- Tuesday, September 10, 2019 at 12:05PM - 1:30PM
- SLB Room 122
- Open To The Yale Community
- Add to Calendar:
Data is potential. Datasets are routinely reused and repurposed, combined with other datasets, or fed into machine learning models to yield new inferences or discoveries. Precisely what a given dataset will yield is uncertain: the value—or harm—that a single dataset could produce is difficult, if not impossible, to fully forecast in advance. In health, the routine of reusing and repurposing health-related data is foundational to advancing medical knowledge.
Datafication and machine learning have complicated this routine. From fitness trackers to health apps, patients are increasingly generating health-related data outside of a clinical or research context. Governments, companies, and schools are collecting data on social indicators and identifying correlations with health outcomes. Finally, non-health stakeholders are increasingly involved in analyzing health-related data, interjecting technology companies into health relationships.
Together, this expands the research potential of health-related data, but also exposes healthcare providers and patients to novel risks. A healthcare provider may be ill-equipped to manage and monitor secondary uses of the data they hold, particularly when faced with technology company counterparties. It may be difficult or impossible to prevent the use of patient data in machine learning models that harm patient welfare. Even among healthcare collaborators, data-driven collaborations may struggle to allocate value and liability among its members as unforseen applications materialize. The pervasive uncertainty about the potential uses—and potential harms—of sharing health data threatens to undermine trust in healthcare institutions as stewards of patient well-being.
This talk will explore how healthcare professionals can adapt to the uncertainty of a data-driven world. The first part of the talk will introduce a draft paper on data governance challenges in health, and on how trusts may provide a tool to ameliorate these challenges. The second part of the talk will discuss an in-progress project to help networks of children's hospitals govern data-sharing collaborations.
Keith Porcaro is a co-founder of Digital Public, an organization that builds governance models for data and technology projects. A lawyer and technologist, Keith's research focuses on how technology systems and interfaces mediate legal decision-making, and on training professionals to better understand sociotechnical systems. Prior to Digital Public, Keith was a technology consultant for American legal aid and international development agencies. Keith is an affiliate with Harvard's Berkman Klein Center and Duke's Center on Law and Technology, and an adjunct professor at Georgetown Law, where he teaches a class on technology and criminal justice. Keith has a JD from Duke Law School, and is licensed to practice law in California.