Visual 6:layered approach to clinical decision making, GIM, November 7

A layered approach to clinical decision making: 

Explainability, trust and layers of clinical decision making in pre and current AI LLM era:

Machine layer and AI dominance with humans in the loop:

How useful is AI in the loop of humans and how crucial are humans if placed in the loop of AI?

Analytical scientific and EBM layer: This layer is where our clinical decision making lab appears to be largely engaged in although the other two layers are no less important.

We have already shared something around this layer in our previous demos particularly our two video links shared above.

Human layer: This is the most important layer where clinical decision making actually happens at multiple human stakeholder levels:

We are all apprentices in a craft where no one ever becomes a master.

Ernest Hemingway, The Wild Years

https://www.goodreads.com/work/quotes/1763688-the-wild-years

Human, Scientific and Machine layers :

Anatomy of cognitive layers:

Physiology of cognitive layers in clinical decision making: enter Bloom's taxonomy!

RUAAEC

ApRUAECAp

More here on the bloom game of learning cognition: https://sites.pitt.edu/~super1/lecture/lec54091/001.htm

Bloom's taxonomy image copyright as well as an enlightening write up: https://www.niallmcnulty.com/2019/12/introduction-to-blooms-

Enter decision trees and the machine layers:

Shukratic Conversational dialogue:

[01/11, 19:51]hu1: @⁨hu2 made this decision tree for cough as presenting symptom. While making this chart, I realised how the decision points, point back to other organ system generalised symptoms to go through a similar decision tree to confirm or rule out a differential.

[01/11, 19:53]hu1: @⁨hu2 @⁨hu3 @⁨hu4

If this is something workable, Im planning to make similar decision tree for all other generalised symptoms.

[02/11, 07:17]hu2: Looks very interesting! Well done 👏👏

The only problem with this "yes no," binary,  reductionist , decision making approach is that the patient will end up trying to answer a lot of questions with 50% probability of accuracy and could be very taxing for the patient!

Our approach to decision making is on the other hand synthetic where instead of asking incisive questions (Bloom's level 4) that could be limited by the limitations of current knowledge (aka limited static ontology), we collect all possible event data and try to synthesize a broader picture (Bloom's level 6) after making the necessary reductionist edits using the Bloom's 4 knife!

Apologies if the above appeared inscrutable because currently I'm looking at whatever data flows into me, in the context of my upcoming November 7 presentation that I shall try to share with you ASAP once ready when it could make more sense!

[02/11, 07:18]hu2: A Herculean task, either way, although perhaps could be made more efficient using ML tools but then ML is limited to what currently humans know and that may not be enough, which you may realise if you start sitting in a real OPD and start seeing real patients!

Hu1 is a medical student who is working as our PaJR clinical decision making volunteer and also working as a research assistant in our collaborative project on building a potential "clinical decision making automated user interface" with IIT Hyderabad. He is possibly one of the rare medical students to have presented a clinical decision making scenario in a medical CPD conference within a month of having just entered medical college! More about him here along with our other team members past and present: https://medicinedepartment.blogspot.com/2021/03/medicine-department-training-programs.html?m=1

The single click entry user interface to our regular decision making workflow: 

https://pajr.in/

Workflow in brief: 

https://medicinedepartment.blogspot.com/2025/10/current-elective-roles-in-pajr-workflow.html?m=1

Workflow theory generalized: https://medicinedepartment.blogspot.com/2024/11/technology-end-user-driven-ecosystem.html?m=1

Workflow theory individualized:

https://preprints.jmir.org/preprint/81950/accepted

Workflow theory simplified: 

https://nivarana.org/vital-signs/beyond-prescriptions-a-user-driven-approach-to-navigating-chronic-diseases