Development and validation of a dynamic delirium prediction rule in patients admitted to the Intensive Care Units (DYNAMIC-ICU): A prospective cohort study

Predictive model of risk factors for 28-day mortality in patients with sepsis or sepsis-associated delirium based on the MIMIC-IV database

Research on the severity and prognosis of sepsis with or without progressive delirium is relatively insufficient. We constructed a prediction model of the risk factors for 28-day mortality in patients who developed sepsis or sepsis-associated delirium. The modeling group of patients diagnosed with Sepsis-3 and patients with progressive delirium of related indicators were selected from the MIMIC-IV database. Relevant independent risk factors were determined and integrated into the prediction model. Receiver operating characteristic (ROC) curves and the Hosmer-Lemeshow (HL) test were used to evaluate the prediction accuracy and goodness-of-fit of the model. Relevant indicators of patients with sepsis or progressive delirium admitted to the intensive care unit (ICU) of a 3A hospital in Xinjiang were collected and included in the verification group for comparative analysis and clinical validation of the prediction model. The total length of stay in the ICU, hemoglobin levels, albumin levels, activated partial thrombin time, and total bilirubin level were the five independent risk factors in constructing a prediction model. The area under the ROC curve of the predictive model (0.904) and the HL test result (χ2 = 8.518) indicate a good fit. This model is valuable for clinical diagnosis and treatment and auxiliary clinical decision-making.

Intensive Care Unit Sleep Promotion Bundle: Impact on Sleep Quality, Delirium, and Other Patient Outcomes

BACKGROUND

High-quality sleep is important for optimal patient recovery. Sleep deprivation during hospitalization may lead to poor patient outcomes.

OBJECTIVE

To examine whether implementation of a sleep promotion bundle in the intensive care unit affects rates of delirium and agitation, restraint use, and length of stay.

METHODS

An evidence-based sleep promotion bundle was developed and implemented in 2 intensive care units in a 1025-bed level I trauma teaching hospital. Deidentified data from the electronic health record were obtained for patients hospitalized before and during the intervention. Data included scores on the Confusion Assessment Method for the Intensive Care Unit, Richmond Agitation-Sedation Scale, and Glasgow Coma Scale; restraint use; and hospital and intensive care unit length of stay.

RESULTS

A total of 137 patients during the preintervention period and 149 patients during the intervention period were hospitalized in the intensive care units and met inclusion criteria. A 9-percentage-point decrease in the incidence of delirium from before to during the intervention was found, although it was not statistically significant (P = .07). Significant reductions were found in both intensive care unit (P = .04) and hospital (P = .03) length of stay. A significant decrease was found in Richmond Agitation-Sedation Scale high scores for patients requiring mechanical ventilation (P = .03). No significant differences were found in Richmond Agitation-Sedation Scale low scores, Glasgow Coma Scale scores, or restraint use.

CONCLUSIONS

Critical care nurses are in an optimal position to implement evidence-based sleep promotion measures. Further research on sleep promotion bundles is needed.

Machine learning for the prediction of delirium in elderly intensive care unit patients

PURPOSE

This study aims to develop and validate a prediction model for delirium in elderly ICU patients and help clinicians identify high-risk patients at the early stage.

METHODS

Patients admitted to ICU for at least 24 h and using the Confusion Assessment Method for the Intensive Care Unit (CAM-ICU) in the Medical Information Mart for Intensive Care-IV (MIMIC-IV) database (76,943 ICU stays from 2008 to 2019) were considered. Patients with a positive delirium test in the first 24 h and under 65 years of age were excluded. Two prediction models, machine learning extreme gradient boosting (XGBoost) and logistic regression (LR) model, were developed and validated to predict the onset of delirium.

RESULTS

Of the 18,760 patients included in the analysis, 3463(18.5%) were delirium positive. A total of 22 significant predictors were selected by LASSO regression. The XGBoost model demonstrated superior performance over the LR model, with the Area Under the Receiver Operating Characteristic (AUC) values of 0.853 (95% confidence interval [CI] 0.846-0.861) and 0.831 (95% CI 0.815-0.847) in the training and testing datasets, respectively. Moreover, the XGBoost model outperformed the LR model in both calibration and clinical utility. The top five predictors associated with the onset of delirium were sequential organ failure assessment (SOFA), infection, minimum platelets, maximum systolic blood pressure (SBP), and maximum temperature.

CONCLUSION

The XGBoost model demonstrated good predictive performance for delirium among elderly ICU patients, thus assisting clinicians in identifying high-risk patients at the early stage and implementing targeted interventions to improve outcome.

Impact of delirium intervention on cognitive load among nurses in the intensive care unit: A multi-centre cluster randomized controlled trial

BACKGROUND

High cognitive load in nurses is a common problem in the intensive care unit (ICU). However, it remains unclear what different types of cognitive load the ICU nurses have experienced during the implementation of delirium interventions.

AIM

To describe the characteristics and explore the effect of implementing a delirium intervention on the cognitive load of nurses working in the ICU.

