Healthcare, intended as a complex domain, is affected among others by the so-called financial toxicity, which describes the problems patients can have with the cost of medical care. This term was coined a decade ago to attract the attention of policymakers, practitioners, and citizens to the rising costs of cancer treatments and the resulting financial burden and distress on patients and their families.
This phenomenon is mostly widespread in those National Healthcare Systems (NHS) not based on universalistic access to care and when people do not have several health costs not covered. This often causes financial problems and may lead to debt and bankruptcy for people dealing with chronic diseases (e.g., cancer, diabetes, etc.). This issue is at the forefront of policymakers and experts. To counteract this phenomenon, we are trying to understand if a possible positive response for patients can come from using AI diagnostic tools.
Some primary analyses demonstrated that AI-based tools and their machine-learning algorithms can be implemented to predict individual risks of financial toxicity based on clinical, demographic, and patient-reported data. More in details, the use of AI for challenging healthcare financial toxicity needs for the identification of patients at risk who are missed by traditional risk assessment protocols as well as the expense and the indirect costs of mitigation strategies. This led to identifying the risk for diseases or other adverse health outcomes by defining if patients can benefit from diagnostic, monitoring, or treatments that they would not have received.
AI-based predictive analysis can be based on different data, such as age, gender, body mass index, number of children, smoking habits, and geolocation. These solutions come with interesting implications in terms of resource allocation, which can be better organized prioritizing the limited healthcare resources thanks to advanced and smart analysis of disease transmission risks based on specific spatiotemporal and socio-economic patterns. Such analysis must be done before starting treatment according to a risk assessment logic useful for both nation-based institutions and insurance companies. Moreover, this kind of assessment can facilitate the healthcare costs also for patients, making them able to take an informed decision about their medical path and the related expenses. In doing so AI-based tools must be appropriately validated, taking into consideration that in any case insurance coverage for such patients may not be easily provided, before can be fully considered real support for a shared decision-making concerning high-risk patient populations to align treatment decisions with their spending possibilities, and preferences.
However, the implementation of such AI-based tools can also have some issues, related, among others, to those patients who lack clear risk factors or symptoms and may not receive insurance coverage for confirmatory tests.
Even though the use of AI tools for challenging healthcare financial toxicity is in its infancy, their progressive advancement can contribute to making healthcare a real learning system soon.
