Improving the Patient Experience in Cancer
Last fall, the Carevive team partnered up with Fierce Pharma to produce a webinar titled Improving the Patient Experience in Cancer geared to individuals across the life science value chain who see the benefit of patient-reported outcomes (PRO) in cancer care, including HEOR, Patient-Centered Outcomes and Real-World Data experts. Ethan Basch, MD (a leading PRO researcher and oncologist), Stacie Hudgens (CEO of COS, a Carevive partner with expertise in PRO research and analysis) and Carevive experts, will illustrate how life science stakeholders can use these data to answer important questions about the experience of real-world patients with cancer.
There were hundreds of attendees and numerous questions that couldn’t be addressed during the Q&A segment of the webinar. Below is the summary of the most frequently asked questions answered by the panel of experts.
Quick Links
How can clinicians improve patient care using the platform?
Madelyn Herzfeld: Hopefully we covered that today, from end to end, getting the patient’s input when making a treatment decision and making sure it’s a shared decision, having the best data you can to make the best possible decision for a patient, and then monitoring patients over time, as Ethan has demonstrated in his pivotal clinical trials. We all know that keeping track of patients and managing them in a proactive manner leads to better outcomes.
What type of data are you collecting, and what is unique about it?
Mordecai Kramer: Our novel SMART Data™ set consists of a combination of data extracted from the EMR to paint a contextual picture of the patient. Data fields such as basic demographics, age, sex, race, ethnicity, diagnosis, and types of medications that the patient is receiving. We combine that with the longitudinal patient-reported outcomes holistic data that we’re collecting about the patient experience week in and week out, such as symptoms, quality of life, physical function, healthcare resource utilization, and a number of other important things, to paint that overall picture.
And this combination of data is really unique in the space right now, and hopefully as you all learned today, it can provide some really interesting insights, particularly as we follow patients over an exceptionally long span of time while they’re on treatment.
What are challenges to clinician adoption of the platform?
Madelyn Herzfeld: We are breaking through, in terms of implementation science, how to remotely monitor patients. We’re sensitive today to health systems that are dealing with Covid and the hardships that the oncology care teams have faced during this pandemic; we are assisting them in thinking about the future and transformation, and how can we continue to improve cancer care.
I would say the biggest hardship right now is around thoughtful planning of resources. That is something that Carevive has worked really, really hard on over the past year, being really thoughtful about how we can guide sites on implementation, design the implementation process, and be there to see them successfully through it. We’ve made significant investments in the implementation framework.
Do you share data back with clinicians?
Tanya Wildes: Yes, on the individual level, the clinician sees the data input by their patient on a dashboard in graphical form. The clinicians also have access to the aggregated de-identified data, both for local quality improvement initiatives and their own research questions as well.
Why is your adherence so high?
Mordecai Kramer: Adherence is high for a number of reasons. One, we know that patients with cancer want to be extremely engaged in their care and they want to share information about themselves so they can help other people in the ecosystem, and that makes them incredibly remarkable in that way. And we’re appreciative of all of them who contribute their data to us in this way and we are proud to be the custodians of that data.
In addition, the way the platform was designed, and you got a taste of that, through Maddy’s slide, is to promote constant feedback between clinicians and patients. When patients are reporting symptoms at any point in time that are moderate to severe, and certainly very severe, the care team is alerted to that and can respond to the patient and help them mitigate that symptom. As Ethan’s data has shown, this can help prevent avoidable hospitalizations and ER visits.
Even when patients only report mild or moderate symptoms, the Carevive platform sends them a customized PDF that gives them some self-care tips to hopefully mitigate the symptom on their own. These tips remind them that, if they’re still in distress, they can call their care team. It reminds them of the phone numbers to do that. We’re constantly promoting this ongoing dialogue that patients find helpful and keeps them engaged week after week.
Can these types of data be included in an FDA submission?
Stacie Hudgens: As I mentioned earlier, the FDA guidance and draft guidance on the use of real-world data to inform real-world evidence are out. I believe it is critical to note that the data will be reviewed, as suggested by the FDA, for evidence to support specific products for regulatory decision-making in terms of efficacy and safety.
The data sources and data uses are the two highlights. As Carevive pushes forward with extremely specific data usages and aligns the scientific investigation to regulatory decision-making, I believe it will be highly informative.
Are the real-world data that you are seeing consistent with clinical trial results?
Ethan Basch: There have been clinical trials looking at the impact of patient-reported outcomes on clinical endpoints that show the improvements that I described: decreased hospitalizations, improved quality of life, longer survival, etc.
