Optimized diagnosis-based comorbidity measures for all-cause mortality prediction in a national population-based ICU population

Explainable machine learning model for prediction of 28-day all-cause mortality in immunocompromised patients in the intensive care unit: a retrospective cohort study based on MIMIC-IV database

OBJECTIVES

This study aimed to develop and validate an explainable machine learning (ML) model to predict 28-day all-cause mortality in immunocompromised patients admitted to the intensive care unit (ICU). Accurate and interpretable mortality prediction is crucial for clinical decision-making and optimal allocation of critical care resources for this vulnerable patient population.

METHODS

We utilized retrospective clinical data from the MIMIC-IV (version 2.2) database, encompassing ICU admissions at Beth Israel Deaconess Medical Center from 2008 to 2019. Eligible immunocompromised patients, including those with primary immunodeficiencies and chronic acquired conditions, such as hematological malignancies, solid tumors, and organ transplantation, were selected. Data were randomly split into training (80%) and testing (20%) cohorts. Ten ML models (logistic regression, XGBoost, LightGBM, AdaBoost, Random Forest, Gradient Boosting, Gaussian Naive Bayes, Complement Naive Bayes, Multilayer Perceptron, and Support Vector Machine) were developed and evaluated using area under the receiver operating characteristic curve (AUROC), area under the precision-recall curve (AUPRC), sensitivity, specificity, accuracy, and F1 score. Model explainability was achieved through SHapley Additive exPlanations (SHAP), and decision curve analysis (DCA) assessed clinical utility. In addition, Cox proportional hazards regression was conducted to evaluate the impact of predictive factors on time-to-event outcomes.

RESULTS

Among the evaluated models, the Support Vector Machine (SVM) demonstrated the highest AUROC of 0.863 (95% CI 0.834-0.890) and a notable AUPRC of 0.678 (95% CI 0.624-0.736). Key predictive factors consistently identified across multiple ML models included 24-h urine output, blood urea nitrogen (BUN) levels, presence of metastatic solid tumors, Charlson Comorbidity Index (CCI), and international normalized ratio (INR). SHAP analyses provided detailed insights into how these features influenced model predictions.

CONCLUSIONS

The explainable ML models based on various artificial intelligence methods demonstrated promising clinical applicability in predicting 28-day mortality risk among immunocompromised ICU patients. Factors such as urine output, BUN, metastatic solid tumors, CCI, and INR significantly contributed to prediction outcomes and may serve as important predictors in clinical practice.

Stress hyperglycemia ratio and 30-day mortality among critically ill patients with acute heart failure: analysis of the MIMIC-IV database

BACKGROUND

The association between the stress hyperglycemia ratio (SHR) and short-term prognosis of acute heart failure (AHF), particularly among those admitted to the intensive care unit (ICU), has not been elucidated. This study aimed to investigate the association between the SHR and adverse outcomes among critically ill patients with AHF and provide a reference for glycemic management range in these patients.

METHODS

We extracted the clinical data of patients from the MIMIC-IV (v3.0) database. The association between the SHR and short-term prognosis was analyzed using the Kaplan‒Meier survival curve, Cox regression, and subgroup analysis. Important features were identified utilizing machine learning methods. Furthermore, the association between the dynamic SHR level and mortality was explored using restricted cubic splines and Cox regression.

RESULTS

A total of 994 patients were included. Patients with the highest SHR (Quartile 4) had a higher risk of 30-day mortality (HR = 2.14; 95% CI = 1.32-3.45; P = 0.002) and in-hospital mortality (HR = 2.22; 95% CI = 1.27-3.88; P = 0.005) than those in Quartile 2 (as reference). The results of machine learning methods revealed the SHR was an important predictor for 30-day mortality of patients with critical AHF. Restricted cubic splines indicated a J-shaped association between the dynamic SHR level and mortality, and the cut-off values were 0.84 and 1.07.

CONCLUSION

The SHR was significantly associated with 30-day mortality and in-hospital mortality among patients with critical AHF. The SHR may be a useful indicator for the glycemic management of patients with AHF in the ICU.

The long-term conditional mortality rate in older ICU patients compared to the general population

BACKGROUND

Understanding how preexisting comorbidities may interact with a critical illness is important for the assessment of long-term survival probability of older patients admitted to the ICU.

