...

Part I: The why

I remember long ago, on the day of my Internal Medicine residency interview at Yale, the Chair of Medicine at the time – Ralph Horwitz – a charismatic general internist, spoke facetiously of the “intellectual bankruptcy” of specialty medicine. The message, I take it, was two-fold. The first point was that it is inherently easier to focus on a single organ system than to manage the myriad complexities of multiple diseases, many of which may not obey such tidy boundaries. But the second point, which might be a more important one, is that patients need a strong general internist (or primary care provider), who is able to make decisions that may impact multiple diseases, all the while weighing relative priorities and respecting patient wishes. It goes without saying that patients are not the sum of a disparate collection of diseases.

Since that time, the practice of internal medicine has (fortunately and unfortunately) become more and more complex. New effective treatments keep on emerging and, remarkably, have begun to show benefits across disease boundaries. For example, new type 2 diabetes medications also mitigate cardiovascular and renal risk. As expected, the burden of keeping up with the literature is even more oppressive. Whereas my own field, cardiology, was the only one with an established set of practice guidelines in the early 2000’s, there are now guidelines for almost every major disease.

Understandably but regrettably, rather than nearing the ideal of a strong central decision-maker, we are going in the other direction with sharper and sharper divisions between clinical domains and more and more decision-makers per patient. When these different providers are in the same system, one can at least have a common record, though it can be difficult to wade through notes (i.e., unstructured data) to extract relevant details. But if they do not share an EHR, it becomes even more difficult as one provider must expend considerable effort to understand what others are doing.

One of the goals of our platform is to bring the software-driven management of as many disease conditions as possible to a single location for a single team. This ensures that impact of all relevant diseases are considered for any medication or diagnostic choice and that patient preferences consistently inform every decision. Moreover the inherent learning nature of our platform (more on that in a future post on machine learning) sees each of these adjacent conditions as relevant features that may guide iterative optimization. Finally, as it maximizes the number of conditions managed by a single team, it begins to resemble the ideal of a strong central-decision maker, well-versed at a near-specialist level for a large group of diseases.

At our practice we are focusing initially on cardiovascular disease, but we are comfortably managing many adjacent conditions as we feel that many of the boundaries across disciplines are artificial and unhelpful. In a few months time we will be broadening our offering to include a seamless integration of primary and cardiovascular care.

All this sounds great … in a future post we will address the “how”.

Rahul D.