Background The pivotal role of the gastrointestinal (GI) tract in sepsis is well recognized. This study aimed to evaluate the associations between defecation frequency as a basic assessment of GI function and the clinical outcomes of intensive care unit patients with suspected sepsis. Methods This retrospective, single-center study included patients suspected of having sepsis. The number of defecations and consecutive days without defecation during the 72 hours preceding the suspected infection were assessed. The primary outcome was 30-day all-cause mortality. Multivariate regression analysis adjusting for potential confounders was employed to establish the associations between GI function and clinical outcomes. Results The final analysis included 1,306 patients with a median age of 56.2 years (interquartile range [IQR], 39.6-69.1); 919 (70.4%) were male, and the median Acute Physiology and Chronic Health Evaluation II score was 22.0 (IQR, 17.0-27.0). The median Sequential Organ Failure Assessment score at the time of suspected infection was 5.0 (IQR, 3.0-7.0). Mortality rates were 20.3%, 28.0%, and 34.3% for patients with 0-2, 3-5, and >5 defecations, respectively (P<0.001). There was a strong correlation between the number of defecations and mortality (r=0.7, P=0.01). In multivariate analyses, each defecation was independently associated with increased mortality (adjusted odds ratio [aOR], 1.07; 95% CI, 1.01-1.12; P=0.01), while each consecutive day without a defecation was associated with reduced mortality (aOR, 0.83; 95% CI, 0.73-0.96; P=0.01). Conclusions A higher number of defecations in the 72 hours preceding suspected sepsis is associated with increased 30-day all-cause mortality, suggesting a potential association with GI tract dysfunction.

Background: The exponential growth in computing power and the increasing digitization of information have substantially advanced the machine learning (ML) research field. However, ML algorithms are often considered "black boxes," and this fosters distrust. In medical domains, in which mistakes can result in fatal outcomes, practitioners may be especially reluctant to trust ML algorithms. Objective: The aim of this study is to explore the effect of user-interface design features on intensivists' trust in an ML-based clinical decision support system. Methods: A total of 47 physicians from critical care specialties were presented with 3 patient cases of bacteremia in the setting of an ML-based simulation system. Three conditions of the simulation were tested according to combinations of information relevancy and interactivity. Participants' trust in the system was assessed by their agreement with the system's prediction and a postexperiment questionnaire. Linear regression models were applied to measure the effects. Results: Participants' agreement with the system's prediction did not differ according to the experimental conditions. However, in the postexperiment questionnaire, higher information relevancy ratings and interactivity ratings were associated with higher perceived trust in the system (P<.001 for both). The explicit visual presentation of the features of the ML algorithm on the user interface resulted in lower trust among the participants (P=.05). Conclusions: Information relevancy and interactivity features should be considered in the design of the user interface of ML-based clinical decision support systems to enhance intensivists' trust. This study sheds light on the connection between information relevancy, interactivity, and trust in human-ML interaction, specifically in the intensive care unit environment.

PURPOSE:Laryngeal and tracheal injuries are known complications of endotracheal intubation. Endotracheal tubes (ETTs) with subglottic suction devices (SSDs) are commonly used in the critical care setting. There is concern that herniation of tissue into the suction port of these devices may lead to tracheal injury resulting in serious clinical consequences such as tracheal stenosis. We aimed to describe the type and location of tracheal injuries seen in intubated critically ill patients and assess injuries at the suction port as well as in-hospital complications associated with those injuries. METHODS:We conducted a prospective observational study of 57 critically ill patients admitted to a level 3 intensive care unit who were endotracheally intubated and underwent percutaneous tracheostomy. Investigators performed bronchoscopy and photographic evaluation of the airway during the percutaneous tracheostomy procedure to evaluate tracheal and laryngeal injury. RESULTS:Forty-one (72%) patients intubated with ETT with SSD and sixteen (28%) patients with standard ETT were included in the study. Forty-seven (83%) patients had a documented airway injury ranging from hyperemia to deep ulceration of the mucosa. A common tracheal injury was at the site of the tracheal cuff. Injury at the site of the subglottic suction device was seen in 5/41 (12%) patients. There were no in-hospital complications. CONCLUSIONS:Airway injury was common in critically ill patients following endotracheal intubation, and tracheal injury commonly occurred at the site of the endotracheal cuff. Injury occurred at the site of the subglottic suction port in some patients although the clinical consequences of these injuries remain unclear.

