Advancing In-Hospital Clinical Deterioration Prediction Models.
Ontology highlight
ABSTRACT: Background
Early warning systems lack robust evidence that they improve patients' outcomes, possibly because of their limitation of predicting binary rather than time-to-event outcomes.Objectives
To compare the prediction accuracy of 2 statistical modeling strategies (logistic regression and Cox proportional hazards regression) and 2 machine learning strategies (random forest and random survival forest) for in-hospital cardiopulmonary arrest.Methods
Retrospective cohort study with prediction model development from deidentified electronic health records at an urban academic medical center.Results
The classification models (logistic regression and random forest) had statistical recall and precision similar to or greater than those of the time-to-event models (Cox proportional hazards regression and random survival forest). However, the time-to-event models provided predictions that could potentially better indicate to clinicians whether and when a patient is likely to experience cardiopulmonary arrest.Conclusions
As early warning scoring systems are refined, they must use the best analytical methods that both model the underlying phenomenon and provide an understandable prediction.
SUBMITTER: Jeffery AD
PROVIDER: S-EPMC6141236 | biostudies-literature | 2018 Sep
REPOSITORIES: biostudies-literature
Publications
Advancing In-Hospital Clinical Deterioration Prediction Models.
Jeffery Alvin D AD Dietrich Mary S MS Fabbri Daniel D Kennedy Betsy B Novak Laurie L LL Coco Joseph J Mion Lorraine C LC
American journal of critical care : an official publication, American Association of Critical-Care Nurses 20180901 5
<h4>Background</h4>Early warning systems lack robust evidence that they improve patients' outcomes, possibly because of their limitation of predicting binary rather than time-to-event outcomes.<h4>Objectives</h4>To compare the prediction accuracy of 2 statistical modeling strategies (logistic regression and Cox proportional hazards regression) and 2 machine learning strategies (random forest and random survival forest) for in-hospital cardiopulmonary arrest.<h4>Methods</h4>Retrospective cohort s ...[more]
Similar Datasets
Project description:To analyze the available literature on the performance of artificial intelligence-generated clinical models for the prediction of serious life-threatening events in non-ICU adult patients and evaluate their potential clinical usage.Data sourcesThe PubMed database was searched for relevant articles in English literature from January 1, 2000, to January 23, 2022. Search terms, including artificial intelligence, machine learning, deep learning, and deterioration, were both controlled terms and free-text terms.Study selectionWe performed a systematic search reporting studies that showed performance of artificial intelligence-based models with outcome mortality and clinical deterioration.Data extractionTwo review authors independently performed study selection and data extraction. Studies with the same outcome were grouped, namely mortality and various forms of deterioration (including ICU admission, adverse events, and cardiac arrests). Meta-analysis was planned in case sufficient data would be extracted from each study and no considerable heterogeneity between studies was present.Data synthesisIn total, 45 articles were included for analysis, in which multiple methods of artificial intelligence were used. Twenty-four articles described models for the prediction of mortality and 21 for clinical deterioration. Due to heterogeneity of study characteristics (patient cohort, outcomes, and prediction models), meta-analysis could not be performed. The main reported measure of performance was the area under the receiver operating characteristic (AUROC) (n = 38), of which 33 (87%) had an AUROC greater than 0.8. The highest reported performance in a model predicting mortality had an AUROC of 0.935 and an area under the precision-recall curve of 0.96.ConclusionsCurrently, a growing number of studies develop and analyzes artificial intelligence-based prediction models to predict critical illness and deterioration. We show that artificial intelligence-based prediction models have an overall good performance in predicting deterioration of patients. However, external validation of existing models and its performance in a clinical setting is highly recommended.
| S-EPMC9423015 | biostudies-literature
Project description:IntroductionFalls remain one of the most prevalent adverse events in hospitals and are associated with substantial negative health impacts and costs. Approaches to assess patients' fall risk have been implemented in hospitals internationally, ranging from brief screening questions to multifactorial risk assessments and complex prediction models, despite a lack of clear evidence of effect in reducing falls in acute hospital environments. The increasing digitisation of hospital systems provides new opportunities to understand and predict falls using routinely recorded data, with potential to integrate fall prediction models into real-time or near-real-time computerised decision support for clinical teams seeking to mitigate fall risk. However, the use of non-traditional approaches to fall risk prediction, including machine learning using integrated electronic medical records, has not yet been reviewed relative to more traditional fall prediction models. This scoping review will summarise methodologies used to develop existing hospital fall prediction models, including reporting quality assessment.Methods and analysisThis scoping review will follow the Arksey and O'Malley framework and its recent advances, and will be reported using Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews recommendations. Four electronic databases (CINAHL via EBSCOhost, PubMed, IEEE Xplore and Embase) will be initially searched for studies up to 12 November 2020, and searches may be updated prior to final reporting. Additional studies will be identified by reference list review and citation analysis of included studies. No restriction will be placed on the date or language of identified studies. Screening of search results and extraction of data will be performed by two independent reviewers. Reporting quality will be assessed by the adherence to the Transparent Reporting of a multivariable prediction model for Individual Prognosis Or Diagnosis.Ethics and disseminationEthical approval is not required for this study. Findings will be disseminated through peer-reviewed publication and scientific conferences.
