Diagnosing patients with the help of statistical physics
Doctors are there to diagnose and treat people. But sometimes a diagnosis can’t be made or doctors differ in opinion. Luckily, Alireza Mashaghi Tabari and his research team have developed a new framework to solve medical diagnostic problems. This framework can also be applied to many other research fields, such as cancer and autoimmunity research. In Toward First Principle Medical Diagnostics: On the Importance of Disease-Disease and Sign-Sign Interactions, they describe the results.
In clinical medicine, there are diagnostic criteria and agreed-upon definitions for diseases, but doctors often lack consensus flowcharts for diagnosing a patient. A patient with a given set of complaints will be handled differently by doctors, who often rely on their intuition and experience.
Therefore, Mashaghi developed a new conceptual framework to study medical diagnostic problems. He used different medical findings, including reported symptoms, observed clinical signs and collected lab data. ‘We show that it is possible to develop consensus flowcharts using our methods. For this aim, we also need new clinical research to generate the information required for this methodology, so we expect our studies will motivate new clinical research.’
Single disease assumption
One important ingredient of their approach is that it puts down the commonly held single disease assumption. When a patient visits a doctor, a common assumption is that either a single disease will be identified or that the information gathered during that visit is not enough to make a diagnosis. This is not a valid assumption in many cases. ‘Many disease processes, in particular those that are age-dependent, evolve simultaneously but emerge at different moments during our lifetime. These diseases may interact and affect each other’s progress. We use the information in such interactions to make a diagnosis even before the consensus diagnostic criteria are fulfilled.’
Linking medical diagnostics with physics
In their research, Alireza and his team are linking statistical physics and medical diagnostics for the first time. ‘Despite challenges, we are making steady progress in this research and published two more research articles since our first paper was published in July 2017. We received wonderful comments from experts and I remember that one of our reviewers praised our work by saying “this is something that has the potential to have a big impact on how we diagnose patients in the real-world.”’
Wide-ranging applications
Not only can the new methods be applied to medical problems, but also to much more wide-ranging diagnostic problems. For instance, the problem of assigning a state to a biological cell or a complex electronic device can be solved. ‘In particular, assignment of state to a cell is a major challenge in immunology and cancer biology and it has complicated developing therapies for cancer and autoimmunity. We predict that our approach will be generically applied to a wide range of problems in medicine, science and technology.’
