Dataset Information

Prediction of Pediatric Critical Care Resource Utilization for Disaster Triage.

ABSTRACT:

Objectives

Pediatric protocols to guide allocation of limited resources during a disaster lack data to validate their use. The 2011 Pediatric Emergency Mass Critical Care Task Force recommended that expected duration of critical care be incorporated into resource allocation algorithms. We aimed to determine whether currently available pediatric illness severity scores can predict duration of critical care resource use.

Design

Retrospective cohort study.

Setting

Seattle Children's Hospital.

Patients

PICU patients admitted 2016-2018 for greater than or equal to 12 hours (n = 3,206).

Interventions

None.

Measurements and main results

We developed logistic and linear regression models in two-thirds of the cohort to predict need for and duration of PICU resources based on Pediatric Risk of Mortality-III, Pediatric Index of Mortality-3, and serial Pediatric Logistic Organ Dysfunction-2 scores. We tested the predictive accuracy of the models with the highest area under the receiver operating characteristic curve (need for each resource) and R (duration of use) in a validation cohort of the remaining one of three of the sample and among patients admitted during one-third of the sample and among patients admitted during surges of respiratory illness. Pediatric Logistic Organ Dysfunction score calculated 12 hours postadmission had higher predictive accuracy than either Pediatric Risk of Mortality or Pediatric Index of Mortality scores. Models incorporating 12-hour Pediatric Logistic Organ Dysfunction score, age, Pediatric Overall Performance Category, Pediatric Cerebral Performance Category, chronic mechanical ventilation, and postoperative status had an area under the receiver operating characteristic curve = 0.8831 for need for any PICU resource (positive predictive value 80.2%, negative predictive value 85.9%) and area under the receiver operating characteristic curve = 0.9157 for mechanical ventilation (positive predictive value 85.7%, negative predictive value 89.2%) within 7 days of admission. Models accurately predicted greater than or equal to 24 hours of any resource use for 78.9% of patients and greater than or equal to 24 hours of ventilation for 83.1%. Model fit and accuracy improved for prediction of resource use within 3 days of admission, and was lower for noninvasive positive pressure ventilation, vasoactive infusions, continuous renal replacement therapy, extracorporeal membrane oxygenation, and length of stay.

Conclusions

A model incorporating 12-hour Pediatric Logistic Organ Dysfunction score performed well in estimating how long patients may require PICU resources, especially mechanical ventilation. A pediatric disaster triage algorithm that includes both likelihood for survival and for requiring critical care resources could minimize subjectivity in resource allocation decision-making.

SUBMITTER: Killien EY

PROVIDER: S-EPMC7199765 | biostudies-literature | 2020 Aug

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Prediction of Pediatric Critical Care Resource Utilization for Disaster Triage.

Killien Elizabeth Y EY Mills Brianna B Errett Nicole A NA Sakata Vicki V Vavilala Monica S MS Rivara Frederick P FP Kissoon Niranjan N King Mary A MA

Pediatric critical care medicine : a journal of the Society of Critical Care Medicine and the World Federation of Pediatric Intensive and Critical Care Societies 20200801 8

<h4>Objectives</h4>Pediatric protocols to guide allocation of limited resources during a disaster lack data to validate their use. The 2011 Pediatric Emergency Mass Critical Care Task Force recommended that expected duration of critical care be incorporated into resource allocation algorithms. We aimed to determine whether currently available pediatric illness severity scores can predict duration of critical care resource use.<h4>Design</h4>Retrospective cohort study.<h4>Setting</h4>Seattle Chil ...[more]

PMID: 32345932

Similar Datasets

Project description:ImportanceDuring the COVID-19 pandemic, many US states issued or revised pandemic preparedness plans guiding allocation of critical care resources during crises. State plans vary in the factors used to triage patients and have faced criticism from advocacy groups due to the potential for discrimination.ObjectiveTo analyze the role of comorbidities and long-term prognosis in state triage procedures.Design, setting, and participantsThis cross-sectional study used data gathered from parallel internet searches for state-endorsed pandemic preparedness plans for the 50 US states, District of Columbia, and Puerto Rico (hereafter referred to as states), which were conducted between November 25, 2021, and June 16, 2023. Plans available on June 16, 2023, that provided step-by-step instructions for triaging critically ill patients were categorized for use of comorbidities and prognostication.Main outcomes and measuresPrevalence and contents of lists of comorbidities and their stated function in triage and instructions to predict duration of postdischarge survival.ResultsOverall, 32 state-promulgated pandemic preparedness plans included triage procedures specific enough to guide triage in clinical practice. Twenty of these (63%) included lists of comorbidities that excluded (11 of 20 [55%]) or deprioritized (8 of 20 [40%]) patients during triage; one state's list was formulated to resolve ties between patients with equal triage scores. Most states with triage procedures (21 of 32 [66%]) considered predicted survival beyond hospital discharge. These states proposed different prognostic time horizons; 15 of 21 (71%) were numeric (ranging from 6 months to 5 years after hospital discharge), with the remaining 6 (29%) using descriptive terms, such as long-term.Conclusions and relevanceIn this cross-sectional study of state-promulgated critical care triage policies, most plans restricted access to scarce critical care resources for patients with listed comorbidities and/or for patients with less-than-average expected postdischarge survival. This analysis raises concerns about access to care during a public health crisis for populations with high burdens of chronic illness, such as individuals with disabilities and minoritized racial and ethnic groups.

