Identifying and characterizing high-risk clusters in a heterogeneous ICU population with deep embedded clustering

Artificial intelligence and machine learning: Definition of terms and current concepts in critical care research

With increasing computing power, artificial intelligence (AI) and machine learning (ML) have prospered, which facilitate the analysis of large datasets, especially those found in critical care. It is important to define these terminologies, to inform a standardized approach to critical care research. This manuscript hopes to clarify these terms with examples from medical literature. Three major components that are required for a successful ML implementation: (i) reliable dataset, (ii) ML algorithm, and (iii) unbiased model evaluation, are discussed. A reliable dataset can be structured or unstructured with limited noise, outliers, and missing values. ML, a subset of AI, is typically focused on supervised or unsupervised learning tasks in which the output is based on inputs and derived from iterative pattern recognition algorithms, while AI is the overall ability of a machine to "think" or mimic human behavior; and to analyze data free from human influence. Even with successful implementation, advanced AI and ML algorithms have faced challenges in adoption into practice, mainly due to their lack of interpretability, which hinders trust, buy-in, and engagement from clinicians. Consequently, traditional algorithms, such as linear and logistic regression, that may have reduced predictive power but are highly interpretable, continue to be widely used.

Deep embedded clustering generalisability and adaptation for integrating mixed datatypes: two critical care cohorts

We validated a Deep Embedded Clustering (DEC) model and its adaptation for integrating mixed datatypes (in this study, numerical and categorical variables). Deep Embedded Clustering (DEC) is a promising technique capable of managing extensive sets of variables and non-linear relationships. Nevertheless, DEC cannot adequately handle mixed datatypes. Therefore, we adapted DEC by replacing the autoencoder with an X-shaped variational autoencoder (XVAE) and optimising hyperparameters for cluster stability. We call this model "X-DEC". We compared DEC and X-DEC by reproducing a previous study that used DEC to identify clusters in a population of intensive care patients. We assessed internal validity based on cluster stability on the development dataset. Since generalisability of clustering models has insufficiently been validated on external populations, we assessed external validity by investigating cluster generalisability onto an external validation dataset. We concluded that both DEC and X-DEC resulted in clinically recognisable and generalisable clusters, but X-DEC produced much more stable clusters.

Explainable hierarchical clustering for patient subtyping and risk prediction

We present a pipeline in which machine learning techniques are used to automatically identify and evaluate subtypes of hospital patients admitted between 2017 and 2021 in a large UK teaching hospital. Patient clusters are determined using routinely collected hospital data, such as those used in the UK's National Early Warning Score 2 (NEWS2). An iterative, hierarchical clustering process was used to identify the minimum set of relevant features for cluster separation. With the use of state-of-the-art explainability techniques, the identified subtypes are interpreted and assigned clinical meaning, illustrating their robustness. In parallel, clinicians assessed intracluster similarities and intercluster differences of the identified patient subtypes within the context of their clinical knowledge. For each cluster, outcome prediction models were trained and their forecasting ability was illustrated against the NEWS2 of the unclustered patient cohort. These preliminary results suggest that subtype models can outperform the established NEWS2 method, providing improved prediction of patient deterioration. By considering both the computational outputs and clinician-based explanations in patient subtyping, we aim to highlight the mutual benefit of combining machine learning techniques with clinical expertise.

Identifying individuals with intellectual disability who access mental health support and are at high risk for adverse clinical outcomes: cohort study

BACKGROUND

People with intellectual disability often experience aggressive challenging behaviour and mental health issues. It can be difficult to identify those who are at higher risk of adverse clinical outcomes when in clinical care.

AIMS

To characterise potential subgroups in adults with intellectual disability referred to mental health services in those presenting with aggressive behaviour or common mental disorders (CMDs).

METHOD

There were 836 adults (≥18 years) with intellectual disability and a record of aggressive challenging behaviour, and 205 patients with intellectual disability and CMDs, who were seen in specialist mental health services over a 5-year period. Cluster analysis was used to define patient characteristics associated with clinical outcome.

RESULTS

Distinct patient groups with differentiated profiles were observed in people with intellectual disability displaying aggressive challenging behaviour, and in those presenting with CMDs. Characteristics of the aggressive behaviour group who experienced adverse outcomes included being <30 years old, being male, more mentions of aggression and agitation in their clinical record, a diagnosis of pervasive developmental disorder and prescription of psychotropic medication. Characteristics of the CMD cluster that experienced adverse clinical outcomes were being older, being a White male, having a mild intellectual disability and physical health concerns.

