Methodologic Flaws, Wrong Answers, and Right Questions: Pediatric Office Emergencies

To the Editor. In their study of the frequency of medical emergencies in pediatric offices in a small, rural state, Heath et al1reached a questionable conclusion that may potentially harm pediatricians and their patients. Their statement that “serious medical emergencies are rare events in the primary care pediatric office” goes beyond their data and appears to reflect several methodologic and analytic flaws: 1. In the retrospective portion of the study, the authors did not identify what was defined as an emergency in the survey of participants. Was an emergency self-identified by the clinician, one of several predefined diagnostic categories (such as status asthmaticus or shock), or based on other criteria? Clearly, how an emergency is defined can lead to substantially different findings and conclusions, but we are never told what definition the authors chose. 2. In the retrospective study phase, there is a distinct possibility of recall bias in asking pediatricians to remember office emergencies that occurred up to 1 year ago. Recall (or anamnestic) bias is a well-known phenomenon in which inaccuracies frequently occur in the recall of distant events.2 3. Regarding the prospective study component, assessing the frequency of office emergencies based on “the need to use the emergency kit” is fundamentally flawed. As the authors themselves report (Table 2), most office practices already had many of the equipment and medication items in the kit, so they probably used their own items, rather than opening the kit. Basing emergency frequency on kit use would thus result in an falsely low estimate. 4. In the “Methods” section, another fatal …