Skip to content

AI in COVID-19 – Lessons Learned and Working Towards Synergy

“The essence of synergy is to value differences- to respect them, to build on strengths, and to compensate for weaknesses.” 

Stephen Covey, author of The Seven Habits of Highly Effective People

It is now over one year since the early weeks of the COVID-19 pandemic and its surrealistic and apocalyptic images: deaths without families and cities in lockdowns around the world.

Many efforts have been made by data scientists to deploy machine and deep learning to battle this novel coronavirus, and several of the meta-analyses to date in assessing clinical utility of these efforts have been disappointing. The deficiencies are universal yet predictable: poor-quality datasets, design biases, low reproducibility, lack of external validation, and inadequate clinical utility of models.

While these meta-analyses are appropriate and insightful in critiquing these efforts of machine and deep learning for COVID-19, there remains inadequate discussion about the increasing gap between clinical medicine relevance and data science evaluation as an equally important element in these works.

In addition, we should look at the positive aspects of these less-than-perfect works: institutions with more willingness to share data, data scientists with more desire to contribute their expertise in the face of a global health emergency, and clinicians increasingly aware of the artificial intelligence tools that can potentially have clinical utility. We need to celebrate these studies that promulgated from a genuine desire to improve our collective ability and capability to overcome this challenging global viral menace.

What remains missing is a call for a higher plane of clinician-data scientist collaboration to achieve not only symbiosis, but synergy (the former is co-existence while the latter is about “the whole is greater than the sum of its parts”).

On the clinician side, more education and appreciation for current era data science and artificial intelligence can expedite the timeline to this synergy. On the data scientist side, stronger efforts in collaborating with clinicians from the starting design phase of the projects and continuing this mutual effort can improve not only the clinical relevance of the projects but perhaps even the data science aspects as well. There is asymmetric focus on the data science and not sufficient assessment of the coupling to the clinical domain.

The Edupreneur Business Model Canvas

One type of physician entrepreneur is an edupreneur. If we use the definition that entrepreneurship is the pursuit of opportunity with scarce or uncontrolled resources with the goal

Show Buttons
Hide Buttons