ICU Resources Research Articles

Helmet c-PAP Versus NIPPV in Association with Early Respiratory Physiotherapy and Mobilization for Treating SARS-CoV-2 ARDS: A Case-control Prospective Single-center Study.

Introduction: Early noninvasive respiratory support (NIRS) is correlated with a success rate of 60-75% in patients experiencing SARS-CoV-2 ARDS. We conducted a prospective case-control study to assess differences in outcomes between Helmet c-PAP (H-c-PAP) and noninvasive positive pressure ventilation (NIPPV). Methods: All patients with SARS-CoV-2 ARDS, treated with H-c-PAP or NIPPV between October 2021 and April 2022 were sampled. We recorded: demographics, comorbidities, clinical, respiratory, sepsis, NIRS parameters, and outcomes. A "NIRS team" followed the patients in respiratory support supplying them with early and timely intensive physiotherapy i-PKT as well. The Cox's proportional hazard model was applied for multivariate analyses. Results: 368 patients were admitted to our hospital medical ward. 85 patients were treated with H-c-PAP and 145 underwent NIPPV. 138 patients needing oxygen supplementation alone were excluded. The two groups were homogeneously distributed and ICU admission rates were lower in the H-c-PAP one (9.4 vs 11% P = .001) while mortality was higher in the NIPPV group (22.7 vs 9.4%, P = .001). The two multivariate models, that had overall mortality as primary outcome, identified age, H-c-PAP daily, i-PKT and ICU admission as independent variables impacting on the outcome. Age was no longer a significant independent predictor after the inclusion of elderly patients (age >80). The third model showed daily i-PKT could prevent ICU admission whereas the length of NIRS was inversely proportional to outcome. Conclusions: A "NIRS multidisciplinary team" made it possible to adopt an early and timely combination of NIRS and i-PKT resulting in the saving of both patient lives and ICU resources.

Epidemiology of Emergency Department Visits for Children With Clinically Significant Cardiovascular Disease.

The aim of the study is to determine the epidemiology, cost, and factors associated with hospital admission, deterioration if hospitalized, and mortality for children with a history of clinically significant cardiovascular disease (CVD) presenting to pediatric emergency departments (EDs). Using the Pediatric Health Information System, we performed a retrospective analysis of ED encounters of children ≤17 years old with clinically significant CVD between 2016 and 2021. Patients were included if they had a cardiovascular complex chronic condition, defined by ICD diagnosis, and procedure codes. We assessed the primary diagnosis, admission rate, ICU transfer rate (as a marker of disease progression), mortality, resource utilization, and costs. We conducted multivariable analyses to identify risk factors for admission, ICU transfer, and mortality. There were 201,551 ED visits (mean 33,592 ± 3354 per year) among 129,938 children with clinically significant CVD. Most ED encounters had a primary diagnosis of a circulatory (21.1%) or respiratory (19.7%) illness. Seventy-six percent of visits had at least one blood test or imaging study conducted. The overall admission rate was 59.7%, with 28.7% admitted to the ICU, and 6.2% transferred to the ICU after the first 24 hours. The median costs for encounters resulting in admission were $13,605 in US 2023 dollars. In multivariable analyses, younger age, a greater number of noncardiac complex chronic conditions, and CVD type were associated with increased odds of admission, ICU transfer after 24 hours, and mortality (all P < 0.05). ED visits for children with clinically significant CVD lead to substantial resource utilization, including frequent hospitalization, ICU level of care, and costs. This baseline data aids in the development of prospective studies to inform the appropriate ED management for children with clinically significant CVD.

Routine ICU Surveillance after Brain Tumor Surgery: Patient Selection Using Machine Learning.

