122 research outputs found
AirPredict: a wearable sensor-based app to track particulate matter exposure and respiratory health
No full textAir pollution poses a significant threat to public health, with Particulate Matter (PM) being one of the most harmful pollutants, especially for those suffering of chronic respiratory diseases. In this work, we propose AirPredict, a digital health mobile application designed to monitor personal PM exposure and respiratory outcomes for asthma patients. By integrating data from wearable sensors, the platform accurately assesses inhaled pollutant doses and estimates individual PM exposure, while users log essential clinical data daily offering a one-in-all solution. The evaluation in a 14-day beta session with an asthma patient demonstrated the platform's intuitive nature and positive user experience. The application's user-friendly interface empowers individuals to make informed decisions to minimize exposure and enhance their quality of life
A Bayesian Framework to Identify Type 1 Diabetes Physiological Models Using Easily Accessible Patient Data
No full textMathematical physiological models of type 1 diabetes (T1D) glucose-insulin dynamics have been of great help in designing and preliminary assessing new algorithm for glucose control. Derivation of models at the individual level is however difficult because of identifiability issues. Recently, fitting these models against data of real patients with T1D has been made possible by both the use of Bayesian estimation techniques and the availability of individual datasets including plasma glucose and insulin concentration samples gathered in clinical protocols. The aim of this work is to make a step further and develop a methodology able to estimate the parameters of T1D physiological models using easily accessible data only, i.e. continuous glucose monitoring (CGM) sensor, carbohydrate intakes (CHO), and exogenous insulin infusion (I) data. The methodology is tested on synthetic data of 100 patients generated by a composite model of glucose-insulin dynamics. To solve identifiability problems, a Bayesian approach numerically implemented by Markov Chain Monte Carlo (MCMC) has been used to obtain point estimates and confidence intervals of model unknown parameters exploiting a priori knowledge available from the literature. Results show goodness of model fit and acceptable precision of parameter estimates. The methodology is also successful in reconstructing of 'non-accessible' glucose-insulin fluxes, i.e. glucose rate of appearance and plasma insulin. These preliminary results encourage further development of this framework and its assessment in more challenging setups
New approaches to determine insulin dose in type 1 diabetes treatment using continuous glucose monitoring data
No full textType 1 Diabetes is a chronic metabolic disorder that is normally treated by subcutaneous administration of exogenous insulin several times per day, by manual injections or by a pump, with individual doses determined using empirical guidelines which exploit knowledge of current blood glucose level assessed by the patient using fingerprick devices. The recent advent of long-lasting (weeks) minimally-invasive continuous glucose monitoring (CGM) sensors, and the growing field of therapeutic applicability granted to them by regulatory agencies in both EU and US, has stimulated investigation on new guidelines to determine insulin dosing exploiting glucose trend information. In this work we first assess, in an in silico clinical trial, the relative performance of three popular methods to determine the size of the insulin bolus at meal using CGM-based glucose trend information. Then we devise, and assess in the same population of virtual subjects, a new method based on a neural network modelling approach, which permits a personalization of the therapy. Preliminary results show that the new method can potentially improve glucose control, encouraging further development of the research
Algorithm based on the "dynamic risk" concept allows ahead-of-time suggestion of corrective insulin boluses in type 1 diabetes therapies employing continuous glucose monitors
No full textAdvanced diabetes management using artificial intelligence and continuous glucose monitoring sensors
No full textWearable continuous glucose monitoring (CGM) sensors are revolutionizing the treatment of type 1 diabetes (T1D). These sensors provide in real-time, every 1–5 min, the current blood glucose concentration and its rate-of-change, two key pieces of information for improving the determination of exogenous insulin administration and the prediction of forthcoming adverse events, such as hypo-/hyper-glycemia. The current research in diabetes technology is putting considerable effort into developing decision support systems for patient use, which automatically analyze the patient’s data collected by CGM sensors and other portable devices, as well as providing personalized recommendations about therapy adjustments to patients. Due to the large amount of data collected by patients with T1D and their variety, artificial intelligence (AI) techniques are increasingly being adopted in these decision support systems. In this paper, we review the state-of-the-art methodologies using AI and CGM sensors for decision support in advanced T1D management, including techniques for personalized insulin bolus calculation, adaptive tuning of bolus calculator parameters and glucose prediction
ReplayBG: A novel in-silico framework to retrospectively assess new therapy guidelines for type 1 diabetes management
No full textDR-CIB: an Algorithm for the Preventive Administration of Corrective Insulin Boluses in Type 1 Diabetes based on Dynamic Risk Concept and Patient-Specific Timing
No full textObjective:
Several strategies to reduce the duration of post-prandial hyperglycemia in type 1 diabetes (T1D) open-loop therapy have been developed in the recent years. Although these heuristics proved to be valid options accounting for continuous glucose monitoring (CGM) trend, they present some limitations and lack of personalization, calling for a more efficient solution.
