Collaborative Database Research Articles

Conservation Action Tracker: A tool to identify and monitor conservation actions for tree species.

The GlobalTree Portal, hosted by Botanic Gardens Conservation International, provides access to information on the approximately 58,000 tree species worldwide. Included in the GlobalTree Portal is the Conservation Action Tracker, a dynamic and collaborative database to identify and monitor conservation actions for tree species globally. The Conservation Action Tracker collates conservation action information at the species level, including species recovery/action plans, ex situ collections, propagation protocols, in situ management, species protection policy, and education/awareness campaigns. To date, the Conservation Action Tracker contains conservation action information for 4126 tree species, including 2161 threatened species, of which 659 are classified as Vulnerable, 783 as Endangered, and 719 as Critically Endangered. It covers conservation action information for at least one tree species in every country; however, more information is needed for 89% of Vulnerable, 87% of Endangered, and 77% of Critically Endangered tree species. Monitoring species conservation actions can support the prioritization and scaling up of conservation practices by sharing knowledge, increasing collaboration, enabling the identification of conservation gaps, and making the information available to be used by decision-makers. Tracking conservation actions at the species level is, therefore, essential to guide future conservation efforts. Increasing the amount of data in the Conservation Action Tracker will improve the tool's ability to guide future conservation efforts and avoid the extinction of tree species.

Improving the Prognostic Evaluation Precision of Hospital Outcomes for Heart Failure Using Admission Notes and Clinical Tabular Data: Multimodal Deep Learning Model.

Clinical notes contain contextualized information beyond structured data related to patients' past and current health status. This study aimed to design a multimodal deep learning approach to improve the evaluation precision of hospital outcomes for heart failure (HF) using admission clinical notes and easily collected tabular data. Data for the development and validation of the multimodal model were retrospectively derived from 3 open-access US databases, including the Medical Information Mart for Intensive Care III v1.4 (MIMIC-III) and MIMIC-IV v1.0, collected from a teaching hospital from 2001 to 2019, and the eICU Collaborative Research Database v1.2, collected from 208 hospitals from 2014 to 2015. The study cohorts consisted of all patients with critical HF. The clinical notes, including chief complaint, history of present illness, physical examination, medical history, and admission medication, as well as clinical variables recorded in electronic health records, were analyzed. We developed a deep learning mortality prediction model for in-hospital patients, which underwent complete internal, prospective, and external evaluation. The Integrated Gradients and SHapley Additive exPlanations (SHAP) methods were used to analyze the importance of risk factors. The study included 9989 (16.4%) patients in the development set, 2497 (14.1%) patients in the internal validation set, 1896 (18.3%) in the prospective validation set, and 7432 (15%) patients in the external validation set. The area under the receiver operating characteristic curve of the models was 0.838 (95% CI 0.827-0.851), 0.849 (95% CI 0.841-0.856), and 0.767 (95% CI 0.762-0.772), for the internal, prospective, and external validation sets, respectively. The area under the receiver operating characteristic curve of the multimodal model outperformed that of the unimodal models in all test sets, and tabular data contributed to higher discrimination. The medical history and physical examination were more useful than other factors in early assessments. The multimodal deep learning model for combining admission notes and clinical tabular data showed promising efficacy as a potentially novel method in evaluating the risk of mortality in patients with HF, providing more accurate and timely decision support.

THE INVESTIGATION OF TRANSFERABILITY OF DEFINITIONS OF CARDIOMETABOLIC AND VASCULAR AGING PHENOTYPES ACROSS MULTIPLE COHORTS

