A deep learning approach for generating intracranial pressure waveforms from extracranial signals routinely measured in the intensive care unit.

Intracranial pressure (ICP) is commonly monitored to guide treatment in patients with serious brain disorders such as traumatic brain injury and stroke. Established methods to assess ICP are resource intensive and highly invasive. We hypothesized that ICP waveforms can be computed noninvasively from three extracranial physiological waveforms routinely acquired in the Intensive Care Unit (ICU): arterial blood pressure (ABP), photoplethysmography (PPG), and electrocardiography (ECG). We evaluated over 600 h of high-frequency (125 Hz) simultaneously acquired ICP, ABP, ECG, and PPG waveform data in 10 patients admitted to the ICU with critical brain disorders. The data were segmented in non-overlapping 10-s windows, and ABP, ECG, and PPG waveforms were used to train deep learning (DL) models to re-create concurrent ICP. The predictive performance of six different DL models was evaluated in single- and multi-patient iterations. The mean average error (MAE) ± SD of the best-performing models was 1.34 ± 0.59 mmHg in the single-patient and 5.10 ± 0.11 mmHg in the multi-patient analysis. Ablation analysis was conducted to compare contributions from single physiologic sources and demonstrated statistically indistinguishable performances across the top DL models for each waveform (MAE±SD 6.33 ± 0.73, 6.65 ± 0.96, and 7.30 ± 1.28 mmHg, respectively, for ECG, PPG, and ABP; p = 0.42). Results support the preliminary feasibility and accuracy of DL-enabled continuous noninvasive ICP waveform computation using extracranial physiological waveforms. With refinement and further validation, this method could represent a safer and more accessible alternative to invasive ICP, enabling assessment and treatment in low-resource settings.

IntraCranial pressure prediction AlgoRithm using machinE learning (I-CARE): Training and Validation Study.

OBJECTIVES: Elevated intracranial pressure (ICP) is a potentially devastating complication of neurologic injury. Developing an ICP prediction algorithm to help the clinician adjust treatments and potentially prevent elevated ICP episodes. DESIGN: Retrospective study. SETTING: Three hundred thirty-five ICUs at 208 hospitals in the United States. SUBJECTS: Adults patients from the electronic ICU (eICU) Collaborative Research Database was used to train an ensemble machine learning model to predict the ICP 30 minutes in the future. Predictive performance was evaluated using a left-out test dataset and externally evaluated on the Medical Information Mart for Intensive Care-III (MIMIC-III) Matched Waveform Database. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Predictors included age, assigned sex, laboratories, medications and infusions, input/output, Glasgow Coma Scale (GCS) components, and time-series vitals (heart rate, ICP, mean arterial pressure, respiratory rate, and temperature). Each patient ICU stay was divided into successive 95-minute timeblocks. For each timeblock, the model was trained on nontime-varying covariates as well as on 12 observations of time-varying covariates at 5-minute intervals and asked to predict the 5-minute median ICP 30 minutes after the last observed ICP value. Data from 931 patients with ICP monitoring in the eICU dataset were extracted (46,207 timeblocks). The root mean squared error was 4.51 mm Hg in the eICU test set and 3.56 mm Hg in the MIMIC-III dataset. The most important variables driving ICP prediction were previous ICP history, patients' temperature, weight, serum creatinine, age, GCS, and hemodynamic parameters. CONCLUSIONS: IntraCranial pressure prediction AlgoRithm using machinE learning, an ensemble machine learning model, trained to predict the ICP of a patient 30 minutes in the future based on baseline characteristics and vitals data from the past hour showed promising predictive performance including in an external validation dataset.

Association between peripheral perfusion index and postoperative acute kidney injury in major noncardiac surgery patients receiving continuous vasopressors: a post hoc exploratory analysis of the VEGA-1 trial.