And those have also been looked at, not just in the prospective clinical research setting, but also in large population data. A group in Ontario Canada did a population study in more than 128,000 patients who were routinely providing PROs in oncology practices across the province. They found similar impact on quality of life, survival, and hospitalization. I would say that these findings, or the positive impact of digital monitoring with PROs, have been seen both in very controlled clinical research settings and also in a real-world general population health setting.
Are the insights we’re finding specific to community-academic cancer centers or both?
Ethan Basch: Benefits have been seen in both. Prospective studies looking at the impact of PROs have been done in tertiary cancer centers (hospital-based practices) and also in community oncology practices. The findings tend to be durable across contexts.
Is Carevive an EMR platform for an MD office to use?
Madelyn Herzfeld: We are independent of the electronic health record. We do build tight integrations with the electronic health record. The most sophisticated integration we have is with Cerner Corporation, and they’re also investors in our company. The integrations are two-fold. One is a data integration: we seamlessly pass clinical data from the electronic health record into our platform, so there’s no need for dual entry. There are also workflow integrations. What we mean by that is, we are smart-enabled, so you see our dashboards within the contexts of the health record. We’re also able to send alerts from the patients to the message center. We have configurations for alerting that, if significant enough, an alert will be sent into the electronic health record message center, similar to how a significant lab value behaves. And then we also auto-document back into the electronic health record, any intervention that was done for the symptom.
Do you find that healthcare professionals are making effective use of the data you collect, or is there still work to be done for them to utilize the data effectively? How do you see them using it effectively in the real world?
Madelyn Herzfeld: We’re just starting work with our sites on deep implementation planning. I’ll give you an example of one site who’s extremely interested in looking at time to symptom resolution. Using the data, they continuously monitor say 50 patients who are on the platform. They see how their symptoms are progressing or resolving and they see the management strategies that have been put in place that accelerated the symptoms or kept them from resolving. The reports are created and then sent back to the administrators and to the nursing managers to use for continuous practice improvement. I would say in real time, you could and should be looking at the data to continuously improve practice.
How does the onboarding process look? Are patients enrolled in office, or does it happen after the fact, as a follow-up?
Madelyn Herzfeld: What we recommend, and a lot of this came out of Ethan’s work and a lot of the early champions in this, is that it is typically introduced in the office by the care team. It’s important that the patient knows that the care team believes it is important for the patient to participate. Once they are enrolled, obviously the follow up and monitoring in between visits happens remotely.
Previous studies have shown that clinicians don’t always respond to symptom alerts. Are you measuring whether or not there is a response to alerts?
Mordecai Kramer: We are being very diligent about tracking the alert workflows. It’s especially important for the sites, as they’re using the Carevive platform as part of their standard of care, to be able to use the platform as designed and be able to provide a response to patients in a timely fashion and help them manage their symptoms better. We’re able to track the flow of information as it comes from the patient to the clinician, and of course, that’s all time-stamped.
We also know when the care team is responding to the patient, what that response is, whether it involved prescribing a new drug, eliminating a drug that they’re taking, or some other recommendation. We’re able to track what happens to the patient afterward and be able to see if the symptom resolved over time or if there were additional alerts that were triggered and ongoing management.
We have all this data, and it tells an incredibly unique and interesting story. It is extremely helpful in the direct daily care setting for these patients.
Is there some sort of concierge or patient navigator service for cancer patients? The question is on behalf of a patient that is struggling to understand the treatment process and struggling with the logistical challenges associated with treatment.
Madelyn Herzfeld: It is something we’re looking at. People are spread pretty thin right now across the cancer team. At Carevive, sort of as a strategy, we’re looking at care management services that we can augment at our sites.
Tanya Wildes: That [navigator service] is typically associated with the cancer center. I’m not familiar with anything that is larger because much of the navigation really is institution-specific about where scans might be held and things like that. But it does sound like a gap and a need there.
How can smaller hospital centers with limited resources implement ePROs?
Ethan Basch: I think that electronic patient-reported outcomes for digital monitoring can be implemented in almost any setting. The technology itself is straightforward. I think that the challenge is the modifications in workflow at the practice and having personnel who can respond to the information. This is a technology that really enhances something that we already do as a part of practice. Symptom monitoring is foundational in cancer care. Any practice, no matter what it’s size, is dealing with patient symptoms. They may call on the phone, or if there is some connected portal system, they can send an electronic message or they discuss symptoms at the visits. There’s somebody in the practice, or multiple people, who address those things. If a patient calls on the phone with a symptom, there’s somebody who checks that message and gets back to the patient and manages it.