Objective: According to the European guidelines for the prevention of cardiovascular diseases (CVD) (2021), Russia was identified as a country with a very high CV risk. We aimed to investigate the applicability of the standardized definitions of vascular aging phenotypes (VP) in Russian population and align them with the existing reference datasets. Design and method: In 2012-2013 1,600 residents of St. Petersburg (age 21-67) were randomly examined as a part of the multicenter epidemiological study. Anthropometry, blood pressure measurement, blood biochemistry analysis were taken. Subjects completed questionnaires describing their lifestyle, current health condition and medication usage. 618 patients (42% men) from initial cohort with a mean age of 48 [37;55] years were randomly subsampled for carotid-femoral pulse wave velocity assessment (Sphygmocor). The VP were defined using the ‘Reference Values for Arterial Stiffness’ Collaboration database which included 16,867 measurements from eight European countries populations15 to 97 years. According to the criteria, patients with overt CVD, diabetes mellitus and the use of antihypertensive, lipid-lowering therapy were excluded from the analysis of VP. Results: The Russian cohort had significantly higher prevalence of CVD (11,4% compared to 6,9% in the Nreference population, p=5.65x10-5), intake of antihypertensives (28,1% compared to 14,3% in the reference, p<1x10-16) and lipid-lowering therapy (6,1% compared to 4,1%, p=0.025, Figure 1A). Additionally, our cohort had significantly higher prevalence of optimal BP (40,2% vs 28%, p=4.6x10-7) and lower prevalence of first-degree (p=0.048) and second degree (p=0.00084) hypertension (Figure 1B). Comparison of the prevalence of VP between the European and Russian cohorts showed a significant decrease of the early vascular aging (EVA) phenotype with age in the Russian population (p=0.011). The prevalence of the supernormal vascular aging phenotype (SUPERNOVA) in the group of 60-69 years old was higher in the Russian population (p=0.0035) Conclusions: A higher prevalence of CVD was demonstrated in the study cohort despite lower hypertension burden and higher CV drugs intake compared to the European cohort. Multidirectional trajectories of EVA and SUPERNOVA VP with age were detected in Russian population, potentially indicating the presence of the clearing selection pressure at the older age (Agreement No 075-15-2022-301).

Triglyceride-glucose index as a potential predictor for in-hospital mortality in critically ill patients with intracerebral hemorrhage: a multicenter, case-control study.

The correlation between the triglyceride-glucose index (TyG) and the prognosis of ischemic stroke has been well established. This study aims to assess the influence of the TyG index on the clinical outcomes of critically ill individuals suffering from intracerebral hemorrhage (ICH). Patients diagnosed with ICH were retrospectively retrieved from the Medical Information Mart for Intensive Care (MIMIC-IV) and the eICU Collaborative Research Database (eICU-CRD). Various statistical methods, including restricted cubic spline (RCS) regression, multivariable logistic regression, subgroup analysis, and sensitivity analysis, were employed to examine the relationship between the TyG index and the primary outcomes of ICH. A total of 791 patients from MIMIC-IV and 1,113 ones from eICU-CRD were analyzed. In MIMIC-IV, the in-hospital and ICU mortality rates were 14% and 10%, respectively, while in eICU-CRD, they were 16% and 8%. Results of the RCS regression revealed a consistent linear relationship between the TyG index and the risk of in-hospital and ICU mortality across the entire study population of both databases. Logistic regression analysis revealed a significant positive association between the TyG index and the likelihood of in-hospital and ICU death among ICH patients in both databases. Subgroup and sensitivity analysis further revealed an interaction between patients' age and the TyG index in relation to in-hospital and ICU mortality among ICH patients. Notably, for patients over 60years old, the association between the TyG index and the risk of in-hospital and ICU mortality was more pronounced compared to the overall study population in both MIMIC-IV and eICU-CRD databases, suggesting a synergistic effect between old age (over 60years) and the TyG index on the in-hospital and ICU mortality of patients with ICH. This study established a positive correlation between the TyG index and the risk of in-hospital and ICU mortality in patients over 60years who diagnosed with ICH, suggesting that the TyG index holds promise as an indicator for risk stratification in this patient population.

Molnupiravir increases SARS-CoV-2 genome diversity and complexity: A case-control cohort study.

Molnupiravir, an oral direct-acting antiviral effective in vitro against SARS-CoV-2, has been largely employed during the COVID-19 pandemic, since December 2021. After marketing and widespread usage, a progressive increase in SARS-CoV-2 lineages characterized by a higher transition/transversion ratio, a characteristic signature of molnupiravir action, appeared in the Global Initiative on Sharing All Influenza Data(GISAID) and International Nucleotide Sequence Database Collaboration(INSDC) databases. Here, we assessed the drug effects by SARS-CoV-2 whole-genome sequencing on 38 molnupiravir-treated persistently positive COVID-19 outpatients tested before and after treatment. Seventeen tixagevimab/cilgavimab-treated outpatients served as controls. Mutational analyses confirmed that SARS-CoV-2 exhibits an increased transition/transversion ratio seven days after initiation of molnupiravir. Moreover we observed an increased G->A ratio compared to controls, which was not related to apolipoprotein B mRNAediting enzyme, catalytic polypeptide-like (APOBEC) activity. In addition, we demonstrated for the first time an increased diversity and complexity of the viral quasispecies.

Development and validation of a nomogram for predicting in-hospital mortality of intensive care unit patients with liver cirrhosis.

