Turning on the lights for health data

Electrical sockets completely changed modern life. Standardized "application programming interfaces" in health care and the life sciences will do the same for health data.

Like Comment
Read the paper

Imagine that you’re living at the turn of the 20th Century and the dawn of electricity. Not only are Edison light bulbs replacing your gas lamps and candles, but people are using electricity for more than light—they’re jury-rigging flatirons and chafing dishes to the new Edison bulb screw-in sockets hanging from their ceilings. A 1904 tally by the Southern California Edison Company of “appliances” counted just 500 in the entire region. Today, it’s unfathomable that you would need to unscrew a lightbulb in order to iron your clothes, or that there wouldn’t be ubiquitous 110v AC wall sockets to plug in anything ranging from a refrigerator to a Wi-Fi-connected security camera. 

In most cases, electricity is generated miles away from our houses, yet we trivially consume it via well-designed and well-understood sockets and plugs. We aren’t even aware if our power supply switches from a coal-fired power plant to a hydroelectric dam, because it doesn’t matter—our television uses the same plug in the wall, regardless.

 Comparably, your iPhone or Android phone “plugs into” a wealth of data, enriching (or distracting from) our lives with apps that help us navigate traffic, share vacation photos, make grocery lists, or send donations to our favorite philanthropy. These “plugs,” called APIs for Application Programming Interface, are analogous to conventional plugs because they define a standardized interface with which we interact with a data source such as Google Maps.

 Today, when it comes to our biomedical and health data, we live in the equivalent of Southern California circa 1904. To power a health app, we mostly have to tie into clunky interfaces or hardwire our app to connections that may be one way in one hospital, and completely different in another. Patient data is still exchanged with fax machines because we haven’t embraced a standard socket for health data. 

 In our NPJ Digital Medicine Commentary we advocate for APIs in the life sciences and healthcare because the next wave of innovation in our field will involve consuming data and transforming it into something useful, whether that is a patient relaying information about her surgical incision to her surgeon, or a predictive model that helps clinicians anticipate a recurring cancer. Of course, just having APIs is not a panacea. Just as the world changed when people harnessed electricity to do all sorts of things using appliances and devices, the world of health will change when we can do the same with data. This means software engineering will become an important part of healthcare and biomedical research. We will need to consider how we bring this new category of people into the profession—and professional responsibilities— of healthcare.

Importantly, a few days after we published this paper, the Office of the National Coordinator for Health Information Technology (ONC) released the “final rule”  implementing the interoperability and information-blocking provisions of the 21st Century Cures Act (an overwhelmingly bipartisan piece of legislation passed in November 2016). Embracing APIs is not only a better way to move data to apps to trigger improved health. Now, it’s also the law.

(image credit: Bruno /Germany from Pixabay)

Erich S. Huang, MD, PhD

Chief Data Officer for Quality, Director of Duke Forge, Duke Health