Cost Analysis on Intensive Care Unit Costs Based on the Length of Stay

Machine Learning Model for Risk Prediction of Prolonged Intensive Care Unit in Patients Receiving Intra-aortic Balloon Pump Therapy during Coronary Artery Bypass Graft Surgery

This study aimed to construct machine learning models and predict prolonged intensive care units (ICU) stay in patients receiving perioperative intra-aortic balloon pump (IABP) therapy during cardiac surgery. 236 patients were divided into the normal (≤ 14 days) and prolonged (> 14 days) ICU groups based on the 75th percentile of ICU duration across the entire cohort. Seven machine learning models were trained and validated. The Shapley Additive explanations (SHAP) method was employed to illustrate the effects of the features. 94 patients (39.83%) experienced prolonged ICU stay. The XGBoost model outperformed other models in predictive performance, as evidenced by its highest area under the receiver operating characteristic curve (training: 0.92; validation: 0.73). The SHAP analysis identified tracheotomy, albumin, Sv1, and cardiac troponin T as the top four risk variables. The XGBoost model predicted risk variables for prolonged ICU stay in patients, possibly contributing to improving perioperative management and reducing ICU duration.

Management of External Ventricular Drains for Neuromonitoring and Traumatic Brain Injury Treatment in Pediatric Patients Outside of Intensive Care Units: A Single-Institution Retrospective Study

BACKGROUND

Most pediatric hospitals manage patients who require external ventricular drains (EVDs) exclusively within pediatric intensive care units (PICUs) because of institutional protocols. Our institution commonly manages patients with EVDs on the neurotrauma floor (NTF). We evaluated whether this practice results in more EVD-associated complications.

METHODS

A retrospective cohort study at our Level 1 pediatric trauma center identified all trauma patients ≤18 years old who received an EVD in 2018-2023. Demographics, presenting characteristics, in-hospital management, and EVD management details were recorded. The primary outcome was EVD-related complication events.

RESULTS

Of the 81 patients who had EVDs placed after neurotrauma, 45 had their EVD managed exclusively in the PICU (PICU-EVD) and 36 had their EVD for some time while on the NTF (NTF-EVD). The groups were similar in sex (p = 0.87) and age (p = 0.054). PICU-EVD patients underwent fewer neurosurgeries (55.6% vs. 77.8%, p = 0.04) but spent more time on ventilators (10.6 ± 8.7 days vs. 6.4 ± 4.8, p = 0.02) and in the PICU (11.8 ± 9.0 days vs. 8.4 ± 5.9, p = 0.02). Total hospital stay was similar between groups (p = 0.44). NTF-EVD patients were on the drain longer (9.0 ± 7.4 days vs. 13.1 ± 9.1, p = 0.03), including 5.9 days on the NTF. Four EVD-related complications occurred overall: 2 accidental dislodgements and 2 cerebrospinal fluid leaks. EVD complication rates were similar on the NTF and PICU (2.2% vs. 8.3%, p = 0.21). All complications occurred late in the hospital course and were minor. A Poisson regression model comparing complication rates between PICU-only and NTF management (433 vs. 441 catheter days, respectively) found a complication rate of 6.8 per 1000 catheter days in the NTF group versus 2.3 per 1000 catheter days in the PICU-only group, yielding a rate ratio of 2.95 (95% confidence interval 0.29-30.4, p = 0.35). However, this difference was not statistically significant.

CONCLUSION

Our center routinely discharges patients from the PICU to the NTF with EVDs in place. This practice may be associated with no increased risk or rate of EVD-related complications compared to PICU-only management.

LEVEL OF EVIDENCE

IV.

Frequency, financial impact, and factors associated with cost outliers in intensive care units: a cohort study in Belgium

OBJECTIVE

This study aimed to explore the association between high outliers and intensive care unit admissions and to identify the factors contributing to high intensive care unit costs.

METHODS

This retrospective cohort study used data from 17 Belgian hospitals from 2018 and 2019. The study focused on the 10 most frequently admitted diagnosis-related groups in the intensive care unit. The dataset included medical discharge summaries and cost per stay from the hospital perspective.

