Effect of Machine Learning on Anaesthesiology Clinician Prediction of Postoperative Complications: The Perioperative ORACLE Randomised Clinical Trial.

Background: Anaesthesiology clinicians can implement risk mitigation strategies if they know which patients are at greatest risk for postoperative complications. Although machine learning models predicting complications exist, their impact on clinician risk assessment is unknown. Methods: This single-centre randomised clinical trial enrolled patients age ≥18 undergoing surgery with anaesthesiology services. Anaesthesiology clinicians providing remote intraoperative telemedicine support reviewed electronic health records with (assisted group) or without (unassisted group) also reviewing machine learning predictions. Clinicians predicted the likelihood of postoperative 30-day all-cause mortality and postoperative acute kidney injury within 7 days. Area under the receiver operating characteristic curve (AUROC) for the clinician predictions was determined. Results: Among 5,071 patient cases reviewed by 89 clinicians, the observed incidence was 2% for postoperative death and 11% for acute kidney injury. Clinician predictions agreed with the models more strongly in the assisted versus unassisted group (weighted kappa 0.75 versus 0.62 for death [difference 0.13, 95%CI 0.10-0.17] and 0.79 versus 0.54 for kidney injury [difference 0.25, 95%CI 0.21-0.29]). Clinicians predicted death with AUROC of 0.793 in the assisted group and 0.780 in the unassisted group (difference 0.013, 95%CI -0.070 to 0.097). Clinicians predicted kidney injury with AUROC of 0.734 in the assisted group and 0.688 in the unassisted group (difference 0.046, 95%CI -0.003 to 0.091). Conclusions: Although there was evidence that the models influenced clinician predictions, clinician performance was not statistically significantly different with and without machine learning assistance. Further work is needed to clarify the role of machine learning in real-time perioperative risk stratification. Trial Registration: ClinicalTrials.gov NCT05042804.

Effect of Telemedicine Support for Intraoperative Anaesthesia Care on Postoperative Outcomes: The TECTONICS Randomised Clinical Trial.

Background: Novel applications of telemedicine can improve care quality and patient outcomes. Telemedicine for intraoperative decision support has not been rigorously studied. Methods: This single centre randomised clinical trial ( clinicaltrials.gov NCT03923699 ) of unselected adult surgical patients was conducted between July 1, 2019 and January 31, 2023. Patients received usual care or decision support from a telemedicine service, the Anesthesiology Control Tower (ACT). The ACT provided real-time recommendations to intraoperative anaesthesia clinicians based on case reviews, machine-learning forecasting, and physiologic alerts. ORs were randomised 1:1. Co-primary outcomes of 30-day all-cause mortality, respiratory failure, acute kidney injury (AKI), and delirium were analysed as intention-to-treat. Results: The trial completed planned enrolment with 71927 surgeries (35956 ACT; 35971 usual care). After multiple testing correction, there was no significant effect of the ACT vs. usual care on 30-day mortality [641/35956 (1.8%) vs 638/35971 (1.8%), risk difference 0.0% (95% CI -0.2% to 0.3%), p=0.96], respiratory failure [1089/34613 (3.1%) vs 1112/34619 (3.2%), risk difference -0.1% (95% CI -0.4% to 0.3%), p=0.96], AKI [2357/33897 (7%) vs 2391/33795 (7.1%), risk difference -0.1% (-0.6% to 0.4%), p=0.96], or delirium [1283/3928 (32.7%) vs 1279/3989 (32.1%), risk difference 0.6% (-2.0% to 3.2%), p=0.96]. There were no significant differences in secondary outcomes or in sensitivity analyses. Conclusions: In this large RCT of a novel application of telemedicine-based remote monitoring and decision support using real-time alerts and case reviews, we found no significant differences in postoperative outcomes. Large-scale intraoperative telemedicine is feasible, and we suggest future avenues where it may be impactful.

Perioperative Care of Patients Using Wearable Diabetes Devices.

The increasing prevalence of diabetes mellitus has been accompanied by a rapid expansion in wearable continuous glucose monitoring (CGM) devices and insulin pumps. Systems combining these components in a "closed loop," where interstitial glucose measurement guides automated insulin delivery (AID, or closed loop) based on sophisticated algorithms, are increasingly common. While these devices' efficacy in achieving near-normoglycemia is contributing to increasing usage among patients with diabetes, the management of these patients in operative and procedural environments remains understudied with limited published guidance available, particularly regarding AID systems. With their growing prevalence, practical management advice is needed for their utilization, or for the rational temporary substitution of alternative diabetes monitoring and treatments, during surgical care. CGM devices monitor interstitial glucose in real time; however, there are potential limitations to use and accuracy in the perioperative period, and, at the present time, their use should not replace regular point-of-care glucose monitoring. Avoiding perioperative removal of CGMs when possible is important, as removal of these prescribed devices can result in prolonged interruptions in CGM-informed treatments during and after procedures, particularly AID system use. Standalone insulin pumps provide continuous subcutaneous insulin delivery without automated adjustments for glucose concentrations and can be continued during some procedures. The safe intraoperative use of AID devices in their hybrid closed-loop mode (AID mode) requires the CGM component of the system to continue to communicate valid blood glucose data, and thus introduces the additional need to ensure this portion of the system is functioning appropriately to enable intraprocedural use. AID devices revert to non-AID insulin therapy modes when paired CGMs are disconnected or when the closed-loop mode is intentionally disabled. For patients using insulin pumps, we describe procedural factors that may compromise CGM, insulin pump, and AID use, necessitating a proactive transition to an alternative insulin regimen. Procedure duration and invasiveness is an important factor as longer procedures increase the risk of stress hyperglycemia, tissue malperfusion, and device malfunction. Whether insulin pumps should be continued through procedures, or substituted by alternative insulin delivery methods, is a complex decision that requires all parties to understand potential risks and contingency plans relating to patient and procedural factors. Currently available CGMs and insulin pumps are reviewed, and practical recommendations for safe glycemic management during the phases of perioperative care are provided.