The exponential growth in computing power and increasing digitization of information have advanced the machine learning (ML) research field substantially. However, ML algorithms are often considered “black boxes”, and this fosters distrust. In medical domains, in which mistakes can result in fatal outcomes, practitioners may be especially reluctant to trust ML algorithms. To explore the effect of user-interface design features on intensivists’ trust in a ML-based clinical decision support system. Forty-seven physicians from critical care specialties were presented three patient cases of bacteremia in the setting of an ML-based simulation system. Three conditions of the simulation were tested according to combinations of information relevancy and interactivity. Participants’ trust in the system was assessed by their agreement with the system’s diagnoses and a post-experiment questionnaire. Linear regression models were applied to measure the effects Participants’ agreement with the system’s diagnoses did not differ according to the experimental conditions. However, in the post-experiment questionnaire, higher information relevancy ratings and interactivity ratings were associated with higher perceived trust in the system (P < 0.001 for both). The explicit visual presentation of the features of the ML algorithm on the user-interface resulted in lower trust by the participants (P < 0.05). : Information relevancy and interactivity features should be considered in the design of user interface of ML-based clinical decision support systems, to enhance intensivists’ trust. This study sheds light on the connection between information relevancy, interactivity, and trust in human–ML interaction, specifically in the intensive care unit environment. Non- clinical

Background Optimal end-of-life care requires identifying patients that are near the end of life. The extent to which attending physicians and trainee physicians agree on the prognoses of their patients is unknown. We investigated agreement between attending and trainee physician on the surprise question: "Would you be surprised if this patient died in the next 12 months?", a question intended to assess mortality risk and unmet palliative care needs. Methods This was a multicentre prospective cohort study of general internal medicine patients at 7 tertiary academic hospitals in Ontario, Canada. General internal medicine attending and senior trainee physician dyads were asked the surprise question for each of the patients for whom they were responsible. Surprise question response agreement was quantified by Cohen's kappa using Bayesian multilevel modeling to account for clustering by physician dyad. Mortality was recorded at 12 months. Results Surprise question responses encompassed 546 patients from 30 attending-trainee physician dyads on academic general internal medicine teams at 7 tertiary academic hospitals in Ontario, Canada. Patients had median age 75 years (IQR 60-85), 260 (48%) were female, and 138 (25%) were dependent for some or all activities of daily living. Trainee and attending physician responses agreed in 406 (75%) patients with adjusted Cohen's kappa of 0.54 (95% credible interval 0.41 to 0.66). Vital status was confirmed for 417 (76%) patients of whom 160 (38% of 417) had died. Using a response of "No" to predict 12-month mortality had positive likelihood ratios of 1.84 (95% CrI 1.55 to 2.22, trainee physicians) and 1.51 (95% CrI 1.30 to 1.72, attending physicians), and negative likelihood ratios of 0.31 (95% CrI 0.17 to 0.48, trainee physicians) and 0.25 (95% CrI 0.10 to 0.46, attending physicians). Conclusion Trainee and attending physician responses to the surprise question agreed in 54% of cases after correcting for chance agreement. Physicians had similar discriminative accuracy; both groups had better accuracy predicting which patients would survive as opposed to which patients would die. Different opinions of a patient's prognosis may contribute to confusion for patients and missed opportunities for engagement with palliative care services.

