Screening Test

A screening test is a procedure applied to an asymptomatic population to identify individuals who may have a disease, condition, or risk factor that warrants further diagnostic evaluation. Screening tests are designed to be simple, rapid, and acceptable to the population, and they are evaluated by their sensitivity, specificity, and predictive values in the target population.

Screening programs aim to detect disease at an early stage when treatment is more effective, reducing morbidity and mortality. Population level screening for conditions such as breast cancer (mammography), cervical cancer (cytology or HPV testing), and newborn metabolic disorders saves lives when evidence supports that early detection improves outcomes. However, screening also carries risks: false positives cause anxiety and unnecessary procedures, false negatives provide false reassurance, and overdiagnosis can lead to treatment of conditions that would never have caused harm. Evaluating whether a screening program does more good than harm requires careful analysis of the test's performance characteristics in the target population.

A screening test identifies individuals who are probably affected (screen positive) for further diagnostic testing, and those who are probably unaffected (screen negative) who require no further action. The test does not provide a definitive diagnosis; it stratifies the population into higher and lower risk groups.

The performance of a screening test is characterized by four metrics. Sensitivity is the proportion of truly affected individuals who test positive. Specificity is the proportion of truly unaffected individuals who test negative. Positive predictive value (PPV) is the proportion of screen positive individuals who actually have the condition. Negative predictive value (NPV) is the proportion of screen negative individuals who are truly unaffected.

The choice of threshold for a positive result involves a tradeoff between sensitivity and specificity. Lowering the threshold increases sensitivity (catching more true cases) but decreases specificity (increasing false positives). The optimal threshold depends on the disease, the consequences of missing a case versus false alarm, and the availability of follow up diagnostic procedures.

In a 1968 report for the World Health Organization, Wilson and Jungner set out ten classic principles for screening, including whether the condition is an important health problem, whether there is an accepted treatment for patients with recognized disease, whether facilities for diagnosis and treatment are available, whether there is a recognizable latent or early symptomatic stage, and whether the cost of case finding (including diagnosis and treatment) is economically balanced in relation to possible expenditure on medical care as a whole. Later frameworks build on these principles; the UK National Screening Committee, for example, appraises programs against criteria for the condition, the test, the intervention, the screening program and implementation, including that the benefit to individuals should outweigh harms such as overdiagnosis, false positives and false reassurance.

Implement a screening program when the target condition is prevalent and serious enough to warrant detection, when there is an effective test that is acceptable to the population, when early treatment confers better outcomes than treatment after symptoms develop, and when the benefits of screening outweigh the harms (including overdiagnosis, false positives, and cost). Do not screen for conditions where early detection does not improve outcomes, where effective treatment is unavailable, or where the test lacks adequate sensitivity and specificity for the target population.

A common error is assuming that detecting a disease earlier automatically improves outcomes, without evidence from randomized trials or rigorous observational studies. Earlier diagnosis lengthens survival measured from diagnosis even when death is not delayed (lead time bias), so survival from diagnosis is a misleading measure of benefit. Another mistake is evaluating a screening test in a high prevalence clinical population and applying its predictive values to a low prevalence general population, overestimating the PPV. Researchers sometimes neglect to consider the harms of screening, including the psychological burden of false positive results and the risks of follow up procedures. Finally, failing to consider the population's baseline risk leads to indiscriminate screening that is not cost effective.

A national screening program for colorectal cancer using fecal immunochemical testing (FIT) targets adults aged 50 to 75. The test has a sensitivity of 79% and specificity of 94% for detecting colorectal cancer. In the screened population, where the prevalence of undiagnosed colorectal cancer is approximately 0.3%, the PPV is about 4%, meaning that approximately 1 in 26 positive FIT results leads to a cancer diagnosis at colonoscopy. Despite this low PPV, the program is considered cost effective because FIT is inexpensive and noninvasive, and randomized trials of an older stool test, the guaiac fecal occult blood test, found that screening reduced colorectal cancer mortality by an estimated 16%.