Delirium misdiagnosis risk in psychiatry: a machine learning-logistic regression predictive algorithm

Evaluation of a Natural Language Processing Approach to Identify Diagnostic Errors and Analysis of Safety Learning System Case Review Data: Retrospective Cohort Study

BACKGROUND

Diagnostic errors are an underappreciated cause of preventable mortality in hospitals and pose a risk for severe patient harm and increase hospital length of stay.

OBJECTIVE

This study aims to explore the potential of machine learning and natural language processing techniques in improving diagnostic safety surveillance. We conducted a rigorous evaluation of the feasibility and potential to use electronic health records clinical notes and existing case review data.

METHODS

Safety Learning System case review data from 1 large health system composed of 10 hospitals in the mid-Atlantic region of the United States from February 2016 to September 2021 were analyzed. The case review outcome included opportunities for improvement including diagnostic opportunities for improvement. To supplement case review data, electronic health record clinical notes were extracted and analyzed. A simple logistic regression model along with 3 forms of logistic regression models (ie, Least Absolute Shrinkage and Selection Operator, Ridge, and Elastic Net) with regularization functions was trained on this data to compare classification performances in classifying patients who experienced diagnostic errors during hospitalization. Further, statistical tests were conducted to find significant differences between female and male patients who experienced diagnostic errors.

RESULTS

In total, 126 (7.4%) patients (of 1704) had been identified by case reviewers as having experienced at least 1 diagnostic error. Patients who had experienced diagnostic error were grouped by sex: 59 (7.1%) of the 830 women and 67 (7.7%) of the 874 men. Among the patients who experienced a diagnostic error, female patients were older (median 72, IQR 66-80 vs median 67, IQR 57-76; P=.02), had higher rates of being admitted through general or internal medicine (69.5% vs 47.8%; P=.01), lower rates of cardiovascular-related admitted diagnosis (11.9% vs 28.4%; P=.02), and lower rates of being admitted through neurology department (2.3% vs 13.4%; P=.04). The Ridge model achieved the highest area under the receiver operating characteristic curve (0.885), specificity (0.797), positive predictive value (PPV; 0.24), and F1-score (0.369) in classifying patients who were at higher risk of diagnostic errors among hospitalized patients.

CONCLUSIONS

Our findings demonstrate that natural language processing can be a potential solution to more effectively identifying and selecting potential diagnostic error cases for review and therefore reducing the case review burden.

Novel opportunities for clinical pharmacy research: development of a machine learning model to identify medication related causes of delirium in different patient groups

The advent of artificial intelligence (AI) technologies has taken the world of science by storm in 2023. The opportunities of this easy to access technology for clinical pharmacy research are yet to be fully understood. The development of a custom-made large language model (LLM) (DELSTAR) trained on a wide range of internationally recognised scientific publication databases, pharmacovigilance sites and international product characteristics to help identify and summarise medication related information on delirium, as a proof-of-concept model, identified new facilitators and barriers for robust clinical pharmacy practice research. This technology holds great promise for the development of much more comprehensive prescribing guidelines, practice support applications for clinical pharmacy, increased patient and prescribing safety and resultant implications for healthcare costs. The challenge will be to ensure its methodologically robust use and the detailed and transparent verification of its information accuracy.

Delirium in psychiatric settings: risk factors and assessment tools in patients with psychiatric illness: a scoping review

BACKGROUND

Delirium is a common disorder affecting patients' psychiatric illness, characterized by a high rate of underdiagnosis, misdiagnosis, and high risks. However, previous studies frequently excluded patients with psychiatric illness, leading to limited knowledge about risk factors and optimal assessment tools for delirium in psychiatric settings.

OBJECTIVES

The scoping review was carried out to (1) identify the risk factors associated with delirium in patients with psychiatric illness; (2) synthesize the performance of assessment tools for detecting delirium in patients with psychiatric illness in psychiatric settings.

DESIGN

Scoping review.

DATA SOURCES

PubMed, Web of Science, and Embase were searched to identify primary studies on delirium in psychiatric settings from inception to Dec 2023 inclusive. Two independent reviewers screened eligible studies against inclusion criteria. A narrative synthesis of the included studies was conducted.

