Study Design
Design and methodological insights from RESONANCE: a US-based registry on the natural history and impact of recurrent pericarditis JoAnn Clair* ALLISON CURTIS Vidhya Parameswaran Paul C. Cremer Sushil A. Luis Michael S. Garshick Ajit Raisinghani Brittany Weber JoAnn Clair Allan L. Klein John F. Paolini
Recurrent pericarditis (RP) is a rare and persistent autoinflammatory disease characterized by serial episodes of chest pain. Treatment often entails years of symptom-suppressive pharmacotherapy (e.g. NSAIDs/colchicine, corticosteroids, and IL-1 pathway inhibition) before resolution of autoinflammation is achieved. Real-world data which capture the long-term RP patient experience are lacking, and registries are susceptible to certain types of bias due to their observational and often retrospective data collection. These biases may include selection (from non-random enrollment), temporal (from inconsistent data collection points), observer (from variability in clinician-recorded data), confounding (from unmeasured factors affecting outcomes), and non-response (from differential participant engagement) biases.
The REgiStry Of the NAtural history of recurreNt periCarditis in pEdiatric and adult patients (RESONANCE), the largest multi-center US-based observational registry, was developed to quantify trends in RP disease burden, management, and outcomes to inform treatment selection and optimize care. The RESONANCE design anticipates and addresses common biases found in registries to ensure the validity and reliability of its findings. An ambispective design combines retrospective (from index acute pericarditis episode or ≤1 year pre-enrollment) and prospective data (collected for up to 5 years) into one observation period (Figure).
RESONANCE incorporates robust data collection and post-data collection protocols for missing data, sensitivity analyses to assess dropout patterns, regression/stratification to adjust for confounders, and propensity score matching to compare responders vs. non-responders.
By combining pre- and post-data collection strategies, the impact of identified biases may be mitigated, improving RESONANCE’s accuracy, generalizability, and utility in clinical and research applications.
