I am planning to condacting a diagnostic accuracy systematic review and meta-analysis. In order to calculate pooled sensitivity and specificity. I need to know the number of TP, TN, FP, and, FN which are rarely reported in published articles. so how can I calculate the number of TP, TN, FP, FN from just the sample size, sensitivity, and specificity reported in a study?
Dear Binyam Girma Sisay:
That is a usual problem we usually face when we intend to perform a meta-analysis of diagnostic tests. Though STARD initiative encourages for providing 2x2 table, many papers do not contain such simple and valuable information.
First option could be asking to the authors of the different papers to provide 2x2 table. Unfortunately, my experience is that many of them do not answer.
To perform the 2x2 Table, you would need: sample size, values of sensitivity and specificity and the prevalence of the disease. Prevalence can be calculated thorough predictive values.
The reference posted down provides a simple calculator to estimate 2x2 table. You only must fill the “Diagnostic Test” table and all the valuable information is automatically provided in “Observed Contingency Table”.
It also gives you the accuracy of the estimates (essentially, when it delivers whole numbers without decimals, the accuracy will be good. The obtained accuracy is provided in “Fisher Exact Test” table.
To calculate prevalence from the predictives values the spanish and free sotfware Epidat is useful ( Route: Methods, diagnostic tests, simple test, predictive values).
Best regards.
https://statpages.info/ctab2x2.html
Dear Ángel Herráiz-Adillo
Thank you very much for your elaborated answer, I find it very helpful.
I have come across articles that didn’t report the prevalence, 2*2 table, and predictive values. In these situations what are the possible steps that I can take besides contact the authors? Thank you in advance for your help.
Dear Binyam Girma Sisay:
If such information is missed, I think it is completely impossible to perform the 2x2 table. Thus, such studies cannot be included in the metaanalysis (but they can be included in the systematic review). At this point, a raise concern is if papers lacking data to perform the 2x2 table are correlated with the accuracy of the studies or they are merely a random sample (it could be that studies with low quality reported their results with poor detail) . In that case, though necessary, elimination of these studies could result in 1) bias of the estimation of the overall test performance and 2) wider 95%CI in the overall estimates for the mataanalysis (as the sample size decreases). Studying publication bias could help to know the possible bias of such missed tests. In addition, Quadas tool will help to evaluate the quality of the included articles.
I recomend the doctoral thesis of Anne Rutjes "Sources of bias and variation in diagnostic accuracy studies, 2005" about these topics.
Best regards.
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Meta-Analysis