| S-EPMC8438847 | biostudies-literature
Project description:ObjectiveImproved understanding of the deteriorating patient in the pre-hospital setting may result in earlier recognition and response. Considering the effects of undetected deterioration are profound, it is fundamental to report the prevalence of pre-hospital clinical deterioration to advance our understanding. The present study investigated the prevalence of pre-hospital clinical deterioration and adverse events (AEs) within 3 days of the pre-hospital episode of care.MethodsThis retrospective cohort study was based on pre-hospital incidents involving adult patients attended by Queensland Ambulance Service between 1 January 2018 and 31 December 2020. Due to lacking a standardised definition of pre-hospital clinical deterioration, established early warning scores (NEWS, MEWS and Q-ADDS) were calculated from pre-hospital vital signs to identify clinical deterioration. Linked hospital data were used to identify the occurrence of an AE.ResultsSome degree of physiological derangement was initially observed in over half of the patients, and pre-hospital clinical deterioration was seen in 2.7%-4% of patients. The prevalence of AEs was 3.2%. Patients that experienced an AE were more likely to be male, elderly, suffering from a medical (non-trauma) condition, and had a greater burden of disease. Concerningly, almost 50% of patients that suffered an AE did not meet escalation thresholds of NEWS, MEWS or Q-ADDS.ConclusionsThe present study found the prevalence of pre-hospital clinical deterioration and AEs subsequent to pre-hospital episodes of care to be low. Future research should prioritise using standardised criteria to define pre-hospital clinical deterioration and evaluate the performance of early warning scores.
| S-EPMC11744426 | biostudies-literature
Project description:BackgroundEarly recognition of severe COVID-19 is essential for timely patient triage.AimsTo report clinical and laboratory findings and patient outcomes at a tertiary hospital in Melbourne, Australia.MethodsThis is a retrospective study of adult inpatients with COVID-19 admitted to Northern Health from March to September 2020. Data were extracted from electronic medical records.ResultsKey admission data were available for 182 patients (median age 67.0 years (interquartile range, 47.9-83.1); 51.1% female). Fifty-six (30.8%) were from residential care. One hundred and seventeen (64.3%) patients were assigned Goals of Patient Care (GOPC) A or B and 65 (35.7%) GOPC C or D. Comorbidities were present in 135 patients (74.2%). 63.2% of patients received antibiotics, 6.6% had antivirals, 45.6% received systemic glucocorticoid and 3.3% had tocilizumab. Fifty-six (30.8%) developed clinical deterioration (24 requiring ventilation, 21 receiving critical care, 34 died). Overall, inhospital clinical deterioration was significantly associated with older age (P < 0.001), history of diabetes (P = 0.038), lower lymphocyte count (P = 0.002) and platelet count (P = 0.004), higher neutrophil-to-lymphocyte ratio (P = 0.002), elevated fibrinogen (P = 0.004), higher serum ferritin (P = 0.027) and C-reactive protein (CRP; P = 0.002). The accuracy of the 4C Deterioration model was moderate, with an area under the curve (AUC) of 0.79 (95% confidence interval (CI), 0.68-0.90) compared with an AUC of 0.77 (95% CI, 0.76-0.78) in the original validation cohort.ConclusionsIn the present study, high neutrophil-to-lymphocyte ratio, abnormal d-dimer, high serum CRP and ferritin appear to be useful prognostic markers.
