**Background:** Large-scale data-sharing strategies are valuable for improving quality of care, and pediatric patient registries are particularly important due to the limited number of patients with specific conditions. In Canada, all 17 PICUs are participating in the creation of a national database called the Canadian Pediatric Intensive Care Consortium (CanPICC). The objective of this study was to develop national consensus-based common data elements for inclusion in this database.
**Methods:** The study used a modified Delphi consensus approach with two phases. Phase 1 involved a focus group of physicians and nurses who generated an initial list of data elements based on literature review, local PICU databases, existing national registries, and clinical expertise. The initial survey included 125 items across six domains (demographics, PICU admission characteristics, severity of illness scores, PICU therapies, PICU adverse events, and outcomes). After item reduction and pilot testing, the final survey contained 117 data elements. Phase 2 was a prospective multicentre iterative survey conducted from March to June 2021. Participants were health care professionals, allied health professionals, caregivers, and other stakeholders from all 17 Canadian PICUs. Three rounds of surveys were conducted, each one month apart. In rounds 1 and 2, participants rated each data element on a five-point scale (mandatory, important, optional, remove, not my area of expertise). Consensus was defined as ≥70% of responses in the same category. In round 3, response options were modified to a four-point scale (include, exclude, add to a module, not my area of expertise), with consensus maintained at 70%.
**Key Results:** Of 86 invited participants, 68 (79%) agreed to participate. Response rates were 91% (62/68) in round 1, 90% (60/68) in round 2, and 81% (55/68) in round 3. Participants were predominantly physicians (63%), with representation from all provinces. After three rounds, 72 data elements were included by consensus from six domains: patient demographics (e.g., age, sex, race, gender, postal code), PICU admission (e.g., admission/discharge dates, diagnosis, chronic conditions), severity of illness (e.g., PRISM3 score, mechanical ventilation in first 6 hours, cardiac arrest in first 2 hours), PICU interventions (e.g., invasive/noninvasive ventilation, ECMO, VAD, renal replacement therapy, CPR), quality assurance (e.g., central line-associated bloodstream infection, ventilator-associated pneumonia, unplanned extubation, delirium), and outcomes (e.g., death in PICU, mode of death, organ donation, disposition location). Notably, race, gender, and home region (first three digits of postal code) were included, but minority status, indigenous status, primary language, and ethnicity were not. The most highly ranked optional modules were ventilation, sepsis, and traumatic brain injury.
**Clinical Implications:** The selected core data elements will provide standardized and synthesized data for research, benchmarking, and quality improvement initiatives for critically ill children in Canada. The exclusion of many social determinants of health highlights a potential gap in understanding health inequities. The study provides a methodological framework that can be applied to the creation of other data registries. Ongoing work includes harmonizing definitions and creating a governance structure for data validation, access, and use.