Triage in major incidents: development and external validation of novel machine learning-derived primary and secondary triage tools.

BACKGROUND: Major incidents (MIs) are an important cause of death and disability. Triage tools are crucial to identifying priority 1 (P1) patients-those needing time-critical, life-saving interventions. Existing expert opinion-derived tools have limited evidence supporting their use. This study employs machine learning (ML) to develop and validate models for novel primary and secondary triage tools. METHODS: Adults (16+ years) from the UK Trauma Audit and Research Network (TARN) registry (January 2008-December 2017) served as surrogates for MI victims, with P1 patients identified using predefined criteria. The TARN database was split chronologically into model training and testing (70:30) datasets. Input variables included physiological parameters, age, mechanism and anatomical location of injury. Random forest, extreme gradient boosted tree, logistic regression and decision tree models were trained to predict P1 status, and compared with existing tools (Battlefield Casualty Drills (BCD) Triage Sieve, CareFlight, Modified Physiological Triage Tool, MPTT-24, MSTART, National Ambulance Resilience Unit Triage Sieve and RAMP). Primary and secondary candidate models were selected; the latter was externally validated on patients from the UK military's Joint Theatre Trauma Registry (JTTR). RESULTS: Models were internally tested in 57 979 TARN patients. The best existing tool was the BCD Triage Sieve (sensitivity 68.2%, area under the receiver operating curve (AUC) 0.688). Inability to breathe spontaneously, presence of chest injury and mental status were most predictive of P1 status. A decision tree model including these three variables exhibited the best test characteristics (sensitivity 73.0%, AUC 0.782), forming the candidate primary tool. The proposed secondary tool (sensitivity 77.9%, AUC 0.817), applicable via a portable device, includes a fourth variable (injury mechanism). This performed favourably on external validation (sensitivity of 97.6%, AUC 0.778) in 5956 JTTR patients. CONCLUSION: Novel triage tools developed using ML outperform existing tools in a nationally representative trauma population. The proposed primary tool requires external validation prior to consideration for practical use. The secondary tool demonstrates good external validity and may be used to support decision-making by healthcare workers responding to MIs.

Comparative analysis of major incident triage tools in children: a UK population-based analysis.

INTRODUCTION: Triage is a key principle in the effective management of major incidents. There is currently a paucity of evidence to guide the triage of children. The aim of this study was to perform a comparative analysis of nine adult and paediatric triage tools, including the novel 'Sheffield Paediatric Triage Tool' (SPTT), assessing their ability in identifying patients needing life-saving interventions (LSIs). METHODS: A 10-year (2008-2017) retrospective database review of the Trauma Audit Research Network (TARN) Database for paediatric patients (<16 years) was performed. Primary outcome was identification of patients receiving one or more LSIs from a previously defined list. Secondary outcomes included mortality and prediction of Injury Severity Score (ISS) >15. Primary analysis was conducted on patients with complete prehospital physiological data with planned secondary analyses using first recorded data. Performance characteristics were evaluated using sensitivity, specificity, undertriage and overtriage. RESULTS: 15 133 patients met TARN inclusion criteria. 4962 (32.8%) had complete prehospital physiological data and 8255 (54.5%) had complete first recorded physiological data. The majority of patients were male (69.5%), with a median age of 11.9 years. The overwhelming majority of patients (95.4%) sustained blunt trauma, yielding a median ISS of 9 and overall, 875 patients (17.6%) received at least one LSI. The SPTT demonstrated the greatest sensitivity of all triage tools at identifying need for LSI (92.2%) but was associated with the highest rate of overtriage (75.0%). Both the Paediatric Triage Tape (sensitivity 34.1%) and JumpSTART (sensitivity 45.0%) performed less well at identifying LSI. By contrast, the adult Modified Physiological Triage Tool-24 (MPTT-24) triage tool had the second highest sensitivity (80.8%) with tolerable rates of overtriage (70.2%). CONCLUSION: The SPTT and MPTT-24 outperform existing paediatric triage tools at identifying those patients requiring LSIs. This may necessitate a change in recommended practice. Further work is needed to determine the optimum method of paediatric major incident triage, but consideration should be given to simplifying major incident triage by the use of one generic tool (the MPTT-24) for adults and children.

Epidemiology, morbidity and mortality in Behçet's disease: a cohort study using The Health Improvement Network (THIN).