Project description:OBJECTIVES:To evaluate hospital-level variability in resource utilization and mortality in children with new leukemia who require ICU support, and identify factors associated with variation. DESIGN:Retrospective cohort study. SETTING:Children's hospitals contributing to the Pediatric Health Information Systems administrative database from 1999 to 2011. PATIENTS:Inpatients less than 25 years old with newly diagnosed acute lymphocytic leukemia or acute myeloid leukemia requiring ICU support (n = 1,754). INTERVENTIONS, MEASUREMENTS, AND MAIN RESULTS:Evaluated exposures included leukemia type, year of diagnosis, and hospital-wide proportion of patients with public insurance. The main outcome was hospital mortality. Wide variability existed in the ICU resources used across hospitals. Combined acute lymphocytic leukemia and acute myeloid leukemia mortality varied by hospital from 0% (95% CI, 0-14.8%) to 42.9% (95% CI, 17.7-71.1%). A mixed-effects model with a hospital-level random effect suggests significant variation across hospitals in mortality (p = 0.007). When including patient and hospital factors as fixed effects into the model, younger age, acute myeloid leukemia versus acute lymphocytic leukemia diagnosis, leukemia diagnosis prior to 2005, hospital-wide proportion of public insurance patients, and hospital-level proportion of leukemia patients receiving ICU care are significantly associated with mortality. The variation across hospitals remains significant with all patient factors included (p = 0.021) but is no longer significant after adjusting for the hospital-level factors proportion of public insurance and proportion receiving ICU care (p = 0.48). CONCLUSIONS:Wide hospital-level variability in ICU resource utilization and mortality exists in the care of children with leukemia requiring ICU support. Hospital payer mix is associated with some mortality variability. Additional study into how ICU support could be standardized through clinical practice guidelines, impact of payer mix on hospital resources allocation to the ICU, and subsequent impact on patient outcomes is warranted.

Project description:BackgroundAn increase in the incidence of central venous catheter (CVC)-associated deep venous thrombosis (CADVT) has been reported in pediatric patients over the past decade. At the same time, current screening guidelines for venous thromboembolism risk have low sensitivity for CADVT in hospitalized children. This study utilized a multimodal deep learning model to predict CADVT before it occurs.MethodsChildren who were admitted to intensive care units (ICUs) between December 2015 and December 2018 and with CVC placement at least 3 days were included. The variables analyzed included demographic characteristics, clinical conditions, laboratory test results, vital signs and medications. A multimodal deep learning (MMDL) model that can handle temporal data using long short-term memory (LSTM) and gated recurrent units (GRUs) was proposed for this prediction task. Four benchmark machine learning models, logistic regression (LR), random forest (RF), gradient boosting decision tree (GBDT) and a published cutting edge MMDL, were used to compare and evaluate the models with a fivefold cross-validation approach. Accuracy, recall, area under the ROC curve (AUC), and average precision (AP) were used to evaluate the discrimination of each model at three time points (24 h, 48 h and 72 h) before CADVT occurred. Brier score and Spiegelhalter's z test were used measure the calibration of these prediction models.ResultsA total of 1830 patients were included in this study, and approximately 15% developed CADVT. In the CADVT prediction task, the model proposed in this paper significantly outperforms both traditional machine learning models and existing multimodal deep learning models at all 3 time points. It achieved 77% accuracy and 90% recall at 24 h before CADVT was discovered. It can be used to accurately predict the occurrence of CADVT 72 h in advance with an accuracy of greater than 75%, a recall of more than 87%, and an AUC value of 0.82.ConclusionIn this study, a machine learning method was successfully established to predict CADVT in advance. These findings demonstrate that artificial intelligence (AI) could provide measures for thromboprophylaxis in a pediatric intensive care setting.

Dataset Information

Prediction of Pediatric Critical Care Resource Utilization for Disaster Triage.

Objectives

Design

Setting

Patients

Interventions

Measurements and main results

Conclusions

Publications

Prediction of Pediatric Critical Care Resource Utilization for Disaster Triage.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets