Overview
The Smart and Connected Health program aims to accelerate the development and use of innovative approaches that partner technology-based solutions with biobehavioral health research. The end-goal is a health care system that is proactive, person-centered, and focused on well-being, rather than reactive, hospital-centered, and disease-focused. This interagency program aims to develop next-generation health care solutions by funding high-risk, high-reward efforts in a variety of areas, including information science, technology, behavior, cognition, sensors, robotics, bioimaging, and engineering.
Due to recent technological advances (e.g. high-throughput computing), medicine is on the threshold of a sector-wide transformation. These innovations have the potential to accelerate discovery, facilitate the delivery of high-quality healthcare, improve patient outcomes, decrease costs, and address the complexity of challenging health problems such as cancer, heart disease, and diabetes. Realizing the promise of disruptive transformation in health and health care will require well-coordinated, multidisciplinary approaches that draw from social, behavioral, economic, and computer science disciplines, as well as engineering, medicine, and biology. The Smart and Connected Health initiative is therefore designed to encourage multidisciplinary approaches and greater collaboration between academic, industry, and nonprofit sectors. It also aims to establish better linkages between fundamental science, clinical practice and technology development, and deployment and use.
Smart and Connected Health projects in the HCIRB portfolio:
“Active Learning for Medical Applications”
(Nikos Papanikolopoulos, University of Minnesota)
Cancer is a major health concern and one of the leading causes of death in the U.S. and around the world. Recent advances in big data analytics, particularly machine learning, could greatly impact the domain of computer-aided cancer diagnosis, enabling improved cancer staging and treatment. Convolutional Neural Networks offer a promising pathway to achieve some degree of automation in identifying cancerous cases in image data. This study aims to 1) explore the underlying discriminative features hidden in images that may be different than those used by human experts, in order to improve the accuracy of diagnosis; and 2) focus on algorithms to minimize the amount of data required to train the neural network without sacrificing performance and generalization.
“Collaborative Research: Multiscale Modeling and Intervention for Improving Long-Term Medication”
(Elizabeth Barnes, University of Virginia)
Adherence to long-term endocrine therapy is crucial for survivors of hormone receptor-positive breast cancer who are prescribed these daily medications to prevent cancer recurrence. However, rates of adherence to these medications are low. The aims of this study are to 1) develop a system consisting of sensor-rich smartphones, wireless medication event monitoring systems, wireless beacons, and wearable sensors to collect in situ data on medication adherence; 2) deploy a Multiscale Modeling and Intervention (MMI) system for breast cancer survivors to model relationships between adherence and multiscale factors, identify patterns associated with medication-taking behavior, and develop interventions; and 3) demonstrate a proof-of-concept for MMI through a human subjects study, with subjects receiving multiscale interventions.
“MYPHA: Automatically Generating Personalized Accounts of In-patient Hospitalization”
(Barbara DiEugenio, University of Illinois At Chicago)
Patients who are engaged in their self-care after hospitalization have better outcomes; however, discharge instructions are often unsuccessful in engaging patients. This study plans to integrate nurse and physician documentation from the Electronic Health Record with the patient's needs and preferences to provide patients with a concise summary of their hospitalization, tailored to their current activation level. Specifically, the study pursues four aims: 1) to investigate socio-cultural patterns of patient perceptions of illness via a socio-linguistic analysis of patient interviews; 2) to integrate multi-faceted information about a patient, as provided by physicians and nurses; 3) to develop MyPHA, a novel application program that automatically generates and delivers personalized patient summaries on a tablet. 4) to evaluate MyPHA through formative evaluation with all stakeholders and a summative evaluation with patients.
“Changegradients: Promoting Adolescent Health Behavior Change with Clinically Integrated Sample-efficient Policy Gradient Methods”
(Elizabeth Ozer, University of California, San Francisco)
This project has two specific aims: 1) design, develop, and iteratively refine a policy-based reinforcement learning behavior change system for preventive adolescent health, and 2) investigate the impact of a clinically integrated sample-efficient policy gradient-based behavior change system on adolescent risk behavior, specifically alcohol use. The project will culminate with an investigation of adolescents’ use of the CHANGEGRADIENTS system and its impact on their alcohol use and self-efficacy to engage in healthy behaviors and avoid risky substance use. It is anticipated that CHANGEGRADIENTS will provide a testbed for a broad range of health behavior change research, including cancer-related behaviors and risk factors.
“Intelligent Information Sharing: Advancing Teamwork in Complex Care”
(Lee Michael Sanders, Stanford University)
The goal of this research is to improve health outcomes for children with cancer and other complex chronic conditions by developing intelligent interactive systems to help care teams use integrated care plans that promote shared decision-making. Specifically, the project will 1) develop novel multi-agent representations and algorithms; 2) identify clear definitions and effective strategies for patients and providers on shared goal setting; and 3) develop GoalKeeper, a technological system to support care plan use and care coordination.
Contact
