Dataset Information

Predicting Intensive Care Transfers and Other Unforeseen Events: Analytic Model Validation Study and Comparison to Existing Methods.

ABSTRACT:

Background

COVID-19 has led to an unprecedented strain on health care facilities across the United States. Accurately identifying patients at an increased risk of deterioration may help hospitals manage their resources while improving the quality of patient care. Here, we present the results of an analytical model, Predicting Intensive Care Transfers and Other Unforeseen Events (PICTURE), to identify patients at high risk for imminent intensive care unit transfer, respiratory failure, or death, with the intention to improve the prediction of deterioration due to COVID-19.

Objective

This study aims to validate the PICTURE model's ability to predict unexpected deterioration in general ward and COVID-19 patients, and to compare its performance with the Epic Deterioration Index (EDI), an existing model that has recently been assessed for use in patients with COVID-19.

Methods

The PICTURE model was trained and validated on a cohort of hospitalized non-COVID-19 patients using electronic health record data from 2014 to 2018. It was then applied to two holdout test sets: non-COVID-19 patients from 2019 and patients testing positive for COVID-19 in 2020. PICTURE results were aligned to EDI and NEWS scores for head-to-head comparison via area under the receiver operating characteristic curve (AUROC) and area under the precision-recall curve. We compared the models' ability to predict an adverse event (defined as intensive care unit transfer, mechanical ventilation use, or death). Shapley values were used to provide explanations for PICTURE predictions.

Results

In non-COVID-19 general ward patients, PICTURE achieved an AUROC of 0.819 (95% CI 0.805-0.834) per observation, compared to the EDI's AUROC of 0.763 (95% CI 0.746-0.781; n=21,740; P<.001). In patients testing positive for COVID-19, PICTURE again outperformed the EDI with an AUROC of 0.849 (95% CI 0.820-0.878) compared to the EDI's AUROC of 0.803 (95% CI 0.772-0.838; n=607; P<.001). The most important variables influencing PICTURE predictions in the COVID-19 cohort were a rapid respiratory rate, a high level of oxygen support, low oxygen saturation, and impaired mental status (Glasgow Coma Scale).

Conclusions

The PICTURE model is more accurate in predicting adverse patient outcomes for both general ward patients and COVID-19 positive patients in our cohorts compared to the EDI. The ability to consistently anticipate these events may be especially valuable when considering potential incipient waves of COVID-19 infections. The generalizability of the model will require testing in other health care systems for validation.

SUBMITTER: Cummings BC

PROVIDER: S-EPMC8061893 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:INTRODUCTION: The discharge of patients from the intensive care unit (ICU) to a hospital ward is a common transition of care that is associated with error and adverse events. Risk stratification tools may help identify high-risk patients for targeted interventions, but it is unclear if proper tools have been developed. METHODS: We searched Ovid EMBASE, Ovid MEDLINE, CINAHL, PUBMED and Cochrane Central Register of Controlled Trials from the earliest available date through March 2013, plus reference lists and citations of all studies included in the systematic review. Cohort studies were selected that described the derivation, validation or clinical impact of tools for predicting medical emergency team activation, ICU readmission or mortality following patient discharge from the ICU. Data were extracted on the study design, setting, population, sample size, tool (components, measurement properties) and outcomes. RESULTS: The literature search identified 9,926 citations, of which eight studies describing eight tools met the inclusion criteria. Reported outcomes included ICU readmission (n=4 studies), hospital mortality (n=3 studies) and both ICU readmission and hospital mortality (n=1 studies). Seven of the tools were comprised of distinct measurable component variables, while one tool used subjective scoring of patient risk by intensive care physicians. The areas under receiver operator curves were reported for all studies and ranged from 0.66 to 0.92. A single study provided a direct comparative analysis between two tools. We did not find any studies evaluating the impact of risk prediction on processes and outcomes of care. CONCLUSIONS: Eight risk stratification tools for predicting severe adverse events following patient discharge from ICU have been developed, but have undergone limited comparative evaluation. Although risk stratification tools may help clinician decision-making, further evaluation of the existing tools' effects on care is required prior to clinical implementation.

Project description:BackgroundTransfers from one intensive care unit (ICU) to another ICU are associated with increased length of intensive care and hospital stay. Inter-hospital ICU transfers are carried out for three main reasons: clinical transfers, capacity transfers and repatriations. The aim of the study was to show that different ICU transfers differ in risk-adjusted mortality rate with repatriations having the least risk.ResultsObservational cohort study of adult patients transferred between Swedish ICUs during 3 years (2016-2018) with follow-up ending September 2019. Primary and secondary end-points were survival to 30 days and 180 days after discharge from the first ICU. Data from 75 ICUs in the Swedish Intensive Care Register, a nationwide intensive care register, were used for analysis (89% of all Swedish ICUs), covering local community hospitals, district general hospitals and tertiary care hospitals. We included adult patients (16 years or older) admitted to ICU and subsequently discharged by transfer to another ICU. Only the first admission was used. Exposure was discharge to any other ICU (ICU-to-ICU transfer), whether in the same or in another hospital. Transfers were grouped into three predefined categories: clinical transfer, capacity transfer, and repatriation. We identified 15,588 transfers among 112,860 admissions (14.8%) and analysed 11,176 after excluding 4112 repeat transfer of the same individual and 300 with missing risk adjustment. The majority were clinical transfers (62.7%), followed by repatriations (21.5%) and capacity transfers (15.8%). Unadjusted 30-day mortality was 25.0% among capacity transfers compared to 14.5% and 16.2% for clinical transfers and repatriations, respectively. Adjusted odds ratio (OR) for 30-day mortality were 1.25 (95% CI 1.06-1.49 p = 0.01) for capacity transfers and 1.17 (95% CI 1.02-1.36 p = 0.03) for clinical transfers using repatriation as reference. The differences remained 180 days post-discharge.ConclusionsThere was a large proportion of ICU-to-ICU transfers and an increased odds of dying for those transferred due to other reasons than repatriation.