Development, evaluation and validation of machine learning algorithms to detect atypical and asymptomatic presentations of Covid-19 in hospital practice.

BACKGROUND: Diagnostic methods for Covid-19 have improved, both in speed and availability. Because of atypical and asymptomatic carriage of the virus and nosocomial spread within institutions, timely diagnosis remains a challenge. Machine learning models trained on blood test results have shown promise in identifying cases of Covid-19. AIMS: To train and validate a machine learning model capable of differentiating Covid-19 positive from negative patients using routine blood tests and assess the model's accuracy against atypical and asymptomatic presentations. DESIGN AND METHODS: We conducted a retrospective analysis of medical admissions to our institution during March and April 2020. Participants were categorized into Covid-19 positive or negative groups based on clinical, radiological features or nasopharyngeal swab. A machine learning model was trained on laboratory parameters and validated for accuracy, sensitivity and specificity and externally validated at an unconnected establishment. RESULTS: An Ensemble Bagged Tree model was trained on data collected from 405 patients (212 Covid-19 positive) producing an accuracy of 81.79% (95% confidence interval (CI) 77.53-85.55%), the sensitivity of 85.85% (CI 80.42-90.24%) and specificity of 76.65% (CI 69.49-82.84%). Accuracy was preserved for atypical and asymptomatic subgroups. Using an external data set for 226 patients (141 Covid-19 positive) accuracy of 76.82% (CI 70.87-82.08%), sensitivity of 78.38% (CI 70.87-84.72%) and specificity of 74.12% (CI 63.48-83.01%) was achieved. CONCLUSION: A machine learning model using routine laboratory parameters can detect atypical and asymptomatic presentations of Covid-19 and might be an adjunct to existing screening measures.

Diuretic lounge and the impact on hospital admissions for treatment of decompensated heart failure.

BACKGROUND: Heart failure is a prevalent condition associated with frequent and costly hospital admissions. Hospitalizations are primarily related to worsening fluid retention and often require admission for decongestion with intravenous diuretics. OBJECTIVE: To assess the safety of an outpatient intravenous diuresis service for heart failure patients, and its impact on emergency admissions and the cost of treatment. METHODS: We conducted a prospective observational cohort registry study on patients referred to the diuretic lounge at our acute hospital between May 2017 and April 2018. RESULTS: We analysed 245 patients treated in the diuretic lounge, of which 190 (77.6%) avoided hospitalization or any adverse events during the 60 days of follow up (77.6% vs. 22.4%; P < 0.001). The diuretic lounge service resulted in a significant decrease in emergency heart failure admissions compared to the previous 12 months (823 vs. 715 per annum; 68.6 ± 10.1 vs. 59.6 ± 14 per month; P = 0.04), and a numerical reduction in readmission rates (17.3% vs. 16.2%). The 13.1% decrease in admissions lead to financial savings of £315 497 per annum and £2921 per admission avoided. During the same time period, at the other acute hospital site in our trust, where no diuretic lounge service is available, the number of admissions did not significantly change (457 vs. 450 per annum; 37.5 ± 7.0 vs. 38.1 ± 7.6 per month; P = 0.81). CONCLUSION: Ambulatory administration of intravenous diuretics reduces emergency admissions and is a safe and cost-effective alternative to treat acute decomposition in heart failure patients.

Intracerebral haematoma masquerading as ventricular standstill.

An 82 year old man was referred to the emergency room by his general practitioner for a right frontoparietal headache. The preceding day he had tripped and fallen, hitting the back of his head on the floor. Computed tomography showed a cortical contre coup haematoma. In view of ventricular standstill noted on ECG, a temporary pacing wire was inserted and a dual chamber permanent pacemaker was subsequently implanted. Intracerebral bleeding was treated conservatively and the patient made a good recovery. All patients admitted with head injury and sinus bradycardia or sinus arrest should be nursed at 15 degrees to 30 degrees with instructions to avoid the head up and supine positions. Furthermore, brain CT should be promptly recorded to assess for intracerebral haematoma and raised intracranial pressure and, if they are confirmed, these patients with cardiovascular compromise should benefit from close collaboration between neurosurgeon and cardiologist. Urgent pacing should be considered for all patients with head injury who experience symptomatic bradycardia or ventricular standstill.