En Garde!

May 20, 2019

Every Thursday night, I find myself facing an opponent brandishing a fencing sabre on a narrow strip, 14 meters long by 2 meters wide, called a piste. Hooked up to an electrical circuit, wearing high-tensile fencing whites and metallic jackets, we face each other. My adversary’s intent is oblique, hidden in shadows behind the mesh of a fencing mask—does he expect to make a “first intention” attack? Or is he inviting me to attack, allowing him to score a touch with a “second intention” parry, then riposte? Once the judge commands “fence!” initiative and intention may change many times in the course of a single touch between us.

Fencing is a fluid, “open-skilled” sport in which one must constantly modulate between offense and defense in response to an opponent who is also constantly sliding between these modes. On top of that, good fencers are excellent at waiting for a split second (“preparing”) and deciding whether to attack or defend based on cues from their opponent. The South Korean “K-Sabre” style emphasizes speed and deep lunging, while the Eastern European “old school” emphasizes small footwork and precise bladework. However, you can’t get away with pursuing just one approach. Competitive fencers must be multidimensional, capable of changing styles and able to go on offense or defense instantly. This dynamism is the challenge and joy of being “on piste”—arriving at the correct combination of offensive and defensive actions over a 5-, 10-, or 15-touch bout to emerge with a win.

Also? Getting to (legally) chase someone with a sword is fun, too.

The sport of fencing offers a great metaphor for thinking about data and data science. In the last decade the armamentarium for actively doing useful things with data has grown tremendously. Data munging, MapReduce, deep neural networks, APIs, and remote procedure calls give us the ability to rapidly reconfigure, move, and refine information from raw data. An excellent article from the Harvard Business Review characterizes this activity as “data offense.” During a recent visit to Alphabet in Mountain View, California, the Google folks described to us their architectures for rapidly turning unstructured data into structured and analyzable form in real time. Their corporate culture is literally engineered around continuously acquiring, transforming, and using data as it flows through our lives (for better and worse).

Erich_lunge.JPG
Forge Co-Director Erich Huang, MD, PhD, on the attack. Photo courtesy the author.

Such new capabilities contrast with a more traditional approach to data, one whose static and defensive nature is reflected in the vocabulary that describes its components: “data warehouse,” “database,” “electronic data capture.” They are artifacts of an era defined by a preoccupation with accumulation, storage, and asset protection rather than with forward-looking, continuous, active data analysis.

Like fencers, data scientists must be able both to go on the offensive and to properly defend themselves, depending on the context of the bout. Yet we are only now beginning to learn data offense. Data offense is helping us understand more about the molecular characteristics of injury recovery in surgical patients, making predictions to take better care of Medicare patients, and providing insights that allow us to improve complex clinical documentation workflows. But we still struggle with legacy infrastructure and a hoary culture of health data left over from a static and defensive era.

And “activity” no longer need be relegated solely to the domain of offense. There are new capabilities that focus on dynamic “defense in depth,” where layers of data defense actively communicate with one another rather than operating as static and independent defensive lines. Major cloud providers such as Google Cloud Platform, Microsoft Azure, and Amazon Web Services provide artificial intelligence services that alert administrators to suspicious network occurrences, unknown vulnerabilities, or inattentive user behavior.

As an academic medical center, we have twin obligations for data: to go on data offense to make maximal use of data to improve our community’s health while also appropriately protecting those data from misuse and upholding our ethical and regulatory obligations. We have to do both and be good at both—just as a fencer succeeds only by using both offensive and defensive actions.

In order to accomplish this we need new coaching and intense development of new fundamentals and novel techniques that allow us to go on data offense. These are unfamiliar skills for an industry that has been mostly defensive. Acquiring them will require different mindsets, entirely novel mental patterns, and new “muscle memory.” And we cannot afford to be caught flat-footed in this transition. The world is changing, and major shifts in how healthcare will be delivered require us to play both data offense and defense with panache.

There are moments—still rare—on the strip where I can feel the space and time between me and my opponent and how I can use them to score a touch. This comes from being mentally ready, and having physical mastery of actions with which I can quickly obtain an “attack in preparation” or take a parry at the correct distance for an effective riposte. Good fencers draw from a large array of such defensive and offensive actions; world-class fencers from an even larger array, together with the ability to read their opponent's tendencies and cues. As a top-flight health system, we must master this space and time—we must be Olympic fencers.


See more blog posts from Dr. Huang