METHODS

A cluster-randomized controlled clinical trial was conducted. Six ICUs were randomized in a 1:1 ratio, and eligible nurses from these units provided either a delirium bundle intervention in addition to usual care (27 nurses) or usual care alone. An instrument was used to measure different types of cognitive load (MDT-CL), assessing intrinsic, extraneous and germane cognitive load. The repeated measures analysis of variance was used to detect between-group differences.

RESULTS

Among these nurses, significant between-group differences were identified in terms of their overall (P < 0.001), intrinsic (P < 0.001) and extraneous (P < 0.001) cognitive load. There was no significant change observed in the germane cognitive load (P = 0.489) in the delirium intervention group.

CONCLUSION

It is important to understand how the implementation of a delirium intervention affects different types of cognitive load in nurses, in order that tailored strategies can be applied to reduce cognitive load in ICU nurses.

Description of an individualised delirium intervention in intensive care units for critically ill patients delivered by an artificial intelligence-assisted system: using the TIDieR checklist

Background

Delirium is a preventable and reversible complication for intensive care unit (ICU) patients, which can be linked to negative outcomes. Early intervention to cope with the risk factors of delirium is necessary. Yet no specific description of the Artificial Intelligence Assisted Prevention and Management for Delirium (AI-AntiDelirium) following the Template for Intervention Description and Replication (TIDieR) checklist was reported. This is the first study to describe a detailed process for the development of an evidence-based delirium intervention.

Aims

To describe an individualised delirium intervention which is delivered by an artificial intelligence-assisted system in the ICU for critically ill patients.

Methods and results

The TIDieR checklist improved the description of ICU delirium interventions, including several key features for improved implementation of the intervention. This descriptive research describes the AI-assisted ICU delirium interventions for improving cognitive load and adherence of nurses and reducing ICU delirium incidence. Following the TIDieR checklist, we standardised the flow chart of ICU delirium assessment tools; formed an evaluation sheet of ICU delirium risk factors; and translated the evidence-based ABCDEF bundle intervention into practice. Therefore, nurses and researchers would benefit from replicating the interventions for clinical use or experimental research.

Conclusions

The TIDieR checklist provided a systematic approach for reporting the complex ICU delirium interventions delivered in a clinical interventional trial, which contributes to the nursing practice policy for the standardisation of interventions.

A cluster-randomized controlled trial of a nurse-led artificial intelligence assisted prevention and management for delirium (AI-AntiDelirium) on delirium in intensive care unit: Study protocol

BACKGROUND

Delirium is a common complication among intensive care unit (ICU) patients that is linked to negative clinical outcomes. However, adherence to the Clinical Practice Guidelines for the Prevention and Management of Pain, Agitation/Sedation, Delirium, Immobility, and Sleep Disruption in Adult Patients in the ICU (PADIS guidelines), which recommend the use of the ABCDEF bundle, is sub-optimal in routine clinical care. To address this issue, AI-AntiDelirium, a nurse-led artificial intelligence-assisted prevention and management tool for delirium, was developed by our research team. Our pilot study yielded positive findings regarding the use of AI-AntiDelirium in preventing patient ICU delirium and improving activities of daily living and increasing intervention adherence by health care staff.

METHODS

The proposed large-scale pragmatic, open-label, parallel-group, cluster randomized controlled study will assess the impact of AI-AntiDelirium on the incidence of ICU delirium and delirium-related outcomes. Six ICUs in two tertiary hospitals in China will be randomized in a 1:1 ratio to an AI-AntiDelirium or a PADIS guidelines group. A target sample size of 1,452 ICU patients aged 50 years and older treated in the ICU for at least 24 hours will be included. The primary outcome evaluated will be the incidence of ICU delirium and the secondary outcomes will be the duration of ICU delirium, length of ICU and hospital stay, ICU and in-hospital mortality rates, patient cognitive function, patient activities of daily living, and ICU nurse adherence to the ABCDEF bundle.

DISCUSSION

If this large-scale trial provides evidence of the effectiveness of AI-AntiDelirium, an artificial intelligence-assisted system tool, in decreasing the incidence of ICU delirium, length of ICU and hospital stay, ICU and in-hospital mortality rates, patient cognitive function, and patient activities of daily living while increasing ICU nurse adherence to the ABCDEF bundle, it will have a profound impact on the management of ICU delirium in both research and clinical practice.

CLINICAL TRIAL REGISTRATION

ChiCTR1900023711 (Chinese Clinical Trial Registry).

Impact of Clinical Decision Support System Assisted prevention and management for Delirium on guideline adherence and cognitive load among Intensive Care Unit nurses (CDSSD-ICU): Protocol of a multicentre, cluster randomized trial

BACKGROUND

Adherence to the delirium bundle intervention is sub-optimal in routine practice, and inappropriate use of the instructional design of interventions may result in higher cognitive load among nurses. It remains unclear whether the Clinical Decision Support System (CDSS) Assisted Prevention and Management for Delirium (CDSS-AntiDelirium) results in the improvement of adherence to delirium intervention and the reduction of extraneous cognitive load, as well as improving adherence to delirium intervention, among nurses in the intensive care unit (ICU).

