Thread by @nelsonSpinto

Accelerated innovation using #BigData and #DataScience in medicine is sometimes just a matter of making the “adjacent possible” happen. The concept of “adjacent possible” in innovation postulates that you need to have the right tools and concepts available to combine.

Mini-🧵👇

In “Where Good Ideas Come From” @stevenbjohnson writes: “What kind of environment creates good ideas? The simplest way to answer it is this: innovative environments are better at helping their inhabitants explore tge adjacent possible, because they expose a wide and diverse...

... sample of spare parts -mechanical or conceptual- and they encourage novel ways of recombining those parts.” It doesn’t mean that people can’t have ideas that are “ahead of their time”, that is, outside of the adjacent possible, several generations down the innovation lane, it

just means that those ideas are unlikely to succeed if the right spare parts are missing. Johnson uses the example of video streaming online. If YouTube had been created in 1995 when a 2 minute video took 1 hour to download, the company would have failed. A decade later, in the

right environment, it was revolutionary. So in the #DataScience in medicine, let’s not focus on the stuff that is far ahead of our time. There won’t be an #AI to “solve cancer” anytime soon (see the demise of heathcare #IBMWatson as an example). Or a closed-loop #AI that will

obviate the need for diagnostic thinking by human clinicians. However, our “adjacent possible” in health #DataScience is huge, and growing! The tools and spare parts are all there: huge biomedical datasets being collected, curated, and merged; machine learning algorithms that

embrace the non-linear compexity of biological systems and allow us to explore the information space for meaningful patterns in ways we couldn’t even a decade ago; the computational power to perform these analyses not in a few universities or private companies with the resources

but in thousands of #DataScience labs around the world; and, of course, a new generation of tech-savvy, data-embracing, machine-speaking healthcare professionals and data science collaborators. Our adjacent possible space is huge. Now, let’s create the environment to combine

those spare parts in meaningful ways. Let’s bring the right clinical problems and the right data and the right computational tools and the right people (clinicians, data scientist, informaticians, implementation specialist) together and let them combine a thousand ways. Let’s

create a primordial soup for #DataScience in medicine and lets give it the right time it needs to combine, evolve, and flourish. Lets not develop and implement bad systems, with narrow scope, biased approaches, and without the right combination of people and tools behind it.

We’re in a privileged time in #DataScience. We’re like a bunch of fatty acids and proteins waiting to combine in the primordial soup (or maybe just the YouTube guys in ‘05). Give it time and magic will happen. We owe it to our patients to not get fooled with tricks and gimmicks.

(11 tweet thread and only 1 typo that I can find... score!)

@threadreaderapp unroll