I was searching for sensitivity and specificity testing for apoB. But I couldn’t find reliable references. But Relative Risk Ratios (RRR) for apoB is available. Mentioning RRR will be a surrogate for Sensitivity and Specificity?
Dear Rukman,
No, those are not quite the same things in analytical terminology. The "RRR" is more of a confidence interval for a hypothesis. "Sensitivity" in any assay usually means the lower limit of detection by a particular method. That can be developed from a calibration curve. "Specificity" means the degree to which other analytes might produce a similar result. That can be tested by doing the test on possible interfering compounds.
Best regards,
Steven
Hi,
Very nice question! If I understand the question correctly you are concerned with the relative risk (which is a ratio). If you think of a 2*2 contingency table, the four live cells are integer numbers (from up left (NW) to right and down (SE)): a, b, c, d. Sensitivity is the joint probability (test positive and have the disease or condition). Specificity is the joint probability (test negative and do not have the disease or condition). You can calculate absolute and relative risks using the data in the live cells. You can also calculate all sorts of conditional probabilities such as pr(test positive|disease) and so on and use Bayes' theorem. That is, you can calculate the pr(disease|test positive), pr(no disease|test negative, and so on.
The false-negative and false-positive rates can be calculated accordingly as well (the cell c is the false negative; cell b is the false positive). Finally, if you calculate the RR or the odds ratio, OR, from the data you can calculate the confidence intervals about these quantities. As the other commentator suggested, the confidence intervals are no substitutes for the specificity and sensitivity of the test. Take a look at G Gigerenzed's book Calculated Risks, 2002 Simon and Schuster NY, for a very pleasant reading of these issues.
Hope that this help. All the best.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Probability
- Probability Theory
- Reliability