RESULTS

A final set of 36 articles meeting the inclusion criteria, two main themes were extracted: risk factors associated with delirium in patients with psychiatric illness and assessment tools for detecting delirium in psychiatric settings. The risk factors associated with delirium primarily included advanced age, physical comorbid, types of psychiatric illness, antipsychotics, anticholinergic drug, Electroconvulsive therapy, and the combination of lithium and Electroconvulsive therapy. Delirium Rating Scale-Revised-98, Memorial Delirium Assessment Scale, and Delirium Diagnostic Tool-Provisional might be valuable for delirium assessment in patients with psychiatric illness in psychiatric settings.

CONCLUSIONS

Delirium diagnosis in psychiatric settings is complex due to the overlapping clinical manifestations between psychiatric illness and delirium, as well as their potential co-occurrence. It is imperative to understand the risk factors and assessment methods related to delirium in this population to address diagnostic delays, establish effective prevention and screening strategies. Future research should focus on designing, implementing, and evaluating interventions that target modifiable risk factors, to prevent and manage delirium in patients with psychiatric illness.

Postoperative delirium prediction after cardiac surgery using machine learning models

OBJECTIVE

Postoperative delirium (POD) is a common postoperative complication in elderly patients, especially those undergoing cardiac surgery, which seriously affects the short- and long-term prognosis of patients. Early identification of risk factors for the development of POD can help improve the perioperative management of surgical patients. In the present study, five machine learning models were developed to predict patients at high risk of delirium after cardiac surgery and their performance was compared.

METHODS

A total of 367 patients who underwent cardiac surgery were retrospectively included in this study. Using single-factor analysis, 21 risk factors for POD were selected for inclusion in machine learning. The dataset was divided using 10-fold cross-validation for model training and testing. Five machine learning models (random forest (RF), support vector machine (SVM), radial based kernel neural network (RBFNN), K-nearest neighbour (KNN), and Kernel ridge regression (KRR)) were compared using area under the receiver operating characteristic curve (AUC-ROC), accuracy (ACC), sensitivity (SN), specificity (SPE), and Matthews coefficient (MCC).

RESULTS

Among 367 patients, 105 patients developed POD, the incidence of delirium was 28.6 %. Among the five ML models, RF had the best performance in ACC (87.99 %), SN (69.27 %), SPE (95.38 %), MCC (70.00 %) and AUC (0.9202), which was far superior to the other four models.

CONCLUSION

Delirium is common in patients after cardiac surgery. This analysis confirms the importance of the computational ML models in predicting the occurrence of delirium after cardiac surgery, especially the outstanding performance of the RF model, which has practical clinical applications for early identification of patients at risk of developing POD.

Delirium pathophysiology in cancer: neurofilament light chain biomarker - narrative review

Background Delirium is a debilitating disorder with high prevalence near the end of life, impacting quality of life of patients and their relatives. Timely recognition of delirium can lead to prevention and/or better treatment of delirium. According to current hypotheses delirium is thought to result from aberrant inflammation and neurotransmission, with a possible role for neuronal damage. Neurofilament light chain (NfL) is a protein biomarker in body fluids that is unique to neurons, with elevated levels when neurons are damaged, making NfL a viable biomarker for early detection of delirium. This narrative review summarises current research regarding the pathophysiology of delirium and the potential of NfL as a susceptibility biomarker for delirium and places this in the context of care for patients with advanced cancer.Results Six studies were conducted exclusively on NfL in patients with delirium. Three of these studies demonstrated that high plasma NfL levels preoperatively predict delirium in older adult patients postoperatively. Two studies demonstrated that high levels of NfL in intensive care unit (ICU) patients are correlated with delirium duration and severity. One study found that incident delirium in older adult patients was associated with increased median NfL levels during hospitalisation.Conclusions Targeted studies are required to understand if NfL is a susceptibility biomarker for delirium in patients with advanced cancer. In this palliative care context, better accessible matrices, such as saliva or urine, would be helpful for repetitive testing. Improvement of biological measures for delirium can lead to improved early recognition and lay the groundwork for novel therapeutic strategies.

Identification and information management of cognitive impairment of patients in acute care hospitals: An integrative review

Objectives

Recognition of the cognitive status of patients is important so that care can be tailored accordingly. The objective of this integrative review was to report on the current practices that acute care hospitals use to identify people with cognitive impairment and how information about cognition is managed within the healthcare record as well as the approaches required and recommended by policies.