RESULTS

A total of 39,279 hospital stays were analyzed, 11,124 of which were intensive care unit admissions; additionally, 2,500 of these stays were high outliers. The proportion of high outliers was significantly greater in the intensive care unit group, and admission to the intensive care unit was significantly associated with high outliers in the multivariate analyses. Factors associated with high intensive care unit outliers included the medical diagnosis-related group category, patients from nursing homes, intensive care unit stay duration exceeding 4 days, and specific technical procedures (measurement of intracranial pressure, continuous hemofiltration, and mechanical ventilation).

CONCLUSION

Admission to the intensive care unit increases the likelihood of being classified as an outlier, thus significantly impacting hospital costs. This study identified factors that can be used to predict intensive care unit outliers, which can enable adjustments to diagnosis-related group-based funding for intensive care units.

Adenomyosis among patients undergoing postpartum hysterectomy for uncontrollable uterine bleeding: A multicenter, observational, retrospective, cohort study on histologically-based prevalence and clinical characteristics

OBJECTIVE

To assess the prevalence of adenomyosis at pathologic examination, and its association with obstetric complications, peripartum maternal clinical characteristics and neonatal birth weight in patients undergoing postpartum hysterectomy due to postpartum hemorrhage (PPH).

METHODS

A multicenter, observational, retrospective, cohort study was carried out including all women who underwent postpartum hysterectomy due to PPH at gestational week 23+0 or later, between January 2010 and May 2023. Patients were categorized into two groups based on the presence of adenomyosis at pathologic examination, and were compared for obstetric complications, peripartum maternal clinical characteristics, and neonatal birth weight.

RESULTS

The histologically-based prevalence of adenomyosis in patients undergoing postpartum hysterectomy due to PPH was 39.4%. Adenomyosis was associated with a longer hospitalization time (regression coefficient: 4.43 days, 95% CI: 0.34-8.52, P = 0.034) and a higher risk of hypertensive disorders (OR: 5.82, 95% CI: 1.38-24.46, P = 0.016), threatened preterm labor (OR: 3.34, 95% CI: 1.08-10.31, P = 0.036), urgent/emergency C-section (OR: 24.15, 95% CI: 2.60-223.96, P = 0.005), postpartum maternal complications (OR: 4.96, 95% CI: 1.48-16.67, P = 0.012), maternal intensive care unit admission (OR: 3.56, 95% CI: 1.05-12.05, P = 0.041), and low birth weight neonates (OR: 3.8, 95% CI: 1.32-11.02, P = 0.013).

CONCLUSION

In patients undergoing postpartum hysterectomy due to PPH, adenomyosis is a highly prevalent condition among, and is associated with adverse obstetric, maternal, and neonatal outcomes.

Cost comparisons and factors related to cost per stay in intensive care units in Belgium

BACKGROUND

Given the variability of intensive care unit (ICU) costs in different countries and the importance of this information for guiding clinicians to effective treatment and to the organisation of ICUs at the national level, it is of value to gather data on this topic for analysis at the national level in Belgium. The objectives of the study were to assess the total cost of ICUs and the factors that influence the cost of ICUs in hospitals in Belgium.

METHODS

This was a retrospective cohort study using data collected from the ICUs of 17 Belgian hospitals from January 01 to December 31, 2018. A total of 18,235 adult ICU stays were included in the study. The data set was a compilation of inpatient information from analytical cost accounting of hospitals, medical discharge summaries, and length of stay data. The costs were evaluated as the expenses related to the management of hospital stays from the hospital's point of view. The cost from the hospital perspective was calculated using a cost accounting analytical methodology in full costing. We used multivariate linear regression to evaluate factors associated with total ICU cost per stay. The ICU cost was log-transformed before regression and geometric mean ratios (GMRs) were estimated for each factor.

RESULTS

The proportion of ICU beds to ward beds was a median [p25-p75] of 4.7% [4.4-5.9]. The proportion of indirect costs to total costs in the ICU was 12.1% [11.4-13.3]. The cost of nurses represented 57.2% [55.4-62.2] of direct costs and this was 15.9% [12.0-18.2] of the cost of nurses in the whole hospital. The median cost per stay was €4,267 [2,050-9,658] and was €2,160 [1,545-3,221] per ICU day. The main factors associated with higher cost per stay in ICU were Charlson score, mechanical ventilation, ECMO, continuous hemofiltration, length of stay, readmission, ICU mortality, hospitalisation in an academic hospital, and diagnosis of coma/convulsions or intoxication.

CONCLUSIONS

This study demonstrated that, despite the small proportion of ICU beds in relation to all services, the ICU represented a significant cost to the hospital. In addition, this study confirms that nursing staff represent a significant proportion of the direct costs of the ICU. Finally, the total cost per stay was also important but highly variable depending on the medical factors identified in our results.

A risk nomogram for predicting prolonged intensive care unit stays in patients with chronic obstructive pulmonary disease

BACKGROUND

Providing intensive care is increasingly expensive, and the aim of this study was to construct a risk column line graph (nomograms)for prolonged length of stay (LOS) in the intensive care unit (ICU) for patients with chronic obstructive pulmonary disease (COPD).

METHODS

This study included 4,940 patients, and the data set was randomly divided into training (n = 3,458) and validation (n = 1,482) sets at a 7:3 ratio. First, least absolute shrinkage and selection operator (LASSO) regression analysis was used to optimize variable selection by running a tenfold k-cyclic coordinate descent. Second, a prediction model was constructed using multifactorial logistic regression analysis. Third, the model was validated using receiver operating characteristic (ROC) curves, Hosmer-Lemeshow tests, calibration plots, and decision-curve analysis (DCA), and was further internally validated.

RESULTS

This study selected 11 predictors: sepsis, renal replacement therapy, cerebrovascular disease, respiratory failure, ventilator associated pneumonia, norepinephrine, bronchodilators, invasive mechanical ventilation, electrolytes disorders, Glasgow Coma Scale score and body temperature. The models constructed using these 11 predictors indicated good predictive power, with the areas under the ROC curves being 0.826 (95%CI, 0.809-0.842) and 0.827 (95%CI, 0.802-0.853) in the training and validation sets, respectively. The Hosmer-Lemeshow test indicated a strong agreement between the predicted and observed probabilities in the training (χ2 = 8.21, p = 0.413) and validation (χ2 = 0.64, p = 0.999) sets. In addition, decision-curve analysis suggested that the model had good clinical validity.

CONCLUSION

This study has constructed and validated original and dynamic nomograms for prolonged ICU stay in patients with COPD using 11 easily collected parameters. These nomograms can provide useful guidance to medical and nursing practitioners in ICUs and help reduce the disease and economic burdens on patients.

Economic Burden of ICU-Hospitalized COVID-19 Patients: A Systematic Review and Meta-Analysis

The impact of the coronavirus disease 2019 (COVID-19) pandemic on the global economy is far-reaching and difficult to assess accurately. We aimed to systematically determine the magnitude of the costs and the economic burden of intensive care for hospitalized COVID-19 patients since the onset of the pandemic by means of a systematic review. We conducted a PRISMA 2020-compliant (protocol: PROSPERO CRD42022348741) systematic review by searching PubMed, EMBASE, and Web of Science for relevant literature. We included studies that presented costs based on a primary partial economic evaluation. Using the Consolidated Health Economic Evaluation Reporting Standards checklist and the population, intervention, control, and outcome criteria, we established the risk of bias in studies at the individual level. Daily cost per ICU admission and total cost per ICU patient of the original studies extracted. A random effect model was adopted for meta-analysis whenever possible. Of the 1,635 unique records identified, 14 studies related to ICU-hospitalized costs due to COVID-19 were eligible for inclusion. Included studies represented 93,721 hospitalized COVID-19 patients. Regarding total direct medical costs, the lowest cost per patient at ICU was observed in Turkey ($2,984.78 ± 2,395.93), while the highest was in Portugal ($51,358.52 ± 30,150.38). The Republic of Korea reported the highest length of stay of 29.4 days (±17.80), and the lowest is observed in India for nine days (±5.98). Our findings emphasize COVID-19's significance on health-economic outcomes. Limited research exists on the economic burden of COVID-19 in the ICU. Further studies on cost estimates can enhance data clarity, enabling informed analysis of healthcare costs and aiding efficient patient care organization by care providers and policymakers.

Mortality Rates in Early versus Late Intensive Care Unit Readmission

Background

ICU readmission is associated with poor outcomes. Few studies have directly compared the outcomes of early versus late readmissions, especially in Saudi Arabia.

Objective

To compare the outcomes between early and late ICU readmissions, mainly with regards to hospital mortality.

