RRML - Evaluation of Diagnostic Tests: Receiver Operating Characteristic Curves

ISSN online: 2284-5623

ISSN-L: 1841-6624

Impact factor (2019): 0.945

Rejection rate (2020): 75%

Română English

Clarivate analytics (ISI) Impact factor

Advanced search

Top 10 downloaded articles
- May 2021 -
Prognostic value of serum lact... 9
Vaginal candidiasis in Konya a... 5
Romanian Review of Laboratory ... 4
Prevalence of ∆F508 cystic f... 4
Polymorphisms in autophagy gen... 3
Rapid loss of HBs antigen in p... 3
Reliability of Friedewald form... 3
Frequency, Distribution and Ge... 3
Predictors associated with inc... 3
Pseudohyponatremia and falsely... 3

Log in

Concept, Design & Programming
Dr. Adrian Man

Nr. 19(3)/2011

Evaluation of Diagnostic Tests: Receiver Operating Characteristic Curves

Cristian Baicus, Adriana Hristea, Anda Baicus


Receiver operator characteristic (ROC) curves are used in order to assess the accuracy of diagnostic tests whose results are continuous numeric variables. This curve is a graph of the sensitivity (or true positive rate) on the Y-axis as a function of 1-specificity (the false positive rate) on the Y-axis. ROC curves have multiple utilizations: 1. The comparison of more tests for the same disease (bigger the area under the ROC curve (AUROC), better the test; an AUROC of 1 means a perfect test, while an AUROC of 0.5 means a useless test; 2. The choice of a cut-off point (in case of a test with a big AUROC, one can choose the closest point to the upper left corner of the graph, in order to have both a good sensitivity and a good specificity); 3. It shows how, for the same diagnostic test, there is a negotiation between sensitivity and specificity so that, for a cut-off with a very good sensitivity it will be a weak specificity and the reverse.

  PDF Download full text PDF
(353 KB)
How to cite
Baicus C, Hristea A, Baicus A. Evaluation of Diagnostic Tests: Receiver Operating Characteristic Curves. Rev Romana Med Lab. 2011;19(3):303-6