OBJECTIVES: The epidemiology of Behçet's disease (BD) has not been well characterized in the UK. Evidence on the risk of cardiovascular disease, thromboembolic disease and mortality in patients with BD compared with the general population is scarce. METHODS: We used a large UK primary care database to investigate the epidemiology of BD. A retrospective matched cohort study was used to assess the following outcomes: risk of cardiovascular, thromboembolic disease and mortality. Controls were selected at a 1:4 ratio (age and gender matched). Cox proportional hazard models were used to derive adjusted hazard ratios (aHR). RESULTS: The prevalence of BD was 14.61 (95% CI 13.35-15.88) per 100 000 population in 2017. A total of 1281 patients with BD were compared with 5124 age- and gender-matched controls. There was significantly increased risk of ischaemic heart disease [aHR 3.09 (1.28-7.44)], venous thrombosis [aHR 4.80 (2.42-9.54)] and mortality [aHR 1.40 (1.07-1.84)] in patients with BD compared with corresponding controls. Patients with BD were at higher risk of pulmonary embolism compared with corresponding controls at baseline [adjusted odds ratio 4.64 (2.66-8.09), P < 0.0001]. The majority of patients with pulmonary embolism and a diagnosis of BD had pulmonary embolism preceding the diagnosis of BD, not after (87.5%; n = 28/32). CONCLUSION: BD has a higher prevalence than previously thought. Physicians should be aware of the increased risk of developing ischaemic heart disease, stroke/transient ischaemic attack and deep venous thrombosis in patients with BD at an earlier age compared with the general population. Risk of embolism in patients with BD might vary across the disease course.

Risk of incident circulatory disease in patients treated for differentiated thyroid carcinoma with no history of cardiovascular disease.

CONTEXT: The incidence of differentiated thyroid cancer (DTC) is increasing, yet the prognosis is favourable and long-term survival is expected. Exogenous TSH suppression has been used for many years to prevent DTC recurrence and may be associated with increased risks of circulatory diseases. DESIGN: Risks of circulatory disease in patients treated for DTC were compared to randomly matched patients without DTC (controls) up to a 1:5 ratio using age, sex, body mass index (BMI) and smoking as the matching parameters in a population-based, open cohort study using The Health Improvement Network. PATIENTS: A total of 3009 patients treated for DTC with no pre-existing cardiovascular disease were identified and matched to 11 303 controls, followed up to median of 5 years. RESULTS: A total of 1259 incident circulatory events were recorded during the observation period. No difference in the risk of ischaemic heart disease (IHD) (adjusted hazards ratio [aHR]: 1.04, 95% CI: 0.80-1.36) or heart failure (HF) (aHR: 1.27, 95% CI: 0.89-1.81) was detected. The risk of atrial fibrillation (AF) and stroke was significantly higher in patients with DTC (aHR: 1.71, 95% CI: 1.36-2.15 and aHR: 1.34, 95% CI: 1.05-1.72, respectively). In a sensitivity analysis limited to newly diagnosed patients with DTC, only the risk of AF was consistently elevated (aHR: 1.86, 95% CI: 1.33-2.60). CONCLUSIONS: The increased risk of AF in patients who have undergone treatment for DTC but without pre-existing CVD may warrant periodic screening for this arrhythmia. Whereas no evidence of increased risk of IHD or HF was observed, the increased risk of stroke/TIA warrants further investigation.

Finding our way through phenotypes.

Despite a large and multifaceted effort to understand the vast landscape of phenotypic data, their current form inhibits productive data analysis. The lack of a community-wide, consensus-based, human- and machine-interpretable language for describing phenotypes and their genomic and environmental contexts is perhaps the most pressing scientific bottleneck to integration across many key fields in biology, including genomics, systems biology, development, medicine, evolution, ecology, and systematics. Here we survey the current phenomics landscape, including data resources and handling, and the progress that has been made to accurately capture relevant data descriptions for phenotypes. We present an example of the kind of integration across domains that computable phenotypes would enable, and we call upon the broader biology community, publishers, and relevant funding agencies to support efforts to surmount today's data barriers and facilitate analytical reproducibility.

Type 1 diabetes mellitus and risk of incident epilepsy: a population-based, open-cohort study.

