Health Screening

SCREENING

Screening is performed to identify the presence of a disease or a risk factor for a disease, typically among asymptomatic persons (those who do not already manifest symptoms of disease). In this way, a disease, or risk factors for a disease, can be detected early, allowing either early treatment or prevention, including preventing further spread of communicable or transmissible diseases. Screening tests are widely used by clinicians as part of the periodic health examination, as well as by public health officials. Examples of screening tests are as varied as blood tests to detect lead poisoning in young children, blood tests to detect the human immunodeficiency virus (HIV), measuring blood pressure to detect high blood pressure, mammography to detect breast cancer, sigmoidoscopy and colonoscopy to detect cancers of the rectum and colon, and questionnaires to identify persons with alcohol or other drug problems.

Table 1

Several factors determine the usefulness of a screening test for use with any individual person. The first is the accuracy of the test itself, specifically its sensitivity and specificity. Sensitivity is the probability that a person with the disease or risk factor will test positive. Specificity is the probability that a person without the disease or risk factor will test negative. Sensitivity and specificity are illustrated in Table 1. The sensitivity of a screening test is determined by the number of true positives divided by the total number with disease (or TP/[TP+FN]). The specificity is the number of true negatives divided by the total number with no disease (or TN/[FP+TN]).

Because there is often some overlap in the distributions of test results among people with and without disease (i.e., some people without disease will have test results in the disease range, and some people with disease will have test results in the no disease range), a test's sensitivity and specificity usually trade-off against one another. As the sensitivity increases the specificity usually decreases, and vice versa. A screening test that identifies almost all people with a disease (high sensitivity) may also produce more false positives among those people without the disease who may have borderline results (results near the cut-off value defined for the test). Conversely, a screening test that correctly identifies almost all people without the disease (high specificity) usually misses more people who truly have the disease (false negatives). Tables 2 and 3 represent the characteristics of two hypothetical screening tests when applied to a sample of 100,000 people with a true prevalence of disease of 10 percent (e.g., a relatively common disease). Table 2 is for a test with a sensitivity of 95

Table 2

percent and a specificity of 65 percent. Table 3 is for a test with 65 percent sensitivity and 95 percent specificity. The test with high sensitivity (Table 2) identifies more people who truly have the disease and misses fewer people who truly have the disease (false negatives). However, this test incorrectly classifies more than three people without the disease (false positives) for every one person it correctly identifies with the disease. In contrast, the test with high specificity (Table 3) incorrectly classifies many fewer nondiseased people as having the disease (false positives) but misses more truly diseased people (false negatives).

In addition to the accuracy of the test itself, another important factor is how well the test is implemented. Errors may be introduced that depend on who is performing the test or on variations in the way the test is performed. For example, not all radiologists are equally proficient at reading mammograms and not all laboratories will get the same result when measuring cholesterol levels from the same blood sample. Therefore, the test characteristics that are initially reported for a screening test often represent a best case scenario—the best that a test can be expected to perform. As a result, it is also important to evaluate test accuracy in the real world settings where the tests are being used.

The usefulness of a screening test also depends upon the probability that the individual being tested has the disease or risk factor of interest. This is termed the "prior probability" of disease. This issue is illustrated by comparing Table 2 and Table 4. Table 4 represents the same hypothetical test shown in Table 2, but applied to a

Table 3

sample in which the true prevalence of disease is less common, only 1 percent instead of 10 percent. As shown, screening tests are more useful when they are used on people who are more likely to have the disease than people who are less likely to have the disease. When a screening test is used in a sample with a lower prior probability of disease, even more false positives are identified. In this example (Table 4), the test has incorrectly identified more than 36 people who do not really have the disease (false positives) for every one person correctly identified with the disease (true positives).

Prior probability is taken into account in calculating the predictive value of a test. The predictive value of a positive test is the probability that someone who tests positive truly has the disease. For the examples shown in Tables 2, 3, and 4, the predictive values of a positive test are 23 percent, 59 percent, and 3 percent, respectively. As can be seen, the predictive value of a positive test is increased when tests with higher specificity are used in samples of people with a higher prevalence of the disease.

Finally, the usefulness of a screening test depends on the existence of an effective and feasible treatment. This may include treatment for the disease or risk factor detected, and/or an intervention to prevent further spread of the problem to others, such as removing lead paint from homes or genetic counseling. If there are no feasible and effective responses to the results of a screening test (e.g., the result wouldn't change anything) then there is no reason to perform the test.

These issues are of particular concern for screening asymptomatic or healthy people. All

Table 4

testing involves risks. These risks might be acceptable to the small number of persons who turn out to have the disease. However, the risks of side-effects from the screening tests themselves, or from an incorrect or ambiguous diagnosis and the subsequent testing that an incorrect initial test result requires, may not be acceptable to the much larger number of people who do not have the disease or risk factor of interest. In addition, there are economic costs to screening large numbers of asymptomatic people to identify a small number of people with disease. Therefore, clinicians, patients, and public health professionals must weigh the risks and benefits when deciding to use a screening test for any individual or population.

THOMAS N. ROBINSON

(SEE ALSO: Assessment of Health Status; Blood Lead; Blood Lipids; Breast Cancer; Cancer; Cholesterol Test; Colorectal Cancer; Diabetes; HIV/AIDS; Mammography; PAP Smear; Periodic Health Examination; Prevention; Preventive Medicine; Serological Markers; VDRL Test)

Content licensed from:

Author Info: THOMAS N. ROBINSON, The Gale Group Inc., Macmillan Reference USA, New York, Gale Encyclopedia of Public Health, 2002

This feature is for informational purposes only and should not be used to replace the care and information received from your healthcare provider. Please consult a healthcare professional with any health concerns you may have.