CONCLUSIONS

People with intellectual disability who experience adverse clinical outcomes can be identified with a cluster analysis approach of common features, but differ by clinical presentation. This could be used not only to stratify this clinically heterogeneous population in terms of response to interventions, but also improve precision in the development of tailored interventions.

Predictive Modeling Using Artificial Intelligence and Machine Learning Algorithms on Electronic Health Record Data: Advantages and Challenges

The rapid adoption of electronic health record (EHR) systems in US hospitals from 2008 to 2014 produced novel data elements for analysis. Concurrent innovations in computing architecture and machine learning (ML) algorithms have made rapid consumption of health data feasible and a powerful engine for clinical innovation. In critical care research, the net convergence of these trends has resulted in an exponential increase in outcome prediction research. In the following article, we explore the history of outcome prediction in the intensive care unit (ICU), the growing use of EHR data, and the rise of artificial intelligence and ML (AI) in critical care.

A new approach to predicting mortality in dialysis patients using sociodemographic features based on artificial intelligence

One of the main problems that affect patients in dialysis therapy who are on the waiting list to receive a kidney transplant is predicting their survival time if they do not receive a transplant. This paper proposes a new approach to survival prediction based on artificial intelligence techniques combined with statistical methods to study the association between sociodemographic factors and patient survival on the waiting list if they do not receive a kidney transplant. This new approach consists of a first stage that uses the clustering techniques that are best suited to the data structure (K-Means, Mini Batch K-Means, Agglomerative Clustering and K-Modes) used to identify the risk profile of dialysis patients. Later, a new method called False Clustering Discovery Reduction is performed to determine the minimum number of populations to be studied, and whose mortality risk is statistically differentiable. This approach was applied to the OPTN medical dataset (n = 44,663). The procedure started from 11 initial clusters obtained with the Agglomerative technique, and was reduced to eight final risk populations, for which their Kaplan-Meier survival curves were provided. With this result, it is possible to make predictions regarding the survival time of a new patient who enters the waiting list if the sociodemographic profile of the patient is known. To do so, the predictive algorithm XGBoost is used, which allows the cluster to which it belongs to be predicted and the corresponding Kaplan-Meier curve to be associated with it. This prediction process is achieved with an overall Multi-class AUC of 99.08 %.

Clustering analysis of geriatric and acute characteristics in a cohort of very old patients on admission to ICU

PURPOSE

The biological and functional heterogeneity in very old patients constitutes a major challenge to prognostication and patient management in intensive care units (ICUs). In addition to the characteristics of acute diseases, geriatric conditions such as frailty, multimorbidity, cognitive impairment and functional disabilities were shown to influence outcome in that population. The goal of this study was to identify new and robust phenotypes based on the combination of these features to facilitate early outcome prediction.

METHODS

Patients aged 80 years old or older with and without limitations of life-sustaining treatment and with complete data were recruited from the VIP2 study for phenotyping and from the COVIP study for external validation. The sequential organ failure assessment (SOFA) score and its sub-scores taken on admission to ICU as well as demographic and geriatric patient characteristics were subjected to clustering analysis. Phenotypes were identified after repeated bootstrapping and clustering runs.

RESULTS

In patients from the VIP2 study without limitations of life-sustaining treatment (n = 1977), ICU mortality was 12% and 30-day mortality 19%. Seven phenotypes with distinct profiles of acute and geriatric characteristics were identified in that cohort. Phenotype-specific mortality within 30 days ranged from 3 to 57%. Among the patients assigned to a phenotype with pronounced geriatric features and high SOFA scores, 50% died in ICU and 57% within 30 days. Mortality differences between phenotypes were confirmed in the COVIP study cohort (n = 280).

CONCLUSIONS

Phenotyping of very old patients on admission to ICU revealed new phenotypes with different mortality and potential need for anticipatory intervention.

Patient and hospital characteristics predict prolonged emergency department length of stay and in-hospital mortality: a nationwide analysis in Korea

BACKGROUND

Prolonged emergency department length of stay (EDLOS) in critically ill patients leads to increased mortality. This nationwide study investigated patient and hospital characteristics associated with prolonged EDLOS and in-hospital mortality in adult patients admitted from the emergency department (ED) to the intensive care unit (ICU).

