“A journey of a thousand miles begins with a single step.”

Lao Tzu, founder of Taoism and Chinese philosopher

The following are the basic steps required to achieve a vision of a center of excellence for artificial intelligence at a healthcare institution. These steps can be executed in series, but preferably in parallel to expedite this very long journey.

Convening a coalition for AI in healthcare

The current timing is good to excellent for gathering a group of AI clinical champions, as interest in this domain has escalated in the past few years. It is good to form an internal group of AI champions, as well as an external group of such advocates. This step can be initiated by organizing a monthly program (weekly is probably too frequent while bimonthly or quarterly becomes difficult to maintain momentum) of AI in clinical medicine with updates on collaborative projects as well as articles.

Improving data and information technology infrastructure

Many health systems do not have adequate data and/or information technology strategy and/or infrastructure, so even a straightforward machine learning project can result in many hours of a data scientist’s time being spent just on data access and curation. Ideally, an organization can attain a stage 7 validation on the HIMSS Maturity Model of medical record adoption, as well as analytics, which indicates a sufficient level of digital transformation.

Building an ecosystem for the AI center

A center of AI in medicine in a health system needs to have a dyadic coupling to a robust ecosystem as its resource for data science students and data scientists, clinicians with this area of interest, and other stakeholders such as investors, entrepreneurs, and others. This can be a traditional “hub-and-spokes” model or other similar variation that will provide the AI center with an increasing amount of resources as it matures.

Initiating an educational agenda for everyone

It is imperative that the center of AI in medicine is the epicenter of local and perhaps regional educational efforts of AI in medicine. The educational portfolio can include the aforementioned monthly meeting with journal article reviews and guest speakers, short courses on AI in medicine for not only clinicians but all stakeholders in the health system, and the American Board of AI in Medicine (ABAIM) monthly introductory courses.

Designing clinician-first AI projects

There is a growing schism between data science and clinical medicine, so most data science projects have little clinical relevance and/or impact. This renders these efforts meaningless intellectual exercises in data science. A more productive approach is to go to the clinicians first and seek relevant clinical inquiries from them directly, thus transforming these into real world data science projects. As the ideas emanate from the clinicians, they are more apt to stay involved.

In addition to the center of artificial intelligence in healthcare discussion, many other topics will be discussed at our in-person AIMed Global Summit, taking place on May 24-26 of this year at the Westin St Francis in San Francisco. Representatives of many centers of AI in medicine will be participating at this meeting, in addition to the diverse attendees. Find more information here.