IMPROVING THE DIAGNOSIS AND MANAGEMENT OF COW’S MILK ALLERGY: A FRAMEWORK TO SUPPORT PERSONALIZED, EVIDENCE-BASED RECOMMENDATIONS

Abstract

Cow’s milk allergy (CMA) is a common and potentially life-threatening condition affecting children worldwide. Yet, major uncertainties persist in its diagnosis and treatment due to limitations in the available evidence, resulting in wide variability in clinical practice. Furthermore, the management is particularly complex, as it involves multiple diagnostic strategies and treatment options that must be tailored to an individual’s risk of being allergic. Consequently, decision-making should be informed by interconnected bodies of evidence spanning multiple steps of patient care. Clinical practice guidelines for CMA must therefore address these uncertainties and practice variability by providing clear, actionable recommendations that support individualized decision-making while accounting for patient values and variations across settings. This sandwich thesis aimed to support the development of rigorous recommendations for the diagnosis and treatment of CMA as part of the World Allergy Organization initiative to update the previous DRACMA guidelines. It is structured around four manuscripts. The first two are systematic reviews evaluating the benefits and harms of oral immunotherapy and specialized formulas, compared to allergen avoidance, in children with CMA. The third manuscript translates the synthesized evidence into recommendations using the GRADE Evidence-to-Decision framework adapted for multiple interventions’ comparison. It provides a ranked list of options with remarks and considerations to guide decisions based on specific patients’ preferences and circumstances. The fourth manuscript addresses previously identified challenges by introducing a conceptual framework to support model-based linkage of evidence bodies in guidelines. It outlines pre-modeling considerations, audience-specific decision-aids, and practical strategies for estimating pre-test probabilities to be linked with testing accuracy and downstream management effects. Collectively, this work presents an integrated approach that combines evidence synthesis, guideline development methodology, and conceptual work to facilitate model-based linkage of evidence bodies to inform decision-making. While developed in the context of CMA, the proposed approaches and tools can be adapted for use in other clinical domains. In conclusion, the contributions of this thesis aim to optimize the care of people with CMA and provide a preliminary framework to support the development of patient centered recommendations for guideline questions that reflect the interconnected and multidimensional nature of clinical decision-making.