Method:
We developed DR-CIB, a novel algorithm for post-prandial corrective insulin bolus (CIB) suggestion based on a preventive trigger threshold exploiting the “risk of hyperglycemia” and a personalized CIB timing retrieved from patients’ specific glucose-insulin dynamics. DR-CIB has been assessed on a dataset consisting of 49 daily CGM traces recorded in real-life conditions using ReplayBG, a novel digital twinning tool that allows a retrospective assessment of alternative insulin therapies using real data. As comparators we evaluated state-of-the-art approaches proposed by Aleppo (AL), Bruttomesso (BR), and Ziegler (ZI). Efficacy of glucose control was quantified by temporal, risk, and hyperglycemic event metrics.
Result:
Compared to literature methods, DR-CIB significantly reduces time spent in hyperglycemia when compared to AL and BR (33.52% vs 39.76% and 36.32%, respectively); significantly reduces daily injected insulin (5.97U vs 7.5U), glycemia risk index (37.78 vs 40.78) and time spent in hypoglycemia (75th percentile from 10.23% to 1.74%) when compared to ZI, resulting overall in a safer solution.
Conclusion:
We proposed DR-CIB, a dynamic risk-based algorithm which allow preventive actions for ahead-in-time management of hyperglycemic events and overcome some literature limitations proposing a patient-specific timing for CIB. DR-CIB proved to be a valid alternative to the most recent heuristic literature guidelines reducing the time spent in hyperglycemia and the hyperglycemic events duration without increasing the time below hypoglycemic threshold
Assessing Personal Exposure to Airborne Particulate Matter with Wearable Sensors and Ventilation Rate Models
No full textAir pollution is a major contributor to global morbidity and mortality. Accurate assessment of individual's exposure to air pollution is important to quantify the impact of air pollution on human health. Historically, human exposure to air pollution has been quantified using pollutant concentrations from fixed air quality monitoring stations. This approach does not consider the subject's activities and the differences between indoor and outdoor air pollution; however, these limitations can be overcome using wearable sensors. In this work, we propose a new approach to measure personal exposure to airborne Particulate Matter (PM) that consists in using a wearable/portable air quality sensor to measure air quality at the subject's location, a wearable Heart Rate (HR) sensor to collect HR timeseries, and a ventilation rate (VE) model to estimate the volume of inhaled air per minute (L/min) based on HR and other subject's covariates. Finally, VE and PM timeseries are combined to estimate the inhaled pollutant doses over time, as a measure of personal exposure. To model VE as a function of HR, 4 literature models are considered. The estimates obtained with the 4 models are compared in 3 representative subjects. Initial data analysis showed that the 4 models may drive to statistically significant differences in exposure estimates, thus the choice of the model can be a critical aspect of this approach. Regardless of the model used, timeseries of inhaled PM revealed significant daily variations in pollutant exposure, highlighting the importance of methodologies for accurate personal exposure assessment
MULTI-MODAL MOBILE PLATFORM FOR THE INVESIGATION OF CLINICAL DISORDERS WITH GLYCAEMIC DISARRAYS
No full textDetection of Self-Reported Stress Level from Wearable Sensor Data Using Machine Learning and Deep Learning-Based Classifiers: Is It Feasible?
No full textSocial, emotional and psychological state of a person are strictly related to their physical and mental health. Wellbeing can be disrupted by numerous factors, such as an extremely dynamic life that lead individuals to be more prone to stress. When a person is stressed, the body reacts in different ways, e.g., with headaches, sweating, heart palpitations. Some of these alterations can be quantified and measured with portable devices, such as smartwatches and wristbands, which potentially could be exploited for developing automatic stress measuring systems and prediction. In this work, we describe four different approaches to the problem by developing machine learning and deep-learning based pipelines for the detection of stress using wearable sensor data. The work was conducted on the SMILE dataset, which included features extracted from 60-minute sequences of electrocardiogram, galvanic skin response and skin temperature collected in 45 subjects. Results are less encouraging than expected, with accuracy and F1 score that reach maximum 0.57 and 0.62 respectively. The obtained results evidence the difficulties in modeling data in the wild to build a reliable stress detection algorithm. Further research studies are needed to demonstrate the feasibility of this tool
- …