BACKGROUND: The peripheral perfusion index is the ratio of pulsatile to nonpulsatile static blood flow obtained by photoplethysmography and reflects peripheral tissue perfusion. We investigated the association between intraoperative perfusion index and postoperative acute kidney injury in patients undergoing major noncardiac surgery and receiving continuous vasopressor infusions. METHODS: In this exploratory post hoc analysis of a pragmatic, cluster-randomised, multicentre trial, we obtained areas and cumulative times under various thresholds of perfusion index and investigated their association with acute kidney injury in multivariable logistic regression analyses. In secondary analyses, we investigated the association of time-weighted average perfusion index with acute kidney injury. The 30-day mortality was a secondary outcome. RESULTS: Of 2534 cases included, 8.9% developed postoperative acute kidney injury. Areas and cumulative times under a perfusion index of 3% and 2% were associated with an increased risk of acute kidney injury; the strongest association was observed for area under a perfusion index of 1% (adjusted odds ratio [aOR] 1.32, 95% confidence interval [CI] 1.00-1.74, P=0.050, per 100%∗min increase). Additionally, time-weighted average perfusion index was associated with acute kidney injury (aOR 0.82, 95% CI 0.74-0.91, P<0.001) and 30-day mortality (aOR 0.68, 95% CI 0.49-0.95, P=0.024). CONCLUSIONS: Larger areas and longer cumulative times under thresholds of perfusion index and lower time-weighted average perfusion index were associated with postoperative acute kidney injury in patients undergoing major noncardiac surgery and receiving continuous vasopressor infusions. CLINICAL TRIAL REGISTRATION: NCT04789330.

Negativity and Positivity in the ICU: Exploratory Development of Automated Sentiment Capture in the Electronic Health Record.

OBJECTIVES: To develop proof-of-concept algorithms using alternative approaches to capture provider sentiment in ICU notes. DESIGN: Retrospective observational cohort study. SETTING: The Multiparameter Intelligent Monitoring of Intensive Care III (MIMIC-III) and the University of California, San Francisco (UCSF) deidentified notes databases. PATIENTS: Adult (≥18 yr old) patients admitted to the ICU. MEASUREMENTS AND MAIN RESULTS: We developed two sentiment models: 1) a keywords-based approach using a consensus-based clinical sentiment lexicon comprised of 72 positive and 103 negative phrases, including negations and 2) a Decoding-enhanced Bidirectional Encoder Representations from Transformers with disentangled attention-v3-based deep learning model (keywords-independent) trained on clinical sentiment labels. We applied the models to 198,944 notes across 52,997 ICU admissions in the MIMIC-III database. Analyses were replicated on an external sample of patients admitted to a UCSF ICU from 2018 to 2019. We also labeled sentiment in 1,493 note fragments and compared the predictive accuracy of our tools to three popular sentiment classifiers. Clinical sentiment terms were found in 99% of patient visits across 88% of notes. Our two sentiment tools were substantially more predictive (Spearman correlations of 0.62-0.84, p values < 0.00001) of labeled sentiment compared with general language algorithms (0.28-0.46). CONCLUSION: Our exploratory healthcare-specific sentiment models can more accurately detect positivity and negativity in clinical notes compared with general sentiment tools not designed for clinical usage.

Development of a core outcome set for ventilation trials in neurocritical care patients with acute brain injury: protocol for a Delphi consensus study of international stakeholders.

INTRODUCTION: There is little consensus and high heterogeneity on the optimal set of relevant clinical outcomes in research studies regarding extubation in neurocritical care patients with brain injury undergoing mechanical ventilation. The aims of this study are to: (1) develop a core outcome set (COS) and (2) reach consensus on a hierarchical composite endpoint for such studies. METHODS AND ANALYSIS: The study will include a broadly representative, international panel of stakeholders with research and clinical expertise in this field and will involve four stages: (1) a scoping review to generate an initial list of outcomes represented in the literature, (2) an investigator meeting to review the outcomes for inclusion in the Delphi surveys, (3) four rounds of online Delphi consensus-building surveys and (4) online consensus meetings to finalise the COS and hierarchical composite endpoint. ETHICS AND DISSEMINATION: This study received ethical approval from the French Society of Anesthesia and Critical Care Medicine Institutional Review Board (SFAR CERAR-IRB 00010254-2023-029). The study results will be disseminated through communication to stakeholders, publication in a peer-reviewed journal, and presentations at conferences. TRIAL REGISTRATION NUMBER: This study is registered with the Core Outcome Measures in Effectiveness Trials (COMET) Initiative.

Development of a Machine Learning Model of Postoperative Acute Kidney Injury Using Non-Invasive Time-Sensitive Intraoperative Predictors.

