Roadmap AIMIG

ES draft version /

• Introduction:
  • The global project is very large and must be broken down in smaller steps. An incremental development strategy is necessary.
  • The full project will need to be supported by a large community of open minded people sharing know-how in Open Source.
• Main objective:
  • At the current early stage the essential question is to explore the feasibility and usefulness of graphs for medical decisions, making prototypes.
• Strategy:
  • Begin with analysis of the traditional decision process.  Look at "Differential Diagnoses" documents and attempt to translate their experience in a visual graph.s  Make a minimal prototype which demonstrate how decision logic can be managed as graphs.
  • Mile stone:
  • From an application point of view, the above represent a minimal coherent prototype, but many technical challenges need solutions in order to demonstrate the feasibility in a nice way..
  • Such a meaningful prototype will become eligible for grants and sponsors opening the the way to further developments, shared in Open Source.
    
  • In order to motivate medical expert we need good visual presentation of graphs. A kind of  "chicken  and egg" problem.
• Main tasks:
• Software developments:
• Developers begin to work on a first minimal version of human-machine interface.
• Medical knowledge base:
• Medical experts, particularly with teaching experience, who will draw their knowledge directly in graph mode, at least in a limited medical domain.
• Patient record:
• Draw experimental patient records as graphs with particular attention to the relations Observation(s) ---> Problem(s) ---> Action(s)
• Patient information evolves step by step when more information become available..
• Decision support:
• Integration between the graphs of patient records and knowledge base.
  
• At any time in a consultation, the main issue is  " Given what is already known about the patient up to now, what should be done next ? ", i.e. the relative priorities of recommendations for typically more explorations as questions, examinations, as well later recommendations for explainable treatments.
• Road map:
1. Basis graph editor:
2. Using the graph editor make an experimental graph knowledge base in at least one medical domain.
   
3. Basic consultation prototype:
4. basic example of decision support
5. Integration of the above into a meaningful "proof of concept".
6. Seek resource from sponsors and apply for grants.
7. More developments
8. Mile stones of evaluation of the preliminary results and choices for the next developments.
   
9. .....
   
• ..