The Application of Human Factors Engineering to Reduce Operating Room Turnover in Robotic Surgery.

BACKGROUND: Challenges associated with turnover time are magnified in robotic surgery. The introduction of advanced technology increases the complexity of an already intricate perioperative environment. We applied a human factors approach to develop systematic, data-driven interventions to reduce robotic surgery turnover time. METHODS: Researchers observed 40 robotic surgery turnovers at a tertiary hospital [20 pre-intervention (Jan 2018 to Apr 2018), 20 post-intervention (Jan 2019 to Jun 2019)]. Components of turnover time, including cleaning, instrument and room set-up, robot preparation, flow disruptions, and major delays, were documented and analyzed. Surveys and focus groups were used to investigate staff perceptions of robotic surgery turnover time. A multidisciplinary team of human factors experts and physicians developed targeted interventions. Pre- and post-intervention turnovers were compared. RESULTS: Median turnover time was 67 min (mean: 72, SD: 24) and 22 major delays were noted (1.1/case). The largest contributors were instrument setup (25.5 min) and cleaning (25 min). Interventions included an electronic dashboard for turnover time reporting, clear designation of roles and simultaneous completion of tasks, process standardization of operating room cleaning, and data transparency through monthly reporting. Post-intervention turnovers were significantly shorter (U = 57.5, p = .000) and ten major delays were noted. CONCLUSIONS: Human factors analysis generated interventions to improve turnover time. Significant improvements were seen post-intervention with a reduction in turnover time by a 26 min and decrease in major delays by over 50%. Future opportunities to intervene and further improve turnover time include targeting pre- and post-operative care phases.

Morbidity, mortality, and systems safety in non-operating room anaesthesia: a narrative review.

Non-operating room anaesthesia (NORA) describes anaesthesia delivered outside a traditional operating room (OR) setting. Non-operating room anaesthesia cases have increased significantly in the last 20 yr and are projected to account for half of all anaesthetics delivered in the next decade. In contrast to most other medication administration contexts, NORA is performed in high-volume fast-paced environments not optimised for anaesthesia care. These predisposing factors combined with increasing case volume, less provider experience, and higher-acuity patients increase the potential for preventable adverse events. Our narrative review examines morbidity and mortality in NORA settings compared with the OR and the systems factors impacting safety in NORA. A review of the literature from January 1, 1994 to March 5, 2021 was conducted using PubMed, CINAHL, Scopus, and ProQuest. After completing abstract screening and full-text review, 30 articles were selected for inclusion. These articles suggested higher rates of morbidity and mortality in NORA cases compared with OR cases. This included a higher proportion of death claims and complications attributable to inadequate oxygenation, and a higher likelihood that adverse events are preventable. Despite relatively few attempts to quantify safety concerns, it was possible to find a range of systems safety concerns repeated across multiple studies, including insufficient lighting, noise, cramped workspace, and restricted access to patients. Old and unfamiliar equipment, lack of team familiarity, and limited preoperative evaluation are also commonly noted challenges. Applying a systems view of safety, it is possible to suggest a range of methods to improve NORA safety and performance.

Red Flag/Blue Flag visualization of a common CNN for text classification.

A shallow convolutional neural network (CNN), TextCNN, has become nearly ubiquitous for classification among clinical and medical text. This research presents a novel eXplainable-AI (X-AI) software, Red Flag/Blue Flag (RFBF), designed for binary classification with TextCNN. RFBF visualizes each convolutional filter's discriminative capability. This is a more informative approach than direct assessment of logit contribution, features that overfit to train set nuances on smaller datasets may indiscriminately activate large logits on validation samples from both classes. RFBF enables model diagnosis, term feature verification, and overfit prevention. We present 3 use cases of (1) filter consistency assessment; (2) predictive performance improvement; and (3) estimation of information leakage between train and holdout sets. The use cases derive from experiments on TextCNN for binary prediction of surgical misadventure outcomes from physician-authored operative notes. Due to TextCNN's prevalence, this X-AI can benefit clinical text research, and hence improve patient outcomes.