METHODS

This study (named the CDSSD-ICU) is a multicentre, prospective, cluster randomized controlled clinical trial. A total of six ICUs in two hospitals will be randomized in a 1:1 ratio to receive either the CDSS-AntiDelirium group or the delirium guidelines group. The CDSS-AntiDelirium consists of four modules: delirium assessment tools, risk factor assessment, a nursing care plan, and a nursing checklist module. Each day, nurses will assess ICU patients with the assistance of the CDSS-AntiDelirium. A total of 78 ICU nurses are needed to ensure statistical power. Outcome assessments will be conducted by investigators who are blinded to group assignments. The primary endpoint will be adherence to delirium intervention, the secondary endpoint will be nurses' cognitive load measured using an instrument to assess different types of cognitive load. Repeated measures analysis of variance will be used to detect group differences. A structural equation model will be used to clarify the mechanism of improvement in adherence.

DISCUSSION

Although the CDSS has been widely used in hospitals for disease assessment, management, and recording, the applications thereof in the area of delirium are still in infancy. This study could provide scientific evidence regarding the impact of a CDSS on nurses' adherence and cognitive load and promote its further development in future studies.

CLINICAL TRIAL REGISTRATION

ChiCTR1900023711 (Chinese Clinical Trial Registry).

Inclusion of frailty improved performance of delirium prediction for elderly patients in the cardiac intensive care unit (D-FRAIL): A prospective derivation and external validation study

BACKGROUND

The elderly patients admitted to cardiac intensive care unit (CICU) are at relatively high risk for developing delirium. A simple and reliable predictive model can benefit them from early recognition of delirium followed by timely and appropriate preventive strategies.

OBJECTIVE

To explore the role of frailty in delirium prediction and develop and validate a delirium predictive model including frailty for elderly patients in CICU.

DESIGN

A prospective, observational cohort study.

SETTINGS

CICU at China-Japan Friendship Hospital from March 1, 2022 to August 25, 2022 (derivation cohort); CICU at Beijing Anzhen Hospital affiliated to Capital Medical University from March 14, 2023 to May 8, 2023 (external validation cohort).

PARTICIPANTS

A total of 236 and 90 participants were enrolled in the derivation and external validation cohorts, respectively. Participants in the derivation cohort were assigned into either the delirium (n = 70) or non-delirium group (n = 166) based on the occurrence of delirium.

METHODS

The simplified Chinese version of the Confusion Assessment Method for the Diagnosis of Delirium in the Intensive Care Unit was used to assess delirium twice a day at 8:00-10:00 and 18:00-20:00 until the onset of delirium or discharge from the CICU. Frailty was assessed using the FRAIL scale during the first 24 h in the CICU. Other possible risk factors were collected prospectively through patient interviews and medical records review. After processing missing data via multiple imputations, univariate analysis and bootstrapped forward stepwise logistic regression were performed to select optimal predictors and develop the models. The models were internally validated using bootstrapping and evaluated comprehensively via discrimination, calibration, and clinical utility in both the derivation and external validation cohorts.

RESULTS

The study developed D-FRAIL predictive model using FRAIL score, hearing impairment, Acute Physiology and Chronic Health Evaluation-II score, and fibrinogen. The area under the receiver operating characteristic curve (AUC) was 0.937 (95% confidence interval [CI]: 0.907-0.967) and 0.889 (95%CI: 0.840-0.938) even after bootstrapping in the derivation cohort. Inclusion of frailty was demonstrated to improve the model performance greatly with the AUC increased from 0.851 to 0.937 (p < 0.001). In the external validation cohort, the AUC of D-FRAIL model was 0.866 (95%CI: 0.782-0.907). Calibration plots and decision curve analysis suggested good calibration and clinical utility of the D-FRAIL model in both the derivation and external validation cohorts.

CONCLUSIONS

For elderly patients in the CICU, FRAIL score is an independent delirium predictor and the D-FRAIL model demonstrates superior performance in predicting delirium.

Dynamic Delirium Prediction in the Intensive Care Unit using Machine Learning on Electronic Health Records

Delirium is a syndrome of acute brain failure which is prevalent amongst older adults in the Intensive Care Unit (ICU). Incidence of delirium can significantly worsen prognosis and increase mortality, therefore necessitating its rapid and continual assessment in the ICU. Currently, the common approach for delirium assessment is manual and sporadic. Hence, there exists a critical need for a robust and automated system for predicting delirium in the ICU. In this work, we develop a machine learning (ML) system for real-time prediction of delirium using Electronic Health Record (EHR) data. Unlike prior approaches which provide one delirium prediction label per entire ICU stay, our approach provides predictions every 12 hours. We use the latest 12 hours of ICU data, along with patient demographic and medical history data, to predict delirium risk in the next 12-hour window. This enables delirium risk prediction as soon as 12 hours after ICU admission. We train and test four ML classification algorithms on longitudinal EHR data pertaining to 16,327 ICU stays of 13,395 patients covering a total of 56,297 12-hour windows in the ICU to predict the dynamic incidence of delirium. The best performing algorithm was Categorical Boosting which achieved an area under receiver operating characteristic curve (AUROC) of 0.87 (95% Confidence Interval; C.I, 0.86-0.87). The deployment of this ML system in ICUs can enable early identification of delirium, thereby reducing its deleterious impact on long-term adverse outcomes, such as ICU cost, length of stay and mortality.