METHODS

We conducted a retrospective cohort study using data from the National Emergency Department Information System. Prolonged EDLOS was defined as an EDLOS of ≥ 6 h. We constructed multivariate logistic regression models of patient and hospital variables as predictors of prolonged EDLOS and in-hospital mortality.

RESULTS

Between 2016 and 2019, 657,622 adult patients were admitted to the ICU from the ED, representing 2.4% of all ED presentations. The median EDLOS of the overall study population was 3.3 h (interquartile range, 1.9-6.1 h) and 25.3% of patients had a prolonged EDLOS. Patient characteristics associated with prolonged EDLOS included night-time ED presentation and Charlson comorbidity index (CCI) score of 1 or higher. Hospital characteristics associated with prolonged EDLOS included a greater number of staffed beds and a higher ED level. Prolonged EDLOS was associated with in-hospital mortality after adjustment for selected confounders (adjusted odds ratio: 1.18, 95% confidence interval: 1.16-1.20). Patient characteristics associated with in-hospital mortality included age ≥ 65 years, transferred-in, artificially ventilated in the ED, assignment of initial triage to more urgency, and CCI score of 1 or higher. Hospital characteristics associated with in-hospital mortality included a lesser number of staffed beds and a lower ED level.

CONCLUSIONS

In this nationwide study, 25.3% of adult patients admitted to the ICU from the ED had a prolonged EDLOS, which in turn was significantly associated with an increased in-hospital mortality risk. Hospital characteristics, including the number of staffed beds and the ED level, were associated with prolonged EDLOS and in-hospital mortality.

Interpretable Machine Learning for Inpatient COVID-19 Mortality Risk Assessments: Diabetes Mellitus Exclusive Interplay

People with diabetes mellitus (DM) are at elevated risk of in-hospital mortality from coronavirus disease-2019 (COVID-19). This vulnerability has spurred efforts to pinpoint distinctive characteristics of COVID-19 patients with DM. In this context, the present article develops ML models equipped with interpretation modules for inpatient mortality risk assessments of COVID-19 patients with DM. To this end, a cohort of 156 hospitalised COVID-19 patients with pre-existing DM is studied. For creating risk assessment platforms, this work explores a pool of historical, on-admission, and during-admission data that are DM-related or, according to preliminary investigations, are exclusively attributed to the COVID-19 susceptibility of DM patients. First, a set of careful pre-modelling steps are executed on the clinical data, including cleaning, pre-processing, subdivision, and feature elimination. Subsequently, standard machine learning (ML) modelling analysis is performed on the cured data. Initially, a classifier is tasked with forecasting COVID-19 fatality from selected features. The model undergoes thorough evaluation analysis. The results achieved substantiate the efficacy of the undertaken data curation and modelling steps. Afterwards, SHapley Additive exPlanations (SHAP) technique is assigned to interpret the generated mortality risk prediction model by rating the predictors' global and local influence on the model's outputs. These interpretations advance the comprehensibility of the analysis by explaining the formation of outcomes and, in this way, foster the adoption of the proposed methodologies. Next, a clustering algorithm demarcates patients into four separate groups based on their SHAP values, providing a practical risk stratification method. Finally, a re-evaluation analysis is performed to verify the robustness of the proposed framework.

Machine learning approach identified clusters for patients with low cardiac output syndrome and outcomes after cardiac surgery

Background

Low cardiac output syndrome (LCOS) is the most serious physiological abnormality with high mortality for patients after cardiac surgery. This study aimed to explore the multidimensional data of clinical features and outcomes to provide individualized care for patients with LCOS.

Methods

The electronic medical information of the intensive care units (ICUs) was extracted from a tertiary hospital in South China. We included patients who were diagnosed with LCOS in the ICU database. We used the consensus clustering approach based on patient characteristics, laboratory data, and vital signs to identify LCOS subgroups. The consensus clustering method involves subsampling from a set of items, such as microarrays, and determines to cluster of specified cluster counts (k). The primary clinical outcome was in-hospital mortality and was compared between the clusters.