These electronic systems, all they do is enhance that. They improve communication between visits, such that the patient will be using an application. They’ll report a symptom and then that’s directed to the same person who would be checking the voicemails or checking the patient portal messages. Any practice of any size can do this if they are engaged in symptom management, which I assume every practice dealing with oncology patients is.
I’d say it’s not really a limitation. In fact, I would argue it’s easier to onboard this for a smaller practice than for a larger practice where technology implementation sometimes has more layers of approvals, and implementation complexity.
Tanya Wildes: From the standpoint of helping the workflow, sometimes a patient sitting in my exam room won’t have reported symptoms by phone in advance of a busy clinic day, but then will come in with substantial symptom burden. Had I known about them [those symptoms] earlier, I could have been intervening and getting appropriate workup done in advance of seeing them in the clinic.
And so, as opposed to making the workflow harder, it can actually make it more efficient so that our face-to-face time, which is so valuable, can be spent discussing the imaging results that I’d gotten in advance of seeing them.
Are there liability concerns from healthcare professionals if they receive an alert they do not address in a timely fashion?
Tanya Wildes: It is a particularly good question and I have not personally worked with having a platform in my workflow, so I’m not certain how they ensure that the alerts are observed and informing the patients that the urgent things need to, on the weekends, go to the ER, things like that. I don’t see it as any different as a phone call that, informing us of symptoms that we need to act on urgently as well.
Ethan Basch: These platforms have legal disclaimers on them, the same as patient portal systems do. The patient portals are ubiquitous or used widely in medicine across all disciplines, and patients send in messages just as patients leave voicemails. Messages can be missed, messages can be delayed in their responses, and so there are standard legal disclaimers, to eliminate exposure. For example, “this form of communication cannot be relied upon as a sole means of connecting with your practice. If you have a severe problem requiring attention, call 911 or come to your doctor’s office or the emergency room or call the doctor, etc.” There have not been any cases that I’m aware of where there has been liability related to these kinds of systems. This is just another form of communication in addition to other forms of communication.
Madelyn Herzfeld: That’s a big part of our implementation. Even in the scripts when we’re onboarding, we educate the patient that this is not an emergency service. This is for us to track the patient longitudinally. All over the platform, we reiterate that. Also, the patient can pick which day every week they want to report their symptoms. A lot of clinicians don’t even allow patients to submit surveys on Fridays because they don’t want that information to be stuck over a weekend.
There’s a number of mechanisms that we have put in place to optimize the communications and mitigate risk. But all over the platform, we do explain very deeply that the intention here is not emergency management. If there is an emergency, call 911 or go to the ER.
One concern we have is mitigating alert fatigue coming in the electronic patient-reported outcomes systems. Can you address those concerns?
Madelyn Herzfeld: This is another thing Ethan has offered a lot of guidance to us on. We’re working on configurations to really limit that [alert fatigue]. Looking at severe configurations, if a patient symptom goes from nothing to moderate, we’re looking at asking the patient if this is something they need a call-back on—embedding that into the platform. We’re constantly looking a number of different strategies.
I know Ethan is looking at machine learning, to be able to understand alert management and how do we do it in a scalable way.
Ethan Basch: These alerts from the system, I liken to the pager that I wear. When my pager goes off, it’s disruptive, but I’m glad that it goes off because I need to know what’s happening with my patients, so that I can be reactive. E-alerts are the same way. They are valuable information, but, if there are too many messages that mount up in somebody’s inbox, it can be a burden.
We know from nurses who are the recipients of these messages, that they feel that unless time is allotted for this activity, checking the alerts feels like a burden. Any implementation needs to carve out time for the nurses or the navigators, whoever’s getting the alerts, to address them, whether it’s by calling patients or documentation. It’s a very modest amount of time, but if you don’t clear that time, then it’s just piled on top of everything else.
And particularly during the last two years, we’ve all learned that any new encumbrances on our time as clinicians can really feel like a substantial burden. The problem isn’t necessarily the alerts per se, the alerts are legitimate representations of needs of the patients that should be addressed by the practice. Really, the issue is having adequate staff bandwidth and time to address. And again, that’s somewhere in the range of about five percent of the time of the clinical nurse or navigator who’s overseeing a standardized population of patients in an oncology practice.