| S-EPMC9011432 | biostudies-literature
Project description:In materials science, accurately computing properties like viscosity, melting point, and glass transition temperatures solely through physics-based models is challenging. Data-driven machine learning (ML) also poses challenges in constructing ML models, especially in the material science domain where data is limited. To address this, we integrate physics-informed descriptors from molecular dynamics (MD) simulations to enhance the accuracy and interpretability of ML models. Our current study focuses on accurately predicting viscosity in liquid systems using MD descriptors. In this work, we curated a comprehensive dataset of over 4000 small organic molecules' viscosities from scientific literature, publications, and online databases. This dataset enabled us to develop quantitative structure-property relationships (QSPR) consisting of descriptor-based and graph neural network models to predict temperature-dependent viscosities for a wide range of viscosities. The QSPR models reveal that including MD descriptors improves the prediction of experimental viscosities, particularly at the small data set scale of fewer than a thousand data points. Furthermore, feature importance tools reveal that intermolecular interactions captured by MD descriptors are most important for viscosity predictions. Finally, the QSPR models can accurately capture the inverse relationship between viscosity and temperature for six battery-relevant solvents, some of which were not included in the original data set. Our research highlights the effectiveness of incorporating MD descriptors into QSPR models, which leads to improved accuracy for properties that are difficult to predict when using physics-based models alone or when limited data is available.
| S-EPMC10938832 | biostudies-literature
Project description:BackgroundHospitalized adults whose condition deteriorates while they are in wards (outside the intensive care unit [ICU]) have considerable morbidity and mortality. Early identification of patients at risk for clinical deterioration has relied on manually calculated scores. Outcomes after an automated detection of impending clinical deterioration have not been widely reported.MethodsOn the basis of a validated model that uses information from electronic medical records to identify hospitalized patients at high risk for clinical deterioration (which permits automated, real-time risk-score calculation), we developed an intervention program involving remote monitoring by nurses who reviewed records of patients who had been identified as being at high risk; results of this monitoring were then communicated to rapid-response teams at hospitals. We compared outcomes (including the primary outcome, mortality within 30 days after an alert) among hospitalized patients (excluding those in the ICU) whose condition reached the alert threshold at hospitals where the system was operational (intervention sites, where alerts led to a clinical response) with outcomes among patients at hospitals where the system had not yet been deployed (comparison sites, where a patient's condition would have triggered a clinical response after an alert had the system been operational). Multivariate analyses adjusted for demographic characteristics, severity of illness, and burden of coexisting conditions.ResultsThe program was deployed in a staggered fashion at 19 hospitals between August 1, 2016, and February 28, 2019. We identified 548,838 non-ICU hospitalizations involving 326,816 patients. A total of 43,949 hospitalizations (involving 35,669 patients) involved a patient whose condition reached the alert threshold; 15,487 hospitalizations were included in the intervention cohort, and 28,462 hospitalizations in the comparison cohort. Mortality within 30 days after an alert was lower in the intervention cohort than in the comparison cohort (adjusted relative risk, 0.84, 95% confidence interval, 0.78 to 0.90; P<0.001).ConclusionsThe use of an automated predictive model to identify high-risk patients for whom interventions by rapid-response teams could be implemented was associated with decreased mortality. (Funded by the Gordon and Betty Moore Foundation and others.).
| S-EPMC7787261 | biostudies-literature
Project description: Not available
| S-EPMC10786836 | biostudies-literature
Project description:BackgroundRapid identification of high-risk polytrauma patients is crucial for early intervention and improved outcomes. This study aimed to develop and validate machine learning models for predicting 72 h mortality in adult polytrauma patients using readily available clinical parameters.MethodsA retrospective analysis was conducted on polytrauma patients from the Dryad database and our institution. Missing values pertinent to eligible individuals within the Dryad database were compensated for through the k-nearest neighbor algorithm, subsequently randomizing them into training and internal validation factions on a 7:3 ratio. The patients of our institution functioned as external validation cohorts. The predictive efficacy of random forest (RF), neural network, and XGBoost models was assessed through an exhaustive suite of performance indicators. The SHapley Additive exPlanations (SHAP) and Local Interpretable Model-Agnostic Explanations (LIME) methods were engaged to explain the supreme-performing model. Conclusively, restricted cubic spline analysis and multivariate logistic regression were employed as sensitivity analyses to verify the robustness of the findings.ResultsParameters including age, body mass index, Glasgow Coma Scale, Injury Severity Score, pH, base excess, and lactate emerged as pivotal predictors of 72 h mortality. The RF model exhibited unparalleled performance, boasting an area under the receiver operating characteristic curve (AUROC) of 0.87 (95% confidence interval [CI] 0.84-0.89), an area under the precision-recall curve (AUPRC) of 0.67 (95% CI 0.61-0.73), and an accuracy of 0.83 (95% CI 0.81-0.86) in the internal validation cohort, paralleled by an AUROC of 0.98 (95% CI 0.97-0.99), an AUPRC of 0.88 (95% CI 0.83-0.93), and an accuracy of 0.97 (95% CI 0.96-0.98) in the external validation cohort. It provided the highest net benefit in the decision curve analysis in relation to the other models. The outcomes of the sensitivity examinations were congruent with those inferred from SHAP and LIME.ConclusionsThe RF model exhibited the best performance in predicting 72 h mortality in adult polytrauma patients and has the potential to aid clinicians in identifying high-risk patients and guiding clinical decision-making.
