Data and Statistics
Likelihood Ratios Bedside Bayesian Lite
A Cardily data-literacy explainer on likelihood Ratios Bedside Bayesian Lite, focused on interpretation rather than invented precision.
Metric definition
Positive likelihood ratio is probability of a given test result in diseased patients divided by probability of that result in non-diseased patients. It updates pre-test probability upward when positive.
Negative likelihood ratio does the reverse for reassuring negatives.
Fagan nomograms visualize the update-useful teaching tool in cardiology workshops.
Where the data usually come from
Test accuracy meta-analyses in cardiology supply pooled sensitivity and specificity converted to LRs.
Single-center studies may misestimate LRs if spectrum bias enriches severe disease.
NHANES rarely supplies stress test LRs directly; derived from trial and cohort literature.
How trends can fool you
Applying troponin LRs from early-generation assays to high-sensitivity assays without recalibration.
Using LRs from inpatient MI workups for outpatient chest pain clinics with different prevalence.
Forgetting that sequential LRs multiply only when tests are conditionally independent-often violated.
Equity and subgroup reading skills
Test performance may differ by sex and body habitus-pooled LRs hide subgroup misprediction.
Access to high-quality echo or CMRI changes post-test pathways after stress testing.
Teach LRs as tools reducing over-testing in marginalized groups historically subjected to cascades.
Questions to ask of any chart
Spectrum of patients in accuracy study?
Same assay generation?
Pre-test probability range assumed in example?
Putting it together
Likelihood ratios offer bedside Bayesian lite for Cardily readers linking stress test false-positive logic to quantitative updates.
Practice one worked example with low pre-test probability to build muscle memory.
Educational-not diagnostic instruction.
Educational disclaimer
This article is part of Cardily’s educational Insights library. Cardily content is for educational and informational purposes only and does not constitute medical advice, diagnosis, or treatment. Always consult a qualified healthcare professional for personal medical concerns. In an emergency, call local emergency services.