Appendix A
Statistical Terminology
Sensitivity represents subjects correctly identified by the clinical examination as having the condition of interest as a percentage of all those who truly have the condition of interest (true positive rate). When a test has a very high sensitivity, a negative clinical finding effectively rules out the diagnosis 4.
Specificity defines those who were correctly identified by the clinical examination as not having the condition of interest as a percentage of those who truly did not have the condition of interest (true negative rate). Tests with high specificity are useful in ruling in the diagnosis when the clinical finding is positive 4.
Positive Likelihood ratios (PLR) and Negative likelihood ratios (NLR) are the ultimate indexes of validity which offer the clinician a better method of using the information obtained from a clinical test. They are used to assess the value of each diagnostic test in terms of changing the pre-test probability. The likelihood ratios provided by methodologically sound studies, can enhance clinical decision making at an individual patient level. Likelihood ratios are interpreted as: LRs >10 or < 0.1 generate large, and often conclusive changes from pre- to post-test probability; LRs of 5-10 and 0.2-0.1 generate moderate shifts in pre- to post-test probability; LRs of 2-5 and 0.5-0.2 generate small changes in probability; and LRs of 1-2 and 0.5-1 do not result in important change 2,3,3.
Percentage of agreement between surgical and MRI findings is the number of cases agreed on divided by total number of cases.
The area under a ROC (Receiver Operating Characteristic) curve quantifies the overall accuracy or overall ability of the test to discriminate between those individuals with the disease and those without the disease. ROC curves graphically display sensitivity (true positive) versus 1- specificity (false-positives) across a range of cut-offs. A perfect test has a zero false positives and zero false negatives with a ROC area of 1.00 1. Predictive values of clinical tests are population dependent and vary with disease prevalence and therefore were not included.
References:
(1) Florkowski CM. Sensitivity, specificity, receiver-operating characteristic (ROC) curves and likelihood ratios: communicating the performance of diagnostic tests. Clin.Biochem.Rev. 2008 Aug;29 Suppl 1:S83-7.
(2) Jaeschke R, Guyatt GH, Sackett DL. Users' guides to the medical literature. III. How to use an article about a diagnostic test. B. What are the results and will they help me in caring for my patients? The Evidence-Based Medicine Working Group. JAMA 1994 Mar 2;271(9):703-707.
(3) Radack KL, Rouan G, Hedges J. The likelihood ratio. An improved measure for reporting and evaluating diagnostic test results. Arch.Pathol.Lab.Med. 1986 Aug;110(8):689-693.
(4) Sackett DL, Haynes RB. The architecture of diagnostic research. BMJ 2002 Mar 2;324(7336):539-541.