Development and Validation of a Delirium Risk Prediction Model for Elderly Patients Undergoing Elective Orthopedic Surgery

Purpose

This study aimed to develop and validate a post-operative delirium (POD) nomogram in a population of elderly patients undergoing elective orthopedic surgery.

Patients and Methods

A predictive model was developed based on a training dataset of 474 elderly patients undergoing elective orthopedic surgery from March 2021 to May 2022. POD was identified using the Confusion Assessment Methods (CAM). The least absolute shrinkage and selection operator (LASSO) method was used to screen risk factors, and prediction models were created by combining the outcomes with logistic regression analysis. We employ bootstrap validation for internal validation to examine the model's repeatability. The results were validated using a prospective study on 153 patients operated on from January 2022 to May 2022 at another institution.

Results

The predictors in the POD nomogram included age, the Mini-Mental State Examination(MMSE), sleep disorder, neurological disorders, preoperative serum creatinine (Pre-SCR), and ASA classification. The c-index of the model was 0.928 (95% confidence interval 0.898 ~ 0.957) and the bootstrap validation still achieved a high c-index of 0.912. The c-index of the external validation was 0.921. The calibration curve for the diagnostic probability showed good agreement between prediction by nomogram and actual observation.

Conclusion

By combining preoperative and intraoperative clinical risk factors, we created a POD risk nomogram to predict the probability of POD in elderly patients who undergo elective orthopedic surgery. It could be a tool for guiding individualized interventions.

Deep Learning-Based Recurrent Delirium Prediction in Critically Ill Patients

OBJECTIVES

To predict impending delirium in ICU patients using recurrent deep learning.

DESIGN

Retrospective cohort study.

SETTING

Fifteen medical-surgical ICUs across Alberta, Canada, between January 1, 2014, and January 24, 2020.

PATIENTS

Forty-three thousand five hundred ten ICU admissions from 38,426 patients.

INTERVENTIONS

None.

MEASUREMENTS AND MAIN RESULTS

We used ICU and administrative health data to train deep learning models to predict delirium episodes in the next two 12-hour periods (0-12 and 12-24 hr), starting at 24 hours after ICU admission, and to generate new predictions every 12 hours. We used a comprehensive set of 3,643 features, capturing patient history, early ICU admission information (first 24 hr), and the temporal dynamics of various clinical variables throughout the ICU admission. Our deep learning architecture consisted of a feature embedding, a recurrent, and a prediction module. Our best model based on gated recurrent units yielded a sensitivity of 0.810, a specificity of 0.848, a precision (positive predictive value) of 0.704, and an area under the receiver operating characteristic curve (AUROC) of 0.909 in the hold-out test set for the 0-12-hour prediction horizon. For the 12-24-hour prediction horizon, the same model achieved a sensitivity of 0.791, a specificity of 0.807, a precision of 0.637, and an AUROC of 0.895 in the test set.

CONCLUSIONS

Our delirium prediction model achieved strong performance by applying deep learning to a dataset that is at least one order of magnitude larger than those used in previous studies. Another novel aspect of our study is the temporal nature of our features and predictions. Our model enables accurate prediction of impending delirium in the ICU, which can potentially lead to early intervention, more efficient allocation of ICU resources, and improved patient outcomes.

Predicting Intensive Care Delirium with Machine Learning: Model Development and External Validation

BACKGROUND

Delirium poses significant risks to patients, but countermeasures can be taken to mitigate negative outcomes. Accurately forecasting delirium in intensive care unit (ICU) patients could guide proactive intervention. Our primary objective was to predict ICU delirium by applying machine learning to clinical and physiologic data routinely collected in electronic health records.

METHODS

Two prediction models were trained and tested using a multicenter database (years of data collection 2014 to 2015), and externally validated on two single-center databases (2001 to 2012 and 2008 to 2019). The primary outcome variable was delirium defined as a positive Confusion Assessment Method for the ICU screen, or an Intensive Care Delirium Screening Checklist of 4 or greater. The first model, named "24-hour model," used data from the 24 h after ICU admission to predict delirium any time afterward. The second model designated "dynamic model," predicted the onset of delirium up to 12 h in advance. Model performance was compared with a widely cited reference model.

RESULTS

For the 24-h model, delirium was identified in 2,536 of 18,305 (13.9%), 768 of 5,299 (14.5%), and 5,955 of 36,194 (11.9%) of patient stays, respectively, in the development sample and two validation samples. For the 12-h lead time dynamic model, delirium was identified in 3,791 of 22,234 (17.0%), 994 of 6,166 (16.1%), and 5,955 of 28,440 (20.9%) patient stays, respectively. Mean area under the receiver operating characteristics curve (AUC) (95% CI) for the first 24-h model was 0.785 (0.769 to 0.801), significantly higher than the modified reference model with AUC of 0.730 (0.704 to 0.757). The dynamic model had a mean AUC of 0.845 (0.831 to 0.859) when predicting delirium 12 h in advance. Calibration was similar in both models (mean Brier Score [95% CI] 0.102 [0.097 to 0.108] and 0.111 [0.106 to 0.116]). Model discrimination and calibration were maintained when tested on the validation datasets.