Methods

Following Whittemore & Knafl's five-step method, we systematically searched Medline, CINAHL, and Scopus databases and various grey literature sources. Articles relevant to the programs that have been implemented in acute care hospitals regarding the identification of cognitive impairment and management of cognition information were included. The Mixed Methods Appraisal Tool and AACODS (Authority, Accuracy, Coverage, Objectivity, Date, Significance) Checklist were used to evaluate the quality of the studies. Thematic analysis was used to present and synthesise results. This review was pre-registered on PROSPERO ( CRD42022343577).

Results

Twenty-two primary studies and ten government/industry publications were included in the analysis. Findings included gaps between practice and policy. Although identification of cognitive impairment, transparency of cognition information, and interaction with patients, families, and carers (if appropriate) about this condition were highly valued at a policy level, sometimes in practice, cognitive assessments were informal, patient cognition information was not recorded, and interactions with patients, families, and carers were lacking.

Discussion

By incorporating cognitive assessment, developing an integrated information management system using information technology, establishing relevant laws and regulations, providing education and training, and adopting a national approach, significant improvements can be made in the care provided to individuals with cognitive impairment.

A protocol of an international validation study to assess the clinical accuracy of the eDIS-ICU delirium screening tool

BACKGROUND

Delirium is a common, yet underdiagnosed neuropsychiatric complication of intensive care unit (ICU) admission, associated with significant mortality and morbidity. Delirium can be difficult to diagnose, with gold standard assessments by a trained specialist being impractical and rarely performed. To address this, various tools have been developed, enabling bedside clinicians to assess for delirium efficiently and accurately. However, the performance of these tools varies depending on factors including the assessor's training. To address the shortcomings of current tools, electronic tools have been developed.

AIMS AND OBJECTIVES

The aims of this validation study are to assess the feasibility, acceptability, and generalisability of a recently developed and pilot-tested electronic delirium screening tool (eDIS-ICU) and compare diagnostic concordance, sensitivity, and specificity between eDIS-ICU, Confusion Assessment Method for the ICU (CAM-ICU), and the Diagnostic and Statistical Manual of Mental Disorders - 5th edition (DSM-V) gold standard in diverse ICU settings.

METHODS

Seven hundred participants will be recruited across five sites in three countries. Participants will complete three assessments (eDIS-ICU, CAM-ICU, and DSM-V) twice within one 24-h period. At each time point, assessments will be completed within one hour. Assessments will be administered by three different people at any given time point, with the assessment order and assessor for eDIS-ICU and CAM-ICU randomly allocated. Assessors will be blinded to previous and concurrent assessment results.

RESULTS

The primary outcome is comparing diagnostic sensitivity of eDIS-ICU and CAM-ICU against the DSM-V.

RELEVANCE TO CLINICAL PRACTICE

This protocol describes a definitive validation study of an electronic diagnostic tool to assess for delirium in the ICU. Delirium remains a common and difficult challenge in the ICU and is linked with multiple neurocognitive sequelae. Various challenges to routine assessment mean many cases are still unrecognised or misdiagnosed. An improved ability for bedside clinicians to screen for delirium accurately and efficiently will support earlier diagnosis, identification of underlying cause(s) and timely treatments, and ultimately improved patient outcomes.

CLINICAL TRIAL REGISTRATION NUMBER

This study was prospectively registered on the Australian New Zealand Clinical Trials Registry (ANZCTR) on 8th February 2022 (ACTRN12622000220763).

Towards experimental models of delirium utilizing zebrafish

Delirium is an acute neuropsychiatric condition characterized by impaired behavior and cognition. Although the syndrome has been known for millennia, its CNS mechanisms and risk factors remain poorly understood. Experimental animal models, especially rodent-based, are commonly used to probe various pathogenetic aspects of delirium. Complementing rodents, the zebrafish (Danio rerio) emerges as a promising novel model organism to study delirium. Zebrafish demonstrate high genetic and physiological homology to mammals, easy maintenance, robust behaviors in various sensitive behavioral tests, and the potential to screen for pharmacological agents relevant to delirium. Here, we critically discuss recent developments in the field, and emphasize the developing utility of zebrafish models for translational studies of delirium and deliriant drugs. Overall, the zebrafish represents a valuable and promising aquatic model species whose use may help understand delirium etiology, as well as develop novel therapies for this severely debilitating disorder.

