Presenter Information

John R. Hayes, Pacific University

Presentation Recording

View the presentation slides with presenter audio:

Open Recording

Description

This was a brief discussion on some basics in determining the effectiveness of a screening program. We discussed the definition of “truth” or the reference, the importance of a representative sample, and the meaning of the statistical measures of success. Specifically we discussed the importance of positive (PPV) and negative (NPV) predictive value rather than sensitivity and specificity in describing results. PPV is the probability of a person actually having a problem if the screening test is positive which is of primary interest to the clinician. Sensitivity is the measurement from the context of already knowing a person has a condition which is not the clinical perspective. PPV requires knowledge of the prevalence of the condition. In fact PPV changes as the prevalence varies across demographic populations. Bayes rule can be used to assist in the calculation.

Included in

Optometry Commons

Share

COinS
 
Jun 7th, 3:15 PM Jun 7th, 3:30 PM

Considerations in validating a 3D screening program

This was a brief discussion on some basics in determining the effectiveness of a screening program. We discussed the definition of “truth” or the reference, the importance of a representative sample, and the meaning of the statistical measures of success. Specifically we discussed the importance of positive (PPV) and negative (NPV) predictive value rather than sensitivity and specificity in describing results. PPV is the probability of a person actually having a problem if the screening test is positive which is of primary interest to the clinician. Sensitivity is the measurement from the context of already knowing a person has a condition which is not the clinical perspective. PPV requires knowledge of the prevalence of the condition. In fact PPV changes as the prevalence varies across demographic populations. Bayes rule can be used to assist in the calculation.