Background/Objectives: Routine postoperative ICU admission following brain tumor surgery may not benefit selected patients. The objective of this study was to develop a risk prediction instrument for early (within 24 h) postoperative adverse events using machine learning techniques. Methods: Retrospective cohort of 1000 consecutive adult patients undergoing elective brain tumor resection. Nine events/interventions (CPR, reintubation, return to OR, mechanical ventilation, vasopressors, impaired consciousness, intracranial hypertension, swallowing disorders, and death) were chosen as target variables. Potential prognostic features (n = 27) from five categories were chosen and a gradient boosting algorithm (XGBoost) was trained and cross-validated in a 5 × 5 fashion. Prognostic performance, potential clinical impact, and relative feature importance were analyzed. Results: Adverse events requiring ICU intervention occurred in 9.2% of cases. Other events not requiring ICU treatment were more frequent (35% of cases). The boosted decision trees yielded a cross-validated ROC-AUC of 0.81 ± 0.02 (mean ± CI95) when using pre- and post-op data. Using only pre-op data (scheduling decisions), ROC-AUC was 0.76 ± 0.02. PR-AUC was 0.38 ± 0.04 and 0.27 ± 0.03 for pre- and post-op data, respectively, compared to a baseline value (random classifier) of 0.092. Targeting a NPV of at least 95% would require ICU admission in just 15% (pre- and post-op data) or 30% (only pre-op data) of cases when using the prediction algorithm. Conclusions: Adoption of a risk prediction instrument based on boosted trees can support decision-makers to optimize ICU resource utilization while maintaining adequate patient safety. This may lead to a relevant reduction in ICU admissions for surveillance purposes.

Characteristics and Outcomes of Home-Ventilated Children Undergoing Noncardiac Surgery.

To determine the frequency of children with chronic respiratory failure (CRF) and home ventilator dependence undergoing surgery at a tertiary children's hospital, and to describe periprocedural characteristics and outcomes. We conducted a retrospective cohort study of patients with CRF and home ventilator dependence who underwent noncardiac surgery from January 1, 2013, to December 31, 2019. Descriptive statistics were used to report patient and procedural characteristics. Univariable and multivariable analyses were used to assess for factors associated with 30-day readmission. We identified 416 patients who underwent 1623 procedures. Fifty-one percent of patients used transtracheal mechanical ventilation (trach/vent) support at the time of surgery; this cohort was younger (median age 5.5 vs 10.8 years) and more complex according to American Society of Anesthesiologists status compared with bilevel positive airway pressure-dependent patients. Postoperatively, compared with bilevel positive airway pressure-dependent patients, trach/vent patients were more likely to be admitted to the ICU with longer ICU length of stay (median 5 vs 2 days). Overall 30-day readmission rate was 12% (n = 193). Presence of chronic lung disease (adjusted odds ratio 1.65, 95% confidence interval 1.01-1.69) and trach/vent dependence (adjusted odds ratio 1.65, 95% confidence interval 1.02-2.67) were independently associated with increased odds for readmission. Children with CRF use anesthetic and surgical services frequently and repeatedly. Those with trach/vent dependence have higher hospital and ICU resource utilization. Although overall mortality for these patients is quite low, underlying diagnoses, nuances of technology dependence, and other factors for frequent readmission require further study to optimize resource utilization and outcomes.

Prediction of 30-day mortality for ICU patients with Sepsis-3

BackgroundThere is a growing demand for advanced methods to improve the understanding and prediction of illnesses. This study focuses on Sepsis, a critical response to infection, aiming to enhance early detection and mortality prediction for Sepsis-3 patients to improve hospital resource allocation.MethodsIn this study, we developed a Machine Learning (ML) framework to predict the 30-day mortality rate of ICU patients with Sepsis-3 using the MIMIC-III database. Advanced big data extraction tools like Snowflake were used to identify eligible patients. Decision tree models and Entropy Analyses helped refine feature selection, resulting in 30 relevant features curated with clinical experts. We employed the Light Gradient Boosting Machine (LightGBM) model for its efficiency and predictive power.ResultsThe study comprised a cohort of 9118 Sepsis-3 patients. Our preprocessing techniques significantly improved both the AUC and accuracy metrics. The LightGBM model achieved an impressive AUC of 0.983 (95% CI: [0.980–0.990]), an accuracy of 0.966, and an F1-score of 0.910. Notably, LightGBM showed a substantial 6% improvement over our best baseline model and a 14% enhancement over the best existing literature. These advancements are attributed to (I) the inclusion of the novel and pivotal feature Hospital Length of Stay (HOSP_LOS), absent in previous studies, and (II) LightGBM’s gradient boosting architecture, enabling robust predictions with high-dimensional data while maintaining computational efficiency, as demonstrated by its learning curve.ConclusionsOur preprocessing methodology reduced the number of relevant features and identified a crucial feature overlooked in previous studies. The proposed model demonstrated high predictive power and generalization capability, highlighting the potential of ML in ICU settings. This model can streamline ICU resource allocation and provide tailored interventions for Sepsis-3 patients.