Methods

This retrospective study included unique patients who, within the same hospitalization, were admitted to the ICU, discharged to the general wards, and then readmitted to the ICU of King Saud Medical City, Riyadh, Saudi Arabia, between January 01, 2015, and June 30, 2022. Patients readmitted within 2 calendar days were grouped into the Early readmission group, while those readmitted after 2 calendar days were in the Late readmission group.

Results

A total of 997 patients were included, of which 753 (75.5%) belonged to the Late group. The mortality rate in the Late group was significantly higher than that in the Early group (37.6% vs. 29.5%, respectively; 95% CI: 1%-14.8%; P = 0.03). The readmission length of stay (LOS) and severity score of both groups were similar. The odds ratio of mortality for the Early group was 0.71 (95% CI: 0.51-0.98, P = 0.04); other significant risk factors were age (OR = 1.023, 95% CI: 1.016-1.03; P < 0.001) and readmission LOS (OR = 1.017, 95% CI: 1.009-1.026; P < 0.001). The most common reason for readmission in the Early group was high Modified Early Warning Score, while in the Late group, it was respiratory failure followed by sepsis or septic shock.

Conclusion

Compared with late readmission, early readmission was associated with lower mortality, but not with lower LOS or severity score.

The Impact of Restrictive Transfusion Practices on Hemodynamically Stable Critically Ill Children Without Heart Disease: A Secondary Analysis of the Age of Blood in Children in the PICU Trial

OBJECTIVES

Guidelines recommend against RBC transfusion in hemodynamically stable (HDS) children without cardiac disease, if hemoglobin is greater than or equal to 7 g/dL. We sought to assess the clinical and economic impact of compliance with RBC transfusion guidelines.

DESIGN

A nonprespecified secondary analysis of noncardiac, HDS patients in the randomized trial Age of Blood in Children (NCT01977547) in PICUs. Costs analyzed included ICU stay and physician fees. Stabilized inverse propensity for treatment weighting was used to create a cohort balanced with respect to potential confounding variables. Weighted regression models were fit to evaluate outcomes based on guideline compliance.

SETTING

Fifty international tertiary care centers.

PATIENTS

Critically ill children 3 days to 16 years old transfused RBCs at less than or equal to 7 days of ICU admission. Six-hundred eighty-seven subjects who met eligibility criteria were included in the analysis.

INTERVENTIONS

Initial RBC transfusions administered when hemoglobin was less than 7 g/dL were considered "compliant" or "non-compliant" if hemoglobin was greater than or equal to 7 g/dL.

MEASUREMENTS AND MAIN RESULTS

Frequency of new or progressive multiple organ system dysfunction (NPMODS), ICU survival, and associated costs. The hypothesis was formulated after data collection but exposure groups were masked until completion of planned analyses. Forty-nine percent of patients (338/687) received a noncompliant initial transfusion. Weighted cohorts were balanced with respect to confounding variables (absolute standardized differences < 0.1). No differences were noted in NPMODS frequency (relative risk, 0.86; 95% CI, 0.61-1.22; p = 0.4). Patients receiving compliant transfusions had more ICU-free days (mean difference, 1.73; 95% CI, 0.57-2.88; p = 0.003). Compliance reduced mean costs in ICU by $38,845 U.S. dollars per patient (95% CI, $65,048-$12,641).

CONCLUSIONS

Deferring transfusion until hemoglobin is less than 7 g/dL is not associated with increased organ dysfunction in this population but is independently associated with increased likelihood of live ICU discharge and lower ICU costs.

Cost-benefit analysis of Intensive Care Unit with Activity-Based Costing approach in the era COVID-19 pandemic: A case study from Iran

BACKGROUND

Providing intensive care to acute patients is a vital part of health systems. However, the high cost of Intensive Care Units (ICU) has limited their development, especially in low-income countries. Due to the increasing need for intensive care and limited resources, ICU cost management is important. This study aimed to analyze the cost-benefit of ICU during COVID-19 in Tehran, Iran.

METHODS

This cross-sectional study is an economic evaluation of health interventions. The study was conducted in the COVID-19 dedicated ICU, from the provider's point of view and within one-year horizon. Costs were calculated using a top-down approach and the Activity-Based Costing technique. Benefits were extracted from the hospital's HIS system. Benefit Cost ratio (BCR) and Net Present Value (NPV) indexes were used for cost-benefit analysis (CBA). A sensitivity analysis was performed to evaluate the dependence of the CBA results on the uncertainties in the cost data. Analysis was performed with Excel and STATA software.

