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.

The Value of Triage during Periods of Intense COVID-19 Demand: Simulation Modeling Study.

BACKGROUND: During the COVID-19 pandemic, many intensive care units have been overwhelmed by unprecedented levels of demand. Notwithstanding ethical considerations, the prioritization of patients with better prognoses may support a more effective use of available capacity in maximizing aggregate outcomes. This has prompted various proposed triage criteria, although in none of these has an objective assessment been made in terms of impact on number of lives and life-years saved. DESIGN: An open-source computer simulation model was constructed for approximating the intensive care admission and discharge dynamics under triage. The model was calibrated from observational data for 9505 patient admissions to UK intensive care units. To explore triage efficacy under various conditions, scenario analysis was performed using a range of demand trajectories corresponding to differing nonpharmaceutical interventions. RESULTS: Triaging patients at the point of expressed demand had negligible effect on deaths but reduces life-years lost by up to 8.4% (95% confidence interval: 2.6% to 18.7%). Greater value may be possible through "reverse triage", that is, promptly discharging any patient not meeting the criteria if admission cannot otherwise be guaranteed for one who does. Under such policy, life-years lost can be reduced by 11.7% (2.8% to 25.8%), which represents 23.0% (5.4% to 50.1%) of what is operationally feasible with no limit on capacity and in the absence of improved clinical treatments. CONCLUSIONS: The effect of simple triage is limited by a tradeoff between reduced deaths within intensive care (due to improved outcomes) and increased deaths resulting from declined admission (due to lower throughput given the longer lengths of stay of survivors). Improvements can be found through reverse triage, at the expense of potentially complex ethical considerations.

COVID-19 scenario modelling for the mitigation of capacity-dependent deaths in intensive care.

Managing healthcare demand and capacity is especially difficult in the context of the COVID-19 pandemic, where limited intensive care resources can be overwhelmed by a large number of cases requiring admission in a short space of time. If patients are unable to access this specialist resource, then death is a likely outcome. In appreciating these 'capacity-dependent' deaths, this paper reports on the clinically-led development of a stochastic discrete event simulation model designed to capture the key dynamics of the intensive care admissions process for COVID-19 patients. With application to a large public hospital in England during an early stage of the pandemic, the purpose of this study was to estimate the extent to which such capacity-dependent deaths can be mitigated through demand-side initiatives involving non-pharmaceutical interventions and supply-side measures to increase surge capacity. Based on information available at the time, results suggest that total capacity-dependent deaths can be reduced by 75% through a combination of increasing capacity from 45 to 100 beds, reducing length of stay by 25%, and flattening the peak demand to 26 admissions per day. Accounting for the additional 'capacity-independent' deaths, which occur even when appropriate care is available within the intensive care setting, yields an aggregate reduction in total deaths of 30%. The modelling tool, which is freely available and open source, has since been used to support COVID-19 response planning at a number of healthcare systems within the UK National Health Service.

Towards a decision support tool for intensive care discharge: machine learning algorithm development using electronic healthcare data from MIMIC-III and Bristol, UK.

OBJECTIVE: The primary objective is to develop an automated method for detecting patients that are ready for discharge from intensive care. DESIGN: We used two datasets of routinely collected patient data to test and improve on a set of previously proposed discharge criteria. SETTING: Bristol Royal Infirmary general intensive care unit (GICU). PATIENTS: Two cohorts derived from historical datasets: 1870 intensive care patients from GICU in Bristol, and 7592 from Medical Information Mart for Intensive Care (MIMIC)-III. RESULTS: In both cohorts few successfully discharged patients met all of the discharge criteria. Both a random forest and a logistic classifier, trained using multiple-source cross-validation, demonstrated improved performance over the original criteria and generalised well between the cohorts. The classifiers showed good agreement on which features were most predictive of readiness-for-discharge, and these were generally consistent with clinical experience. By weighting the discharge criteria according to feature importance from the logistic model we showed improved performance over the original criteria, while retaining good interpretability. CONCLUSIONS: Our findings indicate the feasibility of the proposed approach to ready-for-discharge classification, which could complement other risk models of specific adverse outcomes in a future decision support system. Avenues for improvement to produce a clinically useful tool are identified.

Protective mechanical ventilation in United Kingdom critical care units: A multicentre audit.