A retrospective validation of the effective and safe treatment of patients on general care wards with high velocity nasal insufflation therapy utilizing prognostic risk scores during COVID-19

Background Non-invasive positive pressure ventilation (NIPPV) has been a traditional therapy for acute respiratory failure (ARF). However, the use of NIPPV during the COVID-19 pandemic was challenging, while the use of invasive mechanical ventilation produced poor outcomes. An alternative to NIPPV, high velocity nasal insufflation (HVNI) has shown promise in treating ARF effectively. Objective This study evaluated whether HVNI can be used to treat ARF safely on the general care ward (GCW) during COVID-19 pandemic surges. Methods After introducing HVNI therapy to the facility, an evidence-based scoring system, Modified Early Warning Score (MEWS), was used to risk stratify patients and assist in assigning care level. Initial settings, demographic data, patient outcomes, and health care worker (HCW) virus conversion were measured throughout the study. Treatment failure was defined as the need for invasive mechanical ventilation (IMV) or NIPPV after HVNI therapy. MEWS and ROX index were compared retrospectively using the Pearson product-moment correlation coefficient to identify trends. The Welch two sample t-test (desired power of 90% with alpha=0.05) was used for demographic and outcome analysis. Results Two hundred thirty-four patients were treated with HVNI. The GCW failure rate of 18.56% (n=31/167) was lower than the ICU failure rate of 37.31% (n=25/67) but not statistically significant (P 0.175). No elevated risk to patients or HCW was observed. Respiratory rate (GCW 24.85 vs. ICU 30.14; P &lt;0.001), MEWS (GCW 2.34 vs. ICU 3.09; p=0.002), and ROX index (GCW 5.49 vs. ICU 4.68; P 0.002) assessments appear to be adequate predictors of HVNI failure. The Pearson product-moment coefficient comparing MEWS and ROX index identified a moderate negative correlation (-0.434; P &lt;0.001). Discussion HVNI therapy is an effective alternative to NIPPV for treating patients with COVID-19 associated ARF. Using measures such as MEWS and/or ROX, strict patient monitoring, and HCW surveillance, HVNI can be safely utilized on the GCW. This has a direct impact when dealing with patient surges where ICU beds and resources are limited. Additional studies are needed to further delineate these concepts. Keywords: COVID-19, Non-invasive Positive Pressure Ventilation, Acute Respiratory Failure, High Flow Nasal Oxygen, High Velocity Nasal Insufflation, ICU Admission, General Care Ward, MEWS, ROX

Abstract 114: LACROSSE Trial: Lumbar Drainage Compared to External Ventricular Drainage in Aneurysmal Subarachnoid Hemorrhage, a Randomized Control Trial

Introduction: Delayed cerebral ischemia (DCI) is a leading cause of morbidity and mortality in aneurysmal subarachnoid hemorrhage (aSAH). External ventricular drainage (EVD) has been regarded as the gold standard, but lumbar drains (LD) may impact the rate of DCI due to preferential clearing of subarachnoid blood. Recent trials have not directly compared these drains therefore our goal was to determine if LD is superior to EVD for prevention of DCI in aSAH. Methods: A randomized controlled trial was initiated at UC San Diego, the University of Rochester, and Rochester Regional to randomize drain placement to either EVD or LD in aSAH patients. The primary outcome was DCI, defined as an exclusionary change in GCS or new, none-treatment related infarcts on imaging. Secondary outcomes were ICU measures and functional outcomes. Results: 52 patients were enrolled. The rate of DCI was significantly lower in the LD arm (7.4%, 2/27) compared to the EVD arm (36%, 9/25) (p=0.017) with a NNT of 3.5 for LD. There was no significant difference between the two arms with regards to admission characteristics, aneurysm location, treatment type, or total number of vasospasm treatments. LD patients spent fewer days in the ICU (p=0.015), on a ventilator (p=0.004) and had fewer drain days (p=0.003) compared to EVD patients. The proportion of patients with functional independence was significantly higher in the LD group compared to the EVD group 1-month post-hemorrhage (BI p = 0.04; GOS p = 0.04) Conclusions: In this trial, we show that LD is superior to EVD in aSAH patients in the reduction of DCI, is associated with significantly improved clinical outcomes, and reduced utilization of ICU resources.