That said, I will underscore what Maddy said, which is that there’s a lot of ongoing work to reduce the number of alerts and to make the alerts easier to address and to clear. That’s an area of high interest in digital monitoring. But, nonetheless, there will continue to be alerts and so we have to plan for that when we implement.
What are you looking forward to publishing next?
Stacie Hudgens: I’m particularly interested in comparative tolerability for individual treatment pathways in the real-world setting. Having worked in the clinical trial setting for a very long time, the real-world setting offers the ability to look at treatment decision-making along the longitudinal path for a patient.
I’m extremely excited to continue some of the work that Dr. Basch and I have done with Friends of Cancer on comparative tolerability and starting to use real-world data to understand comparative tolerability of treatments that are on market now, but also considering different data uses for real-world evidence generation.
What type of data products do you see developing in the future?
Bruno Lempernesse: With the data we have, we are already serving a large spectrum of users of this type of data, from the provider clinicians who want to know what’s happening in the practice, combining this clinical data and the ePROs data, and also the life science companies who have different objectives and goals using this type of data. We’re also building up a library of syndicated reports for large cancer populations that we’re covering today, but also for rare populations.
And part of what we are also doing is moving from not only using this data to discover what’s happening, but to predict what’s happen. There’s a lot about predictability and using the symptoms, that after seeing a series of symptoms, are we going to be able to predict what’s going to be the next symptoms and the treatment pathway of the patients. We are seeing lot of exciting developments working on these types of data.
How have you seen this data as part of SOC [standard of care] impact clinician workload?
Tanya Wildes: This is just changing the timing of the workflow. The patients are going to be having the symptoms whether they report them to us in real time or wait until their next clinic visit. This is just giving them the opportunity for more timely intervention, kind of just redistributing the workload a little bit differently than it is in standard of care. That’s my perception.
Ethan Basch: I think that’s right. We know that by monitoring symptoms, and becoming aware of those symptoms earlier, clinicians intervene before things become severe or complicated. As we saw in the data, we drive down ER visits. It also means that fewer patients are arriving in clinic in extremis, which tends to tie up a lot of provider time having to arrange for either admission or IV fluids or whatever.
And so, it’s really an upstream shift of staff time. Now you’re spending more time on small incremental issues. All these little alerts are coming in and staff are intervening early in order to avoid very time-consuming, complex, complications if symptoms are not well controlled. Nonetheless, I still advocate for protecting a little bit of nursing time or navigation time to deal with the alerts. Even though probably it evens out at the end of the day, there is a perception of more work by the people who are receiving the alerts. I think it’s important to acknowledge that they’re now going to be spending time doing this, but I think, on balance, staffing is about the same with this set up.
What are you looking for among the geriatric patients in the dataset? What is most important to you?
Tanya Wildes: I wanted to highlight that geriatrics is not just a unique niche. This is two-thirds of the patients with cancer. It is a disease of aging. Outside of this, the International Society of Geriatric Oncology conference is demonstrating accumulating data that, considering aging factors like comorbidities and daily activities, are going to help improve the care of our older cancer patients.
Some things that I’m anxious to do with the platform involve integrating the information that we’re gathering—things like identifying patients who may have some dependence in their daily activities and who don’t really identify any individualized or care partner. These data can help us recognize in advance that a patient is developing increased dependence and needing more help, so we can plan ahead in an anticipatory way.
Seeing the implementation of these geriatric screening factors and how they influence the clinician’s thinking about selecting the right treatment for an older cancer patient—this is helping us mitigate ageism rather than making cancer treatment decisions based on age. [The dataset can guide us in] tailoring treatment to the older individual’s level of fitness.
Have you partnered with any specialty pharmacies for similar data outcomes? And also, as a liaison to the patient or MD?
Madelyn Herzfeld: We have not worked with a specialty pharmacy, but whoever asked that question, we would love to talk further with you about your idea.
Has this platform been used as part of clinical trials to capture AE’s?
Mordecai Kramer: It has not been used to date, although we have submitted a couple of proposals to life science companies who are interested and see the adaptability of the platform for that purpose. From a compliance perspective, it is ready to go. But primarily our focus has been on the real-world patients, to collect data on them, as an underserved market with unmet needs, and be able to compare those data with clinical populations as well.
From a pure platform applicability perspective, it’s certainly within the realm of the possible. I would love to learn more about that and to explore it more.
Madelyn Herzfeld: We’ve been hyper-focused right now on the real-world setting and collecting data for patients who are on standard of care treatments.