Testing diagnostic tests: Sensitivity and specificity
Posted on 25th June 2015 by Sofía Jaramillo
A diagnostic test is a useful tool which lets us feel confident about the diagnosis of a patient’s illness. However, in medical practice we are excessively confident about what our mentors say are the best ways to reach an accurate diagnosis. It is always better to read a little more and find for ourselves the best techniques or examinations which let us be sure in our decision-making pathway.
When we are reading about a diagnostic test we are going to find terms which define their value, such as sensitivity and specificity. In this post I am going to define them in simple words to make them clear and easy to interpret, so after you read this you can put them into practice.
In another post we will discuss predictive values, which complement sensitivity and specificity in the evaluation of a test.
- Sensitivity is a measure of how good a test is in demonstrating whether the patient really has a condition or not (the true positive rate)
- A high sensitivity test means that it detects the presence of a condition with relatively few indicators (depends on the key point evaluated in the test)
- It is expressed in percentage (%)
- Specificity is a measure of negativity for those patients who do not have the investigated condition (the true negative rate)
- A highly specific test means that it really rules out a diagnosis if a patient does not have the indicators
- It is expressed in percentage (%)
Let’s put this into practice
Using Sensitivity and Specificity to compare the dementia screening tests MMSE and Mini-Cog.
According to the results given in the study performed by Borson et al. the Mini-Cog Test is more useful than MMSE in the dementia screening process. Mini-Cog is able to detect dementia with few characteristics of it – memory impairment and visual-motor abnormalities (sensitivity) – and is also specific enough to detect dementia alone (specificity).
The advantages of using sensitivity and specificity in the evaluation of diagnostic tests
Statistically, the advantages of using sensitivity and specificity are:
- They do not alter if the prevalence changes between populations (see predictive values)
- They can be applied in different populations
- They can be compared between different studies with a wide range of inclusion criteria
- They can be used to determine the diagnostic potential of a test
Making a clinical decision based on the most appropriate diagnostic test is very important. Further factors such as their wider pros and cons and their relative affordability should also be taken into account.
Further Reading on S4BE
Borson S, Scanlan J, Brush M, Vitaliano P, Dokmak A. The Mini-Cog: A Cognitive ‘Vital Signs’ Measure for Dementia Screening in Multi-lingual Elderly. Int. J. Geriatr. Psychiatry 15, 1021 – 1027 (2000)
Dorairajan L, Manikandan R. How to appraise a diagnostic test. Indian J Urol [Internet]. Medknow; 2011;27(4):513. Available from: http://dx.doi.org/10.4103/0970-1591.91444
Peat J, Barton B. Medical Statistics. Blackwell Publishing Inc.; 2005 Jan 1; Available from: http://dx.doi.org/10.1002/9780470755945
Thompson M, Van den Bruel A. Diagnostic Tests Toolkit. Wiley-Blackwell; 2011 Oct 7; Available from: http://dx.doi.org/10.1002/9781119951827
No Comments on Testing diagnostic tests: Sensitivity and specificity
There’s a mnemonic commonly used:
SNout: If a test with a high SeNsibility is negative, it rules OUT the disease.
SPin: If a test with a high SPecificity is positive, it rules IN the disease.
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1252824/30th June 2015 at 2:39 pm
Thanks Giordano!30th June 2015 at 2:43 pm