Liver cirrhosis patients admitted to intensive care unit (ICU) have a high mortality rate. To establish and validate a nomogram for predicting in-hospital mortality of ICU patients with liver cirrhosis. We extracted demographic, etiological, vital sign, laboratory test, comorbidity, complication, treatment, and severity score data of liver cirrhosis patients from the Medical Information Mart for Intensive Care IV (MIMIC-IV) and electronic ICU (eICU) collaborative research database (eICU-CRD). Predictor selection and model building were based on the MIMIC-IV dataset. The variables selected through least absolute shrinkage and selection operator analysis were further screened through multivariate regression analysis to obtain final predictors. The final predictors were included in the multivariate logistic regression model, which was used to construct a nomogram. Finally, we conducted external validation using the eICU-CRD. The area under the receiver operating characteristic curve (AUC), decision curve, and calibration curve were used to assess the efficacy of the models. Risk factors, including the mean respiratory rate, mean systolic blood pressure, mean heart rate, white blood cells, international normalized ratio, total bilirubin, age, invasive ventilation, vasopressor use, maximum stage of acute kidney injury, and sequential organ failure assessment score, were included in the multivariate logistic regression. The model achieved AUCs of 0.864 and 0.808 in the MIMIC-IV and eICU-CRD databases, respectively. The calibration curve also confirmed the predictive ability of the model, while the decision curve confirmed its clinical value. The nomogram has high accuracy in predicting in-hospital mortality. Improving the included predictors may help improve the prognosis of patients.

Blood glucose fluctuation and in-hospital mortality among patients with acute myocardial infarction: eICU collaborative research database.

To assess the relationship between glycemic variability, glucose fluctuation trajectory and the risk of in-hospital mortality in patients with acute myocardial infarction (AMI). This retrospective cohort study included AMI patients from eICU Collaborative Research Database. In-hospital mortality of AMI patients was primary endpoint. Blood glucose levels at admission, glycemic variability, and glucose fluctuation trajectory were three main study variables. Blood glucose levels at admission were stratified into: normal, intermediate, and high. Glycemic variability was evaluated using the coefficient of variation (CV), divided into four groups based on quartiles: quartile 1: CV≤10; quartile 2: 10<CV≤20; quartile 3: 20<CV≤30; quartile 4: CV>30. Univariate and multivariate Cox regression models to assess the relationship between blood glucose levels at admission, glycemic variability, glucose fluctuation trajectory, and in-hospital mortality in patients with AMI. 2590 participants were eventually included in this study. There was a positive relationship between high blood glucose level at admission and in-hospital mortality [hazard ratio (HR) = 1.42, 95%confidence interval (CI): 1.06-1.89]. The fourth quartile (CV>30) of CV was associated with increased in-hospital mortality (HR = 2.06, 95% CI: 1.25-3.40). The findings indicated that only AMI individuals in the fourth quartile of glycemic variability, exhibited an elevated in-hospital mortality among those with normal blood glucose levels at admission (HR = 2.33, 95% CI: 1.11-4.87). Additionally, elevated blood glucose level was a risk factor for in-hospital mortality in AMI patients. Glycemic variability was correlated with in-hospital mortality, particularly among AMI patients who had normal blood glucose levels at admission. Our study findings also suggest early intervention should be implemented to normalize high blood glucose levels at admission of AMI.

Cervical Spine Surgery Following COVID-19 Infection: When is it Safe to Proceed?

Retrospective. We utilized the NIH National COVID Cohort Collaborative (N3C) database to characterize the risk profile of patients undergoing spine surgery during multiple time windows following the COVID-19 infection. While the impact of COVID-19 on various organ systems is well documented, there is limited knowledge regarding its effect on perioperative complications following spine surgery or the optimal timing of surgery after an infection. We asked the National COVID Cohort Collaborative for patients who underwent cervical spine surgery. Patients were stratified into those with an initial documented COVID-19 infection within 3 time periods: 0-2 weeks, 2-6 weeks, or 6-12 weeks before surgery. A total of 29,449 patients who underwent anterior approach cervical spine surgery and 46,379 patients who underwent posterior approach cervical spine surgery were included. Patients who underwent surgery within 2 weeks of their COVID-19 diagnosis had a significantly increased risk for venous thromboembolic events, sepsis, 30-day mortality, and 1-year mortality, irrespective of the anterior or posterior approach. Among patients undergoing surgery between 2 and 6 weeks after COVID-19 infection, the 30-day mortality risk remained elevated in patients undergoing a posterior approach only. Patients undergoing surgery between 6 and 12 weeks from the date of the COVID-19 infection did not show significantly elevated rates of any complications analyzed. Patients undergoing either anterior or posterior cervical spine surgery within 2 weeks from the initial COVID-19 diagnosis are at increased risk for perioperative venous thromboembolic events, sepsis, and mortality. Elevated perioperative complication risk does not persist beyond 2 weeks, except for 30-day mortality in posterior approach surgeries. On the basis of these results, it may be warranted to postpone nonurgent spine surgeries for at least 2 weeks following a COVID-19 infection and advise patients of the increased perioperative complication risk when urgent surgery is required.

