Basically: medicine as a whole is already some sort of expert system.
- Data collection and cleanup: Researchers conduct experiments to produce meaningful data and extract conclusions from that data.
This part isn't more automated because we have strict rules that prevent medical data collection and analysis without a clear purpose. Otherwise we'd be able to collect a lot more information to try and extract results from it using more inference-oriented techniques (deep learning and the like).
- Modeling & training: Expert panels produce guidelines from the results of that research. These panels are the "training part" of the system.
As a sibling comment said, replacing these panels with ML-based techniques isn't trivial because the data produced in the previous step is fairly noisy (p-value hacking, difficulty of capturing all the variables, etc.). Furthermore, the techniques that yield best results nowadays also produce them without clear explanations on why they hold, which is not something we are prepare to accept in medicine.
- Execution: Doctors diagnose and treat following said guidelines. In fact, they use decision flows that they themselves call... algorithms!
The main reason why execution is not automated is that we do not have the technology for machines to capture the contextual and communication nuances that doctors pick up on. There can be a world of difference between the exact same statement given by two different patients or even the same patient in two different situations. Likewise, the effect of a doctors' statement can be quite literally the opposite depending on who the patient is and their state of mind. One of the most important aspects of the GP's job is to handle these differences to achieve the best possible outcomes for their patients.
Basically: medicine as a whole is already some sort of expert system.
- Data collection and cleanup: Researchers conduct experiments to produce meaningful data and extract conclusions from that data.
This part isn't more automated because we have strict rules that prevent medical data collection and analysis without a clear purpose. Otherwise we'd be able to collect a lot more information to try and extract results from it using more inference-oriented techniques (deep learning and the like).
- Modeling & training: Expert panels produce guidelines from the results of that research. These panels are the "training part" of the system.
As a sibling comment said, replacing these panels with ML-based techniques isn't trivial because the data produced in the previous step is fairly noisy (p-value hacking, difficulty of capturing all the variables, etc.). Furthermore, the techniques that yield best results nowadays also produce them without clear explanations on why they hold, which is not something we are prepare to accept in medicine.
- Execution: Doctors diagnose and treat following said guidelines. In fact, they use decision flows that they themselves call... algorithms!
The main reason why execution is not automated is that we do not have the technology for machines to capture the contextual and communication nuances that doctors pick up on. There can be a world of difference between the exact same statement given by two different patients or even the same patient in two different situations. Likewise, the effect of a doctors' statement can be quite literally the opposite depending on who the patient is and their state of mind. One of the most important aspects of the GP's job is to handle these differences to achieve the best possible outcomes for their patients.
All that being said, there are companies trying to produce expert systems to help doctors diagnose. See https://infermedica.com/product/infermedica-api for instance.