Print Page   |   Report Abuse   |   Sign In   |   Become a Member
CSME|CAPDA Medico-Legal Summit - Dr. Jonathan Siegel
Share |

Medical and Psychological Testing: Establishing Cut-Off Scores for Improving Diagnostic Accuracy
Dr. Jonathan Siegel, Ed.D., C.Psych

The Problem and the Solution: Introduction
The problem is that there is no test that is 100% accurate in identifying who has a particular medical disease or psychological condition and who does not. When either a medical or psychological practitioner uses a diagnostic test, there is inevitably some probability of error. The solution is to find the test cut-off score that optimizes decision making in spite of this chance of error. A cut-off score is the test value that determines whether a person scores positive or negative on that test, and every cut-off score is associated with sensitivity and specificity values for that test. Sensitivity and specificity values are ratios that yield percentages associated with the diagnostic accuracy of a test for a particular population. Sensitivity is also referred to as the true positive rate; specificity is also referred to as the true negative rate.

The talk employed both a medical and psychological example to illustrate the relationship between cut-off scores and the diagnostic accuracy of tests.

Two Types of Errors: False Positives and False Negatives
Given that tests do not have 100% accuracy, there are two types of errors that can occur: false positive errors and false negative errors. A false positive error occurs when the test identifies a person as having the disease when the person does not really have the disease; a false negative error occurs when the test identifies a person as not having the disease, when in fact the person does have the disease. If the cut-off score is set to reduce false positive errors, what inevitably happens is that the probability of false negative errors increases. And the converse also applies: if a cut-off score is set to reduce false negative errors, what inevitably happens is that the probability of making false positive errors increases. There is always a trade-off. For any particular cut-off score, the higher the false positive rate, the lower the false negative rate; the higher the false negative rate, the lower the false positive rate. The following example will illustrate. If one asserts (based on diagnostic test results) that an individual is exaggerating symptoms when s/he really is not, that type of error would be a false positive. If you want to reduce the chances of making this type of error, the trade-off is that you are more likely to assert that the person is not exaggerating when s/he really is. This is called a false negative. Both errors have a potential significant downside. The upside is that there are statistical methods for calculating test sensitivity and specificity values to best optimize decision making.

Sensitivity and Specificity
Sensitivity is defined by the following question: of all the individuals in the research sample who actually have a disease, how many score positive on the test? And the specificity of a test is defined by asking: of all the individuals in the research sample who actually do not have a disease, how many score negative on the test?

Summary and Conclusion

There are standard statistical methods for evaluating the effectiveness of both medical and psychological tests. The effectiveness of a test, which is a test’s diagnostic accuracy, is
determined by its sensitivity and specificity values. When combined with prevalence rates, a test’s positive and negative predictive values can be calculated. When reviewing diagnostic test results, it is important to note the range of cut-off scores associated with each corresponding sensitivity and specificity value, and then identify the cut-off score that optimizes the diagnostic classification.

Copyright © 2016 Canadian Society of Medical Evaluators

Contact Us
301-250 Consumers Rd, Toronto, ON  M2J 4V6
416 487 4040 | (1) + 888 672 9999 Fax:   416 495 8723
Email:    W

Privacy Statement