PROPER: Development of an Early Pediatric Intensive Care Unit Readmission Risk Prediction Tool

Validity of Pediatric Early Warning Score in Predicting Unplanned Pediatric Intensive Care Unit Readmission

Despite the fact that unscheduled readmission to pediatric intensive care units (PICUs) has significant adverse consequences, there is a need for a predictive tool appropriate for use in the clinical setting. The aim of this study was to assess the ability of the modified Brighton pediatric early warning score (PEWS) to identify children at high risk for early unplanned readmission. In this retrospective cohort study, all patients aged 1 month to 18 years of age discharged from PICUs of two tertiary children's hospitals during the study interval were enrolled. Apart from demographic data, the association between PEWS and early readmission, defined as readmission within 48 hours of discharge, was analyzed by multivariable logistic regression. From 416 patients, 27 patients had early PICU readmission. Patients who experienced readmission were significantly younger than the controls. (≤12 months, 70.4 vs. 39.1%, p  = 0.001) Patients who were admitted from the emergency room (66.7 and 33.3% for emergency department (ED) and floor, respectively, p  = 0.012) had higher risk of early unplanned readmission. PEWS at discharge was significantly higher in patients who experienced readmission (3.07 vs. 0.8, p  < 0.001). A cut-off PEWS of 2, with sensitivity 85.2% and specificity 78.1%, determined the risk of unplanned readmission. Each 1-point increase in the PEWS at discharge significantly increases the risk of readmission (odds ratio [OR] = 3.58, 95% confidence interval [CI]: [2.42-5.31], p  < 0.001). PEWS can be utilized as a useful predictive tool regarding predicting unscheduled readmission in PICU. Further large-scale studies are needed to determine its benefits in clinical practice.

Frequency, Characteristics, and Outcomes of Patients Requiring Early PICU Readmission

OBJECTIVES

Readmission to the PICU is associated with worse outcomes, but factors associated with PICU readmission within the same hospitalization remain unclear. We sought to describe the prevalence of, and identify factors associated with, early PICU readmission.

METHODS

We performed a retrospective analysis of PICU admissions for patients aged 0 to 26 years in 48 tertiary care children's hospitals between January 1, 2016 and December 31, 2019 in the Pediatric Health Information System. We defined early readmission as return to the PICU within 2 calendar days of floor transfer during the same hospitalization. Generalized linear mixed models were used to analyze associations between patient and clinical variables, including complex chronic conditions (CCC) and early PICU readmission.

RESULTS

The results included 389 219 PICU admissions; early PICU readmission rate was 2.5%. Factors with highest odds of early PICU readmission were CCC, with ≥4 CCCs (reference: no CCC[s]) as highest odds of readmission (adjusted odds ratio [95% confidence interval]: 4.2 [3.8-4.5]), parenteral nutrition (2.3 [2.1-2.4]), and ventriculoperitoneal shunt (1.9 [1.7-2.2]). Factors with decreased odds of PICU readmission included extracorporeal membrane oxygenation (0.4 [0.3-0.6]) and cardiopulmonary resuscitation (0.8 [0.7-0.9]). Patients with early PICU readmissions had longer overall length of stay (geometric mean [geometric SD]: 18.2 [0.9] vs 5.0 [1.1] days, P < .001) and increased odds of mortality (1.7 [1.5-1.9]).

CONCLUSIONS

Although early PICU readmissions within the same hospitalization are uncommon, they are associated with significantly worse clinical outcomes. Patients with medical complexity and technology dependence are especially vulnerable.

Development and validation of a nomogram for predicting hospitalization longer than 14 days in pediatric patients with ventricular septal defect-a study based on the PIC database

