Since its emergence in late 2019, the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has caused a pandemic with more than 55 million reported cases and 1.3 million estimated deaths worldwide. While epidemiological and clinical characteristics of COVID-19 have been reported, risk factors underlying the transition from mild to severe disease among patients remain poorly understood. In this retrospective study, we analysed data of 879 confirmed SARS-CoV-2 positive patients admitted to a two-site NHS Trust hospital in London, England, between January 1st and May 26th, 2020, with a majority of cases occurring in March and April. We extracted anonymised demographic data, physiological clinical variables and laboratory results from electronic healthcare records (EHR) and applied multivariate logistic regression, random forest and extreme gradient boosted trees. To evaluate the potential for early risk assessment, we used data available during patients’ initial presentation at the emergency department (ED) to predict deterioration to one of three clinical endpoints in the remainder of the hospital stay: admission to intensive care, need for invasive mechanical ventilation and in-hospital mortality. Based on the trained models, we extracted the most informative clinical features in determining these patient trajectories. Considering our inclusion criteria, we have identified 129 of 879 (15%) patients that required intensive care, 62 of 878 (7%) patients needing mechanical ventilation, and 193 of 619 (31%) cases of in-hospital mortality. Our models learned successfully from early clinical data and predicted clinical endpoints with high accuracy, the best model achieving area under the receiver operating characteristic (AUC-ROC) scores of 0.76 to 0.87 (F1 scores of 0.42–0.60). Younger patient age was associated with an increased risk of receiving intensive care and ventilation, but lower risk of mortality. Clinical indicators of a patient’s oxygen supply and selected laboratory results, such as blood lactate and creatinine levels, were most predictive of COVID-19 patient trajectories. Among COVID-19 patients machine learning can aid in the early identification of those with a poor prognosis, using EHR data collected during a patient’s first presentation at ED. Patient age and measures of oxygenation status during ED stay are primary indicators of poor patient outcomes.

Background: Hundreds of thousands of deaths have already been recorded for patients with the severe acute respiratory syndrome coronavirus (SARS-CoV-2; aka COVID-19). Understanding whether there is a relationship between comorbidities and COVID-19 positivity will not only impact clinical decisions, it will also allow an understanding of how better to define the long-term complications in the groups at risk. In turn informing national policy on who may benefit from more stringent social distancing and shielding strategies. Furthermore, understanding the associations between medications and certain outcomes may also further our understanding of indicators of vulnerability in people with COVID-19 and co-morbidities. Methods: Electronic healthcare records (EHR) from two London hospitals were analysed between 1st January and 27th May 2020. 5294 patients presented to the hospitals in whom COVID status was formally assessed; 1253 were positive for COVID-19 and 4041 were negative. This dataset was analysed to identify associations between comorbidities and medications, separately and two outcomes: (1) presentation with a COVID-19 positive diagnosis, and (2) inpatient death following COVID-19 positive diagnosis. Medications were analysed in different time windows of prescription to differentiate between short-term and long-term medications. All analyses were done with controls (without co-morbidity) matched for age, sex, and number of admissions, and a robustness approach was conducted to only accept results that consistently appear when the analysis is repeated with different proportions of the data. Results: We observed higher COVID-19 positive presentation for patients with hypertension (1.7 [1.3-2.1]) and diabetes (1.6 [1.2-2.1]). We observed higher inpatient COVID-19 mortality for patients with hypertension (odds ratio 2.7 [95% CI 1.9-3.9]), diabetes (2.2 [1.4-3.5]), congestive heart failure (3.1 [1.5-6.4]), and renal disease (2.6 [1.4-5.1]). We also observed an association with reduced COVID-19 mortality for diabetic patients for whom anticoagulants (0.11 [0.03-0.50]), lipid-regulating drugs (0.15 [0.04-0.58]), penicillins (0.20 [0.06-0.63]), or biguanides (0.19 [0.05-0.70]) were administered within 21 days after their positive COVID-19 test with no evidence that they were on them before, and for hypertensive patients for whom anticoagulants (0.08 [0.02-0.35]), antiplatelet drugs (0.10 [0.02-0.59]), lipid-regulating drugs (0.15 [0.05-0.46]), penicillins (0.14 [0.05-0.45]), or angiotensin-converting enzyme inhibitors (ARBs) (0.06 [0.01-0.53]) were administered within 21 days post-COVID-19-positive testing with no evidence that they were on them before. Moreover, long-term antidiabetic drugs were associated with reduced COVID-19 mortality in diabetic patients (0.26 [0.10-0.67]). Conclusions: We provided real-world evidence for observed associations between COVID-19 outcomes and a number of comorbidities and medications. These results require further investigation and replication in other data sets.