RESULTS

The studied ICU had 43 personnel, 14 active beds, a 77% bed occupancy rate, and 3959 occupied bed days. The total costs were $2,372,125.46 USD, of which 70.3% were direct costs. The highest direct cost was related to human resources. The total net income was $1,213,314.13 USD. NPV and BCR were obtained as $-1,158,811.32 USD and 0.511 respectively.

CONCLUSION

Despite operating with a relatively high capacity, ICU has had high losses during the COVID-19. Proper management and re-planning in the structure of human resources is recommended due to its importance in the hospital economy, provision of resources based on needs assessment, improvement of drugs management, reduction of insurance deductions in order to reduce costs and improve ICU productivity.

Predicting length of stay in ICU and mortality with temporal dilated separable convolution and context-aware feature fusion

In healthcare, Intensive Care Unit (ICU) bed management is a necessary task because of the limited budget and resources. Predicting the remaining Length of Stay (LoS) in ICU and mortality can assist clinicians in managing ICU beds efficiently. This study proposes a deep learning method based on several successive Temporal Dilated Separable Convolution with Context-Aware Feature Fusion (TDSC-CAFF) modules, and a multi-view and multi-scale feature fusion for predicting the remaining LoS and mortality risk for ICU patients. In each TDSC-CAFF module, temporal dilated separable convolution is used to encode each feature separately, and context-aware feature fusion is proposed to capture comprehensive and context-aware feature representations from the input time-series features, static demographics, and the output of the last TDSC-CAFF module. The CAFF outputs of each module are accumulated to achieve multi-scale representations with different receptive fields. The outputs of TDSC and CAFF are concatenated with skip connection from the output of the last module and the original time-series input. The concatenated features are processed by the proposed Point-Wise convolution-based Attention (PWAtt) that captures the inter-feature context to generate the final temporal features. Finally, the final temporal features, the accumulated multi-scale features, the encoded diagnosis, and static demographic features are fused and then processed by fully connected layers to obtain prediction results. We evaluate our proposed method on two publicly available datasets: eICU and MIMIC-IV v1.0 for LoS and mortality prediction tasks. Experimental results demonstrate that our proposed method achieves a mean squared log error of 0.07 and 0.08 for LoS prediction, and an Area Under the Receiver Operating Characteristic Curve of 0.909 and 0.926 for mortality prediction, on eICU and MIMIC-IV v1.0 datasets, respectively, which outperforms several state-of-the-art methods.

Explainable time-series deep learning models for the prediction of mortality, prolonged length of stay and 30-day readmission in intensive care patients

Background

In-hospital mortality, prolonged length of stay (LOS), and 30-day readmission are common outcomes in the intensive care unit (ICU). Traditional scoring systems and machine learning models for predicting these outcomes usually ignore the characteristics of ICU data, which are time-series forms. We aimed to use time-series deep learning models with the selective combination of three widely used scoring systems to predict these outcomes.

Materials and methods

A retrospective cohort study was conducted on 40,083 patients in ICU from the Medical Information Mart for Intensive Care-IV (MIMIC-IV) database. Three deep learning models, namely, recurrent neural network (RNN), gated recurrent unit (GRU), and long short-term memory (LSTM) with attention mechanisms, were trained for the prediction of in-hospital mortality, prolonged LOS, and 30-day readmission with variables collected during the initial 24 h after ICU admission or the last 24 h before discharge. The inclusion of variables was based on three widely used scoring systems, namely, APACHE II, SOFA, and SAPS II, and the predictors consisted of time-series vital signs, laboratory tests, medication, and procedures. The patients were randomly divided into a training set (80%) and a test set (20%), which were used for model development and model evaluation, respectively. The area under the receiver operating characteristic curve (AUC), sensitivity, specificity, and Brier scores were used to evaluate model performance. Variable significance was identified through attention mechanisms.