Comparing Open-Access Database and Traditional Intensive Care Studies Using Machine Learning: Bibliometric Analysis Study.

Intensive care research has predominantly relied on conventional methods like randomized controlled trials. However, the increasing popularity of open-access, free databases in the past decade has opened new avenues for research, offering fresh insights. Leveraging machine learning (ML) techniques enables the analysis of trends in a vast number of studies. This study aims to conduct a comprehensive bibliometric analysis using ML to compare trends and research topics in traditional intensive care unit (ICU) studies and those done with open-access databases (OADs). We used ML for the analysis of publications in the Web of Science database in this study. Articles were categorized into "OAD" and "traditional intensive care" (TIC) studies. OAD studies were included in the Medical Information Mart for Intensive Care (MIMIC), eICU Collaborative Research Database (eICU-CRD), Amsterdam University Medical Centers Database (AmsterdamUMCdb), High Time Resolution ICU Dataset (HiRID), and Pediatric Intensive Care database. TIC studies included all other intensive care studies. Uniform manifold approximation and projection was used to visualize the corpus distribution. The BERTopic technique was used to generate 30 topic-unique identification numbers and to categorize topics into 22 topic families. A total of 227,893 records were extracted. After exclusions, 145,426 articles were identified as TIC and 1301 articles as OAD studies. TIC studies experienced exponential growth over the last 2 decades, culminating in a peak of 16,378 articles in 2021, while OAD studies demonstrated a consistent upsurge since 2018. Sepsis, ventilation-related research, and pediatric intensive care were the most frequently discussed topics. TIC studies exhibited broader coverage than OAD studies, suggesting a more extensive research scope. This study analyzed ICU research, providing valuable insights from a large number of publications. OAD studies complement TIC studies, focusing on predictive modeling, while TIC studies capture essential qualitative information. Integrating both approaches in a complementary manner is the future direction for ICU research. Additionally, natural language processing techniques offer a transformative alternative for literature review and bibliometric analysis.

CT Cadaveric dataset for Radiomics features stability assessment in lumbar vertebrae

Radiomics features (RFs) studies have showed limitations in the reproducibility of RFs in different acquisition settings. To date, reproducibility studies using CT images mainly rely on phantoms, due to the harness of patient exposure to X-rays. The provided CadAIver dataset has the aims of evaluating how CT scanner parameters effect radiomics features on cadaveric donor. The dataset comprises 112 unique CT acquisitions of a cadaveric truck acquired on 3 different CT scanners varying KV, mA, field-of-view, and reconstruction kernel settings. Technical validation of the CadAIver dataset comprises a comprehensive univariate and multivariate GLM approach to assess stability of each RFs extracted from lumbar vertebrae. The complete dataset is publicly available to be applied for future research in the RFs field, and could foster the creation of a collaborative open CT image database to increase the sample size, the range of available scanners, and the available body districts.

Repeat Next-Generation Sequencing Testing on Progression in Men With Metastatic Prostate Cancer Can Identify New Actionable Alterations.