A lung cancer risk warning model based on tongue images

Objective: To investigate the tongue image features of patients with lung cancer and benign pulmonary nodules and to construct a lung cancer risk warning model using machine learning methods. Methods: From July 2020 to March 2022, we collected 862 participants including 263 patients with lung cancer, 292 patients with benign pulmonary nodules, and 307 healthy subjects. The TFDA-1 digital tongue diagnosis instrument was used to capture tongue images, using feature extraction technology to obtain the index of the tongue images. The statistical characteristics and correlations of the tongue index were analyzed, and six machine learning algorithms were used to build prediction models of lung cancer based on different data sets. Results: Patients with benign pulmonary nodules had different statistical characteristics and correlations of tongue image data than patients with lung cancer. Among the models based on tongue image data, the random forest prediction model performed the best, with a model accuracy of 0.679 ± 0.048 and an AUC of 0.752 ± 0.051. The accuracy for the logistic regression, decision tree, SVM, random forest, neural network, and naïve bayes models based on both the baseline and tongue image data were 0.760 ± 0.021, 0.764 ± 0.043, 0.774 ± 0.029, 0.770 ± 0.050, 0.762 ± 0.059, and 0.709 ± 0.052, respectively, while the corresponding AUCs were 0.808 ± 0.031, 0.764 ± 0.033, 0.755 ± 0.027, 0.804 ± 0.029, 0.777 ± 0.044, and 0.795 ± 0.039, respectively. Conclusion: The tongue diagnosis data under the guidance of traditional Chinese medicine diagnostic theory was useful. The performance of models built on tongue image and baseline data was superior to that of the models built using only the tongue image data or the baseline data. Adding objective tongue image data to baseline data can significantly improve the efficacy of lung cancer prediction models.

Altered Tryptophan-Kynurenine Pathway in Delirium: A Review of the Current Literature

Delirium, a common form of acute brain dysfunction, is associated with increased morbidity and mortality, especially in older patients. The underlying pathophysiology of delirium is not clearly understood, but acute systemic inflammation is known to drive delirium in cases of acute illnesses, such as sepsis, trauma, and surgery. Based on psychomotor presentations, delirium has three main subtypes, such as hypoactive, hyperactive, and mixed subtype. There are similarities in the initial presentation of delirium with depression and dementia, especially in the hypoactive subtype. Hence, patients with hypoactive delirium are frequently misdiagnosed. The altered kynurenine pathway (KP) is a promising molecular pathway implicated in the pathogenesis of delirium. The KP is highly regulated in the immune system and influences neurological functions. The activation of indoleamine 2,3-dioxygenase, and specific KP neuroactive metabolites, such as quinolinic acid and kynurenic acid, could play a role in the event of delirium. Here, we collectively describe the roles of the KP and speculate on its relevance in delirium.

Boosting Delirium Identification Accuracy With Sentiment-Based Natural Language Processing: Mixed Methods Study

BACKGROUND

Delirium is an acute neurocognitive disorder that affects up to half of older hospitalized medical patients and can lead to dementia, longer hospital stays, increased health costs, and death. Although delirium can be prevented and treated, it is difficult to identify and predict.

OBJECTIVE

This study aimed to improve machine learning models that retrospectively identify the presence of delirium during hospital stays (eg, to measure the effectiveness of delirium prevention interventions) by using the natural language processing (NLP) technique of sentiment analysis (in this case a feature that identifies sentiment toward, or away from, a delirium diagnosis).

METHODS

Using data from the General Medicine Inpatient Initiative, a Canadian hospital data and analytics network, a detailed manual review of medical records was conducted from nearly 4000 admissions at 6 Toronto area hospitals. Furthermore, 25.74% (994/3862) of the eligible hospital admissions were labeled as having delirium. Using the data set collected from this study, we developed machine learning models with, and without, the benefit of NLP methods applied to diagnostic imaging reports, and we asked the question "can NLP improve machine learning identification of delirium?"

RESULTS

Among the eligible 3862 hospital admissions, 994 (25.74%) admissions were labeled as having delirium. Identification and calibration of the models were satisfactory. The accuracy and area under the receiver operating characteristic curve of the main model with NLP in the independent testing data set were 0.807 and 0.930, respectively. The accuracy and area under the receiver operating characteristic curve of the main model without NLP in the independent testing data set were 0.811 and 0.869, respectively. Model performance was also found to be stable over the 5-year period used in the experiment, with identification for a likely future holdout test set being no worse than identification for retrospective holdout test sets.