An Intraoperative Telemedicine Program to Improve Perioperative Quality Measures: The ACTFAST-3 Randomized Clinical Trial.

Importance: Telemedicine for clinical decision support has been adopted in many health care settings, but its utility in improving intraoperative care has not been assessed. Objective: To pilot the implementation of a real-time intraoperative telemedicine decision support program and evaluate whether it reduces postoperative hypothermia and hyperglycemia as well as other quality of care measures. Design, Setting, and Participants: This single-center pilot randomized clinical trial (Anesthesiology Control Tower-Feedback Alerts to Supplement Treatments [ACTFAST-3]) was conducted from April 3, 2017, to June 30, 2019, at a large academic medical center in the US. A total of 26 254 adult surgical patients were randomized to receive either usual intraoperative care (control group; n = 12 980) or usual care augmented by telemedicine decision support (intervention group; n = 13 274). Data were initially analyzed from April 22 to May 19, 2021, with updates in November 2022 and February 2023. Intervention: Patients received either usual care (medical direction from the anesthesia care team) or intraoperative anesthesia care monitored and augmented by decision support from the Anesthesiology Control Tower (ACT), a real-time, live telemedicine intervention. The ACT incorporated remote monitoring of operating rooms by a team of anesthesia clinicians with customized analysis software. The ACT reviewed alerts and electronic health record data to inform recommendations to operating room clinicians. Main Outcomes and Measures: The primary outcomes were avoidance of postoperative hypothermia (defined as the proportion of patients with a final recorded intraoperative core temperature >36 °C) and hyperglycemia (defined as the proportion of patients with diabetes who had a blood glucose level ≤180 mg/dL on arrival to the postanesthesia recovery area). Secondary outcomes included intraoperative hypotension, temperature monitoring, timely antibiotic redosing, intraoperative glucose evaluation and management, neuromuscular blockade documentation, ventilator management, and volatile anesthetic overuse. Results: Among 26 254 participants, 13 393 (51.0%) were female and 20 169 (76.8%) were White, with a median (IQR) age of 60 (47-69) years. There was no treatment effect on avoidance of hyperglycemia (7445 of 8676 patients [85.8%] in the intervention group vs 7559 of 8815 [85.8%] in the control group; rate ratio [RR], 1.00; 95% CI, 0.99-1.01) or hypothermia (7602 of 11 447 patients [66.4%] in the intervention group vs 7783 of 11 672 [66.7.%] in the control group; RR, 1.00; 95% CI, 0.97-1.02). Intraoperative glucose measurement was more common among patients with diabetes in the intervention group (RR, 1.07; 95% CI, 1.01-1.15), but other secondary outcomes were not significantly different. Conclusions and Relevance: In this randomized clinical trial, anesthesia care quality measures did not differ between groups, with high confidence in the findings. These results suggest that the intervention did not affect the targeted care practices. Further streamlining of clinical decision support and workflows may help the intraoperative telemedicine program achieve improvement in targeted clinical measures. Trial Registration: ClinicalTrials.gov Identifier: NCT02830126.

Reducing perioperative red blood cell unit issue orders, returns, and waste using failure modes and effects analysis.

BACKGROUND: Surgical transfusion has an outsized impact on hospital-based transfusion services, leading to blood product waste and unnecessary costs. The objective of this study was to design and implement a streamlined, reliable process for perioperative blood issue ordering and delivery to reduce waste. STUDY DESIGN AND METHODS: To address the high rates of surgical blood issue requests and red blood cell (RBC) unit waste at a large academic medical center, a failure modes and effects analysis was used to systematically examine perioperative blood management practices. Based on identified failure modes (e.g., miscommunication, knowledge gaps), a multi-component action plan was devised involving process changes, education, electronic clinical decision support, audit, and feedback. Changes in RBC unit issue requests, returns, waste, labor, and cost were measured pre- and post-intervention. RESULTS: The number of perioperative RBC unit issue requests decreased from 358 per month (SD 24) pre-intervention to 282 per month (SD 16) post-intervention (p < .001), resulting in an estimated savings of 8.9 h per month in blood bank staff labor. The issue-to-transfusion ratio decreased from 2.7 to 2.1 (p < .001). Perioperative RBC unit waste decreased from 4.5% of units issued pre-intervention to 0.8% of units issued post-intervention (p < .001), saving an estimated $148,543 in RBC unit acquisition costs and $546,093 in overhead costs per year. DISCUSSION: Our intervention, designed based on a structured failure modes analysis, achieved sustained reductions in perioperative RBC unit issue orders, returns, and waste, with associated benefits for blood conservation and transfusion program costs.

