Q&A: At the epicentre of real-world data: a conversation with Winson Cheung
April 21, 2020
Dr. Cheung holds not one, but two positions with Alberta Health Services: Director, Real World Evidence at Cancer Control Alberta (CCA) and Principal Director of the Oncology Outcomes (O2) Initiative. As such, he spends his days analyzing real-world evidence (RWE) and advising stakeholders on its clinical value. A senior medical oncologist and Full Professor of Medicine at the University of Calgary, Dr. Cheung has published over 200 peer-reviewed manuscripts and also runs a weekly gastrointestinal cancer clinic. His academic research focuses on the interplay between patient, physician, and system-level factors that drive practice patterns in the real world.
20Sense: Tell us a little about your position as RWE director for CCA.
Dr. Cheung: I feel very privileged to have this job, given that no other province in Canada has a similar leadership position. Alberta has a long tradition of generating real world data (RWD) and evidence, and this position takes it a step further.
If you had to give an “elevator pitch” to convey the value of RWE, what might you say?
All healthcare stakeholders want the same outcome: a drug that helps people live longer and better. RWE helps us to achieve this objective. While clinical trials continue to have significant value, subjects enrolled in trials only represent 5 to 10% of the population. RWE broadens the evidence to include the general population. For example, RWE provides information on whether a drug is truly effective in a broad population and also offers insights into health-economic outcomes that can impact policy decisions. Stakeholders are increasingly recognizing RWE as an important complement to clinical trials and biomarker studies.
It seems that RWE has become a “hot topic” in Canada. What might account for the trend?
The term “real world evidence” itself has given new life to the concept, which used to be known as health outcomes research. On a more fundamental level, Canada is a large, geographically dispersed country with many different populations. Stakeholders have come to realize that we need evidence on how drugs perform in these real-world populations—all with the aim of getting the right treatment to the right patient at the right time, and at the appropriate cost to payers.
Who should be conducting RWE research?
While RWE is largely supported by industry at the current time, academic and hospital centres should also spearhead more RWE studies. As for payers, I see them in a more advisory capacity—letting researchers know what type of data they need—rather than conducting the studies themselves. Ideally, trained RWE specialists should be leading the design of RWE studies because they have a fuller grasp of the required methodology and potential sources of bias, and can thus generate more powerful evidence.
How might RWD/E help improve access to and use of specialty cancer drugs in Canada?
Many cancers are rare, making it difficult to find enough subjects to run randomized clinical trials. In such cases, RWE can bridge the data gap—for example, by serving as the historical or real-world comparator in a single-arm study. RWE can also answer specific questions such as: “Out of 100 people eligible and prepared to take drug X, how many did not receive it because of age or geographical barriers?” This helps to quantify the actual demand. By the same token, RWE can provide more accurate population-based projections of cost savings than traditional economic modelling that rests heavily on assumptions or extrapolations.
Can patient support programs (PSPs) be harnessed to generate RWE?
PSPs are a great opportunity for RWE in that they have a pre-existing data set that encompasses patients taking a particular drug. One potential downside is that the patient consent forms used in PSPs can sometimes make it hard for researchers to access the data, though this is evolving over time. When PSP developers consult me, I often suggest they create consent forms that enable us to interrogate the data in an anonymized and aggregate fashion. We have learned that the vast majority of patients support this idea.
[Q&A continues below]
Paving the way—RWE from the Taiho Patient Support Program
Can real-world data from PSPs influence practice and policy decisions? Dr. Cheung put this idea to the test in a 2018 study. In the interest of assessing unmet needs in the treatment of metastatic colon cancer, Dr. Cheung and his research team analyzed real-world treatment patterns for trifluridine/tipiracil (Lonsurf), a refractory colon cancer treatment that many patients were seeking. Specifically, the team looked at RWD from the Lonsurf PSP run by Taiho Pharma Canada and from Health Canada’s special access program for the unfunded treatment.
The analysis included over 700 Canadian patients, most of them enrolled in the PSP during a 7-month period in 2018. At the end of the study period, 32.8% remained on treatment. Among those who stopped, reasons included disease progression (51.9%), decision of the treating physician (19.3%), and death (13%). Notably, only 4.4% discontinued treatment due to toxicity and an additional 4.4% withdrew their consent. “The fact that many people remained on treatment for several months told us that the drug had value beyond a clinical trial population,” Dr. Cheung explains.
And did the analysis move the funding needle? “While this particular drug remains unfunded [at time of writing] in most of Canada, this paper has caught the attention of a number of other companies, many of them interested in generating similar data that could inform funding decisions,” says Dr. Cheung, who expects to see “more and more PSPs being harnessed to produce actionable RWE.” The study was published in Current Oncology in 2019.https://current-oncology.com/index.php/oncology/article/view/5107/4121
Are there opportunities to share RWE across institutions and jurisdictions?
I am currently involved in an initiative to pool data across provincial borders, in hopes of generating more robust data that can be generalized across the country. Differences in data entry formats and privacy rules between provinces have created some unique challenges, but we are working through them with the right provincial leaders and the right software to facilitate data merging.
Are there any other ways to increase the efficiency and quality of RWE?
We have been collaborating with artificial intelligence companies to assist with data extraction. For example, one of our partner companies has the capacity to translate written physician reports (unstructured fields) into quantitative data (structured fields). Another company has the technology to extract quantitative data directly from images such CT scans. This saves us from needing to embark on the resource-intensive step of manually collecting and entering data from images and physician notes. While still in the early stages, these collaborations will help us to generate more comprehensive and higher quality data, which is very exciting.
You recently launched the Oncology Outcomes (O2) RWE program. Can you tell us more about what you hope to achieve with the program?
We created the program as a “one stop shop” for oncology RWE, to showcase Alberta’s real-world data sources and to promote our strong analytics team to collaborators. All stakeholders are welcome to approach us. We conduct the studies in-house and, once we obtain all the necessary access and research ethics approvals, share aggregate level results with our collaborators.
What is your “blue sky” vision for RWE in the future?
At present, we have a tendency to look at RWE after a treatment has already reached the market and patients are already using the drug. We have opportunities to use RWE more creatively. We could generate evidence a lot earlier, perhaps even in parallel with clinical trials. RWE could even help guide clinical trial design and inform the commercial potential of a drug. In short, I believe that RWE has tremendous value across the entire lifecycle of a drug—before, during, and after clinical trials.
What might you like to convey to industry with respect to RWE?
The key message is to plan ahead. It is not unusual for me to get requests for RWE a few weeks before it is needed, which is not realistic. High-quality RWE takes time. I tell companies to give me at least a year’s notice and to aim for scientific publication, as unpublished RWE is not perceived as robust. If payers and regulators are to take RWE seriously, we need to apply the same rigour to the process as we do to clinical trials. We must also be prepared for both positive and negative results, just as with clinical trials. All research involves risks, but the rewards can be significant if we end up with data that has tangible meaning for real patients in the real world.