BayesDX

Diagnoses across, findings down. Each row applies one finding to the row above it.

Incidences are per 100,000 and set the starting odds; they are normalised across the columns you have chosen, so they express relative frequency in this patient's setting, not population risk. Hover a sensitivity for its source; type in any cell to supply or override one — manual values are marked blue and carry no citation. A row is applied only once every column has a value. Findings are treated as independent, so correlated findings will overstate the swing. Research prototype, not a clinical decision tool.