Predicting ICU readmission risks in intracerebral hemorrhage patients: Insights from machine learning models using MIMIC databases

BackgroundIntracerebral hemorrhage (ICH) is a stroke subtype characterized by high mortality and complex post-event complications. Research has extensively covered the acute phase of ICH; however, ICU readmission determinants remain less explored. Utilizing the MIMIC-III and MIMIC-IV databases, this investigation develops machine learning (ML) models to anticipate ICU readmissions in ICH patients. MethodsRetrospective data from 2242 ICH patients were evaluated using ICD-9 codes. Recursive feature elimination with cross-validation (RFECV) discerned significant predictors of ICU readmissions. Four ML models—AdaBoost, RandomForest, LightGBM, and XGBoost—underwent development and rigorous validation. SHapley Additive exPlanations (SHAP) elucidated the effect of distinct features on model outcomes. ResultsICU readmission rates were 9.6% for MIMIC-III and 10.6% for MIMIC-IV. The LightGBM model, with an AUC of 0.736 (95% CI: 0.668–0.801), surpassed other models in validation datasets. SHAP analysis revealed hydrocephalus, sex, neutrophils, Glasgow Coma Scale (GCS), specific oxygen saturation (SpO2) levels, and creatinine as significant predictors of readmission. ConclusionThe LightGBM model demonstrates considerable potential in predicting ICU readmissions for ICH patients, highlighting the importance of certain clinical predictors. This research contributes to optimizing patient care and ICU resource management. Further prospective studies are warranted to corroborate and enhance these predictive insights for clinical utilization.

Measuring the Impact of ICU Strain on Mortality, After-Hours Discharge, Discharge Delay, Interhospital Transfer, and Readmission in Australia With the Activity Index.

ICU resource strain leads to adverse patient outcomes. Simple, well-validated measures of ICU strain are lacking. Our objective was to assess whether the "Activity index," an indicator developed during the COVID-19 pandemic, was a valid measure of ICU strain. Retrospective national registry-based cohort study. One hundred seventy-five public and private hospitals in Australia (June 2020 through March 2022). Two hundred seventy-seven thousand seven hundred thirty-seven adult ICU patients. None. Data from the Australian and New Zealand Intensive Care Society Adult Patient Database were matched to the Critical Health Resources Information System. The mean daily Activity index of each ICU (census total of "patients with 1:1 nursing" + "invasive ventilation" + "renal replacement" + "extracorporeal membrane oxygenation" + "active COVID-19," divided by total staffed ICU beds) during the patient's stay in the ICU was calculated. Patients were categorized as being in the ICU during very quiet (Activity index < 0.1), quiet (0.1 to < 0.6), intermediate (0.6 to < 1.1), busy (1.1 to < 1.6), or very busy time-periods (≥ 1.6). The primary outcome was in-hospital mortality. Secondary outcomes included after-hours discharge from the ICU, readmission to the ICU, interhospital transfer to another ICU, and delay in discharge from the ICU. Median Activity index was 0.87 (interquartile range, 0.40-1.24). Nineteen thousand one hundred seventy-seven patients died (6.9%). In-hospital mortality ranged from 2.4% during very quiet to 10.9% during very busy time-periods. After adjusting for confounders, being in an ICU during time-periods with higher Activity indices, was associated with an increased risk of in-hospital mortality (odds ratio [OR], 1.49; 99% CI, 1.38-1.60), after-hours discharge (OR, 1.27; 99% CI, 1.21-1.34), readmission (OR, 1.18; 99% CI, 1.09-1.28), interhospital transfer (OR, 1.92; 99% CI, 1.72-2.15), and less delay in ICU discharge (OR, 0.58; 99% CI, 0.55-0.62): findings consistent with ICU strain. The Activity index is a simple and valid measure that identifies ICUs in which increasing strain leads to progressively worse patient outcomes.

Intensive care unit admissions with and without COVID-19 in Finland from 2017 to 2021: a retrospective register-based study

