Interview with Thomas Wilke: Health Economics and Real-World Evidence Studies

October 20, 2020

In this interview with Thomas Wilke, Principal Scientist at Ingress-Health (a Cytel company), we talk to him about his background and experience in Health Economics, understand the important considerations of real-world evidence studies and the impact of COVID-19 pandemic on the work of the health economics outcomes researchers who work at Ingress and Cytel. We also cover important HEOR topics such as its benefits for market access studies and real-world analytics (RWA) for regulatory submission.

Cytel and Ingress-Health will be contributing to a range of events at Virtual ISPOR EU 2020, on November 16th – November 19th. Our Real-World analytics teams will be collaborating to deliver a number of interactive workshops, issue panels, posters and podiums to showcase their work and share innovative insights in HEOR, evidence generation, knowledge synthesis and decision analysis.

Click below to download our full list of sessions at ISPOR EU

download the  sessions schedule

You have a wealth of experience working both in the industry and academia. Tell us a little about your journey so far.

I was introduced to health economics during my PhD. At that time, I tried to do research on physician-induced demand in the health care sector which I think is still quite an interesting topic. After three years of working at a strategic consulting company, I accepted an offer as a Professor at University of Wismar, Germany. There I founded the IPAM, a university-affiliated institute specialized in retrospective claims data analyses. After several years in business, I felt that IPAM needed a refreshment and internationalization. It was pure luck that I met my colleague Bart Heeg who was, at that time, leading the European business of another health care consultancy. Together we founded and grew Ingress-health, a European health care consultancy, since early 2015. We decided in early 2020 to go with Cytel as it was the natural next step in our development.

Ingress is now a Cytel company. What added advantage do you think this partnership is going to bring to the market?

There are a lot of advantages, but I will outline two areas. First, Ingress-health was a European health care consultancy with strong roots in the HTA area, which means we mainly dealt with payer needs. Cytel has a deep understanding of trial design and trial analytics, as well as of associated regulatory needs. It is obvious that the combination of both trial design and regulatory needs with HTA-focused health economic modelling and RWE studies generates added value. We are now able to offer integrated solutions. If companies think, for example, about an evidence package consisting of an adaptive trial and an additional real-world data collection and wonder about the regulatory and HTA impact of such an evidence package, we are the right partners to help them. Second, with Cytel, we have now become a Global consultancy which is able to do multi-country modelling and RWE studies. That also adds a lot of value.

For over 15 years you have been leading and conducting German and international Real-world Evidence studies. What are the important considerations before initiating a study using real-world data? Can you say a little bit about good versus bad data?

The considerations are not much different from those you need to take into account when starting a clinical study. First, be very clear about the objectives of the study. Is it about treatment patterns, identification of unmet needs and assessment of real-world outcomes, description of healthcare resource use and cost, or even a comparison of different therapies in terms of their outcomes? We might even aim to assess patients’ quality of life or preferences. Depending on the objectives, you choose the right study design.

Second, select your study methodology and database. If it is prospective observational studies, patient surveys or retrospective medical chart reviews, you will need to think about sampling. The methodologically weakest sample we can think of is the “convenience sample” in which study sites by themselves select whom to include in a study. But the good news is that there are better ways to define a study sample. If it is database studies, you need to select the right database. And, even if there is no such thing as generally good data, that definition of the right database always depends on the study objectives.

There are some golden rules for good patient-level data; you can describe patients in terms of their sociodemographic and clinical characteristics. The treatment of patients can be described completely without missing any GP, specialist or hospital care, there is sufficient follow-up for patients, and outcomes of interest can be measured with high validity. For example, If I want to describe the systemic treatment and overall survival of patients with a specific cancer disease, I cannot use a database that does not cover both the inpatient and outpatient treatment, does not provide a reliable follow-up of patients, and does not report mortality.

The COVID-19 pandemic has elevated the challenge of designing and executing clinical trials within a substantially shortened time frame. What does this mean for outcomes researchers as they look to add and aid in developing evidence in the pandemic space?

We need to be careful here. Clinical trials and specifically, double-blind randomized controlled trials, are and should be the backbone of the assessment of the efficacy and safety of a new treatment. COVID-19 cannot be the reason for changing this paradigm. However, there are several reasons why clinical trials are associated with a certain level of uncertainty such as, short follow-up, exclusion of patients who get the treatment in the real world, low sample sizes, impossibility to randomize patients to treatment etc. High-quality, real-world data may provide additional data here. They help to generate comparison groups that we call synthetic control groups. They provide additional data for trial-based outcome extrapolations and are used to measure real-world outcomes in a post-approval setting, specifically if pay-for-performance contracts are closed.

What insights can HEOR studies provide for market access studies? When it comes to use of RWA for regulatory submission, what can it teach us beyond traditional RCTs?

To assess the value of a new treatment, both regulatory and HTA agencies, as well as the clinical community, need to have an understanding of what the current unmet needs are, such as, the real-world treatment patterns, outcomes, cost etc. That is the most important objective of pre-approval real-world studies. We provide comparator data, sometimes even in more formal synthetic control group comparisons. Additionally, we generate data that are needed for health-economic models and, increasingly important, show whether patients would prefer to receive the new treatments. Finally, we reduce uncertainty of agency decisions by setting up studies in a post-approval setting – product registries, pragmatic trials, post-approval database studies etc. Here, the most interesting are the comparative effectiveness ones which compare different treatments.

Your manuscript around a linked data-based comparison of different type 2 diabetes treatments got published recently. What according to you are the benefits and challenges of such a study?

This was a post-approval comparative effectiveness study that aimed to assess the real-world outcomes of two different treatment strategies in type-2 diabetes. We could demonstrate the benefits of one therapy in a very clear way by running propensity score matched comparisons. However, this study was also unique in another sense; we generated several publications based on this study, and in some of them we used the so-called linked data, i.e. data that come from different data sources. In this study, we linked medical chart review data with claims data of the very same patients. The data source we had at hand, after that linking, was unique – the generalizability of claims data combined with clinical data as available in patient charts. Usage of these data was a unique feature of this study.

Learn more about this and access the manuscript here.

Can you tell us about any past/ongoing/upcoming projects that you are really excited about?

About 18 months ago, we did a retrospective claims data study in Germany addressing the real-world treatment of advanced non-small cell lung cancer patients in Germany. The results were somewhat disappointing as we could not detect any positive development of overall survival in this patient population during the last 5 years. And, one of the main reasons was that only a minority of patients received a mutation test which is recommended by the guidelines. To see that patients could live longer, but do not get the proper treatment, is not really something I get excited about. But I realize that it is our duty to identify and publish these data to improve the daily care of patients.

Another project involved a patient preference study in relapsed / refractory multiple myeloma. In a discrete choice experiment study, we could demonstrate that patients are willing to accept slightly worse outcomes associated with one treatment when compared with the other if, they get the chance to still live their lives. That was the case as the new treatment was an oral one in comparison to an over three hours infusion twice a week, which practically means the end of a normal life, including a potential job loss. I was impressed by how clearly the patients could articulate their wishes, and it is our task to listen to them and tell regulatory and HTA bodies, as well as the physicians that all we do has the purpose to help patients. Listening to them is of utmost importance here.

Click below to download our full list of sessions at ISPOR EU.

download the  sessions schedule