Results

A total of 33 variables for 40,083 patients were enrolled for mortality and prolonged LOS prediction and 36,180 for readmission prediction. The rates of occurrence of the three outcomes were 9.74%, 27.54%, and 11.79%, respectively. In each of the three outcomes, the performance of RNN, GRU, and LSTM did not differ greatly. Mortality prediction models, prolonged LOS prediction models, and readmission prediction models achieved AUCs of 0.870 ± 0.001, 0.765 ± 0.003, and 0.635 ± 0.018, respectively. The top significant variables co-selected by the three deep learning models were Glasgow Coma Scale (GCS), age, blood urea nitrogen, and norepinephrine for mortality; GCS, invasive ventilation, and blood urea nitrogen for prolonged LOS; and blood urea nitrogen, GCS, and ethnicity for readmission.

Conclusion

The prognostic prediction models established in our study achieved good performance in predicting common outcomes of patients in ICU, especially in mortality prediction. In addition, GCS and blood urea nitrogen were identified as the most important factors strongly associated with adverse ICU events.

Association between nursing cost and patient outcomes in intensive care units: A retrospective cohort study of Belgian hospitals

INTRODUCTION

Hospitals with better nursing resources report more favourable patient outcomes with almost no difference in cost as compared to those with worse nursing resources. The aim of this study was to assess the association between nursing cost per intensive care unit bed and patient outcomes (mortality, readmission, and length of stay).

METHODOLOGY

This was a retrospective cohort study using data collected from the intensive care units of 17 Belgian hospitals from January 01 to December 31, 2018. Hospitals were dichotomized using median annual nursing cost per bed. A total of 18,235 intensive care unit stays were included in the study with 5,664 stays in the low-cost nursing group and 12,571 in the high-cost nursing group.

RESULTS

The rate of high length of stay outliers in the intensive care unit was significantly lower in the high-cost nursing group (9.2% vs 14.4%) compared to the low-cost nursing group. Intensive care unit readmission was not significantly different in the two groups. Mortality was lower in the high-cost nursing group for intensive care unit (9.9% vs 11.3%) and hospital (13.1% vs 14.6%) mortality. The nursing cost per intensive care bed was different in the two groups, with a median [IQR] cost of 159,387€ [140,307-166,690] for the low-cost nursing group and 214,032€ [198,094-230,058] for the high-cost group. In multivariate analysis, intensive care unit mortality (OR = 0.80, 95% CI: 0.69-0.92, p < 0.0001), in-hospital mortality (OR = 0.82, 95% CI: 0.72-0.93, p < 0.0001), and high length of stay outliers (OR = 0.48, 95% CI: 0.42-0.55, p < 0.0001) were lower in the high-cost nursing group. However, there was no significant effect on intensive care readmission between the two groups (OR = 1.24, 95% CI: 0.97-1.51, p > 0.05).

CONCLUSIONS

This study found that higher-cost nursing per bed was associated with significantly lower intensive care unit and in-hospital mortality rates, as well as fewer high length of stay outliers, but had no significant effect on readmission to the intensive care unit. .

Associations between two nursing workload scales and the cost of intensive care unit nursing staff: A retrospective study of one Belgian hospital

AIMS

The aim of this study was to assess associations between a general nursing funding scale and an intensive care unit specific nursing workload scale and the cost of nursing staff.

BACKGROUND

Nurse staffing represents the most important cost in the intensive care unit, so it is essential to evaluate it accurately. In addition, the assessment of nursing workload is important for the daily management of the intensive care unit and to ensure quality of care.

METHODS

This was a retrospective and quantitative study carried out in the intensive care unit of a Belgian hospital. The extraction of data from the Nursing Activities Score and the Minimum Hospital Summary Nursing Dataset were carried out during 2 periods of 15 days, from 1 June 2018 to 15 June 2018 and from 1 September 2018 to 15 September 2018.

RESULTS

A total of 234 patients were included in the study. A total of 773 Nursing Activities Score and Minimum Hospital Summary Nursing Dataset recordings were analyzed in the study per intensive care unit day. A strong correlation was observed between Nursing Activities Score and Minimum Hospital Summary Nursing Dataset for the entire intensive care unit stay with a rho (95% CI) of .88 (0.83-.93); however, the correlation was moderate per intensive care unit day with a rho of .51 (0.45-0.57). A strong association was observed between the Minimum Hospital Summary Nursing Dataset and the Nursing Activities Score with the costs of intensive care unit nurses with a rho (95% CI) of .78 (0.72-0.86) and .74 (0.65-0.84), respectively.