Results

A total of 1,205 patients were included and divided into three clusters. Cluster 1 (n = 443) was defined as the low-risk group [in-hospital mortality =10.1%, odds ratio (OR) = 1]. Cluster 2 (n = 396) was defined as the medium-risk group [in-hospital mortality =25.0%, OR = 2.96 (95% CI = 1.97–4.46)]. Cluster 3 (n = 366) was defined as the high-risk group [in-hospital mortality =39.2%, OR = 5.75 (95% CI = 3.9–8.5)].

Conclusion

Patients with LCOS after cardiac surgery could be divided into three clusters and had different outcomes.

Cluster analysis integrating age and body temperature for mortality in patients with sepsis: a multicenter retrospective study

It is not clear whether mortality is associated with body temperature (BT) in older sepsis patients. This study aimed to evaluate the mortality rates in sepsis patients according to age and BT and identify the risk factors for mortality. We investigated the clusters using a machine learning method based on a combination of age and BT, and identified the mortality rates according to these clusters. This retrospective multicenter study was conducted at five hospitals in Korea. Data of sepsis patients aged ≥ 18 years who were admitted to the intensive care unit between January 1, 2011 and April 30, 2021 were collected. BT was divided into three groups (hypothermia < 36 °C, normothermia 36‒38 °C, and hyperthermia > 38 °C), and age groups were divided using a 75-year age threshold. Kaplan‒Meier analysis was performed to assess the cumulative mortality over 90 days. A K-means clustering algorithm using age and BT was used to characterize phenotypes. During the study period, 15,574 sepsis patients were enrolled. Overall, 90-day mortality was 20.5%. Kaplan‒Meier survival analyses demonstrated that 90-day mortality rates were 27.4%, 19.6%, and 11.9% in the hypothermia, normothermia, and hyperthermia groups, respectively, in those ≥ 75 years old (Log-rank p < 0.001). Cluster analysis demonstrated three groups: Cluster A (relatively older age and lower BT), Cluster B (relatively younger age and wide range of BT), and Cluster C (relatively higher BT than Cluster A). Kaplan‒Meier curve analysis showed that the 90-day mortality rates of Cluster A was significantly higher than those of Clusters B and C (24.2%, 17.1%, and 17.0%, respectively; Log-rank p < 0.001). The 90-day mortality rate correlated inversely with BT groups among sepsis patients in either age group (< 75 and ≥ 75 years). Clustering analysis revealed that the mortality rate was higher in the cluster of patients with relatively older age and lower BT.

Subtyping Hyperchloremia among Hospitalized Patients by Machine Learning Consensus Clustering

Background and Objectives: Despite the association between hyperchloremia and adverse outcomes, mortality risks among patients with hyperchloremia have not consistently been observed among all studies with different patient populations with hyperchloremia. The objective of this study was to characterize hyperchloremic patients at hospital admission into clusters using an unsupervised machine learning approach and to evaluate the mortality risk among these distinct clusters. Materials and Methods: We performed consensus cluster analysis based on demographic information, principal diagnoses, comorbidities, and laboratory data among 11,394 hospitalized adult patients with admission serum chloride of >108 mEq/L. We calculated the standardized mean difference of each variable to identify each cluster's key features. We assessed the association of each hyperchloremia cluster with hospital and one-year mortality. Results: There were three distinct clusters of patients with admission hyperchloremia: 3237 (28%), 4059 (36%), and 4098 (36%) patients in clusters 1 through 3, respectively. Cluster 1 was characterized by higher serum chloride but lower serum sodium, bicarbonate, hemoglobin, and albumin. Cluster 2 was characterized by younger age, lower comorbidity score, lower serum chloride, and higher estimated glomerular filtration (eGFR), hemoglobin, and albumin. Cluster 3 was characterized by older age, higher comorbidity score, higher serum sodium, potassium, and lower eGFR. Compared with cluster 2, odds ratios for hospital mortality were 3.60 (95% CI 2.33-5.56) for cluster 1, and 4.83 (95% CI 3.21-7.28) for cluster 3, whereas hazard ratios for one-year mortality were 4.49 (95% CI 3.53-5.70) for cluster 1 and 6.96 (95% CI 5.56-8.72) for cluster 3. Conclusions: Our cluster analysis identified three clinically distinct phenotypes with differing mortality risks in hospitalized patients with admission hyperchloremia.