AIMS/HYPOTHESIS: The aim of this research was to explore the relationship between incident epilepsy and type 1 diabetes in British participants. METHODS: Using The Health Improvement Network database, we conducted a retrospective, open-cohort study. Patients who were newly diagnosed with type 1 diabetes mellitus at the age of ≤40 years were identified and followed-up from 1 January 1990 to 15 September 2015. These patients, identified as not suffering from epilepsy at the time of diagnosis, were randomly matched with up to four individuals without type 1 diabetes mellitus, based on age, sex and participating general practice. A Cox regression analysis was subsequently performed using Townsend deprivation index, cerebral palsy, head injury and learning disabilities as model covariates. RESULTS: The study population consisted of a total of 24,610 individuals (4922 with type 1 diabetes and 19,688 controls). These individuals were followed up for a mean of 5.4 years (approximately 132,000 person-years of follow up). Patients with type 1 diabetes were significantly more likely to be diagnosed with epilepsy during the observation period compared with controls (crude HR [95% CI]: 3.02 [1.95, 4.69]). The incidence rate was estimated to be 132 and 44 per 100,000 person-years in patients and controls, respectively. This finding persisted after adjusting for model covariates (adjusted HR [95% CI]: 3.01 [1.93, 4.68]) and was also robust to sensitivity analysis, excluding adult-onset type 1 diabetes mellitus. CONCLUSIONS/INTERPRETATION: Patients with type 1 diabetes are at approximately three-times greater risk of developing epilepsy compared with matched controls without type 1 diabetes. This should be considered when investigating seizure-related disorders in patients with type 1 diabetes mellitus.

Blood-based cardiometabolic phenotypes in atrial fibrillation and their associated risk: EAST-AFNET 4 biomolecule study.

AIMS: Atrial fibrillation (AF) and concomitant cardiometabolic disease processes interact and combine to lead to adverse events, such as stroke, heart failure, myocardial infarction, and cardiovascular death. Circulating biomolecules provide quantifiable proxies for cardiometabolic disease processes. The aim of this study was to test whether biomolecule combinations can define phenotypes in patients with AF. METHODS AND RESULTS: This pre-specified analysis of the EAST-AFNET 4 biomolecule study assigned patients to clusters using polytomous variable latent-class analysis based on baseline concentrations of 13 precisely quantified biomolecules potentially reflecting ageing, cardiac fibrosis, metabolic dysfunction, oxidative stress, cardiac load, endothelial dysfunction, and inflammation. In each cluster, rates of cardiovascular death, stroke, or hospitalization for heart failure or acute coronary syndrome, the primary outcome of EAST-AFNET 4, were calculated and compared between clusters over median 5.1 years follow-up. Findings were independently validated in a prospective cohort of 748 patients with AF (BBC-AF; median follow-up 2.9 years).Unsupervised biomolecule analysis assigned 1586 patients (71 years old, 46% women) into four clusters. The highest risk cluster was dominated by elevated bone morphogenetic protein 10, insulin-like growth factor-binding protein 7, N-terminal pro-B-type natriuretic peptide, angiopoietin 2, and growth differentiation factor 15. Patients in the lowest risk cluster showed low concentrations of these biomolecules. Two intermediate-risk clusters differed by high or low concentrations of C-reactive protein, interleukin-6, and D-dimer. Patients in the highest risk cluster had a five-fold higher cardiovascular event rate than patients in the low-risk cluster. Early rhythm control was effective across clusters (Pinteraction = 0.63). Sensitivity analyses and external validation in BBC-AF replicated clusters and risk gradients. CONCLUSION: Biomolecule concentrations identify cardiometabolic subphenotypes in patients with AF at high and low cardiovascular risk.

The ultra-acute steroid response to traumatic injury: a cohort study.

OBJECTIVE: Trauma-induced steroid changes have been studied post-hospital admission, resulting in a lack of understanding of the speed and extent of the immediate endocrine response to injury. The Golden Hour study was designed to capture the ultra-acute response to traumatic injury. DESIGN: We conducted an observational cohort study including adult male trauma patients <60 years, with blood samples drawn ≤1 h of major trauma by pre-hospital emergency responders. METHODS: We recruited 31 adult male trauma patients (mean age 28 [range 19-59] years) with a mean injury severity score (ISS) of 16 (IQR 10-21). The median time to first sample was 35 (range 14-56) min, with follow-up samples collected 4-12 and 48-72 h post-injury. Serum steroids in patients and age- and sex-matched healthy controls (HCs) (n = 34) were analysed by tandem mass spectrometry. RESULTS: Within 1 h of injury, we observed an increase in glucocorticoid and adrenal androgen biosynthesis. Cortisol and 11-hydroxyandrostendione increased rapidly, whilst cortisone and 11-ketoandrostenedione decreased, reflective of increased cortisol and 11-oxygenated androgen precursor biosynthesis by 11ß-hydroxylase and increased cortisol activation by 11ß-hydroxysteroid dehydrogenase type 1. Active classic gonadal androgens testosterone and 5α-dihydrotestosterone decreased, whilst the active 11-oxygenated androgen 11-ketotestosterone maintained pre-injury levels. CONCLUSIONS: Changes in steroid biosynthesis and metabolism occur within minutes of traumatic injury. Studies that address whether ultra-early changes in steroid metabolism are associated with patient outcomes are now required.