CONCLUSIONS

Our machine learning model that included NLP (ie, sentiment analysis in medical image description text mining) produced valid identification of delirium with the sentiment analysis, providing significant additional benefit over the model without NLP.

Social determinants of health and the prediction of missed breast imaging appointments

BACKGROUND

Predictive models utilizing social determinants of health (SDH), demographic data, and local weather data were trained to predict missed imaging appointments (MIA) among breast imaging patients at the Boston Medical Center (BMC). Patients were characterized by many different variables, including social needs, demographics, imaging utilization, appointment features, and weather conditions on the date of the appointment.

METHODS

This HIPAA compliant retrospective cohort study was IRB approved. Informed consent was waived. After data preprocessing steps, the dataset contained 9,970 patients and 36,606 appointments from 1/1/2015 to 12/31/2019. We identified 57 potentially impactful variables used in the initial prediction model and assessed each patient for MIA. We then developed a parsimonious model via recursive feature elimination, which identified the 25 most predictive variables. We utilized linear and non-linear models including support vector machines (SVM), logistic regression (LR), and random forest (RF) to predict MIA and compared their performance.

RESULTS

The highest-performing full model is the nonlinear RF, achieving the highest Area Under the ROC Curve (AUC) of 76% and average F1 score of 85%. Models limited to the most predictive variables were able to attain AUC and F1 scores comparable to models with all variables included. The variables most predictive of missed appointments included timing, prior appointment history, referral department of origin, and socioeconomic factors such as household income and access to caregiving services.

CONCLUSIONS

Prediction of MIA with the data available is inherently limited by the complex, multifactorial nature of MIA. However, the algorithms presented achieved acceptable performance and demonstrated that socioeconomic factors were useful predictors of MIA. In contrast with non-modifiable demographic factors, we can address SDH to decrease the incidence of MIA.

Liaison psychiatry before and after the COVID-19 pandemic

INTRODUCTION

the COVID-19 pandemic had an impact on hospital admissions. The clinical profiles of patients referred to liaison psychiatry teams (LPT) remained stable over the last few decades. We postulate changes in patient profiles due to the COVID-19 pandemic.

MATERIALS AND METHODS

a total of 384 patients admitted to a tertiary care University Hospital in Madrid (Spain) and referred to LPTs were recruited. Patients referred 5 months before and after the first admission for COVID-19 were included. Clinical and sociodemographic characteristics were collected, and non-parametric hypothesis contrast tests were used to study possible differences between both periods.

RESULTS

patients referred during the pandemic were significantly older (U = 2.006; p = .045), most of them were admitted to medical hospitalization units (χ2 (2) = 5.962; p = 015), and with a different reason for admission. There was an increase in the rate of adjustment disorders (χ2 (1) =7.893; p = 005) and delirium (χ2 (1) =9.413; p = 002), as well as psychiatric comorbidity (χ2 (2) = 9.930; p = .007), and a reduction in the proportion of patients treated for substance misuse (χ2 (5) = 19.152; p = .002). The number of deaths increased significantly (χ2 (1) = 6.611; p = .010). In persons over 65 years inappropriate prescription was significantly lower (χ2 (1) = 8.200; p = .004).

CONCLUSIONS

the pandemic had an impact on the activity of the LPTs due to the change in the clinical profile and evolution of referred patients, maintaining standards of care that are reflected through prescription.

Positive scores on the 4AT delirium assessment tool at hospital admission are linked to mortality, length of stay and home time: two-centre study of 82,770 emergency admissions

BACKGROUND

Studies investigating outcomes of delirium using large-scale routine data are rare. We performed a two-centre study using the 4 'A's Test (4AT) delirium detection tool to analyse relationships between delirium and 30-day mortality, length of stay and home time (days at home in the year following admission).

METHODS

The 4AT was performed as part of usual care. Data from emergency admissions in patients ≥65 years in Lothian, UK (n = 43,946) and Salford, UK (n = 38,824) over a period of $\sim$3 years were analysed using logistic regression models adjusted for age and sex.

