Process Risks in Perioperative Medication Delivery

One anesthesia provider is often responsible for prescribing, formulating, dispensing, administering, and documenting medications in the operating room. Unlike other hospital units, there are few safety interventions. Systems engineering approaches can provide important insights into improving patient safety during medication delivery processes (Kaplan et al., 2013; Reid et al., 2005). This study observed anesthesia medication delivery during 20 anesthetic cases in the OR and interviewed 10 anesthesia providers in a large midatlantic academic hospital using a Systems Engineering Initiative for Patient Safety (SEIPS) framework to identify process risk in perioperative medication delivery (Holden et al., 2013). Anesthesia attendings, fellows, residents, and certified resident nurse anesthetists (CRNAs) were sampled based on who was in the OR during observations and who volunteered for interviews. Interviews were transcribed and coded through a consensus procedure. The medication delivery process was described using a SEIPS-based process map. Tasks were separated based on the anesthesia phase, though the tasks and phases are not linear; e.g: A provider may prepare for the next case during the maintenance phase. Fourteen unique hazards were identified during observations including vial/syringe swaps, mislabeling, contamination, technology malfunction, dropped medications, delayed or incorrect documentation, distractions and interruptions, unavailability of emergency medication, and miscommunication. Most of the hazards were associated with the induction phase. The majority of the mitigation strategies involved actions that need to be completed by the anesthesia provider to reduce the likelihood of errors. The interviews confirmed these safety hazards and mitigation strategies and elucidated others including double or triple checking each vial and syringe, labeling immediately after making a syringe, contemporaneous documentation, labeling all sides of a syringe, thorough chart review, communicating manufacturing changes, preparing medications early in the morning in a quiet OR, creating detailed notes on the anesthesia plan the night before, doing a thorough pre-operative patient examination, and consistent medication organization. The interviews also identified other organizational complexities of perioperative anesthesia medication delivery. Mitigation strategies were also identified- most of which depended on the individual provider to remember to perform specific actions. These were consistent with healthcare’s focus on individuals when assigning blame for error. The individual tasks involved in medication preparation and delivery are complicated by the complexity of a perioperative setting. Anesthesia providers may be distracted by other healthcare providers, noise, alarms, and monitors. While providers try to mitigate these risks, systems level solutions are needed to improve safety. This includes improved workspaces, usable tools and technologies, streamlined processes, and organizational policies that facilitate anesthesia work and create goals that are harmonious rather than conflicting. The care of an anesthetized patient is a process of multiple tasks that are interrelated, so using a process model has advantages over traditional task and tool level analysis to identify risk. For example, a distraction during the labeling task could lead to an incorrect administration later in the case. In addition, modeling the process can better elucidate task and tool misfits. The locked top drawer of the cart is good for storing vials and syringes and compliance with TJC, but it does not afford easy access to emergency medications. Methodologically, SEIPS helped to structure direct observations of clinical work. Direct observation of clinical work is necessary to understand work-as-done, which can be complex, opaque, and variable. Finally, it is becoming increasingly clear both from the perceptions of healthcare practitioners in this study, and detailed examination of the literature that the notion of “error” may not be useful. This further demonstrates the value of direct observation, non-linear systems modelling, and process (rather than clinical outcome) evaluation as being central to improving perioperative medication delivery.