Statistics Definitions > Gold Standard Test
What is a Gold standard Test?
A gold standard test is a best available diagnostic test for determining whether a patient does or does not have a disease or condition. For example, a biopsy can identify breast cancer cells with good accuracy, while an autopsy are usually accurate at identifying the cause of death. It is usually used when an initial screening gives a positive result. People who test negative are not usually given the gold standard tests, because they are often expensive, invasive, or risky.
Gold standard tests mean that diseases and conditions can be correctly classified.
|Disease/Condition||Initial Test||Gold Standard||Breast Cancer||Mammogram||Tissue Biopsy|
|Rabies (in animals)||?||DFA test|
|Tuberculosis||Tuberculin skin or blood test||Lowenstein-Jensen culture|
|Colorectal cancer||Blood in stool sample||Colonoscopy|
|Diabetes||Fasting plasma glucose||Oral GTT|
Assessing new tests
New diagnostic tests are compared against the gold standard. Sensitivity and specificity for these new tests are usually estimated from comparing them to the gold standard. This system is not perfect: no test (even if the best available) is perfect and has some bias (error) attached to it. For example, biopsies vary wildly in accuracy depending on the specific cancer type.
What this means is you’re essentially comparing any new test to a standard that has some error attached it it — which means your new test will also have error. This can be avoided by using a third, “resolver” test, followed by re-testing some people whose results are the same (either positive or negative) on both the new test and the resolver test.
BreastCancer.org. “Study Looks at Accuracy of Breast Biopsy Results”. Retrieved March 20, 2017.
Hakwins, D. et. al. “Some Issues in Resolution using an Imperfect GS.” Statistics in Medicine, 20, (2001), 1987—2001. Retrieved March 20, 2017 from: http://www.stat.umn.edu/hawkins/Resolving_V7_fix.pdf
Nelson, J. “The Autopsy as a Measure of Accuracy of the Death Certificate”. Engl J Med 1985; 313:1263-1269November 14, 1985DOI: 10.1056/NEJM198511143132005. Retrieved MArch 20 from: http://www.nejm.org/doi/full/10.1056/NEJM198511143132005
If you prefer an online interactive environment to learn R and statistics, this free R Tutorial by Datacamp is a great way to get started. If you're are somewhat comfortable with R and are interested in going deeper into Statistics, try this Statistics with R track.Comments are now closed for this post. Need help or want to post a correction? Please post a comment on our Facebook page and I'll do my best to help!