BackgroundAfter the COVID-19 pandemic started, critical care resources were expanded in Finland to manage a possible surge in patients requiring intensive care. The aim of this study was to evaluate the incidence of overall ICU admissions, patient diagnoses, characteristics, and length of stay during the pandemic.MethodsThis retrospective hospital register-based study was conducted in two large and one mid-size Finnish public hospitals. The required data were collected from ICU patient information systems and all adult patients were included. Monthly and yearly incidences with 95% confidence intervals (CI) were counted per 100 000 persons-years by Poisson exact method and compared by incidence rate ratios (IRR).ResultsA total of 4407 admissions to ICUs for any cause occurred during 2020. In 2021, this figure was 4931. During the reference years (2017–2019), the mean number of admissions to ICU was 4781. In 2020 and 2021, the proportion of patients requiring intensive care due to COVID-19 was only 3%. The incidence of all-cause ICU admissions decreased during the lockdown in 2020 when compared to the reference years. Before the start of the lockdown in February 2020, the IRR of all-cause ICU admissions was 1.02 (CI: 0.89 to 1.18). During the lockdown period, however, the IRR of all-cause ICU admissions decreased to 0.78 (CI: 0.67 to 0.90) in March. When the lockdown ended, the incidence rebounded to the same level as before the lockdown. However, in 2021, the incidence of ICU admissions remained at the same level when compared to the reference years. The most prominent changes occurred in the incidence of diseases of the nervous system, which includes epilepsy and seizures and transient cerebral ischemic attacks, in diseases of the respiratory system, and neoplasms.ConclusionsAccording to the findings of this study, the incidence of all-cause ICU admissions decreased after the lockdown was implemented in 2020. Furthermore, the percentage of patients requiring intensive care due to COVID-19 in Finland was only 3% in 2020 and 2021. These findings may serve to help in the planning and allocating of ICU resources during future pandemics.

Comparison of clinical characteristics and healthcare resource use of pediatric chronic and non-chronic critically ill patients in intensive care units: a retrospective national registry study.

Chronic critically ill patients (CCI) in pediatric intensive care unit (PICU) are at risk of negative health outcomes, and account for a considerable amount of ICU resources. This study aimed to (a) describe the prevalence of CCI children, (b) compare their clinical characteristics and ICU resources use with non-CCI children, and (c) identify associated risk factors of CCI. A retrospective national registry study including 2015-2017 data from the eight Swiss PICUs of five tertiary and three regional hospitals, admitting a broad case-mix of medical and surgical patients, including pre- and full-term infants. CCI patients were identified using an adapted definition: PICU length of stay (LOS) ≥8 days and dependence on ≥1 PICU technology. Out of the 12,375 PICU admissions, 982 (8%) were CCI children and compared to non-CCI children, they were younger (2.8 vs. 6.7 months), had more cardiac conditions (24% vs. 12%), and higher mortality rate (7% vs. 2%) (p < 0.001). Nursing workload was higher in the CCI compared to the non-CCI group (22 [17-27]; 21 [16-26] respectively p < 0.001). Factors associated with CCI were cardiac (aOR = 2.241) and neurological diagnosis (aOR = 2.062), surgery (aORs between 1.662 and 2.391), ventilation support (aOR = 2.278), high mortality risk (aOR = 1.074) and agitation (aOR = 1.867). the results confirm the clinical vulnerability and the complexity of care of CCI children as they were defined in our study. Early identification and adequate staffing is required to provide appropriate and good quality care.

Toward future triage regulations: Investigating preferred allocation principles of the German public

BackgroundWhen intensive care capacity is limited, triage may be required. Given that the German government started working on new triage legislation in 2022, the present study investigated the German public's preferences for intensive care allocation in two situations: ex-ante triage (where multiple patients compete for available resources) and ex-post triage (where admitting a new patient to intensive care means withdrawing treatment from another because ICU resources are depleted). MethodIn an online experiment, N=994 participants were presented with four fictitious patients who differed in age and pre- and post-treatment chance of survival. In a series of pairwise comparisons, participants were asked to select one patient for treatment or to opt for random selection. Ex-ante and ex-post triage situations were varied between participants and preferred allocation strategies were inferred from their decisions. ResultsOn average, participants prioritized better post-treatment prognosis ahead of younger age or treatment benefit. Many participants rejected random allocation (on the flip of a coin) or prioritization by worse pre-treatment prognosis. Preferences were similar for ex-ante and ex-post situations. DiscussionAlthough there may be good reasons for deviating from laypeople's preference for utilitarian allocation, the results can help to design future triage policies and accompanying communication strategies.

Trends of using immune checkpoint inhibitors near end of life in limited resources setting: Experience of single cancer center in the MENA region.

