Objectives: High-quality clinical research of resuscitations in a pediatric emergency department is challenging because of the limitations of traditional methods of data collection (chart review, self-report) and the low frequency of cases in a single center. To facilitate valid and reliable research for resuscitations in the pediatric emergency department, investigators from 3 pediatric centers, each with experience completing successful single-center, video-based studies, formed the Videography In Pediatric Emergency Research (VIPER) collaborative.
Methods: Our initial effort was the development of a multicenter, video-based registry and simulation-based testing of the feasibility and reliability of the VIPER registry. Feasibility of data collection was assessed by the frequency of an indeterminate response for all data elements in the registry. Reliability was assessed by the calculation of Cohen κ for dichotomous data elements and intraclass correlation coefficients for continuous data elements.
Results: Video-based data collection was completed for 8 simulated pediatric resuscitations, with at least 2 reviewers per case. Data were labeled as indeterminate by at least 1 reviewer for 18 (3%) of 524 relevant data fields. The Cohen κ for all dichotomous data fields together was 0.81 (95% confidence interval, 0.61-1.0). For all continuous (time-based) variables combined, the intraclass correlation coefficient was 0.88 (95% confidence interval, 0.70-0.96).
Conclusions: Initial simulation-based testing suggests video-based data collection using the VIPER registry is feasible and reliable. Our next step is to assess feasibility and reliability for actual pediatric resuscitations and to complete several prospective, hypothesis-based studies of specific aspects of resuscitative care, including of cardiopulmonary resuscitation, tracheal intubation, and teamwork and communication.