Lung protective ventilation is becoming increasingly used for all critically ill patients being mechanically ventilated on a mandatory ventilator mode. Compliance with the universal application of this ventilation strategy in intensive care units in the United Kingdom is unknown. This 24-h audit of ventilation practice took place in 16 intensive care units in two regions of the United Kingdom. The mean tidal volume for all patients being ventilated on a mandatory ventilator mode was 7.2(±1.4) ml kg-1 predicted body weight and overall compliance with low tidal volume ventilation (≤6.5 ml kg-1 predicted body weight) was 34%. The mean tidal volume for patients ventilated with volume-controlled ventilation was 7.0(±1.2) ml kg-1 predicted body weight and 7.9(±1.8) ml kg-1 predicted body weight for pressure-controlled ventilation (P < 0.0001). Overall compliance with recommended levels of positive end-expiratory pressure was 72%. Significant variation in practice existed both at a regional and individual unit level.

Increasing compliance with low tidal volume ventilation in the ICU with two nudge-based interventions: evaluation through intervention time-series analyses.

OBJECTIVES: Low tidal volume (TVe) ventilation improves outcomes for ventilated patients, and the majority of clinicians state they implement it. Unfortunately, most patients never receive low TVes. 'Nudges' influence decision-making with subtle cognitive mechanisms and are effective in many contexts. There have been few studies examining their impact on clinical decision-making. We investigated the impact of 2 interventions designed using principles from behavioural science on the deployment of low TVe ventilation in the intensive care unit (ICU). SETTING: University Hospitals Bristol, a tertiary, mixed medical and surgical ICU with 20 beds, admitting over 1300 patients per year. PARTICIPANTS: Data were collected from 2144 consecutive patients receiving controlled mechanical ventilation for more than 1 hour between October 2010 and September 2014. Patients on controlled mechanical ventilation for more than 20 hours were included in the final analysis. INTERVENTIONS: (1) Default ventilator settings were adjusted to comply with low TVe targets from the initiation of ventilation unless actively changed by a clinician. (2) A large dashboard was deployed displaying TVes in the format mL/kg ideal body weight (IBW) with alerts when TVes were excessive. PRIMARY OUTCOME MEASURE: TVe in mL/kg IBW. FINDINGS: TVe was significantly lower in the defaults group. In the dashboard intervention, TVe fell more quickly and by a greater amount after a TVe of 8 mL/kg IBW was breached when compared with controls. This effect improved in each subsequent year for 3 years. CONCLUSIONS: This study has demonstrated that adjustment of default ventilator settings and a dashboard with alerts for excessive TVe can significantly influence clinical decision-making. This offers a promising strategy to improve compliance with low TVe ventilation, and suggests that using insights from behavioural science has potential to improve the translation of evidence into practice.

Using 'nudge' principles for order set design: a before and after evaluation of an electronic prescribing template in critical care.

OBJECTIVE: Computerised order sets have the potential to reduce clinical variation and improve patient safety but the effect is variable. We sought to evaluate the impact of changes to the design of an order set on the delivery of chlorhexidine mouthwash and hydroxyethyl starch (HES) to patients in the intensive care unit. METHODS: The study was conducted at University Hospitals Bristol NHS Foundation Trust, UK. Our intensive care unit uses a clinical information system (CIS). All drugs and fluids are prescribed with the CIS and drug and fluid charts are stored within a database. Chlorhexidine mouthwash was added as a default prescription to the prescribing template in January 2010. HES was removed from the prescribing template in April 2009. Both interventions were available to prescribe manually throughout the study period. We conducted a database review of all patients eligible for each intervention before and after changes to the configuration of choices within the prescribing system. RESULTS: 2231 ventilated patients were identified as appropriate for treatment with chlorhexidine, 591 before the intervention and 1640 after. 55.3% were prescribed chlorhexidine before the change and 90.4% after (p<0.001). 6199 patients were considered in the HES intervention, 2177 before the intervention and 4022 after. The mean volume of HES infused per patient fell from 630 mL to 20 mL after the change (p<0.001) and the percentage of patients receiving HES fell from 54.1% to 3.1% (p<0.001). These results were well sustained with time. CONCLUSIONS: The presentation of choices within an electronic prescribing system influenced the delivery of evidence-based interventions in a predictable way and the effect was well sustained. This approach has the potential to enhance the effectiveness of computerised order sets.