Differential Refinement With New Data Chain

validated medical chain v1.0.0 cc-by-sa

Block ID: 84e27811-4b6f-474c-b75a-4555a3825cac

Community-contributed block. PromptDNA makes no guarantee of output quality or fitness for purpose. This block operates in a regulated domain. Nothing generated using this block constitutes professional medical, legal, or financial advice. User assumes all responsibility for use.

This block carries additional risk (risk level: medium) and requires a stronger domain-specific disclaimer when used.

Template

Refine the existing differential in {current_differential} given the new information in {new_data}, as an educational exercise in iterative diagnostic updating. (1) Restate the current differential and each candidate's rough standing before incorporating the new data, so the update is visible and honest rather than a silent leap to a new favorite. (2) Assess what the new information actually is — a test result, a new symptom, a response to treatment, or the passage of time — and how discriminating it is, since a finding that's equally likely across all candidates changes nothing however dramatic it seems. (3) Update each candidate up or down based on how consistent the new data is with it, reasoning in the spirit of Bayesian updating: the prior standing times how well each diagnosis predicts the new finding. (4) Watch specifically for data that should lower the leading candidate, because the discipline of diagnostic updating is looking for disconfirmation of the favorite, not just confirmation. (5) Reconsider whether any previously dismissed possibility is now back in play, or whether a new candidate should be added that the new data suggests. (6) State the revised ranking, what changed and why, and the next most discriminating piece of information to seek. Emphasize this is a reasoning framework to support a clinician's evolving judgment, not a diagnosis, and that real diagnostic reasoning requires a licensed professional evaluating the patient directly.

Refine the existing differential in {current_differential} given the new information in {new_data}, as an educational exercise in iterative diagnostic updating. (1) Restate the current differential and each candidate's rough standing before incorporating the new data, so the update is visible and honest rather than a silent leap to a new favorite. (2) Assess what the new information actually is — a test result, a new symptom, a response to treatment, or the passage of time — and how discriminating it is, since a finding that's equally likely across all candidates changes nothing however dramatic it seems. (3) Update each candidate up or down based on how consistent the new data is with it, reasoning in the spirit of Bayesian updating: the prior standing times how well each diagnosis predicts the new finding. (4) Watch specifically for data that should lower the leading candidate, because the discipline of diagnostic updating is looking for disconfirmation of the favorite, not just confirmation. (5) Reconsider whether any previously dismissed possibility is now back in play, or whether a new candidate should be added that the new data suggests. (6) State the revised ranking, what changed and why, and the next most discriminating piece of information to seek. Emphasize this is a reasoning framework to support a clinician's evolving judgment, not a diagnosis, and that real diagnostic reasoning requires a licensed professional evaluating the patient directly.

Variables

Name	Type	Required	Trust level
current_differential		yes
new_data		yes

diagnostic-updatingbayesian-clinicaliterative-reasoningchain-of-thought

Ratings

0.0

Overall (0)

0.0

Accuracy

0.0

Consistency

0.0

Clarity

0.0

Efficiency

Benchmarks

Not yet self-validated against any benchmark. Automated, evaluative only - not a factor in whether this block was published.

Submitted by James P FounderMod via mcp · 2026-07-18