Clinicians need recommendations for their daily practice. They also need to be able to appraise the trustworthiness of these recommendations. The Grading of Recommendations Assessment, Development and Evaluation (GRADE) framework has been a major advance in this field. In this thesis, I discuss the specific challenges and potential solutions in applying GRADE to pediatric populations. In particular, I outline issues about (1) clarifying the target population in the presence of numerous age-groups; (2) the choice of patient-important outcomes for young children and their parents; (3) the limitations resulting from the evidence available in pediatrics, (4) the difficulty in finding trustworthy baseline risks for all relevant age-group; (5) the dynamics of values and preferences in pediatrics. Finally, since weak recommendations are preference-sensitive, I also discuss potential challenges of shared-decision making with pediatric patients who gradually express their autonomy and capacity of decision.