Current Practice and Tolerance for Risk in Performing Procedural Sedation and Analgesia on Children Who Have Not Met Fasting Guidelines: A Canadian Survey Using a Stated Preference Discrete Choice Experiment

Objectives: The objectives were to explore the tolerance of pediatric emergency medicine (PEM) physicians for risk in choosing when to perform procedural sedation and analgesia (PSA) and to describe adherence to preprocedural fasting guidelines and factors affecting the physicians' decisions. Methods: A survey of Canadian PEM physicians who perform PSA was conducted. Respondents were asked about their PSA practices. Risk tolerance was assessed using an economics-based stated preference elicitation method called a discrete choice experiment (DCE). Using a hypothetical clinical situation of a healthy child needing PSA, three fasting scenarios (ingestion of full meal < 2 hours ago, full meal between 2 and 4 hours ago, liquids less than 1 hour ago) were created. For each fasting scenario, 16-choice questions with varying numeric risks of major and minor adverse events were presented and respondents were asked whether they would administer PSA now or wait for guidelines to be met. In this exploratory study, choice data were analyzed using multivariate regression analysis, and preliminary inferential statistics are presented. Results: A total of 114 (63.6%) completed surveys were returned. Respondents were 54% male, were in practice for a median of 6-10 years, and reported following fasting guidelines 53% of the time. Most had institutional fasting policies (70%), used ketamine (in 78% of PSA by 95% of users), and identified patient variables (prolonged pain, parental concern) as more important than environmental variables (patient flow, use of resources) in deciding when to administer PSA. Both major and minor risks of adverse events affected the choice of whether to do immediate PSA or wait until fasting guidelines were met. A 0.1-percentage-point increase in the probability of a major adverse event reduced physician likelihood of performing immediate PSA before meeting fasting guidelines by 17.6 percentage points. In comparison, a 10-percentage-point increase in the probability of a minor adverse event reduced physician likelihood of performing immediate PSA by 17.2%. Respondents were less likely to perform PSA immediately if a child had a full meal < 2 hours ago versus 2-4 hours ago or liquids < 1 hour ago. Sex, highest level of postgraduate training, frequency of performing PSA, and experience with a prior adverse event did not affect the choice to do immediate PSA. Years in practice affected the decision, with those in practice for 6-10 years more likely to perform immediate PSA than those in practice for shorter or longer. Those who reported having an institutional fasting policy were less likely to perform immediate PSA. Conclusions: These results suggest that fasting guidelines are not strictly adhered to in Canadian pediatric emergency departments (EDs) currently, and there is some willingness of physicians to change their sedation practice in light of evidence from hypothetical surveillance data about risks. On the other hand, some physicians suggest that they will follow guidelines regardless of how low the estimated risk is from surveillance data. An understanding of how physicians respond to evidence about small risks and how the information is best understood by this population is interesting for knowledge translation if evidence-based practice guidelines for procedural sedation in the ED are developed in the future.