CONCLUSIONS

Machine learning models trained with routinely collected electronic health record data accurately predict ICU delirium, supporting dynamic time-sensitive forecasting.

EDITOR’S PERSPECTIVE

Predisposing and Precipitating Factors Associated With Delirium: A Systematic Review

IMPORTANCE

Despite discrete etiologies leading to delirium, it is treated as a common end point in hospital and in clinical trials, and delirium research may be hampered by the attempt to treat all instances of delirium similarly, leaving delirium management as an unmet need. An individualized approach based on unique patterns of delirium pathophysiology, as reflected in predisposing factors and precipitants, may be necessary, but there exists no accepted method of grouping delirium into distinct etiologic subgroups.

OBJECTIVE

To conduct a systematic review to identify potential predisposing and precipitating factors associated with delirium in adult patients agnostic to setting.

EVIDENCE REVIEW

A literature search was performed of PubMed, Embase, Web of Science, and PsycINFO from database inception to December 2021 using search Medical Subject Headings (MeSH) terms consciousness disorders, confusion, causality, and disease susceptibility, with constraints of cohort or case-control studies. Two reviewers selected studies that met the following criteria for inclusion: published in English, prospective cohort or case-control study, at least 50 participants, delirium assessment in person by a physician or trained research personnel using a reference standard, and results including a multivariable model to identify independent factors associated with delirium.

FINDINGS

A total of 315 studies were included with a mean (SD) Newcastle-Ottawa Scale score of 8.3 (0.8) out of 9. Across 101 144 patients (50 006 [50.0%] male and 49 766 [49.1%] female patients) represented (24 015 with delirium), studies reported 33 predisposing and 112 precipitating factors associated with delirium. There was a diversity of factors associated with delirium, with substantial physiological heterogeneity.

CONCLUSIONS AND RELEVANCE

In this systematic review, a comprehensive list of potential predisposing and precipitating factors associated with delirium was found across all clinical settings. These findings may be used to inform more precise study of delirium's heterogeneous pathophysiology and treatment.

Prediction models for postoperative delirium after cardiac surgery: Systematic review and critical appraisal

BACKGROUND

Many studies have developed or validated prediction models to estimate the risk of delirium after cardiac surgery, but the quality of the model development and model applicability remain unknown.

OBJECTIVES

To systematically review and critically evaluate currently available prediction models for delirium after cardiac surgery.

DATA SOURCES

PubMed, EMBASE, and MEDLINE were systematically searched. This systematic review was registered in PROSPERO (Registration ID: CRD42021251226).

STUDY SELECTION

Prospective or retrospective cohort studies were considered eligible if they developed or validated prediction models or scoring systems for delirium in the ICU. We included studies involving adults (age ≥18 years) undergoing cardiac surgery and excluded studies that did not validate a prediction model.

DATA EXTRACTION

Data extraction was independently performed by two authors using a standardized data extraction form based on the Critical Appraisal and Data Extraction for Systematic Reviews of Prediction Modeling Studies checklist. Quality of the models was assessed with the Prediction Model Risk of Bias Assessment Tool (PROBAST).

DATA SYNTHESIS

Of 5469 screened studies, 13 studies described 10 prediction models. The postoperative delirium incidence varied from 11.3 % to 51.6 %. The most frequently used predictors were age and cognitive impairment. The reported areas under the curve or C-statistics were between of 0.74 and 0.91 in the derivation set. The reported AUCs in the external validation set were between 0.54 and 0.90. All the studies had a high risk of bias, mainly owing to poor reporting of the outcome domain and analysis domain; 10 studies were of high concern regarding applicability.

CONCLUSIONS

The current models for predicting postoperative delirium in the ICU after cardiac surgery had a high risk of bias according to the PROBAST. Future studies should focus on improving current prediction models or developing new models with rigorous methodology.

Recipe for primary prevention of delirium in hospitalized older patients

Delirium is an acute fluctuating syndrome characterized by a change in consciousness, perception, orientation, cognition, sleep-wake rhythm, psychomotor skills, and the mood and feelings of a patient. Delirium and delirium prevention remain a challenge for healthcare professionals, especially nurses who form the basis of patient care. It also causes distress for patients, their caregivers and healthcare professionals. However, delirium is preventable in 30-40% of cases. The aim of this article is to summarize the delirium risk models, delirium screening tools, and (non-pharmacological) delirium prevention strategies. A literature search of review articles supplemented by original articles published in PubMed, Cinahl, and Cochrane between 1 January 2000 and 31 December 2020 was carried out. Among the older patients, delirium is a common condition with major consequences in terms of mortality and morbidity, but prevention is possible. Despite the fact that delirium risk models, delirium screening scales and non-pharmacological prevention are available for the development of a hospital delirium prevention programme, such a programme is still not commonly used on a daily basis.

Risk factors for postoperative delirium in children with congenital heart disease: a prospective nested case-control study

OBJECTIVES

To study the risk factors for postoperative delirium (POD) in children with congenital heart disease.