Postoperative Delirium Severity and Recovery Correlate With Electroencephalogram Spectral Features.

BACKGROUND: Delirium is an acute syndrome characterized by inattention, disorganized thinking, and an altered level of consciousness. A reliable biomarker for tracking delirium does not exist, but oscillations in the electroencephalogram (EEG) could address this need. We evaluated whether the frequencies of EEG oscillations are associated with delirium onset, severity, and recovery in the postoperative period. METHODS: Twenty-six adults enrolled in the Electroencephalography Guidance of Anesthesia to Alleviate Geriatric Syndromes (ENGAGES; ClinicalTrials.gov NCT02241655) study underwent major surgery requiring general anesthesia, and provided longitudinal postoperative EEG recordings for this prespecified substudy. The presence and severity of delirium were evaluated with the confusion assessment method (CAM) or the CAM-intensive care unit. EEG data obtained during awake eyes-open and eyes-closed states yielded relative power in the delta (1-4 Hz), theta (4-8 Hz), and alpha (8-13 Hz) bands. Discriminability for delirium presence was evaluated with c-statistics. To account for correlation among repeated measures within patients, mixed-effects models were generated to assess relationships between: (1) delirium severity and EEG relative power (ordinal), and (2) EEG relative power and time (linear). Slopes of ordinal and linear mixed-effects models are reported as the change in delirium severity score/change in EEG relative power, and the change in EEG relative power/time (days), respectively. Bonferroni correction was applied to confidence intervals (CIs) to account for multiple comparisons. RESULTS: Occipital alpha relative power during eyes-closed states offered moderate discriminability (c-statistic, 0.75; 98% CI, 0.58-0.87), varying inversely with delirium severity (slope, -0.67; 98% CI, -1.36 to -0.01; P = .01) and with severity of inattention (slope, -1.44; 98% CI, -2.30 to -0.58; P = .002). Occipital theta relative power during eyes-open states correlated directly with severity of delirium (slope, 1.28; 98% CI, 0.12-2.44; P = .007), inattention (slope, 2.00; 98% CI, 0.48-3.54; P = .01), and disorganized thinking (slope, 3.15; 98% CI, 0.66-5.65; P = .01). Corresponding frontal EEG measures recapitulated these relationships to varying degrees. Severity of altered level of consciousness correlated with frontal theta relative power during eyes-open states (slope, 11.52; 98% CI, 6.33-16.71; P < .001). Frontal theta relative power during eyes-open states correlated inversely with time (slope, -0.05; 98% CI, -0.12 to -0.04; P = .002). CONCLUSIONS: Presence, severity, and core features of postoperative delirium covary with spectral features of the EEG. The cost and accessibility of EEG facilitate the translation of these findings to future mechanistic and interventional trials.

Predicting Postoperative Troponin in Patients Undergoing Elective Hip or Knee Arthroplasty: A Comparison of Five Cardiac Risk Prediction Tools.

Background: Elderly patients undergoing hip or knee arthroplasty are at a risk for myocardial injury after noncardiac surgery (MINS). We evaluated the ability of five common cardiac risk scores, alone or combined with baseline high-sensitivity cardiac troponin I (hs-cTnI), in predicting MINS and postoperative day 2 (POD2) hs-cTnI levels in patients undergoing elective total hip or knee arthroplasty. Methods: This study is ancillary to the Genetics-InFormatics Trial (GIFT) of Warfarin Therapy to Prevent Deep Venous Thrombosis, which enrolled patients 65 years and older undergoing elective total hip or knee arthroplasty. The five cardiac risk scores evaluated were the atherosclerotic cardiovascular disease calculator (ASCVD), the Framingham risk score (FRS), the American College of Surgeon's National Surgical Quality Improvement Program (ACS-NSQIP) calculator, the revised cardiac risk index (RCRI), and the reconstructed RCRI (R-RCRI). Results: None of the scores predicted MINS in women. Among men, the ASCVD (C-statistic of 0.66; p=0.04), ACS-NSQIP (C-statistic of 0.69; p=0.01), and RCRI (C-statistic of 0.64; p=0.04) predicted MINS. Among all patients, spearman correlations (r s) of the risk scores with the POD2 hs-cTnI levels were 0.24, 0.20, 0.11, 0.11, and 0.08 for the ASCVD, Framingham, ACS-NSQIP, RCRI, and R-RCRI scores, respectively, with p values of <0.001, <0.001, <0.001, 0.006, and 0.025. Baseline hs-cTnI predicted MINS (C-statistics: 0.63 in women and 0.72 in men) and postoperative hs-cTnI (r s = 0.51, p=0.001). Conclusion: In elderly patients undergoing elective hip or knee arthroplasty, several of the scores modestly predicted MINS in men and correlated with POD2 hs-cTnI.

