Specificity (tests)
EditorInChief: C. Michael Gibson, M.S., M.D. [1]; Assistant Editor(s)InChief: Kristin Feeney, B.S.
Overview
The specificity is a statistical measure of how well a binary classification test correctly identifies the negative cases. It is the probability that a test correctly classifies individuals without preclinical disease as negative. It is a proportional measurement and is often expressed in terms of percentage.
Calculation
For example, given a medical test that determines if a person has a certain disease, the specificity of the test to the disease is the probability that the test indicates `negative' if the person does not have the disease.
That is, the specificity is the proportion of true negatives of all negative cases in the population. It is a parameter of the test.
High specificity is important when the treatment or diagnosis is harmful to the patient mentally and/or physically.^{[1]}
Worked example
 Relationships among terms
Condition (as determined by "Gold standard")  
True  False  
Test outcome 
Positive  True Positive  False Positive (Type I error, Pvalue) 
→ Positive predictive value 
Negative  False Negative (Type II error) 
True Negative  → Negative predictive value  
↓ Sensitivity 
↓ Specificity 
 A worked example
 the Fecal occult blood (FOB) screen test is used in 203 people to look for bowel cancer:
Patients with bowel cancer (as confirmed on endoscopy)  
True  False  ?  
FOB test 
Positive  TP = 2  FP = 18  = TP / (TP + FP) = 2 / (2 + 18) = 2 / 20 ≡ 10% 
Negative  FN = 1  TN = 182  = TN / (TN + FN) 182 / (1 + 182) = 182 / 183 ≡ 99.5%  
↓ = TP / (TP + FN) = 2 / (2 + 1) = 2 / 3 ≡ 66.67% 
↓ = TN / (FP + TN) = 182 / (18 + 182) = 182 / 200 ≡ 91% 
Related calculations
 False positive rate (α) = FP / (FP + TN) = 18 / (18 + 182) = 9% = 1  specificity
 False negative rate (β) = FN / (TP + FN) = 1 / (2 + 1) = 33% = 1  sensitivity
 Power = 1 − β
Hence with large numbers of false positives and few false negatives, a positive FOB screen test is in itself poor at confirming cancer (PPV=10%) and further investigations must be undertaken, it will though pickup 66.7% of all cancers (the sensitivity). However as a screening test, a negative result is very good at reassuring that a patient does not have cancer (NPV=99.5%) and at this initial screen correctly identifies 91% of those who do not have cancer (the specificity).
Definition
A specificity of 100% means that the test recognizes all healthy people as healthy. The maximum is trivially achieved by a test that claims everybody healthy regardless of the true condition. Therefore, the specificity alone does not tell us how well the test recognizes positive cases. We also need to know the sensitivity of the test to the class, or equivalently, the specificities to the other classes.^{[1]}
A test with a high specificity has a low Type I error rate.
Specificity is sometimes confused with the precision or the positive predictive value, both of which refer to the fraction of returned positives that are true positives. The distinction is critical when the classes are different sizes. A test with very high specificity can have very low precision if there are far more true negatives than true positives, and vice versa.<^{[1]}
SPPIN and SNNOUT
SPPIN  SNNOUT  Neither  Nearperfect  

Proposed definition  Sp > 95%  SN > 95%  Both < 95%  Both > 99% 
Example  Many physical dx findings  Ottawa fracture rules^{[2]}  Exercise treadmill test^{[3]}  HIV1/HIV2 4th gen test^{[4]} 
Predictive values:  
10% pretest prob  PPV= 35%
NPV = 99% 
PPV = 64%
NPV = 98% 
PPV = 31%
NPV = 97% 
PPV = 92%
NPV > 99% 
50% pretest prob  PPV = 94%
NPV = 83% 
PPV = 83%
NPV = 94% 
PPV = 80%
NPV = 80% 
PPV = 99%
NPV = 99% 
90% pretest prob  PPV = 98%
NPV = 64% 
PPV = 99%
NPV = 35% 
PPV = 97%
NPV = 31% 
PPV > 99%
NPV = 92% 
Clinical messages  Accept test result when:

Accept test result when:

Accept test result unless:
 
Notes: Green font indicates when results are more likely to be trustable 
Related Chapters
 binary classification
 receiver operating characteristic
 sensitivity (tests)
 statistical significance
 Type I and type II errors
 Selectivity
Online Calculators
References
 ↑ ^{1.0} ^{1.1} ^{1.2} Altman DG, Bland JM (1994). "Diagnostic tests. 1: Sensitivity and specificity". BMJ. 308 (6943): 1552. PMID 8019315.
 ↑ Stiell, Ian. "The Ottawa Rules". University of Ottawa. Retrieved January 5, 2020.
 ↑ Banerjee A, Newman DR, Van den Bruel A, Heneghan C (2012). "Diagnostic accuracy of exercise stress testing for coronary artery disease: a systematic review and metaanalysis of prospective studies". Int J Clin Pract. 66 (5): 477–92. doi:10.1111/j.17421241.2012.02900.x. PMID 22512607. Note that 80% is a rough estimate of sensitivity and specificity.
 ↑ Malloch L, Kadivar K, Putz J, Levett PN, Tang J, Hatchette TF; et al. (2013). "Comparative evaluation of the BioRad Geenius HIV1/2 Confirmatory Assay and the BioRad Multispot HIV1/2 Rapid Test as an alternative differentiation assay for CLSI M53 algorithmI". J Clin Virol. 58 Suppl 1: e85–91. doi:10.1016/j.jcv.2013.08.008. PMID 24342484.
External links
 Sensitivity and Specificity Medical University of South Carolina