How standards support the AI promise to revolutionize healthcare delivery

Just consider the recent decision by Geisinger CEO David Fineberg, MD, to take on a newly created position at Google reporting to the tech giant’s AI lead Jeff Dean. It has been rumored (although not officially confirmed by Dr. Fineberg) that he turned down the job to lead the new company formed by Amazon, Berkshire Hathaway and J.P. Morgan Chase. That job went to surgeon, writer and researcher Atul Gawande, MD. In one of Dr. Fineberg’s last interviews before his announced departure for Google, he suggested he was staying on at Geisinger to take some of the healthcare system’s more innovative programs nationwide. One of those is a program in which patients can opt-in for DNA sequencing as part of their routine care. The combination of genetic information, data collected in the course of care, and patient commentary in focus groups is helping Geisinger predict which patients are more at risk for certain types of diseases and better target the kind of care they need, when and where, including at home. Dr. Fineberg believes half of what is currently done in the hospital will soon be ready for home delivery, with far less risk given the high rate of infections and other hospital-acquired conditions. More than 200,000 patients have already signed up for the Geisinger MyCode Community Health program and are receiving clinically actionable results, including the diagnosis of certain diseases long before symptoms develop.

This work dovetails nicely with Google’s investment in Verily, a life sciences AI business that is building tools to gather and organize health data and translate it into insights that can both prevent and better treat disease. Further, Dr. Fineberg has reportedly been tasked with organizing the tech giant’s fragmented health initiatives, which could expand the company’s ability to gather even more data about patients at home and on the move. For example, Google Fit’s wearable sensors could follow Apple’s lead and support the collection of medical grade data. Google’s NEST technology, meanwhile, could expand beyond thermostats and alarm systems to monitoring patients in the home. With Dr. Fineberg’s proven ability to get patients to opt in and share their most personal health data, it will be interesting to see how he might harness the collective intelligence of Google’s offerings in the health space.

Interestingly, the use of large data sets to better target patients and right size their care is not a unique idea. At the most recent Health Care Supplies Association (HCSA) meeting in Wales, well-known healthcare commentator and writer Roy Lilley called on the National Health Service (NHS) to leverage what he called the best healthcare database in the world. It’s true that the NHS — as a single, albeit highly complex, provider of healthcare for an entire nation — has a lot of data. NHS Digital has been collecting, standardizing and publishing data for the past two decades in a wide range of areas — from primary, secondary and emergency care to mental health, social care and population health. Central to this data is a unique NHS number used to identify patients with their associated health conditions and the care they have received. This number is a standard, albeit used only in the United Kingdom, which could limit its value for more global research. On the other hand, the NHS has recognized the value of global standards for the identification of the medicines and medical devices used in patient care. The NHS Dictionary of Medicines and Devices calls out the use of the Global Trade Item Number (GTIN) in the Department of Health’s Scan4Safety program (See the March 2017 edition of Standard Practices, “Standardize Globally, Change Locally” at https://www.hpnonline.com/standardize-globally-change-locally/). The Dictionary stresses the importance of standards for interoperability and analytics, as well as the primary role of suppliers in establishing GTINs for their products.