An ethnographic study on the impact of a novel telemedicine-based support system in the operating room.

OBJECTIVE: The Anesthesiology Control Tower (ACT) for operating rooms (ORs) remotely assesses the progress of surgeries and provides real-time perioperative risk alerts, communicating risk mitigation recommendations to bedside clinicians. We aim to identify and map ACT-OR nonroutine events (NREs)-risk-inducing or risk-mitigating workflow deviations-and ascertain ACT's impact on clinical workflow and patient safety. MATERIALS AND METHODS: We used ethnographic methods including shadowing ACT and OR clinicians during 83 surgeries, artifact collection, chart reviews for decision alerts sent to the OR, and 10 clinician interviews. We used hybrid thematic analysis informed by a human-factors systems-oriented approach to assess ACT's role and impact on safety, conducting content analysis to assess NREs. RESULTS: Across 83 cases, 469 risk alerts were triggered, and the ACT sent 280 care recommendations to the OR. 135 NREs were observed. Critical factors facilitating ACT's role in supporting patient safety included providing backup support and offering a fresh-eye perspective on OR decisions. Factors impeding ACT included message timing and ACT and OR clinician cognitive lapses. Suggestions for improvement included tailoring ACT message content (structure, timing, presentation) and incorporating predictive analytics for advanced planning. DISCUSSION: ACT served as a safety net with remote surveillance features and as a learning healthcare system with feedback/auditing features. Supporting strategies include adaptive coordination and harnessing clinician/patient support to improve ACT's sustainability. Study insights inform future intraoperative telemedicine design considerations to mitigate safety risks. CONCLUSION: Incorporating similar remote technology enhancement into routine perioperative care could markedly improve safety and quality for millions of surgical patients.

Association of a Perioperative Multicomponent Fall Prevention Intervention With Falls and Quality of Life After Elective Inpatient Surgical Procedures.

Importance: Falls after elective inpatient surgical procedures are common and have physical, emotional, and financial consequences. Close interactions between patients and health care teams before and after surgical procedures may offer opportunities to address modifiable risk factors associated with falls. Objective: To assess whether a multicomponent intervention that incorporates education, home medication review, and home safety assessment is associated with reductions in the incidence of falls after elective inpatient surgical procedures. Design, Setting, and Participants: This prospective propensity score-matched cohort study was a prespecified secondary analysis of data from the Electroencephalography Guidance of Anesthesia to Alleviate Geriatric Syndromes (ENGAGES) randomized clinical trial, which was conducted at a single academic medical center between January 16, 2015, and May 7, 2018. Patients in the intervention group of the present study were enrolled in either arm of the ENGAGES clinical trial. Patients in the control group were selected from the Systematic Assessment and Targeted Improvement of Services Following Yearly Surgical Outcomes Surveys prospective observational cohort study, which created a registry of patient-reported postoperative outcomes at the same single center. The propensity score-matched cohort in the present study included 1396 patients (698 pairs) selected from a pool of 2013 eligible patients. All patients underwent elective surgical procedures with general anesthesia and had a hospital stay of 2 or more days. Data were analyzed from January 2, 2020, to January 11, 2022. Interventions: The multicomponent safety intervention (offered to all patients in the ENGAGES clinical trial) included patient education on fall prevention techniques, home medication review by a geriatric psychiatrist (with communication of recommended changes to the surgeon), a self-administered home safety assessment, and targeted occupational therapy home visits with home hazard removal (offered to patients with a preoperative history of falls). Main Outcomes and Measures: The primary outcome was patient-reported falls within 1 year after an elective inpatient surgical procedure. The secondary outcome was quality of life 1 year after an elective surgical procedure, which was measured using the physical and mental composite summary scores on the Veterans RAND 12-item health survey (score range, 0-100 points, with 0 indicating lowest quality of life and 100 indicating highest quality of life). Results: Among 1396 patients, the median age was 69 years (IQR, 64-75 years), and 739 patients (52.9%) were male. With regard to race, 5 patients (0.4%) were Asian, 97 (6.9%) were Black or African American, 2 (0.1%) were Native Hawaiian or Pacific Islander, 1237 (88.6%) were White, 3 (0.2%) were of other race, and 52 (3.7%) were of unknown race; with regard to ethnicity, 12 patients (0.9%) were Hispanic or Latino, 1335 (95.6%) were non-Hispanic or non-Latino, and 49 (3.5%) were of unknown ethnicity. Adherence to individual intervention components was modest (from 22.9% for completion of the self-administered home safety assessment to 28.2% for implementation of the geriatric psychiatrist's recommended medication changes). Falls within 1 year after surgical procedures were reported by 228 of 698 patients (32.7%) in the intervention group and 225 of 698 patients (32.2%) in the control group. No significant difference was found in falls between the 2 groups (standardized risk difference, 0.4%; 95% CI, -4.5% to 5.3%). After adjusting for preoperative quality of life, patients in the intervention group had higher physical composite summary scores (3.8 points; 95% CI, 2.4-5.1 points) and higher mental composite summary scores (5.7 points; 95% CI, 4.7-6.7 points) at 1 year compared with patients in the control group. Conclusions and Relevance: In this cohort study, a multicomponent safety intervention was not associated with reductions in falls within the first year after an elective surgical procedure; however, an increase in quality of life at 1 year was observed. These results suggest a need for other interventions, such as those designed to increase adherence, to lower the incidence of falls after surgical procedures.

