Background Aim of the present study is to describe characteristics of COVID-19-related deaths and to compare the clinical phenotype and course of COVID-19-related deaths occurring in adults (<65 years) and older adults (≥65 years). Method Medical charts of 3,032 patients dying with COVID-19 in Italy (368 aged < 65 years and 2,664 aged ≥65 years) were revised to extract information on demographics, preexisting comorbidities, and in-hospital complications leading to death. Results Older adults (≥65 years) presented with a higher number of comorbidities compared to those aged <65 years (3.3 ± 1.9 vs 2.5 ± 1.8, p < .001). Prevalence of ischemic heart disease, atrial fibrillation, heart failure, stroke, hypertension, dementia, COPD, and chronic renal failure was higher in older patients (≥65 years), while obesity, chronic liver disease, and HIV infection were more common in younger adults (<65 years); 10.9% of younger patients (<65 years) had no comorbidities, compared to 3.2% of older patients (≥65 years). The younger adults had a higher rate of non-respiratory complications than older patients, including acute renal failure (30.0% vs 20.6%), acute cardiac injury (13.5% vs 10.3%), and superinfections (30.9% vs 9.8%). Conclusions Individuals dying with COVID-19 present with high levels of comorbidities, irrespective of age group, but a small proportion of deaths occur in healthy adults with no preexisting conditions. Non-respiratory complications are common, suggesting that the treatment of respiratory conditions needs to be combined with strategies to prevent and mitigate the effects of non-respiratory complications.

Abstract-Quantification is the name given to a novel machine learning task which deals with correctly estimating the number of elements of one class in a set of examples. The output of a quantifier is a real value; since training instances are the same as a classification problem, a natural approach is to train a classifier and to derive a quantifier from it. Some previous works have shown that just classifying the instances and counting the examples belonging to the class of interest (classify & count) typically yields bad quantifiers, especially when the class distribution may vary between training and test. Hence, adjusted versions of classify & count have been developed by using modified thresholds. However, previous works have explicitly discarded (without a deep analysis) any possible approach based on the probability estimations of the classifier. In this paper, we present a method based on averaging the probability estimations of a classifier with a very simple scaling that does perform reasonably well, showing that probability estimators for quantification capture a richer view of the problem than methods based on a threshold.

No abstract

Evaluation of machine learning methods is a crucial step before application, because it is essential to assess how good a model will behave for every single case. In many real applications, not only the "total" or the "average" of the error of the model is important but it is also important to know how this error is distributed or how well confidence or probability estimations are made. However, many machine learning techniques are good in overall results but have a bad distribution /assessment of the error.In these cases, calibration techniques have been developed as postprocessing techniques which aim at improving the probability estimation or the error distribution of an existing model.In this chapter, we present the most usual calibration techniques and calibration measures. We cover both classification and regression, and we establish a taxonomy of calibration techniques, while then paying special attention to probabilistic classifier calibration.

Following civil unrest in North Africa early in 2011, there was a large influx of migrants in Italy. A syndromic surveillance system was set up in April to monitor the health of this migrant population and respond rapidly to any health emergency. In the first six months, the system produced 67 alerts across all syndromes monitored and four alarms. There were no health emergencies, however, indicating that this migration flow was not associated with an increased risk of communicable disease transmission in Italy.