Increasing Biomarker Testing Rates for Patients With NSCLC
David Waterhouse, MD, MPH, spoke with the Journal of Clinical Pathways about his involvement with OHC’s (Oncology Hematology Care) quality improvement project that increased biomarker testing rates for patients with metastatic non–small cell lung cancer, and shared how the program can be scaled to other practices and other cancers.
Transcript
David Waterhouse, MD, MPH: Hi, I'm Dr David Waterhouse, with OHC, Oncology Hematology Care, in Cincinnati, Ohio, and we are also affiliated with The US Oncology Network and the Sarah Cannon Research Institute.
What led to the idea of implementing a quality improvement project to increase biomarker testing rates for patients with non–small cell lung cancer?
Dr Waterhouse: The project itself actually has quite humble beginnings. During Covid, I was asked by the group Ontada, which is part of the US Oncology Research Network to assist them with a contract piece that they were doing looking at whether or not doctors do next-gen sequencing for patients with lung cancer. They had been approached by a pharma company that was about to launch a biomarker-driven therapy, and they just wanted to know the testing rate. So we looked back at a period of time, looked at over 3,000 charts and tried to use structured data to answer that question, realized we couldn’t. So we just did a chart review of only 300 charts, and the disturbing piece of that was that full and comprehensive testing, including next-gen sequencing and PDL-1 testing, was being done in less than 50% of the patients. Now, no one really wanted to believe that to be true, and as a result of that they went back and hand curated over 3,000 charts and added a few more, since the timeline had changed, and that became what is now known as MYLUNG Protocol A, which was presented at ASCO, and later published, showing that biomarker testing was not as widespread as we thought it was, and they went on then to do what is now known as protocol B, or protocol 2, for MYLUNG, which was a prospective study. We were part of that. There were 10 groups in the nation, and that's been at least published in abstract form at last year's ASCO.
Testing rates had improved among all of 10 practices. The testing rate for any one marker might have been as high as 85%. But if you were talking about next-gen sequencing and PDL-1, the rate across all of these practices was only about 58%. In our group it was 68%, and that was not good enough. And we had been very active in quality improvement projects prior to this, and decided that we would take on the task of improving biomarker testing in patients with metastatic lung cancer. Specifically, non–small cell lung cancer. And right about that same time Pfizer came out with an RFP—request for proposals—that we competed for, and we were awarded a grant to initiate this project. The goal of the project was simply to improve comprehensive biomarker testing in all patients with metastatic non–small cell lung cancer.
What methods were used to standardize biomarker testing rates at OHC?
Dr Waterhouse: We used several methods. The first and most important part was how we did it. We used Lean Sigma Six methodology to do this project. The first thing that we had to address was the fact that we couldn't tell who was being tested and who wasn't. One of the hallmarks of any quality improvement is, you can't fix what you can't measure. And so we had to develop methodologies to audit whether or not a patient had been tested, and how the testing was done, and where were the results coming into the chart. Once we developed methodologies for finding the patients, who were are patients with non–small cell lung cancer, who had metastatic disease and ought to be tested, how would we measure whether the testing was done or not. We then started a campaign to improve the testing rates? We knew our baseline because we had participated in the MYLUNG protocol 2, prospectively measuring every chart. Every chart was hand curated. So we knew that our baseline testing rate was 68%. We wanted to bring that up over 90% was our goal. The first thing we did was education. We took tools that had been produced by Pfizer, by McKesson and The US Oncology Network and others to create educational videos that were shared with all of our providers, and we confirmed that everybody had gone through that training. In addition to training the providers themselves, the physicians and our advanced practice personnel, we also developed training modules for our treatment nurses, our billing people, our front desk people. Basically, we wanted all stakeholders to understand the importance of this testing. The next thing that we did was we created a template to be used within the electronic medical record for all new patient consultations for those who had lung cancer. And so doctors were required now to use that specific template. Built into that template were the NCCN guidelines for the appropriate testing for staging and also for the biomarker testing itself, so that they were given a written reminder embedded within the template note: “These are the tests that you should be doing on the patients.” And along with that we created an order set that mirrored those recommendations. It is much easier to opt out of an order than to write an order. So many of these patients, for example, would have already had a CAT scan of the chest. That's how they came to us. All you had to do was X out the CAT scan of the chest, but we wanted to get a PET scan. So you would, you know, that was already built in, and built into the order set, were both tissue and liquid orders for the testing of biomarkers. So again, the doc would actually have to opt out of doing the testing as opposed to having to enter in the need for testing. Another thing that we had done is we also restricted doctors on their coding for lung cancer. Right now there's over 140 different codes in ICD 10 that can be used to code lung cancer. That created a problem with auditing charts. So we restricted doctors to only 6 codes and those 6 codes would allow us to accurately and predictably code these patients so that we could find them and be able to monitor the progress going forward.
