How Far is Reimbursement for Algorithms in Healthcare? I

Reimbursement is critical for any new technology in healthcare. If the use of a new diagnostic or therapeutic technology is not paid for by insurance, there is very little chance that it will gain widespread adoption. Medical innovation is expensive and beyond the reach of most patients or even medical centers if it is not reimbursed. For decades, companies that have developed new solutions have done so with clear plans to show their benefits and thus secure reimbursements. Most public and private payers have established criteria and processes to grant a decision about reimbursement of new medical innovations. Usually, there has to be a fully demonstrated benefit to the patients or providers for a clinical innovation to receive a positive reimbursement decision. This has to be shown in well-designed trials. We addressed the lack of such studies for AI in Healthcare solutions so far. This promises to slow down the adoption of such technologies.

Although the FDA only requires safety and efficacy metrics to approve AI technologies, health plan require outcome data to issue new codes. Physicians will be very cognizant of the impact of the AI technologies on their incomes and if they see an adverse impact, they will be slow to adopt or will continue with current approaches to protect their incomes. This is important in the context of reimbursement since it is possible that part of the reimbursement for the use of these technologies can go to the physicians and thus driving their willingness to use them.  What about the issue of the threat to physician income in the future? or, at least radiologists in the short-term? I have to say that I don’t see any threat in the near future. Radiologists can review the whole scan  and have a global perspective while an algorithm only reads one area and looks for one type of issue.  A future when multiple algorithms run on a scan to look for any possible abnormality without the need for a radiologist seems far off.