Systemic Treatments and Outcomes for Pediatric Atopic Dermatitis: Setting the Stage for Comparative Effectiveness Research

PROJECT SUMMARY / ABSTRACT Atopic dermatitis, or eczema, affects up to 20% of children and is associated with significant physical, emotional, and social morbidity. Up to one-third of children with atopic dermatitis have moderate-to-severe disease and require systemic medications to achieve adequate disease control. While systemic therapies for

pediatric atopic dermatitis were historically limited and consisted of broadly immunosuppressive medications such as methotrexate, highly efficacious and targeted systemic treatments, such as dupilumab and oral Janus kinase inhibitors, have recently emerged. However, the optimal treatment for any individual child remains

uncertain, as real-world data on the long-term effectiveness and safety of both novel and traditional systemic medications for pediatric atopic dermatitis remain sparse. Thus, comparative effectiveness research (CER) is critically needed to fill these current knowledge gaps. Since new treatments for atopic dermatitis continue to

rapidly emerge and traditional randomized trials are often lengthy and expensive, pragmatic trials and observational studies may be more efficient for CER in pediatric populations and nimbler in adapting to the expanding treatment landscape for atopic dermatitis. The growth of large clinical research networks, such as

the national Patient-Centered Outcomes Research Network (PCORnet), also provides opportunities to leverage real-world data to conduct CER in pediatric atopic dermatitis. However, before such efforts can be undertaken, tools to easily and accurately identify children with atopic dermatitis within clinical research

networks must be developed and methods to utilize observational data for CER in a way that ensures valid inferences must be employed. To meet these needs, this proposal has two specific aims: (1) develop and validate computable phenotypes for pediatric atopic dermatitis, and (2) apply a target trial emulation framework

to compare the effectiveness and safety of dupilumab and methotrexate for atopic dermatitis treatment. Utilizing routinely-collected electronic health data held in the PCORnet common data model, this project will apply machine learning approaches to develop validated computable phenotypes for pediatric atopic

dermatitis, including those distinguishing moderate-to-severe disease from mild disease, using structured data elements. In the target trial emulation comparing new pediatric users of dupilumab or methotrexate for atopic dermatitis, this project will test the hypothesis that dupilumab is associated with fewer topical medication

prescriptions, lower rates of systemic treatment switching or intensification, and lower rates of adverse effects compared to methotrexate. This research will enable the efficient conduct of future pragmatic trials in pediatric atopic dermatitis and provide preliminary data for large, multi-center observational CER studies that will be the

focus of future R01 applications. Ultimately, the findings from this proposal will help to meet the long-term goal of generating an evidence base to guide systemic treatment selection for children with atopic dermatitis.