Canadians are justly proud of our health-care system, but I suspect there are a number of us who have been closely watching the U.S. plan for “precision medicine,” and I sincerely hope some CIOs are among them.
For those still getting up to speed, U.S. President Barak Obama recently unveiled the Precision Medicine Initiative, which will see $215 million spent on collecting and analyzing genetic information from a million American volunteers. The desired outcome of this project is to offer more personalized, effective treatment for a variety of conditions. While I’m surprised this wasn’t called “big health care,” the use of the word “precision” in this case might not be a bad way for IT leaders to think about their approach to providing technology as a way of treating everything that ills the enterprise, too.
From Predictive To Prescriptive
At most conferences I attend, for example, the best experts seem to hope for its what might be called predictive IT. This is usually the use of analytics to offer guidance on what may happen in a given market, or with particular customers, or what could change internally. A few years ago, Information Week profiled Michael Wu, chief scientist for Lithium Technologies, who tried to define predictive analytics a little more clearly: it was about looking at what might happen, with an emphasis on the “might,” because you never really know what you don’t yet know.
Obviously this is leaps and bounds beyond what a lot of IT does today, where even getting a holistic sense of what has already happened can be extremely difficult. However Wu also envisioned what he called “prescriptive” analysis, where data would recommend a course of action and give a sense of what might happen if you chose one route over another.
In contrast, I don’t think anyone is talking about precision analytics today, although I think it might be a better way of articulating the intended outcome of all this information gathering and sorting. Too much of what businesses do — take sales, marketing, HR, whatever you want — seems largely based on guesswork. We think a customer may like this product. We have a hunch this ad campaign will generate demand. We have a feeling this job candidate might be the right fit. Precision may never be ultimately achieved, but it’s what we’re shooting for.
How Precision Analytics Could Start
Precision Medicine is predicated on the notion of proactively sharing available information — genetic data — to improve American health care. The concept of precision analytics might be similar, in that could provoke long-overdue discussions among competitors in industries or across the entire business sector on how they can better collaborate by pooling data they keep in confidential silos today. This has been coming up a lot at security events lately, but there’s probably scope for a lot of other kinds of data too.
Of course, that’s a lot easier said than done, and even if it was possible to get the data achieving “precision” would be a challenge in itself. As the New Yorker pointed out, this is already an issue with Precision Medicine:
As scientists continue to draw connections between DNA data and health outcomes, the problem of interpretability continues to grow. Many doctors are simply not qualified to make sense of genetic tests, or to communicate the results accurately to their patients.
Precision analytics will be no different, and will require a different interrelationship (and interdependency) among IT departments and others within the organization. That’s not a reason not to do it, of course. Gartner’s biomodal IT concept may be one of the models to think about to get closer to precision analytics.
As the Precision Medicine initiative suggests, what is aspirational today can be reality tomorrow. Today, CIOs and their teams may find it difficult to be predictive. They may not see the intent of what they work on to be prescriptive. With greater precision, however, everything else will become easier.
