Tool Overview

ProbID turns pretest thinking into a visible workflow by combining setting, findings, and likelihood ratios into an educational post-test estimate.

Use it for

CAP, VAP, endocarditis, invasive mold, and other syndromes where diagnostic uncertainty matters more than rote recall.

Interactive Tool

ProbID

Choose syndrome, location/setting, and features to estimate post-test probability using likelihood ratios. (Educational aid—not a guideline.)

Select features

Clinical syndrome

Pick location/setting + findings/tests in the catalog. Close it when done.

Location: Primary Care(Pretest 3.0%)

Selected findings/tests

None selected yet. Open the catalog to add.

Stepwise update

Choose findings/tests to see stepwise probability updates.

Fagan nomogram

Multiplying LRs assumes conditional independence. Correlated inputs may overestimate certainty.

Post-test probability

Estimated probability

Pretest 3.0%

3.0%

Combined LR: 1.00

Educational estimate only. Always use clinical context.

Decision layer (MVP)

Harms are auto-estimated from syndrome + selected high-impact findings. Treatment threshold uses: P(treat) = Harm of unnecessary treatment / (Harm of unnecessary treatment + Harm of missed diagnosis).

Harm of missed diagnosis

Harm of unnecessary treatment

What is driving harm?

Baseline missed-diagnosis harm (CAP)10

Source: Metlay et al. ATS/IDSA

Total missed-diagnosis harm10

Treatment threshold: 23%

Observation threshold: 12%

0%Observe ≤ 12%Treat ≥ 23%100%

Observe / monitor(Post-test 3.0%)

No additional high-impact risk modifiers selected; using syndrome baseline harms.

What’s driving it?

No selected findings yet.

Educational content only. Not medical advice. See references & methodology.