Evaluation of NEWS2 response thresholds in a retrospective observational study from a UK acute hospital.

OBJECTIVE: Use of National Early Warning Score 2 (NEWS2) has been mandated in adults admitted to acute hospitals in England. Urgent clinical review is recommended at NEWS2 ≥5. This policy is recognised as requiring ongoing evaluation. We assessed NEWS2 acquisition, alerting at key thresholds and patient outcomes, to understand how response recommendations would affect clinical resource allocation. SETTING: Adult acute hospital in England. DESIGN: Retrospective observational cohort study. PARTICIPANTS: 100 362 consecutive admissions between November 2018 and July 2019. OUTCOME: Death or admission to intensive care unit within 24 hours of a score. METHODS: NEWS2 were assembled as single scores from consecutive 24-hour time frames, (the first NEWS2 termed 'Index-NEWS2'), or as all scores from the admission (termed All-NEWS2). Scores were excluded when a patient was in intensive care, in the presence of a decision not to attempt cardiopulmonary resuscitation, or on day 1 of elective admission. RESULTS: A mean of 4.5 NEWS2 were acquired per patient per day. The outcome rate following an Index-NEWS2 was 0.22/100 patient-days. The sensitivity of outcome prediction at Index-NEWS2 ≥5=0.46, and number needed to evaluate (NNE)=52. At this threshold, a mean of 37.6 alerts/100 patient-days would be generated, occurring in 12.3% of patients on any single day. Threshold changes to increase sensitivity by 0.1, would result in a twofold increase in alert rate and 1.5-fold increase in NNE. Overall, NEWS2 classification performance was significantly worse on Index-scores than All-scores (c-statistic=0.78 vs 0.85; p<0.001). CONCLUSIONS: The combination of low event-rate, high alert-rate and low sensitivity, in patients for cardiopulmonary resuscitation, means that at current NEWS2 thresholds, resource demand would be sufficient to meaningfully compete with other pathways to clinical evaluation. In analyses that epitomise in-patient screening, NEWS2 performance suggests a need for re-evaluation of current response recommendations in this population.

Therapies for Long COVID in non-hospitalised individuals: from symptoms, patient-reported outcomes and immunology to targeted therapies (The TLC Study).

INTRODUCTION: Individuals with COVID-19 frequently experience symptoms and impaired quality of life beyond 4-12 weeks, commonly referred to as Long COVID. Whether Long COVID is one or several distinct syndromes is unknown. Establishing the evidence base for appropriate therapies is needed. We aim to evaluate the symptom burden and underlying pathophysiology of Long COVID syndromes in non-hospitalised individuals and evaluate potential therapies. METHODS AND ANALYSIS: A cohort of 4000 non-hospitalised individuals with a past COVID-19 diagnosis and 1000 matched controls will be selected from anonymised primary care records from the Clinical Practice Research Datalink, and invited by their general practitioners to participate on a digital platform (Atom5). Individuals will report symptoms, quality of life, work capability and patient-reported outcome measures. Data will be collected monthly for 1 year.Statistical clustering methods will be used to identify distinct Long COVID-19 symptom clusters. Individuals from the four most prevalent clusters and two control groups will be invited to participate in the BioWear substudy which will further phenotype Long COVID symptom clusters by measurement of immunological parameters and actigraphy.We will review existing evidence on interventions for postviral syndromes and Long COVID to map and prioritise interventions for each newly characterised Long COVID syndrome. Recommendations will be made using the cumulative evidence in an expert consensus workshop. A virtual supportive intervention will be coproduced with patients and health service providers for future evaluation.Individuals with lived experience of Long COVID will be involved throughout this programme through a patient and public involvement group. ETHICS AND DISSEMINATION: Ethical approval was obtained from the Solihull Research Ethics Committee, West Midlands (21/WM/0203). Research findings will be presented at international conferences, in peer-reviewed journals, to Long COVID patient support groups and to policymakers. TRIAL REGISTRATION NUMBER: 1567490.

