I would like to assess the reliability of different immunoassay methods (nephelometry, EMIT, immune diffusion etc) for C-reactive protein. I felt correlation studies using human blood samples would be the gold standard to compare the methods. Does anyone know if this assumption is correct? And if yes, what would be an acceptable correlation?
You can analyse one parameter by various techniques and find regression factor. So each technology is gold standard itself . The quality of each results is depends on many pre and procedural preparations. Most advanced technology gives maximum specificity & increases sensitivity of assay. We consider these factors for selection of method. Technology deliver best results is gold standard for period of time. None of technology remains gold standard method in current era. For time being you can predict your results but you can not label it as gold standard , it could be but not always.
The "acceptable correlation" depends on the proposed usage of the method. If you use a "quick & dirty" method the correlation is of course low and if you use a high sensitive assay in an appropriate analyzer the correlation should be as high as possible. Human serum is a good source of CRP, its stable for several days at room temperature but if you are doing the experiments over a longer period of time you should freeze aliquots. I am not sure whether the companies use purified CRP for the evaluation of the assays but in my opinion serum is the better choice because its "real life" and you will have all interfering substances reflecting the normal situation.
Andrea:
For quantitative methods I’m agree with Yadav, that the statistical analytic procedure will give you the answer in terms of linearity, accuracy and precision (like regression), but you must to determine which will be the reference method and the reference standar.
The term gold standar is used when you have a method (or procedure) that gives you without error the true/false condition or is the best under reasonable conditions. Then your methods are evaluated in function of the responses in this categories. In your case, PCR is the analytical object, the gold standar must be one analytical method, maybe the PCR would be used like reference standar that it's another meaning.
Be careful when use the term “correlation”, if it’s interpreted like Pearson’s or Spearman’s correlation, because that are not the right procedure to measure the reliability of quantitative methods, you can use the Lin’s concordance correlation coefficient or the intraclass correlation coefficient.
Best regards.
- Badges
- Science method
- Similar topics
- Biological Science
- Immunology
- Immunology Techniques
- Immunoassay