There are limited data available on the real-world patterns of molecular testing in men with advanced prostate cancer. We thus sought to evaluate next-generation sequencing (NGS) testing in the United States, focused on single versus serial NGS testing, the different disease states of testing (hormone-sensitive v castration-resistant, metastatic vs nonmetastatic), tissue versus plasma circulating tumor DNA (ctDNA) assays, and how often actionable data were found on each NGS test. The Prostate Cancer Precision Medicine Multi-Institutional Collaborative Effort clinical-genomic database was used for this retrospective analysis, including 1,597 patients across 15 institutions. Actionable NGS data were defined as including somatic alterations in homologous recombination repair genes, mismatch repair deficiency, microsatellite instability (MSI-high), or a high tumor mutational burden ≥10 mut/MB. Serial NGS testing (two or more NGS tests with specimens collected more than 60 days apart) was performed in 9% (n = 144) of patients with a median of 182 days in between test results. For the second NGS test and beyond, 82.1% (225 of 274) of tests were from ctDNA assays and 76.1% (217 of 285) were collected in the metastatic castration-resistant setting. New actionable data were found on 11.1% (16 of 144) of second NGS tests, with 3.5% (5 of 144) of tests detecting a new BRCA2 alteration or MSI-high. A targeted therapy (poly (ADP-ribose) polymerase inhibitor or immunotherapy) was given after an actionable result on the second NGS test in 31.3% (5 of 16) of patients. Repeat somatic NGS testing in men with prostate cancer is infrequently performed in practice and can identify new actionable alterations not present with initial testing, suggesting the utility of repeat molecular profiling with tissue or blood of men with metastatic castration-resistant prostate cancer to guide therapy choices.

Perioperative and functional outcomes following robot-assisted partial nephrectomy: Descriptive analysis of Indian study group on partial nephrectomy database.

There is an unmet need for high-quality data for Robot-assisted partial nephrectomy (RAPN) in the Indian population. Indian study group on partial nephrectomy (ISGPN) is a consortium of Indian centers contributing to the partial nephrectomy (PN) database. The current study is a descriptive analysis of perioperative and functional outcomes following RAPN. For this study, the retrospective ISGPN database was reviewed, which included patients who underwent RAPN for renal masses at 14 centers across India from September 2010 to September 2022. Demographic, clinical, radiological, perioperative, and functional data were collected and analyzed. Ethics approval was obtained from each of the participating centers. In this study, 782 patients were included, and 69.7% were male. The median age was 53 years (interquartile range [IQR 44-62]), median operative time was 180 min (IQR 133-240), median estimated blood loss was 100 mL (IQR 50-200), mean warm ischemia time was 22.7 min and positive surgical margin rates were 2.5%. The complication rate was 16.2%, and most of them were of minor grade. Trifecta and pentafecta outcomes were attained in 61.4% and 60% of patients, respectively. This is the largest Indian multi-centric study using the Indian Robotic PN Collaborative database to evaluate the outcomes of robot-assisted PN, and has proven its safety and efficacy in the management of renal masses.

120 Outcome Related Thresholds of Near Infrared Spectroscopy Based Cerebrovascular Reactivity Metrics in Moderate Severe TBI: A Canadian High Resolution-TBI (CAHR-TBI) Cohort Study

INTRODUCTION: Dysfunctional cerebrovascular reactivity (CVR)/autoregulation contributes to secondary injury following traumatic brain injury (TBI). Currently, monitoring of CVR relies on intracranial pressure (ICP) monitoring and has known thresholds at which outcomes worsen. Interest has shifted to less invasive near infrared spectroscopic (NIRS) regional cerebral oxygen saturation (rSO2) based measures of CVR. However, threshold levels at which outcomes worsen have not yet been determined for these indices. METHODS: A retrospective multi-institutional cohort study utilizing the CAnadian High Resolution TBI (CAHR-TBI) Research Collaborative database was performed. The cohort included TBI patients with ICP, arterial blood pressure (ABP), and rSO2 monitoring treated in adult intensive care units (ICU). COx (using rSO2 and cerebral perfusion pressure) as well as COx_a (using rSO2 and ABP) were calculated for each subject as the rSO2 based indices of CVR. 2 x 2 tables were created grouping patients by alive/dead and favorable/unfavorable outcomes at various thresholds of COx and COx_a as well as rSO2 itself. Chi-square values were calculated and the threshold producing the highest value was identified as providing the best discriminative value. RESULTS: A total of 129 patients were included. For both COx and COx_a an optimal threshold value of 0.2 was identified for both survival (χ2 = 9.57, p = 0.0020; χ2 = 13.04, p = 0.0003 respectively) and favorable outcomes (χ2 = 5.98, p = 0.0145; χ2 = 8.94, p = 0.0028 respectively) with values above this associated with worse outcomes. Notably, there was no identifiable threshold for raw rSO2 values at which outcomes were identified to worsen. CONCLUSIONS: In this multi-institutional cohort study, COx and COx_a were found to have a uniform threshold of 0.2, above which clinical outcomes worsened. This study lays the groundwork to transition to less invasive means of continuously measuring CVR.

Incidence and Mortality of COVID-19-Associated Invasive Fungal Infections Among Critically Ill Intubated Patients: A Multicenter Retrospective Cohort Analysis.

