: Any workflow changes involving the EHR? 

Deeken: We are integrating this into our Epic system. When we order a new chemo regimen for a patient, it automatically includes which frequency of the surveys that they'll be getting. Part of the implementation work will be training all of our staff about this, so that they can tell their patients about it. Then all the triage and nursing staff, as well as the physicians and nurse practitioners will know what this looks like. We are setting up standard responses to those symptoms, and we will be tracking results along the way.We're also going to look at whether there could be any intermediate role for AI in this. Then it gets triaged up to a nurse or to the physician if things are getting more severe or things aren't working based on that AI algorithm.

HCI: Is there any thought given to the idea of standardizing the ePROs across health systems for research purposes to say: Are we comparing apples to apples as we're trying to do this work?

Deeken: It's a great question. Epic developed a program for cancer within their system. It's called ESYM and they set up 75 different questions. So there's a base that we all can work from within Epic that every health system can tweak to their system. 

But over time, one of the things I think the PCORI and these multiple centers are trying to do is see what works and what doesn't work. Is there a common set of questions that led to better outcomes? What's the core query set that we use for these ePROs moving forward?