6566 Background: Immune checkpoint inhibitors (ICIs) are standard of care nowadays in the treatment of many solid tumors as single agent or in combination. Recent literature suggested increased use of ICI near end of life; our aim was to find the characteristics of the patients who received ICI 30 days or less before their death. Methods: Data of patients with advanced solid tumors who have (ICI) therapy and died with disease were retrospectively retrieved. Patients were categorized into those dying within 30 days of the last dose of ICI and those receiving the last dose of ICI more than 30 days before death. Variables of potential interest were compared between the two groups by the Chi-square and independent sample t-tests as appropriate. Results: A total of 411 patients were included in this analysis; 135 (33%) received the last dose of ICI within 30 days of death and 276 (67%) received their last dose more than 30 days before death. Major cancer types included: lung (n=214; 52%), kidney (n=61; 15%), melanoma (n=44; 11%), head and neck (n=41; 10%), and others (n=14; 3%). Patients’ characteristics: median age = 58, majority were males (310; 75%), 105 (26%) had ECOG PS ≥3, and 147 (36%) had brain metastasis. Patients who received the last dose of ICI within 30 days of death had more frequent ICU admissions (50 (37%) vs. 60 (22%), p=0.001, and were less likely to die under the care of the hospice care service (22 patients (16%) vs. 84 (31%), p=0.001. In addition, patients who received the last dose of ICI within 30 days of death were more likely to have a full code status (89 (66%) vs. 147 (53%), p=0.015, and have received a lower number of doses of ICI treatments (mean number of doses: 4.2 vs. 6.5 doses, p=0.001). We did not observe any significant difference between the two groups regarding type of cancer, age, gender, ECOG PS, presence of brain metastasis, number of hospital admissions, dying inside hospital, and number of emergency room visits. Conclusions: Immune checkpoint inhibitors may be overused near end of life which leads to minimal clinical benefits, less likelihood of dying under hospice care, and utilization of ICU resources.

Association Between Vaccination Status and Outcomes in Patients Admitted to the ICU With COVID-19.

Although COVID-19 vaccines can reduce the need for intensive care unit admission in COVID-19, their effect on outcomes in critical illness remains unclear. We evaluated outcomes in vaccinated patients admitted to the ICU with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infections and the association between vaccination and booster status on clinical outcomes. Retrospective cohort. All patients were admitted to an ICU between January 2021 (after vaccination was available) and July 2022 with a diagnosis of COVID-19 based on a SARS-CoV-2 polymerase chain reaction test in Alberta, Canada. None. The propensity-matched primary outcome of all-cause in-hospital mortality was compared between vaccinated and unvaccinated patients, and vaccinated patients were stratified by booster dosing. Secondary outcomes were mechanical ventilation (MV) duration ICU length of stay (LOS). The study included 3,293 patients: 743 (22.6%) were fully vaccinated (54.6% with booster), 166 (5.0%) were partially vaccinated, and 2,384 (72.4%) were unvaccinated. Unvaccinated patients were more likely to require invasive MV (78.4% vs 68.2%), vasopressor use (71.1% vs 66.6%), and extracorporeal membrane oxygenation (2.1% vs 0.5%). In a propensity-matched analysis, in-hospital mortality was similar (31.8% vs 34.0%, adjusted odds ratio [OR], 1.25; 95% CI, 0.97-1.61), but median duration MV (7.6 vs 4.7 d; p < 0.001) and ICU LOS (6.6 vs 5.2 d; p < 0.001) were longer in unvaccinated compared to fully vaccinated patients. Among vaccinated patients, greater than or equal to 1 booster had lower in-hospital mortality (25.5% vs 40.9%; adjusted OR, 0.50; 95% CI, 0.0.36-0.68) and duration of MV (3.8 vs 5.6 d; p = 0.025). Nearly one in four patients admitted to the ICU with COVID-19 after widespread COVID-19 vaccine availability represented a vaccine-breakthrough case. Mortality risk remains substantial in vaccinated patients and similar between vaccinated and unvaccinated patients after the onset of critical illness. However, COVID-19 vaccination is associated with reduced ICU resource utilization and booster dosing may increase survivability from COVID-19-related critical illness.

Forecasting ICU Census by Combining Time Series and Survival Models.

Retrospective observational study. We developed the algorithm using data from the Medical-Surgical ICU (MSICU) at University Hospital, London, Canada and validated using data from the Critical Care Trauma Centre (CCTC) at Victoria Hospital, London, Canada. Adult patient admissions (7,434) to the MSICU and (9,075) to the CCTC from 2015 to 2021. None. We developed an Autoregressive integrated moving average time series model that forecasts patients arriving in the ICU and a survival model using MODS, NEMS, and other factors to estimate patient LOS. The models were combined to create an algorithm that forecasts ICU census for planning horizons ranging from 1 to 7 days. We evaluated the algorithm quality using several fit metrics. The root mean squared error ranged from 2.055 to 2.890 beds/d and the mean absolute percentage error from 9.4% to 13.2%. We show that this forecasting algorithm provides a better fit when compared with a moving average or a time series model that directly forecasts ICU census. Additionally, we evaluated the performance of the algorithm using data during the global COVID-19 pandemic and found that the error of the forecasts increased proportionally with the number of COVID-19 patients in the ICU. It is possible to develop accurate tools to forecast ICU census. This type of algorithm may be important to clinicians and managers when planning ICU capacity as well as staffing and surgical demand planning over a short time horizon.

