How to close the sickcare AI-DI divide

The digital divide describes those having or not having access to broadband, hardware, software and technology support. It’s long been acknowledged that even as the digital industry exploded out of this country, America lived with a “digital divide.” While this is loosely understood as the gap between those who have access to reliable internet service and those who don’t, the true nature and extent of the divide is often under-appreciated. Internet infrastructure is, of course, an essential element of the divide, but infrastructure alone does not necessarily translate into adoption and beneficial use. Local and national institutions, affordability and access, and the digital proficiency of users, all play significant roles — and there are wide variations across the United States along each of these.

There is also a sickcare artificial intelligence (AI) dissemination and implementation (DI) divide. Infrastucture is one of many barriers.

As with most things American, there are the haves and the have nots. Here's how hospitals are categorized. Generally, the smaller ones lack the resources to implement sickcare AI, particularly rural hospitals which are, increasingly, under stress and closing.

So, how do we close the AI-DI divide? Multisystems solutions involve:

1. Data interoperability
2. Federated learning Instead of bring Mohamed to the mountain, bring the mountain to Mohamed
3. AI as a service
4. Better data literacy
5. IT infrastructure access improvement
6. Making cheaper AI products
7. Incorporating AI into a digital health whole product solution
8. Close the doctor-data scientist divide
9. Democratize data and AI
10. Create business model competition for data by empowering patient data entrepreneurs
11. Teach hospital and practice administrators how to make value based AI vendor purchasing decisions
12. Encourage physician intrapreneurship and avoid the landmines
13. Use no-code or low-code tools to innovate
14. Don't know how to code? No problem. A new technology generates program in 12 coding languages and even translates between them.
15. Companies should look for three things in a no-code platform: it needs a simple interface that makes it easy to get data into the model training process; it needs to automate model selection and training, tasks that would normally be performed by data scientists; and it needs to be simple and easy to deploy with existing processes. 

We are still in the early stages of realizing the full potential of sickcare artificial intelligence. However, if we don't close the AI-DI gaps, a large percentage of patients will never realize the benefits.

Arlen Meyers, MD, MBA is the President and CEO of the Society of Physician Entrepreneurs