This is an article I had for quite a while as a draft. As part of my yearly cleanup, I've published it without finishing it. It might not be finished or have other problems.
Science fiction movies are full of advanced systems for medical analysis and treatment:
- Stargate SG1: The Goa'uld healing device is a box in which you lay, it scans yourself and after a few hours you are just healed. From basically anything, IIRC.
- Elysium: A healing pod which looks a bit like a CT scan.
- Prometheus: A medical robot pod performing a surgery (WARNING: The clip is from a horror movie)
- Star Trek: A hand-held medical scanner (Tricoder?) which scans you for diseases (clip)
I can imagine parts of those really well. Especially I think there is much room for improvement in the analysis by using machine learning.
The answer to "what's wrong with me" has many possible answers and depending on this many different treatments. As it is such a complex problem, I think an expert system is the right approach for it.
In the following, I try to structure some thoughts around it.
Problem overview
What we have:
- The patient: Age, sex, job, social background, ethnicity might be some of the important features everybody has.
- Patient record: A list of tuples
- Tests: (date time, test, test outcome)
- Treatement: (date time, treatment) - I would include pills, surgeries, excercise, ...
- Diseases: Possible answers to "what's wrong with me". Examples of this category
are:
- Genetic ones: Sickle cell anaemia, Huntington's disease, ...
- Viruses: HIV, Hepatitis D, Polio, Cowpox, Measles
- Bacteria: Yersinia pestis (black death), Haverhill fever
Expert Systems
Expert Systems typically consist of 3 parts:
- knowledge base: logical rules
- inference engine
- interface to human