METHODS

A prospective nested case-control study was performed on children with congenital heart disease who underwent surgery in Fuwai Hospital, Chinese Academy of Medical Sciences, from December 2020 to June 2021. The clinical data were compared between the POD group (n=114) and non-POD group (n=102). A multivariate unconditional logistic regression analysis was used to investigate the risk factors for POD in children with congenital heart disease.

RESULTS

The multivariate logistic regression analysis showed that age (OR=0.951, P<0.001), gender (OR=2.127, P=0.049), number of invasive catheters per day (OR=1.490, P=0.017), degree of postoperative pain (OR=5.856, P<0.001), and preoperative parental anxiety level (OR=1.025, P=0.010) were independent risk factors for POD in children with congenital heart disease.

CONCLUSIONS

The risk of POD increases in children with congenital heart disease who are younger, male, have higher number of invasive catheters per day, higher degree of postoperative pain, or higher preoperative parental anxiety level.

Gender Differences and Postoperative Delirium in Adult Patients Undergoing Cardiac Valve Surgery

Background: Postoperative delirium (POD) is common in patients following cardiac surgery. According to studies on non-cardiac surgery, males suffered from higher incidence of POD. However, there is no report about effect of gender differences on POD occurrence in cardiac surgery patients. The aim of this study was to investigate the effect of gender differences on POD occurrence in adult patients after cardiac valve surgery.

Methods: This is a retrospective case-control study. We recorded the clinical data in adult patients who underwent elective cardiac valve surgery from May 2019 to October 2020. POD was assessed by the Confusion Assessment Method for Intensive Care Unit. Univariate analysis was used to screen the potential risk factors. Collinearity analysis was conducted to detect overlapping predictor variables on the outcomes. A multivariate logistic regression with odds ratio (OR) and 95% confidence interval (CI) was used to identify the independent risk factors. The Hosmer-Lemeshow test was performed to show the good calibration of the logistic regression model.

Results: In total, we recorded the perioperative data in 431 adult patients, including 212 males and 219 females. Sixty patients suffered from POD, including 39 males and 21 females. Twenty-one perioperative variables were selected, and 11 were screened by univariate analysis. We did not detect the severe collinearity among the 11 variables. Male gender was identified as a significant risk factor in POD occurrence in patients undergoing cardiac surgery (Adjusted OR: 2.213, 95% CI: 1.049–4.670, P = 0.037). The Hosmer-Lemeshow test demonstrated good calibration of the logistic regression model (χ² = 7.238, P = 0.511). Besides, compared with females, the relationship of male and delirium subtypes was as follows: (1) hyperactive: adjusted OR: 3.384, 95% CI: 1.335–8.580, P = 0.010; (2) hypoactive: adjusted OR: 0.509, 95% CI: 0.147–1.766, P = 0.287. A Stratification analysis by age demonstrated that the males showed higher POD incidence in patients aged younger than 60 years (adjusted OR: 4.384, 95% CI: 1.318–14.586, P = 0.016).

Conclusions: Male gender is an important risk factor in POD occurrence in patients following cardiac surgery. Furthermore, the incidence of hyperactive delirium is higher in males. Besides, the male patients aged younger than 60 years are at high risk of POD. We should pay more attention to the male patients to prevent their POD occurrence.

Delirium in Intensive Care

Purpose of Review

Delirium in the intensive care unit (ICU) has become increasingly acknowledged as a significant problem for critically ill patients affecting both the actual course of illness as well as outcomes. In this review, we focus on the current evidence and the gaps in knowledge.

Recent Findings

This review highlights several areas in which the evidence is weak and further research is needed in both pharmacological and non-pharmacological treatment. A better understanding of subtypes and their different response to therapy is needed and further studies in aetiology are warranted. Larger studies are needed to explore risk factors for developing delirium and for examining long-term consequences. Finally, a stronger focus on experienced delirium and considering the perspectives of both patients and their families is encouraged.

Summary

With the growing number of studies and a better framework for research leading to stronger evidence, the outcomes for patients suffering from delirium will most definitely improve in the years to come.

A Machine Learning Approach for Investigating Delirium as a Multifactorial Syndrome

Delirium is a psycho-organic syndrome common in hospitalized patients, especially the elderly, and is associated with poor clinical outcomes. This study aims to identify the predictors that are mostly associated with the risk of delirium episodes using a machine learning technique (MLT). A random forest (RF) algorithm was used to evaluate the association between the subject’s characteristics and the 4AT (the 4 A’s test) score screening tool for delirium. RF algorithm was implemented using information based on demographic characteristics, comorbidities, drugs and procedures. Of the 78 patients enrolled in the study, 49 (63%) were at risk for delirium, 32 (41%) had at least one episode of delirium during the hospitalization (38% in orthopedics and 31% both in internal medicine and in the geriatric ward). The model explained 75.8% of the variability of the 4AT score with a root mean squared error of 3.29. Higher age, the presence of dementia, physical restraint, diabetes and a lower degree are the variables associated with an increase of the 4AT score. Random forest is a valid method for investigating the patients’ characteristics associated with delirium onset also in small case-series. The use of this model may allow for early detection of delirium onset to plan the proper adjustment in healthcare assistance.

