Suppose there are two types of tests available. Test A: low-cost, but relatively less precise; Test B: high-cost, but more precise. Test B can be considered as the good standard with 100% accuracy. Suppose there is a group of patients with different sex, age, gender and race. The question is, which patient should be recommended to the high-cost Test B after the initial screening with Test A. Another complexity is that the precision of Test A varies from individual to individual, affected by sex, age, gender, and race.

The literature I currently reviewed is always trying to come up with the best strategy at the population level instead of individual patient level. That is a common cutoff point for all the patients regardless of their age, gender, etc. What really interests me is to come up a personalized strategy.

Any idea which literature I should look into?


Since nobody has answered on a week, I'll try a partial answer.

The personalized strategy you propose is already being used. For example, it was used in the widely performed triple trest for pregnant women - although nowadays it is a bit outdated.

What you want to know after the cheap first test (test A) is the conditional probability of illness given the outcome of test A and all other relevant patient circumstances. With that probability you can do a cost-benefit analysis of test B and decide if it's worth.

After triple test, probability of some fetus problems was computed and checked against risk (and cost) of more invasive tests to decide if it was worth taking such tests.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.