Sample size and clinical trials in pediatric resuscitation

Abstract

Ideally, high quality randomized controlled trials would be the foundation of every clinical decision.We especiallywant high quality evidence for all of our decisions in resuscitative care. Every day, in prehospital medicine, emergency medicine, and critical care, we are flooded with new research. It is understandable if we assume that our foundation of evidence is solid. The reality, however, is that the foundation of resuscitation science is more shifting sands than bedrock. There is a disheartening lack of high-quality evidence for the vast majority of resuscitative decisions. In the 2020 update to the American Heart Association’s Guidelines for Cardiopulmonary Resuscitation, only 32% of 491 individual recommendations were based on moderate or high quality evidence, and only 11% of the recommendations were based on randomized trials.1 The reasons for this lack of evidence should be obvious. Resuscitation research is expensive, in both money and time. Randomized controlled trials in resuscitation science typically require hundreds of thousands of dollars (at least) and hundreds of hours to execute. The best clinical trials require single-center, observational studies and infrastructure building, followed by multicenter collaboration. Application for external funding is usually required, with funding rates depressingly low. Moreover, resuscitation clinical trials often produce negative or unimpressively positive results, meaning we need to completemany trials to get 1 that will positively influence care. In pediatric resuscitation research, one factor has led to an even greater lack of high quality evidence—compared with adults, critical illness and injury is mercifully less common in children. Although tens of thousands of children require resuscitation each year, for single centers in the United States, there are rarely enough patients with any form of critical illness to conduct adequately powered clinical trials.2,3 This is a universal challenge in pediatric settings—prehospital, emergency department, and critical care—high quality research is prohibitively expensive because it cannot be done at 1 or 2 centers over a relatively short time period. Federal funding and multicenter designs are almost always required. Compounding the sample size problem is that many, if not most, of the effect sizes studied in resuscitation research are small. Survival is perhaps themost commonoutcome in resuscitate research; however, it is uncommon for even a high quality study to showmore than amodest absolute survival benefit for a given therapy. Survival and other binary outcomes, therefore, often require large sample sizes to detect more