RESULTS

4AT completion rates were 77% in Lothian and 49% in Salford. 4AT scores indicating delirium (≥4/12) were present in 18% of patients in Lothian, and 25% of patients in Salford. Thirty-day mortality with 4AT ≥4 was 5.5-fold greater than the 4AT 0/12 group in Lothian (adjusted odds ratio (aOR) 5.53, 95% confidence interval [CI] 4.99-6.13) and 3.4-fold greater in Salford (aOR 3.39, 95% CI 2.98-3.87). Length of stay was more than double in patients with 4AT scores of 1-3/12 (indicating cognitive impairment) or ≥ 4/12 compared with 4AT 0/12. Median home time at 1 year was reduced by 112 days (Lothian) and 61 days (Salford) in the 4AT ≥4 group (P < 0.001).

CONCLUSIONS

Scores on the 4AT used at scale in practice are strongly linked with 30-day mortality, length of hospital stay and home time. The findings highlight the need for better understanding of why delirium is linked with poor outcomes and also the need to improve delirium detection and treatment.

Longitudinal validation of an electronic health record delirium prediction model applied at admission in COVID-19 patients

OBJECTIVE

To validate a previously published machine learning model of delirium risk in hospitalized patients with coronavirus disease 2019 (COVID-19).

METHOD

Using data from six hospitals across two academic medical networks covering care occurring after initial model development, we calculated the predicted risk of delirium using a previously developed risk model applied to diagnostic, medication, laboratory, and other clinical features available in the electronic health record (EHR) at time of hospital admission. We evaluated the accuracy of these predictions against subsequent delirium diagnoses during that admission.

RESULTS

Of the 5102 patients in this cohort, 716 (14%) developed delirium. The model's risk predictions produced a c-index of 0.75 (95% CI, 0.73-0.77) with 27.7% of cases occurring in the top decile of predicted risk scores. Model calibration was diminished compared to the initial COVID-19 wave.

CONCLUSION

This EHR delirium risk prediction model, developed during the initial surge of COVID-19 patients, produced consistent discrimination over subsequent larger waves; however, with changing cohort composition and delirium occurrence rates, model calibration decreased. These results underscore the importance of calibration, and the challenge of developing risk models for clinical contexts where standard of care and clinical populations may shift.

Drug Treatment for Patients with Postoperative Delirium and Consultation-Liaison Psychiatry in Japan: A Retrospective Observational Study of a Nationwide Hospital Claims Database

BACKGROUND

Delirium is the most commonly experienced disorder in consultation liaisons. There are currently research and guidelines in Japan for delirium treatment. Still, there is no retrospective observational study of consultation-liaison psychiatry (CLP) and antipsychotic-centered drugs. This study aims to examine CLP's effectiveness and drug treatment.

METHODS

Using a Japanese national inpatient database of 2016 and 2017, we investigated the presence or absence of CLP for the treatment of delirium in postoperative delirium patients, the status of drug selection, delirium days, and the average days from surgery to discharge. We examined factors affecting days from surgery to discharge using multiple linear regression analysis.

RESULTS

This study was classified into a CLP group (n = 1,142) and a non-CLP group (n = 11,355). The days from surgery to discharge in the CLP and non-CLP groups was 16.7 and 17.1, respectively (p = 0.3613). There was a significant difference in the delirium days between the CLP and non-CLP groups (8.9 vs. 7.4; p < 0.00001). Haloperidol infusion was frequently used between the days from surgery to first day of delirium. It was prescribed less often than other oral drugs. Multiple regression analysis identified an association between age, men, CCI1-2, CCI ≥3, number of drugs used, days from surgery to first day of delirium, and early CLP (0-2days) with days from surgery to discharge.

CONCLUSIONS

We investigated the effectiveness of CLP and the actual conditions of pharmacotherapy for postoperative delirium. Our findings suggest that early CLP may be associated with shorter days from surgery to discharge.

Developing a machine learning model to identify delirium risk in geriatric internal medicine inpatients

PURPOSE

To develop a machine learning model that predicts delirium risk in geriatric internal medicine inpatients.

METHODS

A prospective cohort study of internal medicine wards in a tertiary care hospital in China. Blinded observers assessed delirium using the Confusion Assessment Method (CAM). The data set was randomly divided into a training set (70%) and a test set (30%). The model was trained on the training set using the decision tree and the five-fold cross-validation, and then the model performance was evaluated on the test set. Under-sampling was used to address the class imbalance. The discriminatory power of the model was measured by the area under the receiver operating characteristic curve (AUC) and F1 score. The data set comprised 740 patients from March 2016 to January 2017.