Violence-related knife injuries in a UK city; epidemiology and impact on secondary care resources.

BACKGROUND: The incidence of knife-related injuries is rising across the UK. This study aimed to determine the spectrum of knife-related injuries in a major UK city, with regards to patient and injury characteristics. A secondary aim was to quantify their impact on secondary care resources. METHODS: Observational study of patients aged 16+ years admitted to a major trauma centre following knife-related injuries resulting from interpersonal violence (May 2015 to April 2018). Patients were identified using Emergency Department and discharge coding, blood bank and UK national Trauma Audit and Research prospective registries. Patient and injury characteristics, outcome and resource utilisation were collected from ambulance and hospital records. FINDINGS: 532 patients were identified; 93% male, median age 26 years (IQR 20-35). Median injury severity score was 9 (IQR 3-13). 346 (65%) underwent surgery; 133 (25%) required intensive care; 95 (17·9%) received blood transfusion. Median length of stay was 3·3 days (IQR 1·7-6·0). In-hospital mortality was 10/532 (1·9%). 98 patients (18·5%) had previous attendance with violence-related injuries. 24/37 females (64·9%) were injured in a domestic setting. Intoxication with alcohol (19·2%) and illicit drugs (17·6%) was common. Causative weapon was household knife in 9%, knife (other/unspecified) in 38·0%, machete in 13·9%, small folding blade (2·8%) and, unrecorded in 36·3%. INTERPRETATION: Knife injuries constitute 12·9% of trauma team workload. Violence recidivism and intoxication are common, and females are predominantly injured in a domestic setting, presenting opportunities for targeted violence reduction interventions. 13·9% of injuries involved machetes, with implications for law enforcement strategies.

Construction and accessibility of a cross-species phenotype ontology along with gene annotations for biomedical research.

Phenotype analyses, e.g. investigating metabolic processes, tissue formation, or organism behavior, are an important element of most biological and medical research activities. Biomedical researchers are making increased use of ontological standards and methods to capture the results of such analyses, with one focus being the comparison and analysis of phenotype information between species. We have generated a cross-species phenotype ontology for human, mouse and zebrafish that contains classes from the Human Phenotype Ontology, Mammalian Phenotype Ontology, and generated classes for zebrafish phenotypes. We also provide up-to-date annotation data connecting human genes to phenotype classes from the generated ontology. We have included the data generation pipeline into our continuous integration system ensuring stable and up-to-date releases. This article describes the data generation process and is intended to help interested researchers access both the phenotype annotation data and the associated cross-species phenotype ontology. The resource described here can be used in sophisticated semantic similarity and gene set enrichment analyses for phenotype data across species. The stable releases of this resource can be obtained from http://purl.obolibrary.org/obo/hp/uberpheno/.

The RICORDO approach to semantic interoperability for biomedical data and models: strategy, standards and solutions.

BACKGROUND: The practice and research of medicine generates considerable quantities of data and model resources (DMRs). Although in principle biomedical resources are re-usable, in practice few can currently be shared. In particular, the clinical communities in physiology and pharmacology research, as well as medical education, (i.e. PPME communities) are facing considerable operational and technical obstacles in sharing data and models. FINDINGS: We outline the efforts of the PPME communities to achieve automated semantic interoperability for clinical resource documentation in collaboration with the RICORDO project. Current community practices in resource documentation and knowledge management are overviewed. Furthermore, requirements and improvements sought by the PPME communities to current documentation practices are discussed. The RICORDO plan and effort in creating a representational framework and associated open software toolkit for the automated management of PPME metadata resources is also described. CONCLUSIONS: RICORDO is providing the PPME community with tools to effect, share and reason over clinical resource annotations. This work is contributing to the semantic interoperability of DMRs through ontology-based annotation by (i) supporting more effective navigation and re-use of clinical DMRs, as well as (ii) sustaining interoperability operations based on the criterion of biological similarity. Operations facilitated by RICORDO will range from automated dataset matching to model merging and managing complex simulation workflows. In effect, RICORDO is contributing to community standards for resource sharing and interoperability.