Profiling Delirium Progression in Elderly Patients via Continuous-Time Markov Multi-State Transition Models

Poor recognition of delirium among hospitalized elderlies is a typical challenge for health care professionals. Considering methodological insufficiency for assessing time-varying diseases, a continuous-time Markov multi-state transition model (CTMMTM) was used to investigate delirium evolution in elderly patients. This is a longitudinal observational study performed in September 2016 in an Italian hospital. Change of delirium states was modeled according to the 4AT score. A Cox model (CM) and a CTMMTM were used for identifying factors affecting delirium onset both with a two-state and three-state model. In this study, 78 patients were enrolled and evaluated for 5 days. Both the CM and the CTMMTM show that urine catheter (UC), aging, drugs, and invasive devices (ID) are risk factors for delirium onset. The CTMMTM model shows that transition from no-delirium/cognitive impairment to delirium was associated with aging (HR = 1.14; 95%CI, 1.05, 1.23) and neuroleptics (HR = 4.3; 1.57, 11.77), dopaminergic drugs (HR = 3.89; 1.2, 12.6), UC (HR = 2.92; 1.09, 7.79) and ID (HR = 1.67; 103, 2.71). These results are confirmed by the multivariable model. Aging, ID, antibiotics, drugs affecting the central nervous system, and absence of moving ability are identified as the significant predictors of delirium. Additionally, it seems that modeling with CTMMTM may show associations that are not directly detectable with the traditional CM.

Delirium screening for patients in the intensive care unit: A prospective validation study of the iCAM-ICU by nurse researchers and bedside nurses in routine practice

BACKGROUND

Daily delirium assessment using the Confusion Assessment Method for the Intensive Care Unit was recommended for patients in the intensive care unit, yet implementation may be difficult because of lack of simple and standardized data collection schemes which may result in low sensitivities when used by bedside nurses.

OBJECTIVE

This study was to validate the diagnostic accuracy of the intelligent Confusion Assessment Method for the Intensive Care Unit (iCAM-ICU) used by both nurse investigators and bedside nurses in Chinese patients in the intensive care unit.

DESIGN

Prospective cohort study.

SETTING

A university affiliated tertiary hospital in China.

PARTICIPANTS

A total of 373 hospitalized patients (181 in phase I and 192 in phase II) in the intensive care units met the inclusion criteria and participated in the study. There were two nursing researchers (phase I) and 24 bedside nurses (phase II) used the iCAM-ICU to assess delirium among patients.

METHODS

Two prospective cohort studies were conducted to validate the diagnostic accuracy of iCAM-ICU on delirium screening when it was used by nurse investigators and bedside nurses in the intensive care unit. Using the Diagnostic and Statistical Manual of Mental Disorders as the gold standard, the sensitivity, specificity, predictive values along with the likelihood ratios were determined to estimate the performance of the iCAM-ICU in patients in the intensive care setting. The Kappa consistency test was examined to determine the inter-rater consistency. Subgroup analysis in terms of different age, level of education, severity of illness and cognitive status were also conducted to evaluate potential variations of the iCAM-ICU performance in different patient groups.

RESULTS

A total of 373 patients were included in the validation studies. In comparing with the gold standard, the sensitivities of the iCAM-ICU demonstrated by the two nurse investigators were 95.2 % and 93.7%, while the specificities of the iCAM-ICU were 93.3% and 93.2%. The Kappa consistency between two nurse investigators was 0.96. The sensitivity and specificity of the iCAM-ICU demonstrated by bedside nurses in intensive care patients were 86.7% and 97.7%, respectively. Subgroup analysis also revealed that the sensitivities and specificities in those different subgroups were acceptable, with all statistics being above 80%.

CONCLUSIONS

The iCAM-ICU, an information technology enabled delirium screening tool, showed highly acceptable accuracy in detecting delirium in the intensive care units. It can assist bedside nurses to detect delirium reliably and identify potential patients with delirium accurately.

REGISTRATION NUMBER

ChiCTR-OCH-13003050.

ICU Delirium-Prediction Models: A Systematic Review

Supplemental Digital Content is available in the text.

Objective:

Summarize performance and development of ICU delirium-prediction models published within the past 5 years.

Data Sources:

Systematic electronic searches were conducted in April 2019 using PubMed, Embase, Cochrane Central, Web of Science, and Cumulative Index to Nursing and Allied Health Literature to identify peer-reviewed studies.

Study Selection:

Eligible studies were published in English during the past 5 years that specifically addressed the development, validation, or recalibration of delirium-prediction models in adult ICU populations.

Data Extraction:

Screened citations were extracted independently by three investigators with a 42% overlap to verify consistency using the CHecklist for critical Appraisal and data extraction for systematic Reviews of prediction Modelling Studies.

Data Synthesis:

Eighteen studies featuring 23 distinct prediction models were included. Model performance varied greatly, as assessed by area under the receiver operating characteristic curve (0.62–0.94), specificity (0.50–0.97), and sensitivity (0.45–0.96). Most models used data collected from a single time point or window to predict the occurrence of delirium at any point during hospital or ICU admission, and lacked mechanisms for providing pragmatic, actionable predictions to clinicians.