Introduction Our goal was to evaluate if traffic-light driven personalized care for COVID-19 was associated with improved survival in acute hospital settings. Methods Discharge outcomes were evaluated before and after prospective implementation of a real-time dashboard with feedback to ward-based clinicians. Thromboembolism categories were “medium-risk” (D-dimer >1000 ng/mL or CRP >200 mg/L); “high-risk” (D-dimer >3000 ng/mL or CRP >250 mg/L) or “suspected” (D-dimer >5000 ng/mL). Cytokine storm risk was categorized by ferritin. Results 939/1039 COVID-19 positive patients (median age 67 years, 563/939 (60%) male) completed hospital encounters to death or discharge by 21st May 2020. Thromboembolism flag criteria were reached by 568/939 (60.5%), including 238/275 (86.6%) of the patients who died, and 330/664 (49.7%) of the patients who survived to discharge, p < 0.0001. Cytokine storm flag criteria were reached by 212 (22.6%) of admissions, including 80/275 (29.1%) of the patients who died, and 132/664 (19.9%) of the patients who survived, p < 0.0001. The maximum thromboembolism flag discriminated completed encounter mortality (no flag: 37/371 [9.97%] died; medium-risk: 68/239 [28.5%]; high-risk: 105/205 [51.2%]; and suspected thromboembolism: 65/124 [52.4%], p < 0.0001). Flag criteria were reached by 535 consecutive COVID-19 positive patients whose hospital encounter completed before traffic-light introduction: 173/535 (32.3% [95% confidence intervals 28.0, 36.0]) died. For the 200 consecutive admissions after implementation of real-time traffic light flags, 46/200 (23.0% [95% confidence intervals 17.1, 28.9]) died, p = 0.013. Adjusted for age and sex, the probability of death was 0.33 (95% confidence intervals 0.30, 0.37) before traffic light implementation, 0.22 (0.17, 0.27) after implementation, p < 0.001. In subgroup analyses, older patients, males, and patients with hypertension ( p ≤ 0.01), and/or diabetes ( p = 0.05) derived the greatest benefit from admission under the traffic light system. Conclusion Personalized early interventions were associated with a 33% reduction in early mortality. We suggest benefit predominantly resulted from early triggers to review/enhance anticoagulation management, without exposing lower-risk patients to potential risks of full anticoagulation therapy.

Background Since its emergence in late 2019, the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has caused a pandemic, with more than 4.8 million reported cases and 310 000 deaths worldwide. While epidemiological and clinical characteristics of COVID-19 have been reported, risk factors underlying the transition from mild to severe disease among patients remain poorly understood. Methods In this retrospective study, we analysed data of 820 confirmed COVID-19 positive patients admitted to a two-site NHS Trust hospital in London, England, between January 1st and April 23rd, 2020, with a majority of cases occurring in March and April. We extracted anonymised demographic data, physiological clinical variables and laboratory results from electronic healthcare records (EHR) and applied multivariate logistic regression, random forest and extreme gradient boosted trees. To evaluate the potential for early risk assessment, we used data available during patients' initial presentation at the emergency department (ED) to predict deterioration to one of three clinical endpoints in the remainder of the hospital stay: A) admission to intensive care, B) need for mechanical ventilation and C) mortality. Based on the trained models, we extracted the most informative clinical features in determining these patient trajectories. Results Considering our inclusion criteria, we have identified 126 of 820 (15%) patients that required intensive care, 62 of 808 (8%) patients needing mechanical ventilation, and 170 of 630 (27%) cases of in-hospital mortality. Our models learned successfully from early clinical data and predicted clinical endpoints with high accuracy, the best model achieving AUC-ROC scores of 0.75 to 0.83 (F1 scores of 0.41 to 0.56). Younger patient age was associated with an increased risk of receiving intensive care and ventilation, but lower risk of mortality. Clinical indicators of a patient's oxygen supply and selected laboratory results were most predictive of COVID-19 patient trajectories. Conclusion Among COVID-19 patients machine learning can aid in the early identification of those with a poor prognosis, using EHR data collected during a patient's first presentation at ED. Patient age and measures of oxygenation status during ED stay are primary indicators of poor patient outcomes.

Background: COVID-19 is a global health emergency. Recent data indicate a 50% mortality rate across UK intensive care units.Methods: A single institution, two-centre retrospective analysis following implementation of a Decision Support tool and real-time data dashboard for early detection of patients requiring personalised enhanced care, focussing on respiratory rate, diastolic blood pressure, oxygenation indices, C-reactive protein, D-dimer and ferritin. Protocols differing from conventional practice included high-dose prophylactic anticoagulation for all COVID-19 positive patients and prescription of antioxidants.Results: By 22/04/2020, 923 patients tested COVID-19 positive. 569 patients (61.7%) were male. The majority presented with advanced disease: interquartile ranges were C-reactive protein 44.9-179mg/L, D-dimer 1070-3802ng/mL, and ferritin 261-1208μg/L. Completed case fatality rates were 25.1% [95% CI 20.0, 30.0] in females, 40.5% [95% CI 35.9, 45.0] in males. 139 patients were admitted to intensive care where current death rates are 16.2% [95% CI 3.8, 28.7] in females, 38.2% [95% CI 28.6, 47.8] inmales with no trends for differences based on ethnicity. A real-time traffic lights dashboard enabled rapid assessment of patients using critical parameters to accelerate adjustments to management protocols. In total 513 (55.6%) of patients were flagged as high risk for thromboembolic disease, exceeding the numbers flagged for respiratory deteriorations (N=391, 42.4%), or cytokine storm (N=68, 7.4%). There was minimal evidence that age was associated with disease severity, but males had higher levels of all dashboard indices, particularly C-reactive protein and ferritin (p<0.0001) which displayed no relationship with age.Conclusions: Survival rates are encouraging. Protocols employed (traffic light-driven personalised care, protocolised early therapeutic anticoagulation based on D-dimer >1,000ng/mL and/or CRP>200 mg/L, personalised ventilatory strategies and antioxidants) are recommended to other units. Males are at greater risk of severe disease, most likely as the obligate SARS-CoV-2 receptor is encoded by the Xchromosome, and require especially close, and early attention.