Background: Ventricular septal defect is a common congenital heart disease. As the disease progresses, the likelihood of lung infection and heart failure increases, leading to prolonged hospital stays and an increased likelihood of complications such as nosocomial infections. We aimed to develop a nomogram for predicting hospital stays over 14 days in pediatric patients with ventricular septal defect and to evaluate the predictive power of the nomogram. We hope that nomogram can provide clinicians with more information to identify high-risk groups as soon as possible and give early treatment to reduce hospital stay and complications. Methods: The population of this study was pediatric patients with ventricular septal defect, and data were obtained from the Pediatric Intensive Care Database. The resulting event was a hospital stay longer than 14 days. Variables with a variance inflation factor (VIF) greater than 5 were excluded. Variables were selected using the least absolute shrinkage and selection operator (Lasso), and the selected variables were incorporated into logistic regression to construct a nomogram. The performance of the nomogram was assessed by using the area under the receiver operating characteristic curve (AUC), Decision Curve Analysis (DCA) and calibration curve. Finally, the importance of variables in the model is calculated based on the XGboost method. Results: A total of 705 patients with ventricular septal defect were included in the study. After screening with VIF and Lasso, the variables finally included in the statistical analysis include: Brain Natriuretic Peptide, bicarbonate, fibrinogen, urea, alanine aminotransferase, blood oxygen saturation, systolic blood pressure, respiratory rate, heart rate. The AUC values of nomogram in the training cohort and validation cohort were 0.812 and 0.736, respectively. The results of the calibration curve and DCA also indicated that the nomogram had good performance and good clinical application value. Conclusion: The nomogram established by BNP, bicarbonate, fibrinogen, urea, alanine aminotransferase, blood oxygen saturation, systolic blood pressure, respiratory rate, heart rate has good predictive performance and clinical applicability. The nomogram can effectively identify specific populations at risk for adverse outcomes.

High-Frequency Utilization of the PICU

OBJECTIVES

To describe the demographic, clinical, outcome, and cost differences between children with high-frequency PICU admission and those without.

DESIGN

Retrospective, cross-sectional cohort study.

SETTING

United States.

PATIENTS

Children less than or equal to 18 years old admitted to PICUs participating in the Pediatric Health Information System database in 2018.

INTERVENTIONS

None.

MEASUREMENTS AND MAIN RESULTS

We assessed survivors of PICU admissions for repeat PICU admissions within a year of their index visit. Children with greater than or equal to 3 PICU admissions within a year were classified as high-frequency PICU utilization (HFPICU). We compared demographic, clinical, outcome, and cost characteristics between children with HFPICU and those with only an index or two admissions per year (nHFPICU). Of 95,465 children who survived an index admission, 5,880 (6.2%) met HFPICU criteria. HFPICU patients were more frequently younger, technology dependent, and publicly insured. HFPICU patients had longer lengths of stay and were more frequently discharged to a rehabilitation facility or with home nursing services. HFPICU patients accounted for 24.8% of annual hospital utilization costs among patients requiring PICU admission. Time to readmission for children with HFPICU was 58% sooner (95% CI, 56-59%) than in those with nHFPICU with two admissions using an accelerated failure time model. Among demographic and clinical factors that were associated with development of HFPICU status calculated from a multivariable analysis, the greatest effect size was for time to first readmission within 82 days.

CONCLUSIONS

Children identified as having HFPICU account for 6.2% of children surviving an index ICU admission. They are a high-risk patient population with increased medical resource utilization during index and subsequent ICU admissions. Patients readmitted within 82 days of discharge should be considered at higher risk of HFPICU status. Further research, including validation and exploration of interventions that may be of use in this patient population, are necessary.

Risk Factors of Readmission to Pediatric Intensive Care Unit Within 1 Year: A Case-Control Study

BACKGROUND

Research on pediatric intensive care unit (PICU) readmission is lacking in China. This study was conducted to describe the risk factors associated with PICU readmission within 1 year after PICU discharge.

METHODS

This retrospective case-control study included patients aged from 1 month to 16 years who were discharged between January 2018 and May 2020. The case group included readmitted patients with two or more PICU admissions within 1 year during the study period. The control group included survivors with only one PICU admission during the same study period, and the controls were matched on age and sex. Demographic and clinical variables were collected from the electronic administrative database. Risk factors were analyzed by univariate and multivariate analyses.

RESULTS

From January 2018 to May 2020, 2,529 patients were discharged from the PICU, and 103 (4.07%) were readmitted within 1 year. In the univariate analysis, PICU readmission within 1 year was associated with lower weight, the presence of chronic conditions, a higher StrongKids score on admission, length of PICU stay of more than 2 weeks, the presence of dysfunction at discharge, sedation medications use, vasopressor use, and invasive mechanical ventilation in the first PICU stay. Patients had a higher StrongKids score as a surrogate for increased risk of malnutrition. In the multivariate analysis, the factors most significantly associated with PICU readmission within 1 year were the presence of chronic conditions, a higher StrongKids score on admission, and length of PICU stay of more than 2 weeks in the first PICU stay. In the subgroup analysis, compared with the control group, the factors most significantly associated with readmission within 48 h of discharge were the presence of chronic conditions, a higher StrongKids score on admission, and vasopressor use during the first PICU stay. The mortality rate was 8.74% (9/103) in patients with PICU readmission. The overall PICU mortality rate was 7.39% (201/2,721) during the study period.

