Medical tests such as, physical and visual exams, diagnostic imaging, genetic testing, chemical and cellular analysis, relating to clinical chemistry and molecular diagnostics, are typically performed in a medical setting.

Medical tests can be classified by their purposes, including diagnosis, screening or monitoring.

[4][5] Screenings may be performed to monitor disease prevalence, manage epidemiology, aid in prevention, or strictly for statistical purposes.

[6] Examples of screenings include measuring the level of TSH in the blood of a newborn infant as part of newborn screening for congenital hypothyroidism,[7] checking for Lung cancer in non-smoking individuals who are exposed to second-hand smoke in an unregulated working environment, and Pap smear screening for prevention or early detection of cervical cancer.

[citation needed] Especially in the taking of a medical history, there is no clear limit between a detecting or quantifying test versus rather descriptive information of an individual.

[citation needed] The classification of tests into either positive or negative gives a binary classification, with resultant ability to perform bayesian probability and performance metrics of tests, including calculations of sensitivity and specificity.

Most diagnostic tests basically use a reference group to establish performance data such as predictive values, likelihood ratios and relative risks, which are then used to interpret the post-test probability for an individual.

Some medical testing procedures have associated health risks, and even require general anesthesia, such as the mediastinoscopy.

For example, a basic cholesterol test may be indicated (medically appropriate) for a middle-aged person.

Patients may give informed consent to undergo medical tests that will benefit other people.

In addition to considerations of the nature of medical testing noted above, other realities can lead to misconceptions and unjustified expectations among patients.

These include: Different labs have different normal reference ranges; slightly different values will result from repeating a test; "normal" is defined by a spectrum along a bell curve resulting from the testing of a population, not by "rational, science-based, physiological principles"; sometimes tests are used in the hope of turning something up to give the doctor a clue as to the nature of a given condition; and imaging tests are subject to fallible human interpretation and can show "incidentalomas", most of which "are benign, will never cause symptoms, and do not require further evaluation," although clinicians are developing guidelines for deciding when to pursue diagnoses of incidentalomas.