An association between coronavirus disease 2019 (COVID-19)-associated invasive fungal infections (CAIFIs) and high mortality among intubated patients has been suggested in previous research. However, some of the current evidence was derived from small case series and multicenter studies conducted during different waves of the COVID-19 pandemic. We examined the incidence of CAIFIs and their associated mortality using a large, multicenter COVID-19 database built throughout the pandemic. We conducted a retrospective analysis of the National COVID Cohort Collaborative (N3C) database collected from 76 medical centers in the United States between January 2020 and August 2022. Patients were 18 years or older and intubated after severe acute respiratory syndrome coronavirus 2 infection. The primary outcomes were incidence and all-cause mortality at 90 days. To assess all-cause mortality, we fitted Cox proportional hazard models after adjusting for confounders via inverse probability weighting. Out of the 4 916 229 patients with COVID-19 diagnosed during the study period, 68 383 (1.4%) met our cohort definition. The overall incidence of CAIFI was 2.80% (n = 1934/68 383). Aspergillus (48.2%; n = 933/1934) and Candida (41.0%; n = 793/1934) were the most common causative organisms. The incidence of CAIFIs associated with Aspergillus among patients who underwent BAL was 6.2% (n = 83/1328). Following inverse probability weighting, CAIFIs caused by Aspergillus (hazard ratio [HR], 2.0; 95% CI, 1.8-2.2) and Candida (HR, 1.7; 95% CI, 1.5-1.9) were associated with increased all-cause mortality. Systemic antifungals reduced mortality in 17% of patients with CAIFI with Aspergillus and 24% of patients with CAIFI with Candida. The incidence of CAIFI was modest but associated with higher 90-day all-cause mortality among intubated patients. Systemic antifungals modified mortality.

Predicting Successful Weaning from Mechanical Ventilation by Reduction in Positive End-expiratory Pressure Level Using Machine Learning.

Weaning patients from mechanical ventilation (MV) is a critical and resource intensive process in the Intensive Care Unit (ICU) that impacts patient outcomes and healthcare expenses. Weaning methods vary widely among providers. Prolonged MV is associated with adverse events and higher healthcare expenses. Predicting weaning readiness is a non-trivial process in which the positive end-expiratory pressure (PEEP), a crucial component of MV, has potential to be indicative but has not yet been used as the target. We aimed to predict successful weaning from mechanical ventilation by targeting changes in the PEEP-level using a supervised machine learning model. This retrospective study included 12,153 mechanically ventilated patients from Medical Information Mart for Intensive Care (MIMIC-IV) and eICU collaborative research database (eICU-CRD). Two machine learning models (Extreme Gradient Boosting and Logistic Regression) were developed using a continuous PEEP reduction as target. The data is splitted into 80% as training set and 20% as test set. The model's predictive performance was reported using 95% confidence interval (CI), based on evaluation metrics such as area under the receiver operating characteristic (AUROC), area under the precision-recall curve (AUPRC), F1-Score, Recall, positive predictive value (PPV), and negative predictive value (NPV). The model's descriptive performance was reported as the variable ranking using SHAP (SHapley Additive exPlanations) algorithm. The best model achieved an AUROC of 0.84 (95% CI 0.83-0.85) and an AUPRC of 0.69 (95% CI 0.67-0.70) in predicting successful weaning based on the PEEP reduction. The model demonstrated a Recall of 0.85 (95% CI 0.84-0.86), F1-score of 0.86 (95% CI 0.85-0.87), PPV of 0.87 (95% CI 0.86-0.88), and NPV of 0.64 (95% CI 0.63-0.66). Most of the variables that SHAP algorithm ranked to be important correspond with clinical intuition, such as duration of MV, oxygen saturation (SaO2), PEEP, and Glasgow Coma Score (GCS) components. This study demonstrates the potential application of machine learning in predicting successful weaning from MV based on continuous PEEP reduction. The model's high PPV and moderate NPV suggest that it could be a useful tool to assist clinicians in making decisions regarding ventilator management.

Linking preoperative and early intensive care unit data for prolonged intubation prediction.

