In the previous section, we mentioned that the main reason why artificial intelligence applications in health are not adopted as quickly as in other sectors is that clinicians and engineers do not work together, therefore value propositions are not created correctly and artificial intelligence applications are not accepted (1). In this regard, we defined the correct value proposition as “performing tasks and procedures that consume the doctor’s time and break the bond between him and the patient.”
Even in cases where the above basic factors are eliminated, we encounter technical, legal and operational obstacles. The main obstacles are the length and complexity of the processes related to the development, validation and obtaining of the necessary legal permissions for use of applications to be used in diagnosis and treatment. In many cases where all these stages are completed, we face the complexity in using the applications. Another issue is that many applications enter the market every day (2). Most of these products have not yet been approved by institutions such as FDA and CE, and even the approved ones are not yet possible to use together in clinical workflows, at least on a practical level. In other words, the algorithms are not integrated into each other and workflows. It is thought that a possible articulation may involve risks such as the products giving incompatible results for the patient or experiencing incompatibility problems due to different hardware and infrastructure requirements. In this context, platforms that include all the features of cyber-physical systems ranging from connectivity to cognition can provide access to all medical data and all applications in the market and integrate them into existing workflows in the context of the basic value proposition mentioned above.
(1) https://www.linkedin.com/pulse/sa%C4%9Fl%C4%B1kta-yapay-zeka-destekli-hizmet-sunumu-neden-di%C4%9Fer-karakas
(2) At the time of writing, there were approximately 300 applications in the market in the development, validation and approval stages.