RESULTS

The training set included 518 patients; the median (IQR) age was 84 (79-87) years; 364 (70.3%) were men; 71 (13.7%) with delirium. The test set included 222 patients; the median (IQR) age was 84.5 (79-87) years; 163 (73.4%) were men; 30 (13.5%) with delirium. In total, the data set included 740 hospital admissions with a median (IQR) age of 84 (79-87) years, 527 (71.2%) were men, and 101 (13.6%) with delirium. From 32 potential predictors, we included five variables in the predictive model: depression, cognitive impairment, types of drugs, nutritional status, and activity of daily life (ADL). The mean AUC on the training set was 0.967, the AUC and F1 score on the test set was 0.950 and 0.810, respectively. The model achieved 93.3% sensitivity, 94.3% specificity, 71.8% positive predictive value, 98.9% negative predictive value, and 94.1% accuracy on the test set.

CONCLUSION

This machine learning model may allow more precise targeting of delirium prevention and could support clinical decision making in geriatric internal medicine wards.

Psychiatry in the Digital Age: A Blessing or a Curse?

Social distancing and the shortage of healthcare professionals during the COVID-19 pandemic, the impact of population aging on the healthcare system, as well as the rapid pace of digital innovation are catalyzing the development and implementation of new technologies and digital services in psychiatry. Is this transformation a blessing or a curse for psychiatry? To answer this question, we conducted a literature review covering a broad range of new technologies and eHealth services, including telepsychiatry; computer-, internet-, and app-based cognitive behavioral therapy; virtual reality; digital applied games; a digital medicine system; omics; neuroimaging; machine learning; precision psychiatry; clinical decision support; electronic health records; physician charting; digital language translators; and online mental health resources for patients. We found that eHealth services provide effective, scalable, and cost-efficient options for the treatment of people with limited or no access to mental health care. This review highlights innovative technologies spearheading the way to more effective and safer treatments. We identified artificially intelligent tools that relieve physicians from routine tasks, allowing them to focus on collaborative doctor-patient relationships. The transformation of traditional clinics into digital ones is outlined, and the challenges associated with the successful deployment of digitalization in psychiatry are highlighted.

Cognitive Computing-Based CDSS in Medical Practice

Importance. The last decade has witnessed the advances of cognitive computing technologies that learn at scale and reason with purpose in medicine studies. From the diagnosis of diseases till the generation of treatment plans, cognitive computing encompasses both data-driven and knowledge-driven machine intelligence to assist health care roles in clinical decision-making. This review provides a comprehensive perspective from both research and industrial efforts on cognitive computing-based CDSS over the last decade.Highlights. (1) A holistic review of both research papers and industrial practice about cognitive computing-based CDSS is conducted to identify the necessity and the characteristics as well as the general framework of constructing the system. (2) Several of the typical applications of cognitive computing-based CDSS as well as the existing systems in real medical practice are introduced in detail under the general framework. (3) The limitations of the current cognitive computing-based CDSS is discussed that sheds light on the future work in this direction.Conclusion. Different from medical content providers, cognitive computing-based CDSS provides probabilistic clinical decision support by automatically learning and inferencing from medical big data. The characteristics of managing multimodal data and computerizing medical knowledge distinguish cognitive computing-based CDSS from other categories. Given the current status of primary health care like high diagnostic error rate and shortage of medical resources, it is time to introduce cognitive computing-based CDSS to the medical community which is supposed to be more open-minded and embrace the convenience and low cost but high efficiency brought by cognitive computing-based CDSS.

Development and External Validation of a Delirium Prediction Model for Hospitalized Patients With Coronavirus Disease 2019

Background

The coronavirus disease 2019 pandemic has placed unprecedented stress on health systems and has been associated with elevated risk for delirium. The convergence of pandemic resource limitation and clinical demand associated with delirium requires careful risk stratification for targeted prevention efforts.

Objectives

To develop an incident delirium predictive model among coronavirus disease 2019 patients.

Methods

We applied supervised machine learning to electronic health record data for inpatients with coronavirus disease 2019 at three hospitals to build an incident delirium diagnosis prediction model. We validated this model in three different hospitals. Both hospital cohorts included academic and community settings.