Personalized Surgical Transfusion Risk Prediction Using Machine Learning to Guide Preoperative Type and Screen Orders.

BACKGROUND: Accurate estimation of surgical transfusion risk is essential for efficient allocation of blood bank resources and for other aspects of anesthetic planning. This study hypothesized that a machine learning model incorporating both surgery- and patient-specific variables would outperform the traditional approach that uses only procedure-specific information, allowing for more efficient allocation of preoperative type and screen orders. METHODS: The American College of Surgeons National Surgical Quality Improvement Program Participant Use File was used to train four machine learning models to predict the likelihood of red cell transfusion using surgery-specific and patient-specific variables. A baseline model using only procedure-specific information was created for comparison. The models were trained on surgical encounters that occurred at 722 hospitals in 2016 through 2018. The models were internally validated on surgical cases that occurred at 719 hospitals in 2019. Generalizability of the best-performing model was assessed by external validation on surgical cases occurring at a single institution in 2020. RESULTS: Transfusion prevalence was 2.4% (73,313 of 3,049,617), 2.2% (23,205 of 1,076,441), and 6.7% (1,104 of 16,053) across the training, internal validation, and external validation cohorts, respectively. The gradient boosting machine outperformed the baseline model and was the best- performing model. At a fixed 96% sensitivity, this model had a positive predictive value of 0.06 and 0.21 and recommended type and screens for 36% and 30% of the patients in internal and external validation, respectively. By comparison, the baseline model at the same sensitivity had a positive predictive value of 0.04 and 0.144 and recommended type and screens for 57% and 45% of the patients in internal and external validation, respectively. The most important predictor variables were overall procedure-specific transfusion rate and preoperative hematocrit. CONCLUSIONS: A personalized transfusion risk prediction model was created using both surgery- and patient-specific variables to guide preoperative type and screen orders and showed better performance compared to the traditional procedure-centric approach.

Protocol for the perioperative outcome risk assessment with computer learning enhancement (Periop ORACLE) randomized study.

Background: More than four million people die each year in the month following surgery, and many more experience complications such as acute kidney injury. Some of these outcomes may be prevented through early identification of at-risk patients and through intraoperative risk mitigation. Telemedicine has revolutionized the way at-risk patients are identified in critical care, but intraoperative telemedicine services are not widely used in anesthesiology. Clinicians in telemedicine settings may assist with risk stratification and brainstorm risk mitigation strategies while clinicians in the operating room are busy performing other patient care tasks. Machine learning tools may help clinicians in telemedicine settings leverage the abundant electronic health data available in the perioperative period. The primary hypothesis for this study is that anesthesiology clinicians can predict postoperative complications more accurately with machine learning assistance than without machine learning assistance. Methods: This investigation is a sub-study nested within the TECTONICS randomized clinical trial (NCT03923699). As part of TECTONICS, study team members who are anesthesiology clinicians working in a telemedicine setting are currently reviewing ongoing surgical cases and documenting how likely they feel the patient is to experience 30-day in-hospital death or acute kidney injury. For patients who are included in this sub-study, these case reviews will be randomized to be performed with access to a display showing machine learning predictions for the postoperative complications or without access to the display. The accuracy of the predictions will be compared across these two groups. Conclusion: Successful completion of this study will help define the role of machine learning not only for intraoperative telemedicine, but for other risk assessment tasks before, during, and after surgery. Registration: ORACLE is registered on ClinicalTrials.gov: NCT05042804; registered September 13, 2021.

Effect of electroencephalogram-guided anaesthesia administration on 1-yr mortality: follow-up of a randomised clinical trial.

BACKGROUND: Intraoperative EEG suppression duration has been associated with postoperative delirium and mortality. In a clinical trial testing anaesthesia titration to avoid EEG suppression, the intervention did not decrease the incidence of postoperative delirium, but was associated with reduced 30-day mortality. The present study evaluated whether the EEG-guided anaesthesia intervention was also associated with reduced 1-yr mortality. METHODS: This manuscript reports 1 yr follow-up of subjects from a single-centre RCT, including a post hoc secondary outcome (1-yr mortality) in addition to pre-specified secondary outcomes. The trial included subjects aged 60 yr or older undergoing surgery with general anaesthesia between January 2015 and May 2018. Patients were randomised to receive EEG-guided anaesthesia or usual care. The previously reported primary outcome was postoperative delirium. The outcome of the current study was all-cause 1-yr mortality. RESULTS: Of the 1232 subjects enrolled, 614 subjects were randomised to EEG-guided anaesthesia and 618 subjects to usual care. One-year mortality was 57/591 (9.6%) in the guided group and 62/601 (10.3%) in the usual-care group. No significant difference in mortality was observed (adjusted absolute risk difference, -0.7%; 99.5% confidence interval, -5.8% to 4.3%; P=0.68). CONCLUSIONS: An EEG-guided anaesthesia intervention aiming to decrease duration of EEG suppression during surgery did not significantly decrease 1-yr mortality. These findings, in the context of other studies, do not provide supportive evidence for EEG-guided anaesthesia to prevent intermediate term postoperative death. CLINICAL TRIAL REGISTRATION: NCT02241655.

