Positive About What?
Asystole

MC Positive in the building.
"My test says yes. Case closed."
Doctor Data:
Positive about what?

Sensitivity: of those with disease,
How many test positive? Give me that, please.
Specificity: of those without,
How many test negative? Work that out.
You swap the groups, then act surprised;
Your denominator came in a disguise.
Positive predictive value asks something new:
Among the positive tests, how many are true?
Pretest probability sets where we start;
The history and setting both play a part.
Base rates matter when interpreting the sign;
Your strip has two lines. Your argument has nine.
A result can change the chance that disease is there;
How much it changes takes a little more care.
You brought a gold-plated test to the booth;
The colour of the casing didn't certify truth.

Before the test, then after the test,
Update the odds; let the evidence address
The chance of disease in the person in view.
Positive about what? Show me how you knew.
Before the test, then after the test,
Update the odds; let the evidence address
The chance of disease in the person in view.
Positive about what? Show me how you knew.

One thousand people; ten have the disease.
Ninety percent sensitive: nine positives, please.
Nine hundred ninety without it in the crowd;
Ninety percent specific still lets false alarms sound.
Ninety-nine false positives join those nine;
A hundred eight positive tests in the line.
Nine out of a hundred eight: about eight percent.
Your "definitely has it" needs to pay the reasoning rent.
Same test, different prior, different chance at the end;
Your certainty's a sales pitch your numbers won't defend.
Likelihood ratio: weigh the result you got;
Multiply the pretest odds. That's the plot.
Probability over one minus probability: odds.
You skipped that conversion and appealed to the gods.
Post-test odds over one plus those odds brings us back;
Probability's the answer at the end of that track.

Before the test, then after the test,
Update the odds; let the evidence address
The chance of disease in the person in view.
Positive about what? Show me how you knew.
Before the test, then after the test,
Update the odds; let the evidence address
The chance of disease in the person in view.
Positive about what? Show me how you knew.

L-R plus: sensitivity on top,
One minus specificity beneath the fraction bar.
L-R minus: one minus sensitivity,
Over specificity. Know where your numbers are.
A ratio of one leaves the odds where they stood;
A useful result shifts them, as evidence should.
Then weigh the next step, its harms and its gains;
The patient has a context beyond your test's claims.

Before the test, then after the test,
Update the odds; let the evidence address
The chance of disease in the person in view.
Positive about what? Show me how you knew.
Before the test, then after the test,
Update the odds; let the evidence address
The chance of disease in the person in view.
Positive about what? Show me how you knew.

"So a negative means I'm completely clear?"
You missed one of the ten.
"Right. Maybe keep the calculator here."
Now we're getting somewhere.