Prolonged intubation (PI) is a frequently encountered severe complication among patients following cardiac surgery (CS). Solely concentrating on preoperative data, devoid of sufficient consideration for the ongoing impact of surgical, anesthetic, and cardiopulmonary bypass procedures on subsequent respiratory system function, could potentially compromise the predictive accuracy of disease prognosis. In response to this challenge, we formulated and externally validated an intelligible prediction model tailored for CS patients, leveraging both preoperative information and early intensive care unit (ICU) data to facilitate early prophylaxis for PI. We conducted a retrospective cohort study, analyzing adult patients who underwent CS and utilizing data from two publicly available ICU databases, namely, the Medical Information Mart for Intensive Care and the eICU Collaborative Research Database. PI was defined as necessitating intubation for over 24 h. The predictive model was constructed using multivariable logistic regression. External validation of the model's predictive performance was conducted, and the findings were elucidated through visualization techniques. The incidence rates of PI in the training, testing, and external validation cohorts were 11.8%, 12.1%, and 17.5%, respectively. We identified 11 predictive factors associated with PI following CS: plateau pressure [odds ratio (OR), 1.133; 95% confidence interval (CI), 1.111-1.157], lactate level (OR, 1.131; 95% CI, 1.067-1.2), Charlson Comorbidity Index (OR, 1.166; 95% CI, 1.115-1.219), Sequential Organ Failure Assessment score (OR, 1.096; 95% CI, 1.061-1.132), central venous pressure (OR, 1.052; 95% CI, 1.033-1.073), anion gap (OR, 1.075; 95% CI, 1.043-1.107), positive end-expiratory pressure (OR, 1.087; 95% CI, 1.047-1.129), vasopressor usage (OR, 1.521; 95% CI, 1.23-1.879), Visual Analog Scale score (OR, 0.928; 95% CI, 0.893-0.964), pH value (OR, 0.757; 95% CI, 0.629-0.913), and blood urea nitrogen level (OR, 1.011; 95% CI, 1.003-1.02). The model exhibited an area under the receiver operating characteristic curve (AUROC) of 0.853 (95% CI, 0.840-0.865) in the training cohort, 0.867 (95% CI, 0.853-0.882) in the testing cohort, and 0.704 (95% CI, 0.679-0.727) in the external validation cohort. Through multicenter internal and external validation, our model, which integrates early ICU data and preoperative information, exhibited outstanding discriminative capability. This integration allows for the accurate assessment of PI risk in the initial phases following CS, facilitating timely interventions to mitigate adverse outcomes.

Serum phosphate levels and the development of sepsis associated acute kidney injury: evidence from two independent databases.

We aimed to investigate the association between serum phosphate levels and the risk for developing sepsis associated acute kidney injury (SAKI). Septic patients from the Medical Information Mart for Intensive Care IV (MIMIC IV) and the eICU Collaborative Research Database (eICU-CRD) were enrolled. Restricted cubic spline (RCS) was used to visualize the relationship between phosphate levels and the risk of SAKI. Patients were divided into four categories based on their serum phosphate levels. Logistic regression analysis, receiver operating characteristic (ROC) curve and subgroup analysis were performed to evaluate the predictive value of serum phosphate for SAKI. A total of 9,244 and 2,124 patients from the MIMIC IV and eICU-CRD database were included in the final analysis. RCS curve revealed a non-linear correlation between phosphate levels and the risk of SAKI (p for non-linearity <0.05). Each 1 mg/dL increase in phosphate levels was associated with a 1.51 to 1.64-fold increased risk of SAKI (OR 2.51-2.64, p < 0.001) in the MIMIC IV cohort and a 0.29 to 0.38-fold increased risk (OR 1.29-1.38, p < 0.001) in the eICU-CRD cohort. Compared to the normal-low category, hyperphosphatemia and normal-high category were independently associated with an increased risk of SAKI, while hypophosphatemia was independently associated with a decreased risk in the MIMIC IV cohort. A similar trend was observed in the eICU-CRD cohort, but statistical significance disappeared in the hypophosphatemia category and the adjusted model of normal high category. These finding was consistent in subgroup analysis. Elevated serum phosphate, even within the normal range, is an independent risk factor for developing SAKI in septic patients. Abnormal change in serum phosphate levels may be a novel biomarker for early prediction of SAKI occurrence.