MATERIAL AND METHODS

The mortality after a first ICU admission in patients ≥ 55 years old registered in the Swedish Intensive Care Registry was compared to age- and sex-matched individuals from the general population with a landmark after 1 year. The comparison was adjusted for age, sex, and baseline comorbidity using Cox regression.

RESULTS

The 7-year study period included 140 008 patients, of whom 23% were 80 years or older. Patients surviving the first year remained at an increased risk compared to the general population, but much of this difference was attenuated after adjustment for baseline comorbidity (HR, 1.03; 95% CI 1.02-1.04). Excluding cardio-thoracic ICU admissions, the increased risk remained slightly elevated (adjusted HR, 1.15; 95% CI 1.13-1.16). Also, the subgroup ≥ 75 years old surviving the first year returned to a mortality rate comparable to the general population (HR, 0.98; 95% CI 0.96-0.99). Stratified by admission diagnosis an increased mortality rate remained beyond the first year for acute-on-chronic respiratory failure (adjusted HR, 1.47; 95% CI 1.36-1.58) but not for other respiratory causes (adjusted HR, 1.03; 95% CI 0.99-1.07) or admission for septic shock (adjusted HR, 1.04; 95% CI 0.95-1.13). No substantial increased mortality rate was notable beyond the first year for other admission diagnoses.

CONCLUSION

Older ICU patients that survive the first year after an ICU admission return to a mortality rate close to that of the general population having similar baseline comorbidity, but variability is seen depending on the ICU admission diagnosis. Trial registration ClinicalTrials.gov ID: NCT06234709, date 02/01/2024.

Implementation of Medical Hospitalist Care at a Korean Tertiary Hospital: A Retrospective Cross-Sectional Study

Background/Objectives: In March 2018, a tertiary teaching hospital launched a medical hospitalist team. This study presents the clinical characteristics and outcomes of medical hospitalist care and reveals the relationship between them. Methods: This study included 4003 patients first admitted to the hospitalist team via emergency room and then discharged from the hospitalist team between March 2018 and November 2022. The patients were admitted either to the teaching admitter hospitalist team or the hospitalist-led acute medical unit (AMU). Afterward, the patients were either discharged, if possible, within a few days or transferred to ward hospitalists if assigned wards for hospitalist care were available. Results: The patients' mean Charlson Comorbidity Index score was 3.5 and the mean National Early Warning Score was 3.4. Of the admissions, 44.2% of the patients were admitted to the AMU, and 26.8% received an early consultation with a subspecialist. Each hospitalist managed 12.8 patients per month on average. The patients' mean LOS was 14.52 days, 10.5% of patients died during hospitalization, and 13.0% of patients had unscheduled readmission within 1 month. The patients' mean total cost per hospital stay was 572,836 won per day. Admission to the AMU was associated with a lower total cost per hospital stay, but the relationships with mortality, readmission, and LOS were not significant. Conclusions: The study reports on the outcomes of implementing a medical hospitalist care system that combines short-term admission wards with integrated care models to manage complex cases. These findings provide insights into optimizing hospitalist systems for improved patient outcomes.

Ethical issues surrounding appropriate care for older persons in the Intensive Care Unit

Increasing numbers of older patients are being admitted to the Intensive Care Unit (ICU) as the world's population ages. The biological process of ageing, senescence, results in altered ability to maintain normal homeostasis and organ function, including of the cardiovascular, immune, and neuromuscular systems. This contributes towards increased frailty in older patients, associated with functional limitations and increased vulnerability. Although widely defined using chronological age, the concept of "old age" is thus multifactorial, including biological, but also psychological and sociocultural aspects, which should all be taken into account when considering what is appropriate in terms of ICU admission and management. As for all patients, but perhaps particularly in this subgroup, decisions regarding ICU admission and treatment and the withdrawing and withholding of life support must be individualized.

Early prediction of sepsis in intensive care patients using the machine learning algorithm NAVOY® Sepsis, a prospective randomized clinical validation study

PURPOSE

To prospectively validate, in an ICU setting, the prognostic accuracy of the sepsis prediction algorithm NAVOY® Sepsis which uses 4 h of input for routinely collected vital parameters, blood gas values, and lab values.

MATERIALS AND METHODS

Patients 18 years or older admitted to the ICU at Skåne University Hospital Malmö from December 2020 to September 2021 were recruited in the study. A total of 304 patients were randomized into one of two groups: Algorithm group with active sepsis alerts, or Standard of care. NAVOY® Sepsis made silent predictions in the Standard of care group, in order to evaluate its performance without disturbing the outcome. The study was blinded, i.e., study personnel did not know to which group patients were randomized. The healthcare provider followed standard practices in assessing possible development of sepsis and intervening accordingly. The patients were followed-up in the study until ICU discharge.

