Comparing machine learning screening approaches using clinical data and cytokine profiles for COVID-19 in resource-limited and resource-abundant settings.

Accurate screening of COVID-19 infection status for symptomatic patients is a critical public health task. Although molecular and antigen tests now exist for COVID-19, in resource-limited settings, screening tests are often not available. Furthermore, during the early stages of the pandemic tests were not available in any capacity. We utilized an automated machine learning (ML) approach to train and evaluate thousands of models on a clinical dataset consisting of commonly available clinical and laboratory data, along with cytokine profiles for patients (n = 150). These models were then further tested for generalizability on an out-of-sample secondary dataset (n = 120). We were able to develop a ML model for rapid and reliable screening of patients as COVID-19 positive or negative using three approaches: commonly available clinical and laboratory data, a cytokine profile, and a combination of the common data and cytokine profile. Of the tens of thousands of models automatically tested for the three approaches, all three approaches demonstrated > 92% sensitivity and > 88 specificity while our highest performing model achieved 95.6% sensitivity and 98.1% specificity. These models represent a potential effective deployable solution for COVID-19 status classification for symptomatic patients in resource-limited settings and provide proof-of-concept for rapid development of screening tools for novel emerging infectious diseases.

Gallbladder perforation: case series and systematic review.

BACKGROUND: Gallbladder perforation is a serious complication of acute cholecystitis. Its management has evolved considerably since its classification by Niemeier in 1934. This review summarises the evidence surrounding the natural progression of this condition and potential problems with Niemeier's classification, and proposes a management algorithm for the more complex type II perforation. METHODS: Data from a retrospective case series and a systematic review were combined. The case series included all patients with gallbladder perforations from 2004 to 2008 at a British teaching hospital. The systematic review searched for gallbladder perforation using the MEDLINE, Embase, Web of Science and Cochrane Library (2011 Issue 4) databases, as well as recent conference abstracts. The outcome data were analysed using SPSS version 15. No adjustments were made for multiple testing. RESULTS: 198 patients (including 19 patients from the present series) with a mean age of 62.1+/-9.7 years and male gender proportion of 55.4% (range 33.3-76.7%) were included. The most common gallbladder perforations were type II (median 46.2%, range 7.4-83.3%), followed by type I (median 40.6%, range 16.7-70.0%) and type III (median 10.1%, range 0-48.1%). Perforation was associated with cholelithiasis in 86.6% (range 78.9-90.6%) of patients, and the overall median mortality rate was 10.8% (range 0-12.5%). Male gender was weakly associated with mortality (p = 0.089) but age (p = 0.877) and cholelithiasis (p = 0.425) were not. Mortality did not vary significantly with perforation type. CONCLUSIONS: Gallbladder perforation should be reported according to the original Neimeier's classification to avoid heterogeneity in data (e.g. varying rates of perforation types). The algorithm proposed in this study aims to guide the management of complex type II gallbladder perforations to minimise subsequent morbidity and mortality.