Population aging and the increase of chronic conditions incidence and prevalence produce a higher risk of hospitalization or death. This is particularly high for patients with multimorbidity leading to a great consumption of resources. Identifying as soon as possible high-risk patients becomes an important challenge to improve health care service provision and to reduce costs. Nowadays, population health management, based on intelligent models, can be used to assess the risk and identify these "complex" patients. The aim of this study is to validate machine learning algorithms (Naïve Bayes, Cart, C5.0, Conditional Inference Tree, Random Forest, Artificial Neural Network and LASSO) to predict the risk of hospitalization or death starting from administrative and socio-economics data. The study involved the residents in the Local Health Unit of Central Tuscany.

Covid-19 has brought many difficulties in the management of infected and high-risk patients. Telemedicine platforms can really help in this situation, since they allow remotely monitoring Covid-19 patients, reducing the risk for the doctors, without decreasing the efficiency of the therapies and while alleviating patients’ mental issues. In this paper, we present the entire architecture and the experience of using the Tel.Te.Covid19 telemedicine platform. Projected for the treatment of chronic diseases, it has been technologically updated for the management of Covid-19 patients with the support of a group of doctors in the territory when the pandemic arrived, introducing new sensors and functionalities (e.g., the familiar use and video calls). In Tuscany (Central Italy), during the first wave of outbreak, a model for enrolling patients was created and tested. Because of the positive results, the latter has been then adopted in the second current wave. The Tel.Te.Covid19 platform has been used by 40 among general practitioners and doctors of continuity care and about 180 symptomatic patients since March 2020. Both patients and doctors have good opinion of the platform, and no hospitalisations or deaths occurred for the monitored patients, reducing also the impact on the National Healthcare System.

Together with population ageing, the number of people suffering from multimorbidity is increasing, up to more than half of the population by 2035. This part of the population is composed by the highest-risk patients, who are, at the same time, the major users of the healthcare systems. The early identification of this sub-population can really help to improve people’s quality of life and reduce healthcare costs. In this paper, we describe a population health management tool based on state-of-the-art intelligent algorithms, starting from administrative and socio-economic data, for the early identification of high-risk patients. The study refers to the population of the Local Health Unit of Central Tuscany in 2015, which amounts to 1,670,129 residents. After a trade-off on machine learning models and on input data, Random Forest applied to 1-year of historical data achieves the best results, outperforming state-of-the-art models. The most important variables for this model, in terms of mean minimal depth, accuracy decrease and Gini decrease, result to be age and some group of drugs, such as high-ceiling diuretics. Thanks to the low inference time and reduced memory usage, the resulting model allows for real-time risk prediction updates whenever new data become available, giving General Practitioners the possibility to early adopt personalised medicine.

Congenital heart disease, the most frequent malformation at birth, is usually not fatal but leads to multiple hospitalisations and outpatient visits, with negative impact on the quality of life and psychological profile not only of children but also of their families. In this paper, we describe the entire architecture of a system for remotely monitoring paediatric/neonatal patients with congenital heart disease, with the final aim of improving quality of life of the whole family and reducing hospital admissions. The interesting vital parameters for the disease are ECG, heart rate, oxygen saturation, body temperature and body weight. They are collected at home using some biomedical sensors specifically selected and calibrated for the paediatric field. These data are then sent to the smart hub, which proceeds with the synchronisation to the remote e-Health care center. Here, the doctors can log and evaluate the patient’s parameters. Preliminary results underline the sensor suitability for children and infants and good usability and data management of the smart-hub technology (E@syCare). In the clinical trial, some patients from the U.O.C. Paediatric and Adult Congenital Cardiology- Monasterio Foundation are enrolled. They receive a home monitoring kit according to the group they belong to. The trial aims to evaluate the effects of the system on quality of life. Psychological data are collected through questionnaires filled in by parents/caregivers in self-administration via the gateway at the beginning and at the end of the study. Results highlight an overall improvement in well-being and sleep quality, with a consequent reduction in anxious and stressful situations during daily life thanks to telemonitoring. At the same time, users reported a good level of usability, ease of data transmission and management of the devices.

Heart failure patients have become an important challenge for the healthcare system, since they represent a medical, social and economic problem. Early heart failure diagnoses can be very useful to improve patients' quality of life and to reduce the resources consumption, but they can be complex for the general practitioners. Data mining and machine learning techniques can really help in this field. The aim of this study is to validate some machine learning models to identify heart failure patients, starting from administrative data, and to make them transparent and interpretable. Despite the lack of clinical data, not available in Italy, but the most employed for the identification of heart failure patients, the results are comparable with the state-of-the-art ones and the models outperform the performances already obtained in Tuscany.