So, following the education and the creation of the template and the creation of the order sets, we then partnered with a couple of different vendors to also help us build a molecular tumor board. It doesn't do any good to order the tests if you don't know how to interpret that test, and you don't know how to use the results. So we created a molecular tumor board to support the physicians in decision-making once these tests came back. And then we began the project. We did have internal auditing going on throughout the course of the project. There were providers that we had to circle back with and assist them with further training. There were like, you know, whenever you do quality improvement, methodology, it’s an iterative process. You're constantly checking what you’re doing and retooling it and assessing whether or not what you’re doing is working or not working. So there was a constant approach to circling back and saying, Okay, perhaps it was an individual struggling. Perhaps it was a group that was misinterpreting what we're doing. We would go back and correct that. And in doing so, we clearly improved the testing rate, from 68% up to 92.7% in less than a year. So we did develop a methodology that could work. And that’s where we’re going now.
After these measures were implemented, what were the results of the quality improvement project?
Dr Waterhouse: The testing rate improved from a baseline of 68% up to a testing rate of 92.7%. And when I say testing rate, I am talking about comprehensive testing, PDL-1, and next-gen sequencing. We were not proponents of hotspot testing, you know, testing just a single marker. This was a comp this was the rate for comprehensive testing in those patients. But along the way we did find some other issues. We saw that doctors were struggling with staging their patients in the electronic record. If you read a doctor's note, their patient is staged wonderfully, and the treatments that they were getting were very appropriate, based on the stage of the patient. But if you went back and looked at the electronic record, oftentimes, the information that was obtained during staging was not reentered into the electronic record to create a full picture. So that was a problem that we identified. We’re still looking to see now whether or not doctors used the information appropriately. In other words, if a patient did test positive for an actionable biomarker, did the doctor actually use that information in developing the treatment plan for the patients. But did we increase biomarker awareness? I think we clearly did so. And this is going to allow our group the opportunity to scale this going forward.
What aspects of the project do you think were instrumental to its success?
Dr Waterhouse: I think the first aspect is the education piece. I think that everybody understands that biomarker testing is important. That message is being sold universally now. I think that people overestimate their own personal testing rates, and I think institutions overestimate their testing rates. I have heard academic centers say, well, at our center. We do it much better than it is done in the community. Well, I don't believe that's true. I've worked at some of these centers. or least a couple of them, and I don't know that that's necessarily true. I think that some places are better than others. But there's clearly a problem with getting comprehensive testing.
Another important part was simply making it easier for the physician. Physicians are challenged to get everything done during a patient encounter. By creating a written template that they could use in the EHR, we built in a reminder for them, that make sure you get the MRI of the brain. Make sure you do the biomarker testing and other aspects of the care that need to be done to appropriately stage and plan for a patient’s treatment. And by creating an order set we made it much easier for them to order the testing needed, whereas before they would have to enter each order individually, and there was a lot of opportunity there to forget or to omit an order. So I think those were pretty simple things to do, but had a significant impact.
Looking ahead, what impact do you hope the success of the project will have on increasing biomarker testing rates in practices across the country?
Dr Waterhouse: Well, I think going forward, we showed that you can improve the testing rate. That doesn’t mean that we necessarily improved the care. We still need to make sure that the doctors are interpreting the testing correctly, and that the appropriate therapy is given to the patients. And then, taking that a step further, we need to confirm that, in fact, through getting the appropriate targeted therapy, or whatever is the best care that that’s leading to better outcomes. There have been a lot of studies looking at targeted therapies versus standard of care and showing that you can obtain better outcomes. But we'd like to take it a step further and show that this can be scaled. Lung cancer is not the only disease where biomarker testing is important. It may be the one that everybody calls out because there are so many actionable biomarkers. But there are biomarkers from many other diseases. And so internally, we hope to be able to scale this type of methodology to other disease states. And we’re currently in the process of sharing our best practices with other physician groups within the US Oncology and Sarah Cannon networks. Part 3 of the MYLUNG initiative is interventional to improve testing. Some of our methodologies are being used in a pilot project within part 3 of MYLUNG to see if this will scale across other practices. Our methodologies have been shared in multiple venues across The US Oncology Network to allow them the opportunity to take advantage of some fairly simple interventions to see if they can improve their testing rates as well. We think that this is not something that’s unique to US Oncology. We did it inside The US Oncology Network, because we’re affiliated with that network. But the types of interventions that we did can cross other EHRs and can cross other institutions, whether it be a community practice or an academic center. The methodology can be scaled rather easily. And so I think that I’d like to see us go to other disease states, I’d like to see us show that we can use this kind of methodology to scale to other practices. And so those are some of the important next steps going forward.
Is there anything else you’d like to add?
Dr Waterhouse: I’d like to encourage other groups to engage themselves in quality improvement. It’s you know, as I said, I’ve been a clinical trialist, my entire career and very active in drug development. But there’s also a great satisfaction in process improvement. And this work is important. It can be published. It can be shared. And it’s not unique to our practice. Many practices have developed better ways of doing things, and you know, I think all of us can feel comfortable about competing on our service, but we should never compete when it comes to best practices, we should be sharing those.