CONCLUSIONS

Patients with chronic conditions, a higher StrongKids score on admission, and length of PICU stay of more than 2 weeks were at much higher risk for PICU readmission within 1 year. Patients with vasopressor use during the first PICU hospitalization were more likely to be readmitted within 48 h of discharge.

Predictive modelling of hospital readmission: Evaluation of different preprocessing techniques on machine learning classifiers

Hospital readmission is a major cost for healthcare systems worldwide. If patients with a higher potential of readmission could be identified at the start, existing resources could be used more efficiently, and appropriate plans could be implemented to reduce the risk of readmission. Therefore, it is important to predict the right target patients. Medical data is usually noisy, incomplete, and inconsistent. Hence, before developing a prediction model, it is crucial to efficiently set up the predictive model so that improved predictive performance is achieved. The current study aims to analyse the impact of different preprocessing methods on the performance of different machine learning classifiers. The preprocessing applied by previous hospital readmission studies were compared, and the most common approaches highlighted such as missing value imputation, feature selection, data balancing, and feature scaling. The hyperparameters were selected using Bayesian optimisation. The different preprocessing pipelines were assessed using various performance metrics and computational costs. The results indicated that the preprocessing approaches helped improve the model’s prediction of hospital readmission.

Is Unplanned PICU Readmission a Proper Quality Indicator? A Systematic Review and Meta-analysis

CONTEXT

Unplanned PICU readmissions within 48 hours of discharge (to home or a different hospital setting) are considered a quality metric of critical care.

OBJECTIVE

We sought to determine identifiable risk factors associated with early unplanned PICU readmissions.

DATA SOURCES

A comprehensive search of Medline, Embase, the Cochrane Database of Systematic Reviews, and Scopus was conducted from each database's inception to July 16, 2018.

STUDY SELECTION

Observational studies of early unplanned PICU readmissions (<48 hours) in children (<18 years of age) published in any language were included.

DATA EXTRACTION

Two reviewers selected and appraised studies independently and abstracted data. A meta-analysis was performed by using the random-effects model.

RESULTS

We included 11 observational studies in which 128 974 children (mean age: 5 years) were evaluated. The presence of complex chronic diseases (odds ratio 2.42; 95% confidence interval 1.06 to 5.55; I ² 79.90%) and moderate to severe disability (odds ratio 2.85; 95% confidence interval 2.40 to 3.40; I ² 11.20%) had the highest odds of early unplanned PICU readmission. Other significant risk factors included an unplanned index admission, initial admission to a general medical ward, spring season, respiratory diagnoses, and longer initial PICU stay. Readmission was less likely after trauma- and surgery-related index admissions, after direct admission from home, or during the summer season. Modifiable risk factors, such as evening or weekend discharge, revealed no statistically significant association. Included studies were retrospective, which limited our ability to account for all potential confounders and establish causality.

CONCLUSIONS

Many risk factors for early unplanned PICU readmission are not modifiable, which brings into question the usefulness of this quality measure.

Unplanned PICU Readmission in a Middle-Income Country: Who Is at Risk and What Is the Outcome?

OBJECTIVES

To study the rate of unplanned PICU readmission, determine the risk factors and its impact on mortality.

DESIGN

A single-center retrospective cross-sectional study.

SETTING

Tertiary referral PICU in Johor, Malaysia.

PATIENTS

All children admitted to the PICU over 8 years were included. Patients readmitted into PICU after the first PICU discharge during the hospitalization period were categorized into "early" (within 48 hr) and "late" (after 48 hr), and factors linked to the readmissions were identified. The mortality rate was determined and compared between no, early, and late readmission.

INTERVENTIONS

None.

MEASUREMENTS AND MAIN RESULTS

There were 2,834 patients in the study with 70 early and 113 late readmissions. Therefore, the rate of early and late PICU readmission was 2.5% (95% CI, 1.9-3.0%) and 3.9% (95% CI, 3.2-4.7%), respectively. The median length of stay of the second PICU admission for early and late readmissions was 2.7 days (interquartile range, 1.1-7.0 d) and 3.2 days (interquartile range, 1.2-7.5 d), respectively. The majority of early and late readmissions had a similar diagnosis with their first PICU admission. Multivariable multinomial logistic regression revealed a Pediatric Index Mortality 2 score of greater than or equal to 15, chronic cardiovascular condition, and oxygen supplement upon discharge as independent risk factors for early PICU readmission. Meanwhile, an infant of less than 1 year old, having cardiovascular, other congenital and genetic chronic conditions and being discharged between 8 AM and 5 PM was an independent risk factor for late readmission. There was no significant difference in the mortality rate of early (12.9%), late (13.3%), and no readmission (10.7%).

