Skip to content
English
Get help from a human
  • There are no suggestions because the search field is empty.

How does GIDEON calculate the probability for diagnosis or pathogen?

GIDEON’s diagnostic and microbiology modules are driven by a Bayesian probability matrix that estimates the likelihood of each possible disease or pathogen based on user-input data.

GIDEON’s diagnostic and microbiology modules are driven by a Bayesian probability matrix that estimates the likelihood of each possible disease or pathogen based on user-input data.

The underlying calculation follows Bayes’ theorem:

                                              ( P[S/D1] × PD1 )P-D1 [given S] = -----------------------------------------------                              P[S/D1] × PD1 + P[S/D2] × PD2 + ... + P[S/Dn] × PDn

Where:
P = Probability
S = Symptom complex
D1 = Disease 1, D2 = Disease 2, ..., Dn = Disease n

In the microbiology module:
PD = Organism probability
S = Phenotype complex

Two main databases provide input for this formula:

  1. Incidence of diseases by country: determines baseline probabilities of each disease.

  2. Likelihood of symptom by disease: defines how likely each symptom is for each disease.

How GIDEON Uses the Formula

When a user enters a patient’s list of symptoms and the country of acquisition, GIDEON performs the following steps:

  1. Retrieves all diseases known to occur in the selected country.

  2. Ranks these diseases by their relative incidence.

  3. Removes any diseases that are incompatible with the provided symptoms.

  4. Multiplies each remaining disease’s prior probability by the probability of the observed symptoms.

  5. Re-ranks the diseases according to the resulting values, producing a list of the most likely diagnoses.

This approach combines epidemiologic data and clinical information, applying Bayesian reasoning to continuously refine disease or pathogen ranking as more information becomes available.