Arte SheetsLearn. Try. Understand.

Healthcare & Clinical Calculations

Fagan’s Nomogram: Bayes Updates in a Spreadsheet

Calculate illustrative post-test probabilities from a prior, sensitivity and specificity, including valid boundary inputs and impossible-result cases.

From a prior probability to a conditional probability

Fagan’s nomogram provides a graphical way to update pre-test odds with a likelihood ratio. For a binary test, LR+ is sensitivity ÷ (1 − specificity), LR− is (1 − sensitivity) ÷ specificity, and post-test odds are pre-test odds multiplied by the applicable likelihood ratio.

The formulas below express the equivalent Bayes update directly in probabilities. With a 25% prior, 90% sensitivity and 85% specificity, they give 66.67% after a positive result and 3.77% after a negative result. The latter is not zero and does not by itself exclude disease. Results depend on whether the estimated prior and test characteristics apply to the setting; rounding does not remove uncertainty.

Fictional priorAfter positive resultAfter negative result
25%66.67%3.77%
60%90.00%15.00%

Probability of the condition given a positive result

Enter prior probability, sensitivity and specificity as decimals from 0 to 1. The numerator is prior × sensitivity; the denominator is the total probability of a positive test. If that denominator is zero, a positive result has zero probability under the stated inputs and its conditional probability is undefined.

=IF(COUNT(B2:B4)<>3,"Enter all three numeric probabilities",IF(OR(MIN(B2:B4)<0,MAX(B2:B4)>1),"Use probabilities from 0 to 1",IF((B2*B3+(1-B2)*(1-B4))=0,"Positive result impossible under these inputs",ROUND(100*B2*B3/(B2*B3+(1-B2)*(1-B4)),2))))

Probability of the condition given a negative result

The numerator is prior × (1 − sensitivity); the denominator is the total probability of a negative result. This is residual probability of the condition after a negative result, not the negative predictive value, which describes absence of the condition. A zero denominator is reported as an impossible result under the model.

=IF(COUNT(B2:B4)<>3,"Enter all three numeric probabilities",IF(OR(MIN(B2:B4)<0,MAX(B2:B4)>1),"Use probabilities from 0 to 1",IF((B2*(1-B3)+(1-B2)*B4)=0,"Negative result impossible under these inputs",ROUND(100*B2*(1-B3)/(B2*(1-B3)+(1-B2)*B4),2))))

Try the example

Use the original values or change them to test your understanding.

Further reading