| S-EPMC10518974 | biostudies-literature
Project description:IntroductionIn pre-hospital settings, identifying a deteriorating child can be challenging, especially considering that the proportion of paediatric patients with acute illnesses is lower compared with adults. This challenge is exacerbated in pre-hospital settings, where information might be scarce. Physiological alterations indicating changes in a patient's condition can be detected hours preceding a cardiac arrest. Therefore, maintaining continuous monitoring of the patient's clinical condition is crucial to detecting any physiological changes promptly, facilitating early identification of critical illness. This scoping review aims to assess the extent, range and nature of published research related to recognising paediatric out-of-hospital clinical deterioration by pre-hospital staff.Methods and analysisThis scoping review is registered with the Open Science Framework. The review will follow the Joanna Briggs Institute's (JBI) methodology for scoping reviews. A systematic search of relevant databases (MEDLINE, EMBASE, Web of Science, CINAHL and Scopus) will be conducted. In this scoping review, all types of study designs including quantitative and qualitative studies will be considered. The inclusion is limited to English-language studies published between January 1990 and March 2024. Two independent reviewers (AG and SS) will conduct a thorough screening of titles and abstracts against the pre-defined inclusion criteria for the review. For the selected citations, the full texts will undergo detailed assessment by the two reviewers, ensuring alignment with the inclusion criteria. A quality assessment of the included studies will be done using the Mixed Methods Appraisal Tool. The findings will be presented using diagrams or tables, supplemented by narrative summaries following the JBI guidelines.Ethics and disseminationEthical approval is not required. The findings will be disseminated through publication in a peer-reviewed journal and presentation at conferences and/or seminars.
| S-EPMC11261697 | biostudies-literature
Project description:IntroductionConsumer escalation systems allow patients and families to escalate concerns about acute clinical deterioration. Hospital staff can impact upon the success of this process. As part of evaluation processes within a Local Health Network, where a consumer escalation system was introduced in accordance with National requirements, we sought to explore clinicians' understanding and perceptions of consumer escalation.MethodsVoluntary and anonymous staff surveys pre, and post, system introduction. Quantitative data was analysed using descriptive statistics, chi-square independence, and non-parametric independent samples median tests. Qualitative data was evaluated using content analysis and cross-referenced with quantitative responses.ResultsRespondent's (pre: 215; post: 89) area of work varied significantly between survey periods. Most agreed that patients/families have a sound knowledge of a patient's typical health status (pre: 192/215 (89.3%); post 82/88 (93.2%)) and that patients/families should be encouraged to escalate concerns of deterioration to ward staff (pre: 209/212 (98.6%); post: 85/89 (95.5%)). Respondent perceptions of patient/family ability to recognise clinical deterioration varied. Staff agreement towards local response expectations decreased as the degree of clinical requirement increased. Staff concerns of increased workloads (pre: 90/214 (42.1%); post 12/72 (16.7%), p<0.001) and conflict generation (pre: 71/213 (33.3%); post: 7/71 (9.9%), p = 0.001) decreased significantly following system introduction. However, clinician perceptions of positive system effects also decreased (patient-staff rapport pre: 163/213 (76.5%); post: 38/72 (52.8%), p = 0.001; patient centred care pre: 188/214 (87.9%); post: 53/72 (73.6%), p = 0.012; patient safety pre: 173/214 (80.8%); post: 49/72 (68.1%), p = 0.077). Only 53% of respondents (pre: 112/213 (52.6%); post: 48/88 (54.5%)) perceived that patient/family have sufficient confidence to escalate concerns.ConclusionConsumer escalation systems require staff support. Staff perceptions may indicate, and act as, barriers to the operation of consumer escalation processes. Further exploration in identifying and managing staff barriers is crucial to the success of consumer escalation.
| S-EPMC9202900 | biostudies-literature