This series of talks feature Health Delivery research and innovation, collaboration networks, and the next generation of research talent.<\/p>\n
00:00 Introduction
\nSpeakers:
\nBrianna Goodwin, Data Science Researcher, Microsoft Research & Incubations
\nBrooke Loughrin, Global Health Access Incubation Lead, Microsoft Research & Incubations<\/p>\n
05:37 Research Talk: Increasing access to health measurement through mobile-based, noncontact sensing
\nSpeakers:
\nCJ Park, PhD Student & Research Intern, University of Washington
\nDaniel McDuff, Microsoft Research Redmond<\/p>\n
The SARS-CoV-2 (COVID-19) pandemic has acutely highlighted the growing need for technology that supports accessible healthcare. One example is in the number appointments held via telehealth and mobile health platforms, which has greatly increased due to stay-at-home orders and greater burdens on healthcare systems. Experts suggest that particular attention be given to cardiovascular and pulmonary protection during the treatment of many conditions, but physicians lack access to objective measurements of a patient\u2019s condition due to the inability to capture vital signs. In this talk, we will present work on methods that leverage ordinary sensors, such as webcams, to measure physiological signals, such as peripheral blood flow, heart rate, respiration, blood oxygenation, and blood pressure. We will discuss our work in developing state-of-the-art neural models to help us learn more robust representations and achieve performance close to that of medical-grade devices.<\/p>\n
19:04 Research talk: How to build a better digital health system: Lessons learned from 13 countries
\nSpeakers:
\nBill Thies, Sr. Principal Researcher, Microsoft Research India
\nPatricia Moscibrodzki, Research Consultant, London School of Hygiene and Tropical Medicine<\/p>\n
As countries continue to adapt digital tools to their national health systems, it will be important to learn from the experiences gained in other settings. To this end, we examined digital health systems across 13 low- and middle-income countries using tuberculosis (TB) as a case study. In collaboration with the Global Fund and the Stop TB Partnership, we conducted an interview study with 44 stakeholders from government, funding agencies, and implementation partners. We surveyed the technologies used in the TB cascade of care and the associated challenges and best practices. Our findings indicate that while there is a diversity of experiences across environments, there are also unrealized opportunities to share knowledge and tools in support of better health systems. This talk, we will discuss such opportunities along four emergent dimensions: spanning technical building blocks, digital architecture, interaction with people, and interaction with the ecosystem.<\/p>\n
13:44 Research talk: Supporting caregivers in chronic disease care in India
\nSpeakers:
\nKarthik Bhat, PhD Student, Georgia Tech
\nAmanda (Mandi) Hall, Senior User Researcher, Microsoft Health Futures<\/p>\n
Treatment for chronic diseases involves not only clinical care but also several behavior and lifestyle changes to effectively control and live with the conditions. Though such changes are prescribed as at-home self-care for the patient, family caregivers are directly affected and are heavily involved stakeholders in the patient’s at-home care, especially in strongly family-oriented cultural contexts like India. In this talk, we present an internship project carried out in the summer of 2021, where we conducted 20 semi-structured interviews with caregivers of patients with a variety of chronic diseases in India. We sought to understand the roles they play in their loved one’s care and uncover opportunities for interventions that could address their unmet needs in performing their caregiving responsibilities.<\/p>\n
45:29 Research talk: Using interpretable machine learning to understand health and healthcare inequities
\nSpeakers:
\nRich Caruana, Sr. Principal Researcher, Microsoft Research Redmond
\nVanessa Polk, Senior Partner, Microsoft Research & Incubations<\/p>\n
Substantial disparities in healthcare outcomes are linked to race, gender, and socioeconomic status. To correct these disparities, we first need to understand what causes them. As machine learning (ML) and artificial intelligence (AI) methods become more common in healthcare, it is crucial to understand why models become biased and how to mitigate this bias. Fortunately, intelligible ML methods now make it easier to understand what models have learned and provide tools to mitigate some of this bias. In this presentation, we show how interpretable ML methods can be used to better answer the questions of why we see certain inequities in healthcare data.<\/p>\n