An observational study of postoperative handoff standardization failures.

BACKGROUND: Patient handoffs from an operating room (OR) to an intensive care unit (ICU) require precise coordination among surgical, anesthesia, and critical care teams. Although several standardized handoff strategies have been developed, their sustainability remains is poor. Little is known regarding factors that impede handoff standardization. PURPOSE: Our objectives are three-fold: (1) highlight compliance failures with standardized handoffs; (2) identify factors contributing to compliance failures; and (3) develop guidelines for sustainable handoff interventions and processes. METHODS: We used ethnographic data collection methods-general observations, handoff shadowing, and semi-structured clinician interviews-with 84 participants from OR, ICU, and telemedicine teams at a large academic medical center. We conducted thematic analysis supported by inductive and deductive coding using the Systems Engineering Initiative for Patient Safety (SEIPS) framework. RESULTS: Post-operative handoffs can be characterized into four phases: pre-transfer preparation, transfer and setup, report preparation and delivery, and post-transfer care. We identified compliance failures with standardized handoff protocols and associated risk factors within the OR-ICU work system including limited teamwork, absence of handoff-specific tools, and poor clinician buy-in. To improve handoffs, clinicians provided suggestions for developing collaborative Electronic Health Record (EHR)-integrated handoff tools and re-engineering the handoff process. CONCLUSIONS: Compliance failures are prevalent in all handoff phases, leading to poor adherence with standardization. We propose theoretically grounded guidelines for designing "flexibly standardized" bundled handoff interventions for ensuring care continuity in OR to ICU transitions of care.

Use of Machine Learning to Develop and Evaluate Models Using Preoperative and Intraoperative Data to Identify Risks of Postoperative Complications.

Importance: Postoperative complications can significantly impact perioperative care management and planning. Objectives: To assess machine learning (ML) models for predicting postoperative complications using independent and combined preoperative and intraoperative data and their clinically meaningful model-agnostic interpretations. Design, Setting, and Participants: This retrospective cohort study assessed 111 888 operations performed on adults at a single academic medical center from June 1, 2012, to August 31, 2016, with a mean duration of follow-up based on the length of postoperative hospital stay less than 7 days. Data analysis was performed from February 1 to September 31, 2020. Main Outcomes and Measures: Outcomes included 5 postoperative complications: acute kidney injury (AKI), delirium, deep vein thrombosis (DVT), pulmonary embolism (PE), and pneumonia. Patient and clinical characteristics available preoperatively, intraoperatively, and a combination of both were used as inputs for 5 candidate ML models: logistic regression, support vector machine, random forest, gradient boosting tree (GBT), and deep neural network (DNN). Model performance was compared using the area under the receiver operating characteristic curve (AUROC). Model interpretations were generated using Shapley Additive Explanations by transforming model features into clinical variables and representing them as patient-specific visualizations. Results: A total of 111 888 patients (mean [SD] age, 54.4 [16.8] years; 56 915 [50.9%] female; 82 533 [73.8%] White) were included in this study. The best-performing model for each complication combined the preoperative and intraoperative data with the following AUROCs: pneumonia (GBT), 0.905 (95% CI, 0.903-0.907); AKI (GBT), 0.848 (95% CI, 0.846-0.851); DVT (GBT), 0.881 (95% CI, 0.878-0.884); PE (DNN), 0.831 (95% CI, 0.824-0.839); and delirium (GBT), 0.762 (95% CI, 0.759-0.765). Performance of models that used only preoperative data or only intraoperative data was marginally lower than that of models that used combined data. When adding variables with missing data as input, AUROCs increased from 0.588 to 0.905 for pneumonia, 0.579 to 0.848 for AKI, 0.574 to 0.881 for DVT, 0.5 to 0.831 for PE, and 0.6 to 0.762 for delirium. The Shapley Additive Explanations analysis generated model-agnostic interpretation that illustrated significant clinical contributors associated with risks of postoperative complications. Conclusions and Relevance: The ML models for predicting postoperative complications with model-agnostic interpretation offer opportunities for integrating risk predictions for clinical decision support. Such real-time clinical decision support can mitigate patient risks and help in anticipatory management for perioperative contingency planning.

Abnormal preoperative cognitive screening in aged surgical patients: a retrospective cohort analysis.