CONCLUSIONS

A general nursing funding scale in Belgium was strongly correlated with the nursing workload for the whole intensive care unit stay, but this correlation was moderate per intensive care unit day. In contrast, both scales showed a good correlation with intensive care unit nursing costs.

IMPLICATIONS FOR NURSING MANAGEMENT

In Belgium, a general funding scale for nurses does not allow for an assessment of the nursing workload in the intensive care unit. The Nursing Activities Score is strongly correlated with the cost of nursing staff in the intensive care unit. The authors recommend that the Belgian authorities carry out this type of study in several intensive care units in the country and eventually replace the general funding scale for nurses with the Nursing Activities Score.

Predicting Prolonged Length of ICU Stay through Machine Learning

This study aimed to construct machine learning (ML) models for predicting prolonged length of stay (pLOS) in intensive care units (ICU) among general ICU patients. A multicenter database called eICU (Collaborative Research Database) was used for model derivation and internal validation, and the Medical Information Mart for Intensive Care (MIMIC) III database was used for external validation. We used four different ML methods (random forest, support vector machine, deep learning, and gradient boosting decision tree (GBDT)) to develop prediction models. The prediction performance of the four models were compared with the customized simplified acute physiology score (SAPS) II. The area under the receiver operation characteristic curve (AUROC), area under the precision-recall curve (AUPRC), estimated calibration index (ECI), and Brier score were used to measure performance. In internal validation, the GBDT model achieved the best overall performance (Brier score, 0.164), discrimination (AUROC, 0.742; AUPRC, 0.537), and calibration (ECI, 8.224). In external validation, the GBDT model also achieved the best overall performance (Brier score, 0.166), discrimination (AUROC, 0.747; AUPRC, 0.536), and calibration (ECI, 8.294). External validation showed that the calibration curve of the GBDT model was an optimal fit, and four ML models outperformed the customized SAPS II model. The GBDT-based pLOS-ICU prediction model had the best prediction performance among the five models on both internal and external datasets. Furthermore, it has the potential to assist ICU physicians to identify patients with pLOS-ICU risk and provide appropriate clinical interventions to improve patient outcomes.

Critical care nursing service costs: Comparison of the top-down versus bottom-up micro-costing approach in Brazil

AIM

To estimate the nursing service costs using a top-down micro-costing approach and to compare it with a bottom-up micro-costing approach.

BACKGROUND

Accurate data of nursing cost can contribute to reliable resource management.

METHOD

We employed a retrospective cohort design in an adult intensive care unit in São Paulo. A total of 286 patient records were included. Micro-costing analysis was conducted in two stages: a top-down approach, whereby nursing costs were allocated to patients through apportionment, and a bottom-up approach, considering actual nursing care hours estimated by the Nursing Activities Score (NAS).

RESULTS

The total mean cost by the top-down approach was US$1,640.4 ± 1,484.2/patient. The bottom-up approach based on a total mean NAS of 833 ± 776 points (equivalent to 200 ± 86 hr of nursing care) yielded a mean cost of US$1,487.2 ± 1,385.7/patient. In the 268 patients for whom the top-down approach estimated higher costs than the bottom-up approach, the total cost discrepancy was US$4,427.3, while for those costed higher based on NAS, the total discrepancy was US$436.9. The top-down methodology overestimated costs for patients requiring lower intensity of care, while it underestimated costs for patients requiring higher intensity of care (NAS >100).

CONCLUSIONS

The top-down approach may yield higher estimated ICU costs compared with a NAS-based bottom-up approach.

IMPLICATIONS FOR NURSING MANAGEMENT

These findings can contribute to an evidence-based approach to budgeting through reliable costing methods based on actual nursing workload, and to efficient resource allocation and cost management.

The Implementation of Protocol-Based Utilization of Neuromuscular Blocking Agent Using Clinical Variables in Acute Respiratory Distress Syndrome Patients

Supplemental Digital Content is available in the text.

Objectives:

The recent conflicting data on the mortality benefit of neuromuscular blocking agents in acute respiratory distress syndrome and the potential adverse effects of continuous neuromuscular blocking agent necessitates that these medications should be used judiciously with dose reduction in mind. The aims of the study were to improve the process of care by provider education of neuromuscular blocking agent titration and monitoring and to determine the impact of clinical endpoint based neuromuscular blocking agent titration protocol.