CONCLUSIONS

Despite the lack of resources and expertise in lower- and middle-income countries, the rate and factors for PICU readmission are similar to those in high-income countries. However, PICU readmission has no statistically significant association with mortality.

Compassion in Pediatric Healthcare: A Scoping Review

PROBLEM

Compassion has been described as a central construct or essential feature of quality healthcare and is as important to patients' and families' overall healthcare experience as the health interventions and treatments they receive. However, there is little shared understanding of what constitutes compassion, how it is delivered within a pediatric setting, and pediatric patients' and families perspectives and preferences for receiving it.

ELIGIBILITY CRITERIA

Studies that (1) described the nature of the existing literature on compassion in pediatric healthcare; (2) summarized key concepts in the existing evidence base that pertain to compassion in pediatric healthcare; and 3) identified factors that are associated with compassion in pediatric healthcare were eligible for inclusion in this review.

SAMPLE

Twenty-nine papers were included in the review.

RESULTS

Findings revealed several factors are associated with compassion in pediatric healthcare, including continuity of care, communication, and coordination of care. Most notably, identified studies treated compassion in a subsidiary fashion, and this review revealed no studies that provided a patient-informed evidence-based definition of compassion in the pediatric healthcare setting.

CONCLUSION

Future research is required to generate a comprehensive and accurate understanding of the terms 'compassion' and 'compassionate care' when used in the context of pediatric healthcare.

IMPLICATIONS

This research will inform the therapeutic processes and ultimately enable the development of strategies to improve the delivery of compassionate healthcare to pediatric patients.

Incidence and risk factors for readmission to a paediatric intensive care unit

BACKGROUND

Unscheduled readmission to a paediatric intensive care unit can lead to unfavourable patient outcomes. Therefore, determining the incidence and risk factors is important. Previous studies on such readmissions have only focused on the first 48 hours after discharge and described the relative risk factors as unmodifiable.

AIM

To identify the incidence and risk factors of unscheduled readmission to a paediatric intensive care unit within 7 days of discharge.

DESIGN

This was a retrospective observational study.

METHODS

Our study population comprised consecutive patients admitted to the paediatric intensive care unit of our tertiary hospital in Japan in 2012 to 2016. We determined the incidence of unscheduled readmission to the unit within 7 days of discharge and identified potential risk factors using multivariable logistic regression analysis.

RESULTS

Among the 2432 admissions (1472 patients), 60 admissions (2.5%, 44 patients) were followed by ≥1 unscheduled readmission. The median time to readmission was 3.5 days. The most common causes for readmission were respiratory issues and cardiovascular symptoms. The significant risk factors for readmission within 7 days of discharge were unscheduled initial admission (odds ratio [OR]: 3.02; 95% confidence interval [CI:] 1.45-6.31), admission from a general ward (OR: 5.13; 95% CI: 1.75-15.0), and withdrawal syndrome during the initial stay (OR: 3.95; 95% CI: 1.53-10.2).

CONCLUSIONS

The incidence of unscheduled readmission within 7 days was not high (2.5%), and one of the three identified risk factors for readmissions (withdrawal syndrome) is potentially modifiable.

RELEVANCE TO CLINICAL PRACTICE

Appropriate treatment of withdrawal syndrome may reduce readmissions and improve patient outcomes. Although unscheduled initial admission and admission from general ward are not modifiable risk factors, careful discharge judgement and follow up after discharge from paediatric intensive care units for high-risk patients may be beneficial.

Machine Learning Readmission Risk Modeling: A Pediatric Case Study

BACKGROUND

Hospital readmission prediction in pediatric hospitals has received little attention. Studies have focused on the readmission frequency analysis stratified by disease and demographic/geographic characteristics but there are no predictive modeling approaches, which may be useful to identify preventable readmissions that constitute a major portion of the cost attributed to readmissions.

OBJECTIVE

To assess the all-cause readmission predictive performance achieved by machine learning techniques in the emergency department of a pediatric hospital in Santiago, Chile.

