In the United States, artificial intelligence has already entered particularly delicate territory: that of authorizing medical treatments. In six American states – Arizona, New Jersey, Ohio, Oklahoma, Texas and Washington – a pilot program called WISeR allows AI systems to evaluate whether certain treatments may be covered by Medicare, the federal health insurance program aimed primarily at people over 65 and some younger people with disabilities.
The use of AI in healthcare, however, is not an American exclusive. More than 50 Eurasian countries have introduced the use of this technology in the medical field, albeit with a decidedly different approach compared to the USA, since in the Old Continent AI is used as a support tool for medical personnel and not to determine who to offer treatment to and in what way.
The limitations of the USA WISeR program
The WISeR program, acronym for Wasteful and Inappropriate Service Reductionwas launched by the US Department of Health and Human Services, to test whether AI can help reduce expenses considered «useless» within the Medicare program and prevent patients from undergoing procedures judged to be «not necessary».
The mechanism is that of the so-called prior authorization: before carrying out certain treatments, the doctor must upload the clinical information to a portal and explain why the procedure is necessary. The AI, provided by third-party technology companies selected by the Centers for Medicare and Medicaid Services (the program’s watchdog), compares the request to Medicare rules and is expected to provide a response within 72 hours. If the request is rejected, however, a human doctor from the appointed technology company must review the decision. The problem is that, in the opinion of some doctors, the system is producing delays and refusals of certain treatments that are difficult to understand.
For the record, 15 treatments are included in the WISeR program, including epidural steroid injections, some knee procedures and the use of skin substitutes, chosen because they have previously been associated with overuse or increased risks of fraud and waste.
In reality, however, the timing turned out to be very different from expectations. In some cases, authorizations that previously took a day to respond have started to require weeks of waiting. In Washington State, according to a report by Senator Maria Cantwell, patients had to endure an average wait of 15 to 20 days, two to four times longer than before. In some facilities the wait was initially up to a month.
The use of health intelligence in the EU in healthcare
According to a study by the World Health Organization, 50 European and Eurasian countries are also already using AI in some healthcare functions. Assisted diagnostics is the most widespread application: it is reported by 64% of countries, especially in areas such as radiology, ophthalmology and dermatology. France, Portugal, Hungary, Sweden and the Netherlands already use it and intend to continue. Italy and the United Kingdom, however, still report informal use in some structures, while Spain, Poland, Belgium, Serbia and Ukraine are experimenting with various applications.
AI is also used to assist patients through chatbots, indicated by 50% of countries, and to automate logistics and administrative tasks, present in 40% of the systems analyzed. In Slovakia, for example, software used in radiotherapy has automated organ counting, halving the time needed for oncologists to carry out this activity.
France and Spain are among the European countries most advanced in adoption: France has consolidated applications in surgery, diagnostics and patient monitoring, while Spain has developed defined practices for assisted surgery, administrative activities and chatbots.
In Europe, however, there is a certain difficulty in implementing the use of AI on a more extensive and widespread basis, not so much due to an economic issue, but due to the legal barriers currently present. In this case, 86% of the countries interviewed by the WHO indicate legal uncertainty as the most important barrier, ahead of financial accessibility, cited by 78%.