Abstract 4923: A systematic framework for predicting adverse drug reaction signals using medical data

Abstract Prescribing a combination of medications is a common strategy to enhance the effectiveness of disease treatment. While this approach aims to improve therapeutic outcomes, it may lead to unintended adverse drug reactions (ADRs) due to interactions between prescribed drugs and individual factors. Recognizing the importance of ADRs, various machine learning models have been developed. However, these models face limitations, particularly in predicting reactions induced by three or more drugs and in accounting for individual patient factors, such as age and medical history. To address these limitations, we propose a systematic framework that leverages medical data for predicting ADR signals. In the framework, the MIMIC-IV database was systematically preprocessed, and a series of machine learning models were developed that predict the ADR signals. The model predictions were validated using the eICU collaborative research database. The machine learning models process multiple inputs, including information on administered drugs, age, gender, ethnicity, and underlying medical conditions to predict ADR signals. The ADR signals are determined by classifying abnormalities in 20 specific laboratory test values, such as hematocrit, creatinine, and hemoglobin. The machine learning models developed in this study hold promise as a valuable tool for assessing potential risks, such as ADRs, associated with the concurrent use of multiple drugs. Citation Format: Junhyeok Jeon, Eujin Hong, Hyun Uk Kim. A systematic framework for predicting adverse drug reaction signals using medical data [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 4923.

Development and validation a nomogram prediction model for early diagnosis of bloodstream infections in the intensive care unit.

This study aims to develop and validate a nomogram for predicting the risk of bloodstream infections (BSI) in critically ill patients based on their admission status to the Intensive Care Unit (ICU). Patients' data were extracted from the Medical Information Mart for Intensive Care-IV (MIMIC-IV) database (training set), the Beijing Friendship Hospital (BFH) database (validation set) and the eICU Collaborative Research Database (eICU-CRD) (validation set). Univariate logistic regression analyses were used to analyze the influencing factors, and lasso regression was used to select the predictive factors. Model performance was assessed using area under receiver operating characteristic curve (AUROC) and Presented as a Nomogram. Various aspects of the established predictive nomogram were evaluated, including discrimination, calibration, and clinical utility. The model dataset consisted of 14930 patients (1444 BSI patients) from the MIMIC-IV database, divided into the training and internal validation datasets in a 7:3 ratio. The eICU dataset included 2100 patients (100 with BSI) as the eICU validation dataset, and the BFH dataset included 419 patients (21 with BSI) as the BFH validation dataset. The nomogram was constructed based on Glasgow Coma Scale (GCS), sepsis related organ failure assessment (SOFA) score, temperature, heart rate, respiratory rate, white blood cell (WBC), red width of distribution (RDW), renal replacement therapy and presence of liver disease on their admission status to the ICU. The AUROCs were 0.83 (CI 95%:0.81-0.84) in the training dataset, 0.88 (CI 95%:0.88-0.96) in the BFH validation dataset, and 0.75 (95%CI 0.70-0.79) in the eICU validation dataset. The clinical effect curve and decision curve showed that most areas of the decision curve of this model were greater than 0, indicating that this model has a certain clinical effectiveness. The nomogram developed in this study provides a valuable tool for clinicians and nurses to assess individual risk, enabling them to identify patients at a high risk of bloodstream infections in the ICU.

Caregiver Burden Associated With Pediatric Chronic Pain: A Retrospective Study Using the Pediatric Electronic Persistent Pain Outcomes Collaboration Database.

This retrospective, cross-sectional study investigated the nature and extent of burden experienced by caregivers of children and adolescents with chronic pain, and factors associated with increased caregiver burden. The Pediatric Electronic Persistent Pain Outcomes Collaboration database provided prospectively collected data from 1929 families attending 9 pediatric chronic pain services across Australia and New Zealand. Data included demographic information, responses to child pain and functioning measures, caregiver work impairment, and psychosocial functioning. Caregivers of children with chronic pain reported work impairment associated with their child's pain (mean: 15% ± SD 25 absenteeism; 38% ± SD 29 productivity lost), significantly worse than published international population norms (large-scale community survey data), most other caregiver samples of adults and children with other chronic conditions, and adult samples with various pain conditions. Caregivers reported considerable burden in multiple psychosocial functioning domains, particularly leisure functioning, pain-related catastrophizing, and adverse parenting behaviors (with greater pain-related avoidance). Caregiver psychosocial burden was significantly associated with child psychosocial functioning (β = -0.308, P < 0.01), school absenteeism (β = 0.161, P < 0.01), physical disability (β = 0.096, P < 0.05), and pain duration (β = 0.084, P < 0.05), but not pain intensity. Caregiver work productivity loss was significantly associated with school absenteeism (β = 0.290, P < 0.01), child physical disability (β = 0.148, P < 0.01), child health care utilization (β = 0.118, P < 0.05), and worst pain intensity (β = 0.101, P < 0.05). These results highlight the significant and varied impacts experienced by caregivers of children with chronic pain. This work is novel in reporting significant work impairment and confirms psychosocial burden in a larger sample than previous studies.