MATERIALS

An all-cause admissions dataset has been collected along six consecutive years in a pediatric hospital in Santiago, Chile. The variables collected are the same used for the determination of the child's treatment administrative cost.

METHODS

Retrospective predictive analysis of 30-day readmission was formulated as a binary classification problem. We report classification results achieved with various model building approaches after data curation and preprocessing for correction of class imbalance. We compute repeated cross-validation (RCV) with decreasing number of folders to assess performance and sensitivity to effect of imbalance in the test set and training set size.

RESULTS

Increase in recall due to SMOTE class imbalance correction is large and statistically significant. The Naive Bayes (NB) approach achieves the best AUC (0.65); however the shallow multilayer perceptron has the best PPV and f-score (5.6 and 10.2, resp.). The NB and support vector machines (SVM) give comparable results if we consider AUC, PPV, and f-score ranking for all RCV experiments. High recall of deep multilayer perceptron is due to high false positive ratio. There is no detectable effect of the number of folds in the RCV on the predictive performance of the algorithms.

CONCLUSIONS

We recommend the use of Naive Bayes (NB) with Gaussian distribution model as the most robust modeling approach for pediatric readmission prediction, achieving the best results across all training dataset sizes. The results show that the approach could be applied to detect preventable readmissions.

Predictive models for hospital readmission risk: A systematic review of methods

OBJECTIVES

Hospital readmission risk prediction facilitates the identification of patients potentially at high risk so that resources can be used more efficiently in terms of cost-benefit. In this context, several models for readmission risk prediction have been proposed in recent years. The goal of this review is to give an overview of prediction models for hospital readmission, describe the data analysis methods and algorithms used for building the models, and synthesize their results.

METHODS

Studies that reported the predictive performance of a model for hospital readmission risk were included. We defined the scope of the review and accordingly built a search query to select the candidate papers. This query string was used as input for the chosen search engines, namely PubMed and Google Scholar. For each study, we recorded the population, feature selection method, classification algorithm, sample size, readmission threshold, readmission rate and predictive performance of the model.

RESULTS

We identified 77 studies that met the inclusion criteria, out of 265 citations. In 68% of the studies (n = 52) logistic regression or other regression techniques were utilized as the main method. Ten (13%) studies used survival analysis for model construction, while 14 (18%) used machine learning techniques for classification, of which decision tree-based methods and SVM were the most utilized algorithms. Among these, only four studies reported the use of any class imbalance addressing technique, of which resampling is the most frequent (75%). The performance of the models varied significantly among studies, with Area Under the ROC Curve (AUC) values in the ranges between 0.54 and 0.92.

CONCLUSION

Logistic regression and survival analysis have been traditionally the most widely used techniques for model building. Nevertheless, machine learning techniques are becoming increasingly popular in recent years. Recent comparative studies suggest that machine learning techniques can improve prediction ability over traditional statistical approaches. Regardless, the lack of an appropriate benchmark dataset of hospital readmissions makes a comparison of models' performance across different studies difficult.

Going back to the ward-transitioning care back to the ward team

Transition of care from the intensive care unit (ICU) to the ward is usually an indication of the patient's improving clinical status, but is also a time when patients are particularly vulnerable. The transition between care teams poses a higher risk of medical error, which can be mitigated by safe and complete patient handoff and medication reconciliation. ICU readmissions are associated with increased mortality as well as ICU and hospital length of stay (LOS); however tools to accurately predict ICU readmission risk are limited. While there are many mechanisms in place to carefully identify patients appropriate for transfer to the ward, the optimal timing of transfer can be affected by ICU strain, limited resources such as ICU beds, and overall hospital capacity and flow leading to suboptimal transfer times or delays in transfer. The patient and family perspectives should also be considered when planning for transfer from the ICU to the ward. During times of transition, families will meet a new care team, experience uncertainty of future care plans, and adjust to a different daily routine which can lead to increased stress and anxiety. Additionally, a subset of patients, such as those with new technology, require additional multidisciplinary support, education and care coordination which can contribute to longer hospital LOS if not addressed proactively early in the hospitalization while the patient remains in the ICU. In this review article, we describe key components of the transfer from ICU to the ward, discuss current strategies to optimize timing of patient transfers, explore strategies to partner with patients and families during the transfer process, highlight patient populations where additional considerations are needed, and identify future areas of exploration which could improve the care transition from the ICU to the ward.