Multi-event survival analysis through dynamic multi-modal learning for ICU mortality prediction

Background and objective: Survival analysis is widely applied for assessing the expected duration of patient status towards event occurrences such as mortality in healthcare domain, which is generally considered as a time-to-event problem. Patients with multiple complications have high mortality risks and oftentimes require specific intensive care and clinical treatments. The progression of complications is time-varying according to disease development and intrinsic interactions between complications with respect to mortality are uncertain. Classical methods for mortality prediction and survival analysis in critical care, such as risk scoring systems and cause-specific survival models, were not designed for this multi-event survival analysis problem and able to measure the competing risks of death for mutually exclusive events. In addition, multivariate temporal information of complications is not taken into consideration while estimating differentiated mortality risks in the early stage.Methods: In this paper, we propose a novel multi-event survival analysis solution using a tree-based autoregressive survival model of multi-modal electronic health record data. Specifically, we focus on modeling the temporal trajectory of complications and estimating the mortality risk associated with multiple potential complications simultaneously. In dynamic modeling, no assumptions are made for the relationships between time-dependent variables and risk transition over time.Results: Validated with the eICU database, our model achieves a better prediction performance with C-index ranging in 74–80%, compared to state-of-the-art machine learning methods in the literature, for the complications of acute respiratory distress syndrome and cardiovascular disease cases.Conclusions: Our model provides the distinguishable mortality risk curves over time for specific complications and the track of risk development that could potentially support the ICU resource reallocation.

Review of the Utility of Extended Recovery Room Observation after Adenotonsillectomy.

Our institution implemented a post-anesthesia care unit (PACU) extended-stay model (Grey Zone model), where the post-operative level of care for high-risk adenotonsillectomy patients (general care vs. intensive care unit) was decided based on the clinical course of 2-4 h of PACU admission. To assess the correlation between post-tonsillectomy respiratory compromise and the need for respiratory support during an extended stay at PACU. To identify comorbidities associated with a need for intensive care after extended observation. A retrospective cohort study of high-risk children who underwent adenotonsillectomy and were admitted to the Grey Zone following surgery. 274 patients met inclusion criteria. 262 (95.6%) met criteria for general care unit transfer (mean oxygen saturation 94.4 ± 5.1%). Twelve (4.4%) patients were transferred from the PACU to the ICU due to respiratory distress (mean oxygen saturation 86.8 ± 11%). Of the patients admitted to general care, 4 (1.5%) secondarily developed respiratory compromise, requiring escalation of care. Three of these maintained oxygen saturation ≥95% throughout the PACU period. There was no difference between the groups with respect to demographic data, rates of morbid obesity, and severity of obstructive sleep apnea. Neuromuscular disease, chronic lung disease, seizure disorder, and gastrostomy-tube status were more prevalent in those requiring ICU level of care compared to the general care unit. The Grey Zone model accurately identifies patients requiring ICU-level care following adenotonsillectomy, allowing for a safe reduction in the utilization of ICU resources. Due to rare delayed respiratory events, overnight observation in this cohort is recommended. 4 Laryngoscope, 133:3582-3587, 2023.

Deep Learning-Based Recurrent Delirium Prediction in Critically Ill Patients.

To predict impending delirium in ICU patients using recurrent deep learning. Retrospective cohort study. Fifteen medical-surgical ICUs across Alberta, Canada, between January 1, 2014, and January 24, 2020. Forty-three thousand five hundred ten ICU admissions from 38,426 patients. None. We used ICU and administrative health data to train deep learning models to predict delirium episodes in the next two 12-hour periods (0-12 and 12-24 hr), starting at 24 hours after ICU admission, and to generate new predictions every 12 hours. We used a comprehensive set of 3,643 features, capturing patient history, early ICU admission information (first 24 hr), and the temporal dynamics of various clinical variables throughout the ICU admission. Our deep learning architecture consisted of a feature embedding, a recurrent, and a prediction module. Our best model based on gated recurrent units yielded a sensitivity of 0.810, a specificity of 0.848, a precision (positive predictive value) of 0.704, and an area under the receiver operating characteristic curve (AUROC) of 0.909 in the hold-out test set for the 0-12-hour prediction horizon. For the 12-24-hour prediction horizon, the same model achieved a sensitivity of 0.791, a specificity of 0.807, a precision of 0.637, and an AUROC of 0.895 in the test set. Our delirium prediction model achieved strong performance by applying deep learning to a dataset that is at least one order of magnitude larger than those used in previous studies. Another novel aspect of our study is the temporal nature of our features and predictions. Our model enables accurate prediction of impending delirium in the ICU, which can potentially lead to early intervention, more efficient allocation of ICU resources, and improved patient outcomes.