BACKGROUND: Preoperative cognitive dysfunction has been associated with adverse postoperative outcomes. There are limited data characterising the epidemiology of preoperative cognitive dysfunction in older surgical patients. METHODS: This retrospective cohort included all patients ≥65 yr old seen at the Washington University preoperative clinic between January 2013 and June 2018. Cognitive screening was performed using the Short-Blessed Test (SBT) and Eight-Item Interview to Differentiate Aging and Dementia (AD8) screen. The primary outcome of abnormal cognitive screening was defined as SBT score ≥5 or AD8 score ≥2. Multivariable logistic regression was used to identify associated factors. RESULTS: Overall, 21 666 patients ≥65 yr old completed screening during the study period; 23.5% (n=5099) of cognitive screens were abnormal. Abnormal cognitive screening was associated with increasing age, decreasing BMI, male sex, non-Caucasian race, decreased functional independence, and decreased metabolic functional capacity. Patients with a history of stroke or transient ischaemic attack, chronic obstructive pulmonary disease, diabetes mellitus, hepatic cirrhosis, and heavy alcohol use were also more likely to have an abnormal cognitive screen. Predictive modelling showed no combination of patient factors was able to reliably identify patients who had a <10% probability of abnormal cognitive screening. CONCLUSIONS: Routine preoperative cognitive screening of unselected aged surgical patients often revealed deficits consistent with cognitive impairment or dementia. Such deficits were associated with increased age, decreased function, decreased BMI, and several common medical comorbidities. Further research is necessary to characterise the clinical implications of preoperative cognitive dysfunction and identify interventions that may reduce related postoperative complications.

Protocol for the Prognosticating Delirium Recovery Outcomes Using Wakefulness and Sleep Electroencephalography (P-DROWS-E) study: a prospective observational study of delirium in elderly cardiac surgical patients.

INTRODUCTION: Delirium is a potentially preventable disorder characterised by acute disturbances in attention and cognition with fluctuating severity. Postoperative delirium is associated with prolonged intensive care unit and hospital stay, cognitive decline and mortality. The development of biomarkers for tracking delirium could potentially aid in the early detection, mitigation and assessment of response to interventions. Because sleep disruption has been posited as a contributor to the development of this syndrome, expression of abnormal electroencephalography (EEG) patterns during sleep and wakefulness may be informative. Here we hypothesise that abnormal EEG patterns of sleep and wakefulness may serve as predictive and diagnostic markers for postoperative delirium. Such abnormal EEG patterns would mechanistically link disrupted thalamocortical connectivity to this important clinical syndrome. METHODS AND ANALYSIS: P-DROWS-E (Prognosticating Delirium Recovery Outcomes Using Wakefulness and Sleep Electroencephalography) is a 220-patient prospective observational study. Patient eligibility criteria include those who are English-speaking, age 60 years or older and undergoing elective cardiac surgery requiring cardiopulmonary bypass. EEG acquisition will occur 1-2 nights preoperatively, intraoperatively, and up to 7 days postoperatively. Concurrent with EEG recordings, two times per day postoperative Confusion Assessment Method (CAM) evaluations will quantify the presence and severity of delirium. EEG slow wave activity, sleep spindle density and peak frequency of the posterior dominant rhythm will be quantified. Linear mixed-effects models will be used to evaluate the relationships between delirium severity/duration and EEG measures as a function of time. ETHICS AND DISSEMINATION: P-DROWS-E is approved by the ethics board at Washington University in St. Louis. Recruitment began in October 2018. Dissemination plans include presentations at scientific conferences, scientific publications and mass media. TRIAL REGISTRATION NUMBER: NCT03291626.

Protocol for a proof-of-concept observational study evaluating the potential utility and acceptability of a telemedicine solution for the post-anesthesia care unit.

Introduction: The post-anesthesia care unit (PACU) is a clinical area designated for patients recovering from invasive procedures. There are typically several geographically dispersed PACUs within hospitals. Patients in the PACU can be unstable and at risk for complications. However, clinician coverage and patient monitoring in PACUs is not well regulated and might be sub-optimal. We hypothesize that a telemedicine center for the PACU can improve key PACU functions. Objectives: The objective of this study is to demonstrate the potential utility and acceptability of a telemedicine center to complement the key functions of the PACU. These include participation in hand-off activities to and from the PACU, detection of physiological derangements, identification of symptoms requiring treatment, recognition of situations requiring emergency medical intervention, and determination of patient readiness for PACU discharge. Methods and analysis: This will be a single center prospective before-and-after proof-of-concept study. Adults (18 years and older) undergoing elective surgery and recovering in two selected PACU bays will be enrolled. During the initial three-month observation phase, clinicians in the telemedicine center will not communicate with clinicians in the PACU, unless there is a specific patient safety concern. During the subsequent three-month interaction phase, clinicians in the telemedicine center will provide structured decision support to PACU clinicians. The primary outcome will be time to PACU discharge readiness determination in the two study phases. The attitudes of key stakeholders towards the telemedicine center will be assessed. Other outcomes will include detection of physiological derangements, complications, adverse symptoms requiring treatments, and emergencies requiring medical intervention. Registration: This trial is registered on clinicaltrials.gov, NCT04020887 (16 th July 2019).

