Is there any Statistic method which could reflect the diagnose value when the prevalence is lower?

[Dennisky]

Dear all,

Our present study is conducting a novel diagnosis test for predicting a disease by the X-ray. The gold-standard is a minimally invasive procedure can test a tissue sample for the disease.

In our clinical practice, if a special type of characteristic was observed in the X-ray image, we could diagnose the people with the disease (Makes Sure It (Almost)).

However, the special type of characteristic might be obeseverd only 10% in the patient with the disease (This is the estimated proportion). So, the sensitive is much lower in our preliminary experiments (PS: The Specificity is very higher).

But the special type of characteristic in the X-ray image is very valuable for the diagnosis and treatment. We still want to develop the method for diagnosing the disease (the gold-standard is expensive and painful for patient).

Is there any Statistic method or Statistical Indicator which could reflect its impact and value on clinical practice (when the prevalence of special type of characteristic is lower in the patient )?

Thanks a lot!

[SteveDenham · Accepted solution]

You might try a google search on "Bayesian methods for diagnostic tests" to see what other folks in this area may have tried.  That 10% incidence of the Xray variable may make this very difficult to establish a relationship, especially in terms of how many subjects might be needed to get reasonable credible intervals on the probability.

SteveDenham

[Dennisky]

Dear Prof.SteveDenham
Thank your for your suggestion!
We will study the topic about "Bayesian methods for diagnostic tests" for solving the problem.
What are some of the books or papers that are worth reading that you can recommend?
We have searched a book neamed Advanced Bayesian methods for medical test accuracy by Lyle D Broemeling.

Best regards

Dennis