Acute kidney injury (AKI) is a major postoperative complication that lacks established intraoperative predictors. Our objective was to develop a prediction model using preoperative and high-frequency intraoperative data for postoperative AKI. In this retrospective cohort study, we evaluated 77,428 operative cases at a single academic center between 2016 and 2022. A total of 11,212 cases with serum creatinine (sCr) data were included in the analysis. Then, 8519 cases were randomly assigned to the training set and the remainder to the validation set. Fourteen preoperative and twenty intraoperative variables were evaluated using elastic net followed by hierarchical group least absolute shrinkage and selection operator (LASSO) regression. The training set was 56% male and had a median [IQR] age of 62 (51-72) and a 6% AKI rate. Retained model variables were preoperative sCr values, the number of minutes meeting cutoffs for urine output, heart rate, perfusion index intraoperatively, and the total estimated blood loss. The area under the receiver operator characteristic curve was 0.81 (95% CI, 0.77-0.85). At a score threshold of 0.767, specificity was 77% and sensitivity was 74%. A web application that calculates the model score is available online. Our findings demonstrate the utility of intraoperative time series data for prediction problems, including a new potential use of the perfusion index. Further research is needed to evaluate the model in clinical settings.

Pulse contour techniques for perioperative hemodynamic monitoring: A nationwide carbon footprint and cost estimation.

BACKGROUND: The question of environmentally sustainable perioperative medicine represents a new challenge in an era of cost constraints and climate crisis. The French Society of Anaesthesia and Intensive Care (SFAR) recommends stroke volume optimization in high-risk surgical patients. Pulse contour techniques have become increasingly popular for stroke volume monitoring during surgery. Some require the use of specific disposable pressure transducers (DPTs), whereas others can be used with standard DPTs. OBJECTIVE: Quantify and compare the carbon footprint and cost of pulse contour techniques using specific and standard DPTs on a yearly basis and at a national level. METHODS: We estimated the number of high-risk surgical patients monitored every year in France with a pulse contour technique, and the plastic waste, carbon footprint and cost associated with the use of specific and standard DPTs. MAIN FINDINGS: When compared to pulse contour techniques working with a standard DPT, techniques requiring a specific DPT are responsible for an increase in carbon dioxide emission estimated at 65-83 tons/yr and for additional hospital cost estimated at €67 million/yr. If, as recommended by the SFAR, all high-risk surgical patients were monitored, the difference would reach 179-227 tons/yr for the environmental impact and €187 million/yr for the economic impact. CONCLUSION: From an environmental and economic standpoint, pulse contour techniques working with standard DPTs should be recommended for the perioperative hemodynamic monitoring of high-risk surgical patients.

Leveraging observational data to identify targeted patient populations for future randomized trials.

Randomized controlled trials reported in the literature are often affected by poor generalizability, and pragmatic trials have become an increasingly utilized workaround approach to overcome logistical limitations and explore routine interventions demonstrating equipoise in clinical practice. Intravenous albumin, for example, is commonly administered in the perioperative setting despite lacking supportive evidence. Given concerns for cost, safety, and efficacy, randomized trials are needed to explore the clinical equipoise of albumin therapy in this setting, and we therefore present an approach to identifying populations exposed to perioperative albumin to encourage clinical equipoise in patient selection and optimize study design for clinical trials.

Availability of information needed to evaluate algorithmic fairness - A systematic review of publicly accessible critical care databases.

BACKGROUND: Machine learning (ML) may improve clinical decision-making in critical care settings, but intrinsic biases in datasets can introduce bias into predictive models. This study aims to determine if publicly available critical care datasets provide relevant information to identify historically marginalized populations. METHOD: We conducted a review to identify the manuscripts that report the training/validation of ML algorithms using publicly accessible critical care electronic medical record (EMR) datasets. The datasets were reviewed to determine if the following 12 variables were available: age, sex, gender identity, race and/or ethnicity, self-identification as an indigenous person, payor, primary language, religion, place of residence, education, occupation, and income. RESULTS: 7 publicly available databases were identified. Medical Information Mart for Intensive Care (MIMIC) reports information on 7 of the 12 variables of interest, Sistema de Informação de Vigilância Epidemiológica da Gripe (SIVEP-Gripe) on 7, COVID-19 Mexican Open Repository on 4, and eICU on 4. Other datasets report information on 2 or fewer variables. All 7 databases included information about sex and age. Four databases (57%) included information about whether a patient identified as native or indigenous. Only 3 (43%) included data about race and/or ethnicity. Two databases (29%) included information about residence, and one (14%) included information about payor, language, and religion. One database (14%) included information about education and patient occupation. No databases included information on gender identity and income. CONCLUSION: This review demonstrates that critical care publicly available data used to train AI algorithms do not include enough information to properly look for intrinsic bias and fairness issues towards historically marginalized populations.

Intensive Care Unit activity in France from the national database between 2013 and 2019: More critically ill patients, shorter stay and lower mortality rate.