Preoperative Cognitive Abnormality, Intraoperative Electroencephalogram Suppression, and Postoperative Delirium: A Mediation Analysis.

BACKGROUND: Postoperative delirium is a common complication that hinders recovery after surgery. Intraoperative electroencephalogram suppression has been linked to postoperative delirium, but it is unknown if this relationship is causal or if electroencephalogram suppression is merely a marker of underlying cognitive abnormalities. The hypothesis of this study was that intraoperative electroencephalogram suppression mediates a nonzero portion of the effect between preoperative abnormal cognition and postoperative delirium. METHODS: This is a prespecified secondary analysis of the Electroencephalography Guidance of Anesthesia to Alleviate Geriatric Syndromes (ENGAGES) randomized trial, which enrolled patients age 60 yr or older undergoing surgery with general anesthesia at a single academic medical center between January 2015 and May 2018. Patients were randomized to electroencephalogram-guided anesthesia or usual care. Preoperative abnormal cognition was defined as a composite of previous delirium, Short Blessed Test cognitive score greater than 4 points, or Eight Item Interview to Differentiate Aging and Dementia score greater than 1 point. Duration of intraoperative electroencephalogram suppression was defined as number of minutes with suppression ratio greater than 1%. Postoperative delirium was detected via Confusion Assessment Method or chart review on postoperative days 1 to 5. RESULTS: Among 1,113 patients, 430 patients showed evidence of preoperative abnormal cognition. These patients had an increased incidence of postoperative delirium (151 of 430 [35%] vs.123 of 683 [18%], P < 0.001). Of this 17.2% total effect size (99.5% CI, 9.3 to 25.1%), an absolute 2.4% (99.5% CI, 0.6 to 4.8%) was an indirect effect mediated by electroencephalogram suppression, while an absolute 14.8% (99.5% CI, 7.2 to 22.5%) was a direct effect of preoperative abnormal cognition. Randomization to electroencephalogram-guided anesthesia did not change the mediated effect size (P = 0.078 for moderation). CONCLUSIONS: A small portion of the total effect of preoperative abnormal cognition on postoperative delirium was mediated by electroencephalogram suppression. Study precision was too low to determine if the intervention changed the mediated effect.

Practical Training of Anesthesia Clinicians in Electroencephalogram-Based Determination of Hypnotic Depth of General Anesthesia.

BACKGROUND: Electroencephalographic (EEG) brain monitoring during general anesthesia provides information on hypnotic depth. We hypothesized that anesthesia clinicians could be trained rapidly to recognize typical EEG waveforms occurring with volatile-based general anesthesia. METHODS: This was a substudy of a trial testing the hypothesis that EEG-guided anesthesia prevents postoperative delirium. The intervention was a 35-minute training session, summarizing typical EEG changes with volatile-based anesthesia. Participants completed a preeducational test, underwent training, and completed a posteducational test. For each question, participants indicated whether the EEG was consistent with (1) wakefulness, (2) non-slow-wave anesthesia, (3) slow-wave anesthesia, or (4) burst suppression. They also indicated whether the processed EEG (pEEG) index was discordant with the EEG waveforms. Four clinicians, experienced in intraoperative EEG interpretation, independently evaluated the EEG waveforms, resolved disagreements, and provided reference answers. Ten questions were assessed in the preeducational test and 9 in the posteducational test. RESULTS: There were 71 participants; 13 had previous anesthetic-associated EEG interpretation training. After training, the 58 participants without prior training improved at identifying dominant EEG waveforms (median 60% with interquartile range [IQR], 50%-70% vs 78% with IQR, 67%-89%; difference: 18%; 95% confidence interval [CI], 8-27; P < .001). In contrast, there was no significant improvement following the training for the 13 participants who reported previous training (median 70% with IQR, 60%-80% vs 67% with IQR, 67%-78%; difference: -3%; 95% CI, -18 to 11; P = .88). The difference in the change between the pre- and posteducational session for the previously untrained versus previously trained was statistically significant (difference in medians: 21%; 95% CI, 2-28; P = .005). Clinicians without prior training also improved in identifying discordance between the pEEG index and the EEG waveform (median 60% with IQR, 40%-60% vs median 100% with IQR, 75%-100%; difference: 40%; 95% CI, 30-50; P < .001). Clinicians with prior training showed no significant improvement (median 60% with IQR, 60%-80% vs 75% with IQR, 75%-100%; difference: 15%; 95% CI, -16 to 46; P = .16). Regarding the identification of discordance, the difference in the change between the pre- and posteducational session for the previously untrained versus previously trained was statistically significant (difference in medians: 25%; 95% CI, 5-45; P = .012). CONCLUSIONS: A brief training session was associated with improvements in clinicians without prior EEG training in (1) identifying EEG waveforms corresponding to different hypnotic depths and (2) recognizing when the hypnotic depth suggested by the EEG was discordant with the pEEG index.