Objectives The aim of the study was to evaluate the essential and nonessential blood tests ordered on the internal medicine clinical teaching units (CTUs) at Kingston General Hospital. Our aim was to establish a baseline performance measure identifying appropriate use of laboratory tests that could be used to inform improvement over time. Methods For an 8-week period, 14 CTU attending physicians at Kingston General Hospital were surveyed. They were asked for each of their patients, "What blood tests do you consider to be essential for tomorrow morning to maintain appropriate care for this patient?" The following day, blood tests that were ordered were compared with the "essential" list previously given by the attending physicians. Results Of 291 processed blood tests, 148 (51%) had not been considered essential by attending physicians; of the 203 tests considered essential, 60 (30%) were not ordered. Total agreement between "essential" and processed tests was poor (kappa = 0.51; confidence interval, 0.45-0.56). Conclusions Inadequate use of blood tests for CTU patients is common. Quality improvement initiatives should aim to address the lack of observed consensus between attending physicians' views and the ordered tests and to streamline decision-making and the ordering/communication processes. Clinical standards and guidelines regarding ordering of laboratory tests should be clearly defined.

The complexity and criticality of healthcare environments involve many factors, including the humans involved, the environment in which they work, and the tasks they perform. By analyzing the human factors involved, it is possible to improve the tools, processes, and environment layout, ultimately improving healthcare outcomes, such as patient safety. Intuitively, we can identify many human factors that might involve risks and opportunities for improvement. At the same time, with a systematic approach, we believe it is possible to achieve even better results. In the context of healthcare, these improved results could ultimately save lives.

Now that we have assessed the extent of fit between task demands and the environment, on the one hand, and human capabilities, on the other, we are ready to put together our second summary: the fit problems and their severity. Note that while we focus on problems, we should not forget about factors that have a good fit that we want to preserve or facilitate.

Healthcare working environments are complex, and intensive care units (ICUs) are particularly complex due to the influx of data to the healthcare professionals who are providing continuous care to the most critically ill patients. Systems that are designed to work in these environments should take into consideration varied patient conditions, the clinical professionals who use these systems, and the features and performance requirements that will support their efforts to provide care to their patients. We suggest that developing systems that will meet these challenges requires customized design approach, including cognitive system engineering. Until recently, this work domain has been largely ignored by manufacturers of patient monitoring systems. This panel brought together two separate teams who have been using such an approach independently to design new systems for information integration and display in ICU settings. The goals of this panel discussion were to take a close look at the tools and methods that are being used for such a cognitive system engineering approach to the design processes, and to review the recommendations and concepts that are emerging from these processes from each of the two independent teams. This paper summarizes the presentations made during the panel by the two teams regarding updates of ongoing work followed by a lively discussion between panelists and the symposium participants in the audience. Each team had its unique design process that was customized to the specific target ICU, the available resources and goals. The designed systems have original features that evolve from the unique needs of the target unit, yet the designs also share some common features.

This part of the book is designed as a job aid, to support the process you use to analyze adverse events and propose improvements in the system. It is designed as a guide to continuous improvement.

Have you ever experienced the burden of an adverse event or a near-miss in healthcare and wished there was a way to mitigate it? This book walks you through a classic adverse event as a case study and shows you how.It is a practical guide to continuously improving your healthcare environment, processes, tools, and ultimate outcomes, through the discipline of human factors. Using this book, you as a healthcare professional can improve patient safety and quality of care.Adverse events are a major concern in healthcare today. As the complexity of healthcare increases-with technological advances and information overload-the field of human factors offers practical approaches to understand the situation, mitigate risk, and improve outcomes.The first part of this book presents a human factors conceptual framework, and the second part offers a systematic, pragmatic approach. Both the framework and the approach are employed to analyze and understand healthcare situations, both proactively-for constant improvement-and reactively-learning from adverse events.This book guides healthcare professionals through the process of mapping the environmental and human factors; assessing them in relation to the tasks each person performs; recognizing how gaps in the fit between human capabilities and the demands of the task in the environment have a ripple effect that increases risk; and drawing conclusions about what types of changes facilitate improvement and mitigate risk, thereby contributing to improved healthcare outcomes.