RESULTS

NAVOY® Sepsis could predict the development of sepsis, according to the Sepsis-3 criteria, three hours before sepsis onset with high performance: accuracy 0.79; sensitivity 0.80; and specificity 0.78.

CONCLUSIONS

The accuracy, sensitivity, and specificity were all high, validating the prognostic accuracy of NAVOY® Sepsis in an ICU setting, including Covid-19 patients.

Development and validation of a multi-dimensional diagnosis-based comorbidity index that improves prediction of death in men with prostate cancer: Nationwide, population-based register study

Assessment of comorbidity is crucial for confounding adjustment and prediction of mortality in register-based studies, but the commonly used Charlson comorbidity index is not sufficiently predictive. We aimed to develop a multidimensional diagnosis-based comorbidity index (MDCI) that captures comorbidity better than the Charlson Comorbidity index. The index was developed based on 286,688 men free of prostate cancer randomly selected from the Swedish general population, and validated in 54,539 men without and 68,357 men with prostate cancer. All ICD-10 codes from inpatient and outpatient discharges during 10 years prior to the index date were used to define variables indicating frequency of code occurrence, recency, and total duration of related hospital admissions. Penalized Cox regression was used to predict 10-year all-cause mortality. The MDCI predicted risk of death better than the Charlson comorbidity index, with a c-index of 0.756 (95% confidence interval [CI] = 0.751, 0.762) vs 0.688 (95% CI = 0.683, 0.693) in the validation cohort of men without prostate cancer. Men in the lowest vs highest MDCI quartile had distinctively different survival in the validation cohort of men with prostate cancer, with an overall hazard ratio [HR] of 5.08 (95% CI = 4.90, 5.26). This was also consistent within strata of age and Charlson comorbidity index, e.g. HR = 5.90 (95% CI = 4.65, 7.50) in men younger than 60 years with CCI 0. These results indicate that comorbidity assessment in register-based studies can be improved by use of all ICD-10 codes and taking related frequency, recency, and duration of hospital admissions into account.

Characteristics, clinical course and outcome of ventilated patients at a non-surgical intensive care unit in Germany: a single-centre, retrospective observational cohort analysis

OBJECTIVES

The objective of this study was to evaluate epidemiological characteristics, clinical course and outcome of mechanically ventilated non-surgical intensive care unit (ICU) patients, with the aim of improving the strategic planning of ICU capacities.

DESIGN

We conducted a retrospective observational cohort analysis. Data from mechanically ventilated intensive care patients were obtained by investigating electronic health records. The association between clinical parameters and ordinal scale data of clinical course was evaluated using Spearman correlation and Mann-Whitney U test. Relations between clinical parameters and in-hospital mortality rates were examined using binary logistic regression analysis.

SETTING

A single-centre study at the non-surgical ICU of the University Hospital of Frankfurt, Germany (tertiary care-level centre).

PARTICIPANTS

All cases of critically ill adult patients in need of mechanical ventilation during the years 2013-2015 were included. In total, 932 cases were analysed.

RESULTS

From a total of 932 cases, 260 patients (27.9%) were transferred from peripheral ward, 224 patients (24.1%) were hospitalised via emergency rescue services, 211 patients (22.7%) were admitted via emergency room and 236 patients (25.3%) via various transfers. In 266 cases (28.5%), respiratory failure was the reason for ICU admission. The length of stay was higher in non-geriatric patients, patients with immunosuppression and haemato-oncological disease or those in need of renal replacement therapy. 431 patients died, which corresponds to an all-cause in-hospital mortality rate of 46.2%. 92 of 172 patients with presence of immunosuppression (53.5%), 111 of 186 patients (59.7%) with pre-existing haemato-oncological disease, 27 of 36 patients (75.0%) under extracorporeal membrane oxygenation (ECMO) therapy, and 182 of 246 patients (74.0%) undergoing renal replacement therapy died. In logistic regression analysis, these subgroups and older age were significantly associated with higher mortality rates.

CONCLUSIONS

Respiratory failure was the main reason for ventilatory support at this non-surgical ICU. Immunosuppression, haemato-oncological diseases, the need for ECMO or renal replacement therapy and older age were associated with higher mortality.