Results

Among 2907 patients across 6 hospitals, 488 (16.8%) developed delirium. Applying the predictive model in the external validation cohort of 755 patients, the c-index was 0.75 (0.71–0.79) and the lift in the top quintile was 2.1. At a sensitivity of 80%, the specificity was 56%, negative predictive value 92%, and positive predictive value 30%. Equivalent model performance was observed in subsamples stratified by age, sex, race, need for critical care and care at community vs. academic hospitals.

Conclusion

Machine learning applied to electronic health records available at the time of inpatient admission can be used to risk-stratify patients with coronavirus disease 2019 for incident delirium. Delirium is common among patients with coronavirus disease 2019, and resource constraints during a pandemic demand careful attention to the optimal application of predictive models.

Implementation of ABCDEF care bundle in intensive care units: A cross-sectional survey

BACKGROUND

Delirium affects up to 80% of patients in intensive care units (ICUs) and is associated with higher mortality, physical dependence, and health care costs. The 2018 pain, agitation, delirium, immobility, and sleep guideline recommended ABCDEF care bundle for delirium prevention and management. However, limited information is available regarding the adoption of the care bundle in ICUs in Mainland China.

AIMS AND OBJECTIVES

To assess the current implementation of the ABCDEF care bundle for delirium prevention as reported by ICU nurses in Mainland China.

DESIGN

A cross-sectional study was conducted.

METHODS

A cross-sectional online survey using a validated questionnaire about the practices of the ABCDEF care bundle was conducted among 334 registered nurses in 167 ICUs of 65 cities in Mainland China.

RESULTS

Almost 50% of the sampled ICU nurses were unaware of the ABCDEF care bundle, though 86.83% of the surveyed ICUs implemented pain assessments and 95.51% implemented sedation assessments. Nearly half (46.41%) of the surveyed ICUs performed routine spontaneous awaking trials, with 21.26% performing them daily. Spontaneous breathing trials were performed in 38.32% of the surveyed ICUs. Only 47% of the surveyed ICUs routinely monitored patients for delirium. About one-third (38.35%) of the surveyed ICUs were supported by specialist teams that implemented the mobilization programmes. Most ICUs restricted the duration of family visits per day (<0.5 hour: 61.67%; 0.5-2 hours: 23.65%; >2 hours: 3.29%) and only 28.14% of the surveyed ICUs employed dedicated staff to support the families.

CONCLUSIONS

Although most of the surveyed ICUs implemented pain and sedation assessments, many of them did not implement structured delirium assessments. Early mobilization programmes and family participation should be encouraged.

RELEVANCE TO CLINICAL PRACTICE

Promoting the uses of a reliable delirium assessment tool such as Confusion Assessment Method for Intensive Care Unit patients, building an early mobilization team, and engaging family caregivers in the care plan may contribute to improved patients' clinical outcomes.

Risk of dementia and death in very-late-onset schizophrenia-like psychosis: A national cohort study

Knowledge is limited regarding the risks of death and dementia in very-late onset schizophrenia-like psychosis (VLOS). This study aims to scrutinize the associations between VLOS with the risks of death and dementia. Based on a prospective Israeli cohort study with national coverage, 94,120 persons without dementia or schizophrenia diagnoses aged 60 to 90 in 2012 were followed-up for the risks of dementia or death from 2013 to 2017. VLOS was classified as present from the age of the first ICD-9 diagnosis during follow-up, otherwise as absent. Hazard ratios (HR) with confidence intervals (95% CI) were computed with survival models to quantify the associations between VLOS and the risks of death and dementia, without and with adjustment for confounding. Nine sensitivity analyses were computed to examine the robustness of the results. The group with VLOS, compared to the group without, had higher death (n = 61, 18.5% vs. n = 7028, 7.5%, respectively) and dementia (n = 64, 19.5% vs. n = 5962, 6.4%, respectively) rates. In the primary analysis, the group with VLOS compared to the group without had increased risks of death (unadjusted HR = 3.10, 95% CI = 2.36, 4.06, P < .001; adjusted HR = 2.89, 95% CI = 2.15, 3.89; P < .001) and dementia (unadjusted HR = 3.81, 95% CI = 2.90, 4.99, P < .001; adjusted HR = 2.67, 95% CI = 1.82, 3.91; P < .001). The results remained statistically significant (P < .05) in all sensitivity analyses, including among persons without antipsychotic medication. The results may support notions of increased dementia risk and accelerated aging in VLOS, or that VLOS is a prodromal state of dementia.