Conclusions:

Although most ICU delirium-prediction models have relatively good performance, they have limited applicability to clinical practice. Most models were static, making predictions based on data collected at a single time-point, failing to account for fluctuating conditions during ICU admission. Further research is needed to create clinically relevant dynamic delirium-prediction models that can adapt to changes in individual patient physiology over time and deliver actionable predictions to clinicians.

Delirium risk prediction models for intensive care unit patients: A systematic review

OBJECTIVE

To systematically review the delirium risk prediction models for intensive care unit (ICU) patients.

METHODS

A systematic review was conducted. The Cochrane Library, PubMed, Ovid and Web of Science were searched to collect studies on delirium risk prediction models for ICU patients from database establishment to 31 March 2019. Two reviewers independently screened the literature according to the pre-determined inclusion and exclusion criteria, extracted the data and evaluated the risk of bias of the included studies using the CHARMS (CHecklist for critical Appraisal and data extraction for systematic Reviews of prediction Modelling Studies) checklist. A descriptive analysis was used to describe and summarise the data.

RESULTS

A total of six models were included. All studies reported the area under the receiver operating characteristic curve (AUROC) of the prediction models in the derivation and (or) validation datasets as over 0.7 (from 0.75 to 0.9). Five models reported calibration metrics. Decreased cognitive reserve and the Acute Physiology and Chronic Health Evaluation II (APACHE-II) score were the most commonly reported predisposing and precipitating factors, respectively, of ICU delirium among all models. The small sample size, lack of external validation and the absence of or unreported blinding method increased the risk of bias.

CONCLUSION

According to the discrimination and calibration statistics reported in the original studies, six prediction models may have moderate power in predicting ICU delirium. However, this finding should be interpreted with caution due to the risk of bias in the included studies. More clinical studies should be carried out to validate whether these tools have satisfactory predictive performance in delirium risk prediction for ICU patients.

Delirium prediction in the intensive care unit: a temporal approach

The incidence of delirium in intensive care units is high and associated with poor outcomes; therefore, its prediction is desirable to establish preventive treatments. This retrospective study proposes a novel approach for delirium prediction. We analyzed static and temporal data from 10,475 patients admitted to one of 15 intensive care units (ICUs) in Alberta, Canada between January 1, 2014 and June 30, 2016. We tested 168 different combinations of study design parameters and five different predictive models (logistic regression, support vector machines, random forests, adaptive boosting and neural networks). The area under the receiver operating characteristic curve (AUROC) ranged from 0.754 (CI 95% ± 0.018) to 0.852 (± 0.033), with sensitivity and specificity respectively ranging from 0.739 (CI 95% ± 0.047) to 0.840 (CI 95% ± 0.064), and 0.770 (CI 95% ± 0.030) to 0.865 (CI 95% ± 0.038). These results are similar to previous studies; however, our approach allows for continuous updates and short-term prediction horizons which might provide major advantages.

Development and validation of the delirium risk assessment score (DRAS)

PURPOSE

Development and validation of a delirium risk assessment score. Predisposing risk factors for delirium were used, which are easily assessed at hospital admission without additional clinical or laboratory testing.

METHODS

A systematic literature search identified ten risk factors: acute admission, alcohol use > 4 units/day, cognitive impairment, ADL impairment, age > 75 years, earlier delirium, hearing/vision problems, number of medication ≥ 5, number of morbidities > 2 and male. The DRAS was developed in a mixed patient population (N = 842) by the use of univariate and multivariate analyses and -2 log-likelihood calculation to weigh the risk factors. Based on the sensitivity and specificity, a cutoff score was calculated. The validation was performed in 3 cohorts (N = 408, N = 186, N = 365). In cohort 3, the DRAS was compared (AUC, sensitivity and specificity) to 3 instruments (Inouye, Kalisvaart, VMS rules).

RESULTS

The delirium incidence was 31.8%, 20.3%, 15.6% and 15.1%. All risk factors were independently predictive for delirium, except male. The multivariate analyses excluded morbidities. The final DRAS consists of 8 items; acute admission, cognitive impairment, alcohol use (3 points), ADLimpairment/mobilityproblems (2 points), higher age, earlier delirium, hearing/vision problems, and medication (1 point). The total score is 15 points and at a cut-of score of 5 or higher the patient is at risk of developing a delirium. The cutoff was at 5 or more points, AUC: 0.76 (95% CI 0.72-0.79), sensitivity 0.77, specificity 0.60. Validation cohorts AUC was 0.75 (95% CI 0.96-0.81), 0.76 (95% CI 0.70-0.83) and 0.78 (95% CI 0.70-0.87), sensitivity 0.71, 0.67 and 0.89 and specificity 0.70, 0.72 and 0.60. The comparison revealed the highest AUC for the DRAS.

CONCLUSION

Based on an admission interview, the delirium risk can be easily evaluated using the DRAS shortlist score of predisposing risk factors for delirium in older inpatients.