A multicenter study of ICU resource utilization in pediatric, adolescent and young adult patients post CAR-T therapy.

Tisagenlecleucel is associated with remarkable outcomes in treating patients up to the age of 25 years with refractory B-cell acute lymphoblastic leukemia (ALL). Yet, due to unique and potentially life-threatening complications, access remains limited to higher-resource and certified centers. Reports of inequity and related disparities in care are emerging. In this multicenter study of ALL patients admitted for anti-leukemia therapy, who required pediatric intensive care (ICU) support (n = 205), patients receiving tisagenlecleucel (n = 39) were compared to those receiving conventional chemotherapy (n = 166). The median time to ICU transfer was 6 (0-43) versus 1 (0-116) days, respectively (p < 0.0001). There was no difference in the use of vasopressor, ionotropic, sedating, and/or paralytic agents between groups, but use of dexamethasone was higher among tisagenlecleucel patients. Patients receiving tisagenlecleucel were more likely to have cardiorespiratory toxicity (p = 0.0002), but there were no differences in diagnostic interventions between both groups and/or differences in ICU length of stay and/or overall hospital survival. Toxicities associated with tisagenlecleucel are generally reversible, and our findings suggest that resource utilization once admitted to the ICU may be similar among patients with ALL receiving tisagenlecleucel versus conventional chemotherapy. As centers consider improved access to care and the feasibility of tisagenlecleucel certification, our study may inform strategic planning.

SAPS 3 as a 28-day mortality predictor in critically ill COVID-19 patients

Background: The case fatality rate (CFR) of COVID-19 was 8.7% in Indonesia on April 2020. Simplified Acute Physiology Score 3 (SAPS 3) has been used to predict the hospital mortality based on different variables including acute physiologic derangements, current conditions and interventions, and previous health status assess the severity of condition during the first hour of admission to the ICU. We assessed SAPS 3 to predict the outcome and mortality of critical COVID-19 patients in ICU over a period of 28 days.&#x0D; Methodology: This retrospective cohort study consisted of adult patients admitted to ICU with probable or confirmed COVID-19 in our hospital. We recorded the patients SAPS 3 score from the medical record as well as the 28-day mortality. Validity of the SAPS 3 score was done by the Area Under Curve (AUC) measurement and Hosmer-Lemeshow calibration test.&#x0D; Results: The mortality rate of critical COVID-19 patients was 43.8%. The age, intra-hospital location before ICU admission, use of vasoactive drugs (P &lt; 0.0001), focal neurological deficits (P &lt; 0.0001), respiratory failure (P = 0.004), creatinine ≥ 3.5 mg/dL (P = 0.005), and platelets &lt; 50,000 /µL (P = 0.032) were significantly associated with 28-days mortality in the ICU. SAPS 3 showed good discrimination and predictability. The optimal cut-off point was 39 with 70.3% sensitivity and 74.4% specificity.&#x0D; Conclusion: SAPS3 score system was valid in predicting the 28-day mortality of COVID-19 patients in the ICU with good discrimination and calibration value; therefore, it is an important predictor tool for early prognosis screening that will help reduce the strain over the ICU resources.&#x0D; Abbreviations: CFR: Case Fatality Rate; SAPS 3: Simplified Acute Physiology Score 3; COVID-19: The Coronavirus Disease 2019; ICU: Intensive Care Unit; APACHE: Acute Physiology and Chronic Health Evaluation; SPSS: Statistical Package for Social Sciences; GCS: Glasgow Coma Scale; ROC: Receiver Operating Characteristic; PHEIC: Public Health Emergency of International Concern; OR: Odds Ratio;&#x0D; Citation: Sukmono RB, Manggala SK, Priscilla P, Aditianingsih D. SAPS 3 as a 28-day mortality predictor in critically ill COVID-19 patients. Anaesth. pain intensive care 2022;26(5):640-648.&#x0D; DOI: 10.35975/apic.v26i5.1986