Start by adding symptoms / signs and test results. The more symptoms / signs and results you add, the more accurate differential diagnosis will be.
Please note that if % probability of the given diagnosis within the “Possible diagnoses” chart is less than 50%, then, this diagnosis is actually likely absent. On the other hand if the % probability is more than 50%, than this diagnosis is likely present.
If “Use default / normal values for features not provided” is ON, the app/API will assume ALL other, non provided features, results of tests, physical examination are normal. This is can be useful sometimes, when you are generating note / encounter data, and want to generate the documentation documenting “all other negative / normal”. In your apps, you may want to build narrower methods of this kind – allowing user to for example to use default /normal only for physical examination or history.
If “Use default / normal values for features not provided” is OFF, the app/API will assume all other, non provided features, results of tests are “NA“. This is can be useful if you want to write a bot, or guide provider through diagnostic process.