BACKGROUND: Knowledge of the occurrence and outcome of admissions to Intensive Care Units (ICU) over time is important to inform healthcare services planning. This observational study aims at describing the activity of French ICUs between 2013 and 2019. METHODS: Patient admission characteristics, organ dysfunction scores, therapies, ICU and hospital lengths of stay and case fatality were collected from the French National Hospital Database (population-based cohort). Logistic regression models were developed to investigate the association between age, sex, SAPS II, organ failure, and year of care on in-ICU case fatality. FINDINGS: Among 1,594,801 ICU admissions, the yearly ICU admission increased from 3.3 to 3.5 per year per 1000 inhabitants (bed occupancy rate between 83.4 and 84.3%). The mean admission SAPS II was 42 ± 22, with a gradual annual increase. The median lengths of stay in ICU and in hospital were 3 (interquartile range (IQR) = [1-7]) and 11 days (IQR = [6-21]), respectively, with a progressive decrease over time. The in-ICU and hospital mortality case fatalities decreased from 18.0% to 17.1% and from 21.1% to 19.9% between 2013 and 2019, respectively. Male sex, age, SAPS II score, and the occurrence of any organ failure were associated with a higher case fatality rate. After adjustment on age, sex, SAPS II and organ failure, in-ICU case fatality decreased in 2019 as compared to 2013 (adjusted Odds Ratio = 0.87 [95% confidence interval, 0.85-0.89]). INTERPRETATION: During the study, an increasing incidence of ICU admission was associated with higher severity of illness but lower in-ICU case fatality.

Norepinephrine versus phenylephrine for treating hypotension during general anaesthesia in adult patients undergoing major noncardiac surgery: a multicentre, open-label, cluster-randomised, crossover, feasibility, and pilot trial.

BACKGROUND: Intraoperative hypotension is associated with postoperative complications. The use of vasopressors is often required to correct hypotension but the best vasopressor is unknown. METHODS: A multicentre, cluster-randomised, crossover, feasibility and pilot trial was conducted across five hospitals in California. Phenylephrine (PE) vs norepinephrine (NE) infusion as the first-line vasopressor in patients under general anaesthesia alternated monthly at each hospital for 6 months. The primary endpoint was first-line vasopressor administration compliance of 80% or higher. Secondary endpoints were acute kidney injury (AKI), 30-day mortality, myocardial injury after noncardiac surgery (MINS), hospital length of stay, and rehospitalisation within 30 days. RESULTS: A total of 3626 patients were enrolled over 6 months; 1809 patients were randomised in the NE group, 1817 in the PE group. Overall, 88.2% received the assigned first-line vasopressor. No drug infiltrations requiring treatment were reported in either group. Patients were median 63 yr old, 50% female, and 58% white. Randomisation in the NE group vs PE group did not reduce readmission within 30 days (adjusted odds ratio=0.92; 95% confidence interval, 0.6-1.39), 30-day mortality (1.01; 0.48-2.09), AKI (1.1; 0.92-1.31), or MINS (1.63; 0.84-3.16). CONCLUSIONS: A large and diverse population undergoing major surgery under general anaesthesia was successfully enrolled and randomised to receive NE or PE infusion. This pilot and feasibility trial was not powered for adverse postoperative outcomes and a follow-up multicentre effectiveness trial is planned. CLINICAL TRIAL REGISTRATION: NCT04789330 (ClinicalTrials.gov).

Personalized online ensemble machine learning with applications for dynamic data streams.

In this work we introduce the personalized online super learner (POSL), an online personalizable ensemble machine learning algorithm for streaming data. POSL optimizes predictions with respect to baseline covariates, so personalization can vary from completely individualized, that is, optimization with respect to subject ID, to many individuals, that is, optimization with respect to common baseline covariates. As an online algorithm, POSL learns in real time. As a super learner, POSL is grounded in statistical optimality theory and can leverage a diversity of candidate algorithms, including online algorithms with different training and update times, fixed/offline algorithms that are not updated during POSL's fitting procedure, pooled algorithms that learn from many individuals' time series, and individualized algorithms that learn from within a single time series. POSL's ensembling of the candidates can depend on the amount of data collected, the stationarity of the time series, and the mutual characteristics of a group of time series. Depending on the underlying data-generating process and the information available in the data, POSL is able to adapt to learning across samples, through time, or both. For a range of simulations that reflect realistic forecasting scenarios and in a medical application, we examine the performance of POSL relative to other current ensembling and online learning methods. We show that POSL is able to provide reliable predictions for both short and long time series, and it's able to adjust to changing data-generating environments. We further cultivate POSL's practicality by extending it to settings where time series dynamically enter and exit.