Maryland analytics expert describes importance of the social determinants of healthcare at the Conference on Health IT and Analytics
Healthcare data analytics expert Warren D. D’Souza of the University of Maryland Medical System described the importance of social factors, such as access to transportation, on patient outcomes at the Conference on Health IT and Analytics (CHITA) in Washington, D.C. The Nov. 3-4, 2017. The event held in downtown D.C. was organized by the Center for Health Information & Decision Systems (CHIDS), an academic research center within the University of Maryland’s Robert H. Smith School of Business, and in collaboration with the University of Michigan School of Public Health.
“If patients don’t show up for their appointment, no amount of sophistication is going to help you,” D’Souza, the system’s vice president of enterprise data and analytics, said during his keynote address on the social determinants of healthcare outcomes.
In fact, the 14-hospital medical system includes a patient’s proximity to a bus line as one of several variables, both clinical and non-clinical, in its model to judge the risk of an individual’s chances of readmission, he said. D’Souza noted that hospitals systems are now “on the hook” for certain readmissions due to recent federal initiatives that link government payments to the metric, as well as other outcomes.
Other important social determinants used in the model include access to a grocery store, as well as church membership and marital status. The readmissions model assigns each patient a risk score based on the input variables, so that clinicians can focus their efforts on patients most likely to be readmitted.
Overall, the clinical care provided by hospitals is responsible for about 20 percent of a given patient’s heath outcome, half as much as social and economic factors. Individual behaviors like tobacco usage (or lack thereof) and the physical environment are also important determinants, D’Souza said.
“The social determinants of healthcare is a great area where people are still debating who owns this,” he said, adding that the federal incentives are pushing hospitals to consider the non-clinical data, although it is not clear if they are equipped to do so.
Physician buy-in is key to deployment of healthcare analytic models in the real-world. That makes integration into electronic medical records crucial because clinicians are resistant to logging into additional systems, D’Souza said. “Even if you have the best predictive model in the world, it just ain’t going to happen without integration.”
According to D’Souza, automation and machine learning has a role to play in the interpretation of medical images (a task currently performed by radiologists), but warned against overreliance and the dulling of common sense, saying “we’ve lost this notion of how physicians used to interact with patients.”
Anchored by the University of Maryland Medical Center in Baltimore, the University of Maryland Medical System consists of 14 hospitals and 25,000 employees. It admits 120,000 patients annually and earns about $4 billion in operating revenue.
The conference also featured research panels on topics at the intersection of health and IT, such as machine learning, genetics, payment reform and social media. Scholars from universities all over the U.S. presented their findings, as well as academics from the University of Sydney, University of Manchester, Hong Kong University of Science and Technology, National University of Singapore, Harbin Institute of Technology and The Neu-Ulm University of Applied Sciences.
“CHIDS is pioneering work at the intersection of AI, information systems, mHealth and healthcare, and is proud to have convened this event, where experts debated the landscape and impact of current uses and potential future breakthroughs such as deep learning,” says CHITA Chair Ritu Agarwal, CHIDS founding director, Robert H. Smith Dean's Chair of Information Systems and senior associate dean at the Smith School. “When we combine the increasingly large amounts of digitized medical data with state-of-the-art analytical methods, we have unprecedented opportunities to improve health outcomes and reduce costs."
For more information about the conference and the Center for Health Information & Decision Systems (CHIDS), visit: http://www.rhsmith.umd.edu/chids.
Varun Saxena, MBA Candidate 2019, Office of Marketing Communications
About The Center for Health Information and Decision Systems (CHIDS)
The Center for Health Information and Decision Systems (CHIDS) is an academic research center based in the Department of Decision, Operations & Information Technology (DO&IT) at the Robert H. Smith School of Business that collaborates closely with industry, government, and other key health system stakeholders. CHIDS' research seeks to understand how digital technologies can be more effectively deployed to address outcomes such as patient safety, healthcare quality, efficiency in healthcare delivery, and a reduction in health disparities. CHIDS offers the benefit of renowned scholars in healthcare analytics, technology innovation, adoption, implementation, and design. The pool of talent, knowledge and expertise in DO&IT is acknowledged by several publications as a top-5 performer in research production worldwide; the Information Systems group is ranked in the top-10 worldwide by Businessweek and U.S. News and World Report. CHIDS pioneers in the study of digitally enabled health system transformation and is widely known for its thought leadership and research collaborations.