Updated and Initial Validation of Warning Signs: Western Expert-Based Pidcap Delphi Project

ObjectiveEarly recognition of inborn errors of immunity (IEI) is critical to avoid morbidity and irreversible organ damage. Jeffrey Modell Foundation (JMF) warning signs have long guided and still guide clinical suspicion. However, the IEI phenotype has broadened substantially since 1993. We aimed to update and evaluate IEI warning signs using an international Delphi process, generating harmonized suspicion criteria that can serve as the basis for scalable, user-friendly tools and future data-driven technologies to support early detection.Design and MethodsThe Delphi study was carried out in two rounds in 2024/2025. Multidisciplinary IEI experts (including physicians, nurses, and patients from European Society for Immunodeficiencies [ESID], Clinical Immunology Society [CIS], International Nursing Group for Immunodeficiencies [INGID], and International Patient Organisation for Primary Immunodeficiencies [IPOPI] under the umbrella of the JMF) rated and weighted literature-derived candidate signs using a 9-point Likert scale. Consensus was defined as ≥66% of respondents scored 7–9. Dedicated pediatric and adult panels assessed age-specific signs. Outputs included consensus level and diagnostic weight matched with International Classification of Diseases (ICD) codes. To test the feasibility of an automated early-detection system for community-dwelling individuals at high risk of IEI, we ran the algorithm against primary care electronic health records (EHR) of a healthcare district (catchment population of 571,123; 91.9% adults and 8.1% children).ResultsOf 103 invited experts, 96 from 22 countries in Europe, Canada, and the U.S. completed round 1 (93% response rate) and 90 completed round 2 (94%). The pediatric panel (n = 68) reached consensus on 26 of 27 signs, and the adult panel (n = 49) on 17 of 24. High-consensus, high-weight items included ≥3 pneumonias, bronchiectasis, very early onset inflammatory bowel disease, positive family history, and multimorbidity (≥2 signs). When run against primary care EHRs: 14,519 individuals were considered at high-risk of IEI, eligible for considering referral: 14,174 adults (2.7%) and 345 children (0.75%).ConclusionsThis international Delphi provides the first harmonized, consensus-based update of IEI warning signs. The weighted, ICD-10–mapped signs support the implementation of early-detection screening strategies and facilitate federated learning frameworks. The pilot test of potential referrals suggests that the indicator would be feasible, although the potential overburdening of specialized services should be taken into consideration. Further testing is ongoing to validate the implementation strategy.