Design:

We conducted a proof-of-concept historically controlled study of protocol-based intervention standardizing paralytic monitoring and titration using clinical variables. Education of the protocol was provided to ICU staff via bedside teaching and workshops. The primary outcomes were the time to reach goal paralysis and cumulative neuromuscular blocking agent dose. Secondary outcomes included maintenance of deeper sedation (Richmond Agitation and Sedation Scale –5) prior to neuromuscular blocking agent initiation, total time on mechanical ventilation, length of stay, and mortality.

Setting:

Medical ICU at a quaternary academic hospital between March 2019 and June 2020.

Patients:

Adult severe acute respiratory distress syndrome (Pao₂/Fio₂ <150) patients requiring neuromuscular blocking agent for greater than or equal to 12 hours. Eighty-two patients fulfilled inclusion criteria, 46 in the control group and 36 in the intervention group.

Interventions:

Education and implementation of standardized protocol.

Measurements and Main Results:

Compared with the control group, the time to reach goal paralysis in the intervention group was shorter (8.55 ± 9.4 vs 2.63 ± 5.9 hr; p < 0.0001) on significantly lower dose of cisatracurium (total dose 1,897.96 ± 1,241.0 vs 562.72 ± 546.7 mg; p < 0.0001 and the rate 5.84 ± 2.66 vs 1.99 ± 0.95 µg/kg/min; p < 0.0001). Deeper sedation was achieved at the time of initiation of neuromuscular blocking agent in the intervention arm (mean Richmond Agitation and Sedation Scale –3.3 ± 1.9 vs –4.3 ± 1.7; p = 0.015). There was no significant difference in total time on mechanical ventilation, length of ICU stay, length of hospital stay, and mortality between the two groups.

Conclusions:

Implementation of comprehensive education, standardization of sedation prior to neuromuscular blocking agent initiation, integration of clinical variables in determining paralysis achievement, and proper use of peripheral nerve stimulation served as optimal strategies for the titration and monitoring of neuromuscular blocking agent in acute respiratory distress syndrome. This reduced drug utilization while continuing to achieve benefit without causing adverse effects.

Profile of users receiving Speech-Language Therapy service at a Critical Patient Unit

ABSTRACT Objective: to describe the profile of patients treated by Speech-Language therapists in a Critical Patient Unit. Methods: an ex post facto, observational and descriptive study was carried out. Monthly statistical data of patients hospitalized in the period January-December 2018 were analyzed, in the Intensive Care Unit at a public hospital. Data were described from the analysis of frequency and measures of central tendency. The distribution of the variables was determined through the skewness-kurtosis test, considering a significance level of p<0.05. Results: 217 individuals got 868 speech-language therapy services. Men (57.26%), older than 65 years old, required a more frequent intervention. The main medical diagnosis of admission to the unit corresponded to non-specific pathologies (57.14%), respiratory disease (15.21%) and cerebrovascular disease (12.79%). The speech-language therapy functions were related to the evaluation of swallowing (54.31%) and voice (32.4%). In relation to the intervention, the treatment of dysphagia (25.82%) and oral motor functions (25.04%) was predominant in the duties. Functions associated with language, speech and cognition were secondary. Conclusion: the profile of the critical patient and the speech-language therapy work in this field represent a first step to characterize the role of the speech-language therapist in Intensive Medicine teams.

Benchmarking machine learning models on multi-centre eICU critical care dataset

Progress of machine learning in critical care has been difficult to track, in part due to absence of public benchmarks. Other fields of research (such as computer vision and natural language processing) have established various competitions and public benchmarks. Recent availability of large clinical datasets has enabled the possibility of establishing public benchmarks. Taking advantage of this opportunity, we propose a public benchmark suite to address four areas of critical care, namely mortality prediction, estimation of length of stay, patient phenotyping and risk of decompensation. We define each task and compare the performance of both clinical models as well as baseline and deep learning models using eICU critical care dataset of around 73,000 patients. This is the first public benchmark on a multi-centre critical care dataset, comparing the performance of clinical gold standard with our predictive model. We also investigate the impact of numerical variables as well as handling of categorical variables on each of the defined tasks. The source code, detailing our methods and experiments is publicly available such that anyone can replicate our results and build upon our work.