Using RWE to characterize diseases as novel treatments become available: Q&A with Dr. Maria Schneeweiss
As novel therapies come to market, especially in disease areas that aren’t as well characterized, it’s important to first understand the baseline risks associated with the conditions they aim to treat. This helps inform the safety profile of the new interventions, as researchers can rule out whether certain safety events are associated with the medication, or with the disease itself.
A recent real-world evidence (RWE) study published in JAMA Dermatology set out to evaluate the baseline risk of a blood clotting event—venous thromboembolism (VTE)—in patients with chronic inflammatory skin diseases, a key step to build the knowledge base around these conditions as highly efficacious dermatology interventions become available to patients.
We spoke with Maria Schneeweiss, M.D., lead author on the study, about the intended impact of the research, the methodological choices made, and how similar RWE studies can inform the evolving treatment landscape in dermatology. Dr. Schneeweiss is a physician and post-doctoral research fellow at Brigham and Women’s Hospital, working in conjunction with the Department of Dermatology and the Division of Pharmacoepidemiology. She focuses her research on dermatology pharmacoepidemiology.
Responses have been edited for clarity and length.
Q: What was the goal of this study?
A: In this study we aimed to establish the background rate of VTE in patients with five different chronic inflammatory skin diseases: psoriasis, eczema, alopecia, vitiligo, and hidradenitis suppurativa.
This is overdue for two reasons: first, we know that non-skin-related chronic inflammatory diseases, like rheumatoid arthritis or inflammatory bowel disease, have been linked with an elevated clotting risk, likely due to their substantially increased inflammatory state. But we don’t have the same information for skin-related chronic inflammatory diseases. It’s important that we rule out VTE risk for these conditions as well, that way physicians and patients can make well-informed benefit-risk assessments.
Second, there are several new systemic immune-modulating medications on the market that have gained popularity in dermatology. They’re extremely efficacious, and they target the specific part of the immune system that causes the skin disease to occur. But some of the newer agents—in particular, the oral medications with small molecules that target the JAK receptor—have been associated with an increased risk of VTE in patients with rheumatoid arthritis or inflammatory bowel disease (non-skin-related chronic inflammatory diseases). Now that these treatments are coming into the dermatology world, we need to know the background risk of VTE in chronic inflammatory skin conditions in order to know what’s attributable to the disease itself, versus a new medication.
Q: What data sources did you work with, and why were they fit for the purpose of this study?
A: We used real-world data from Optum’s national commercial insurance claims database, which includes data from 2004 through 2019. VTE is a rare event, therefore we needed a large, nationally representative patient population for this study. This database also happens to capture race, which contributed to our decision as race is an important factor in skin diseases.
Q: How did you design the study? What methodological choices did you make, and why?
A: We compared patients with chronic inflammatory skin diseases to patients who were seen by a dermatologist, but were never diagnosed with any of these conditions. We then followed these patients until they either had a VTE event, disenrolled in their health plan, their data stream ended, or they died.
One of the key parts about the design was separating the patient groups. We had to make sure that the patients in the study group did in fact have one of the chronic inflammatory skin conditions, and that the comparator patients truly did not have any of those diseases. Therefore, we required patients in both groups to have been seen by a dermatologist; if a patient is seen by a specialist and the disease has been recorded over more than one visit, we can be fairly certain that the patient has the disease.
We chose a validated, broad outcome—for example, inpatient or outpatient VTE with evidence of treatment—to capture the overall incidence, then refined the secondary outcomes to just inpatient events. To make the groups comparable in all aspects of VTE risk factors, comorbidities, and treatments, we used one-to-one propensity score matching and exclusions.
To test the robustness of the population, we included a positive control, rheumatoid arthritis, as we know there’s an increased risk of VTE in patients with this disease. We applied the same methods to this patient group, and we saw VTE risk double among the rheumatoid arthritis population—which is in line with the current literature and what we were expecting. We also used a negative tracer outcome: skull and face fractures. We assumed patients likely reached the outcome as a result of trauma unrelated to any of these skin conditions, and so expected to see a null result, as both groups should have the same risk of skull/face fracture since we know that this does not relate to having or not having a skin condition.
Q: How did the Aetion Evidence Platform® enable this work?
A: The Aetion Evidence Platform enabled collaboration across our team. This was a big project, and we used a big team. Being able to follow up with team members, to see this study happening live, and to review the outcomes made a huge difference. That’s hard to do with programming codes; it’s not as easy to follow the analyses step-by-step.
The platform is also very user-friendly, so we were able to conduct this study pretty quickly. We could add outcomes and events as we progressed, and the platform was easy to use, generate results and share them, and work as a team.
Q: What were your findings, and how do you interpret them?
A: We found no indication of an increased background risk of VTE events, which is reassuring. Knowing the background risk of VTE can indicate, early on, if a VTE signal among users of a new medication is due to the medication, or due to the underlying skin disease. With newer medications on the horizon for many of these chronic inflammatory skin diseases, the VTE risk is an important potential adverse event that needs to be monitored. By knowing that there’s no increased risk inherent to having the disease, we can be more sure that VTE events are related to the drug, not the disease.
Q: Why is dermatology a therapeutic area that’s well suited to RWE research?
A: As new, extremely efficacious drugs enter the market in dermatology, the treatment landscape is changing. In the early 2000s, for example, new systemic agents for psoriasis became available, and now other chronic inflammatory skin diseases—most recently atopic dermatitis—are getting these drugs as well. One of the things holding physicians and patients back from using them widely is the question of safety. Fortunately, claims data is great at capturing safety, especially hospitalizations and rare events, so there is a great opportunity for RWE research in this area.
Q: What can other RWE researchers take away from this study?
A: Disease epidemiology studies are difficult to do, and they do have their limitations. But when you’re waiting for a drug to come out and for real-world data on the drug to become available, it’s important to know the baseline risks of the disease the drugs aim to treat to support the interpretation of future studies. That information is important for contextualizing further research.