How difficult is it to commercialize health AI solutions? I

In my discussions with leaders at cutting edge companies with FDA-approved products in healthcare AI, I’ve noticed a common thread. Most were surprised by how difficult it’s been to successfully sell FDA-approved solutions in radiology, pathology, ophthalmology, and more. They noted that medical centers aren’t sure how to use the solutions and are therefore cautious about pulling the trigger. They explained that the caution stems from uncertainty about how healthcare AI solutions will fit into their existing technology portfolio, how it will interface with their legacy systems, what their ROI will be and whether they’ll be reimbursed. They also asked, “Is it really necessary?”

Here are some of their comments:

•  “Medical centers are interested but they’re not moving rapidly to buy FDA-approved algorithms to incorporate into their daily operations. Although the data used for FDA approval establishes the safety and efficacy of the algorithms, it doesn’t establish improvements to efficiency or patient outcomes.”

• “We’re finding that medical centers don’t yet have the expertise to operationalize algorithms and selling them algorithms doesn’t seem conducive to their operating model.”
• “Changes to workflow in radiology departments is a barrier to adoption.”
•  “Medical centers are asking for more evidence than what they’ve been shown so farwhen buying these algorithms.”
• “The FDA only cares if [the algorithms] are safe or efficacious in performing the required task, but it’s not currently asking providers to show endpoints such as time savings or improved patient outcomes.” 

   

  As you can see from these comments, AI healthcare developers are running into issues with buyers, and a lot of it is centered around two areas—buyers aren’t sure how to operationalize AI solutions, and the evidence used to gain FDA approval isn’t enough to convince buyers to make a purchase. Often, several committees are involved in making a final decision and each application is reviewed and decided upon in isolation. Key questions include, “what problem does it solve?”, “how much better is it than the current solutions?” and “what value does it add?”

   

  The sale of these solutions to buyers like health systems, payers, and life science companies is complicated and bureaucratic. The sellers need internal champions to navigate clinical and administrative minefields to get the software adopted and paid for. The internal champion is someone in the clinical or operational team at the healthcare customers or someone in the R&D or commercial team in the life science companies who make the case for the organization to buy your solution. They make the case that you solve an important issue for their group and do so in a way that makes its adoption easy. They present to multiple committees and make the clinical and business cases for buying the new technology. The more these people have been equipped with evidence of the benefits of your technology, the more convincingly they can make the case for it and the better the chances are of their company buying it.

   

You often need to create a pipeline of customers and continue to do product demonstrations and provide evidence of the ROI of your solution to potential customers. In healthcare, the rule of thumb is that your customers are looking for at least a 4 or 5 to 1 ROI. That means that they’ll gain $4-5 of new sales or cost savings for every dollar they pay you.