A precise and detailed understanding of patients’ functional and mental well-being, particularly among those with serious or high-risk conditions, is imperative. A thorough assessment based on continuous data collected over long periods of time will allow for improved risk stratification, treatment selection, and monitoring for adverse events. The real-time ability to detect decline or improvement may facilitate tailoring of interventions, which in turn, may reduce resource utilization while enhancing quality of life. As healthcare systems move towards both personalized care and accountable care, providing high quality precision medicine will require efficient and accurate risk stratification and patient monitoring through technology that is innovative and cost-efficient. Importantly, the advent of various mobile applications, wearable devices and physiological sensors have made continuous monitoring of a myriad of physiological and psychological parameters, “digital breadcrumbs”, a reality.
This Research Topic aims to explore the use of digital breadcrumbs to improve clinical diagnosis and decision making. We also welcome contributions focusing on drawbacks and obstacles that may arise in analyzing data collated from remote patient monitoring technologies, such as quality of wearable sensor data, missing data, bias, and causal reasoning. We are hoping that your contributions will further facilitate the “translation” of clinical use of technology from “bench to bedside”. Relevant topics include, but are not limited to:
- Sensor fusion methods, feasibility, and practicality of commercially available and research-grade sensors harnessed in healthcare
- Novel frameworks and methods of digital phenotyping
- Causality and model reliability in digital phenotyping: learning causal structures from wearable sensor data and patients’ daily behavior
- Improving model reliability, fairness, and bias in digital phenotyping
- Bias in sensors technology, such as PPG or pulse oximetry reading in people with darker skin tones
- Exploring and mitigating limitations in patients’ compliance using wearable sensors
- Dealing with imbalanced datasets in remote patient monitoring
- Applications of machine learning and deep learning technologies on wearable sensors to improve monitoring, diagnosis, and prediction of disease.
- Enforcing security and data privacy in applications and IoT architectures used in healthcare industry
A precise and detailed understanding of patients’ functional and mental well-being, particularly among those with serious or high-risk conditions, is imperative. A thorough assessment based on continuous data collected over long periods of time will allow for improved risk stratification, treatment selection, and monitoring for adverse events. The real-time ability to detect decline or improvement may facilitate tailoring of interventions, which in turn, may reduce resource utilization while enhancing quality of life. As healthcare systems move towards both personalized care and accountable care, providing high quality precision medicine will require efficient and accurate risk stratification and patient monitoring through technology that is innovative and cost-efficient. Importantly, the advent of various mobile applications, wearable devices and physiological sensors have made continuous monitoring of a myriad of physiological and psychological parameters, “digital breadcrumbs”, a reality.
This Research Topic aims to explore the use of digital breadcrumbs to improve clinical diagnosis and decision making. We also welcome contributions focusing on drawbacks and obstacles that may arise in analyzing data collated from remote patient monitoring technologies, such as quality of wearable sensor data, missing data, bias, and causal reasoning. We are hoping that your contributions will further facilitate the “translation” of clinical use of technology from “bench to bedside”. Relevant topics include, but are not limited to:
- Sensor fusion methods, feasibility, and practicality of commercially available and research-grade sensors harnessed in healthcare
- Novel frameworks and methods of digital phenotyping
- Causality and model reliability in digital phenotyping: learning causal structures from wearable sensor data and patients’ daily behavior
- Improving model reliability, fairness, and bias in digital phenotyping
- Bias in sensors technology, such as PPG or pulse oximetry reading in people with darker skin tones
- Exploring and mitigating limitations in patients’ compliance using wearable sensors
- Dealing with imbalanced datasets in remote patient monitoring
- Applications of machine learning and deep learning technologies on wearable sensors to improve monitoring, diagnosis, and prediction of disease.
- Enforcing security and data privacy in applications and IoT architectures used in healthcare industry
