Abstract
Cardiac surgery is one of the most complex specialties in medicine, akin to a complex sociotechnical system. Patient outcomes are vulnerable to surgical flow disruptions (SFDs), a source of preventable harm. Healthcare providers' (HCPs) sympathetic activation secondary to emotional states represent an underappreciated source of SFDs. This study's objective was to demonstrate the feasibility of detecting elevated sympathetic nervous system (SNS) activity as a proxy for emotional distress associated with a medication error using heart rate variability (HRV) analysis. After obtaining informed consent, audio/video and HRV data were captured intraoperatively during cardiac surgery from multiple HCPs. Following a critical medication administration error by the anesthesiologist in-training, the attending anesthesiologists' recorded HRV data was analyzed using pyphysio, an open-source signal analysis package, to identify events precipitating this near-miss event. We considered elevated low-frequency/high-frequency (LF/HF) HRV ratio (normal value <2) as a primary indicator of SNS activity and emotional distress. A heightened SNS response by the attending anesthesiologist, observed as an LF/HF ratio value of 3.39, was detected prior to the near-miss event. The attending anesthesiologist confirmed a state of significant SNS activity/distress induced by task-irrelevant environmental factors, which led to a temporarily ineffective mental model. Qualitative analysis of audio/video recordings revealed that SNS activation coincided with an argument over operating room management causing SFD. This preliminary study confirms the feasibility of recognizing potentially detrimental psychophysiological states during cardiac surgery in the wild using HRV analysis. To our knowledge, this is the first case demonstrating SNS activation coinciding with self-reported and observable emotional distress during live surgery using HRV. Irrespective of the HCP's expertise, transient but intense emotional changes may disrupt attention processes leading to SFDs and preventable errors. This work supports the possibility to detect real-time SNS activation, which could enable interventions to proactively mitigate errors. Additional studies on our large database of surgical cases are underway to confirm this observation.