Describe screening for disease.

Screening for Disease

• The fundamental point, which must be stated early: a screening test is not diagnostic. It sorts a population into those who probably have the condition and those who probably do not, and a positive result must always be followed by a diagnostic test
• The types — mass screening of a whole population; high-risk or selective screening of a defined risk group, which is usually more efficient; multiphasic screening, applying several tests at once; and opportunistic (case-finding), testing patients who present for some other reason
• The WILSON and jungner criteria (WHO, 1968) are examined verbatim and group naturally into criteria about the disease, the test and the treatment:
• About the disease — it should be an important health problem; have a recognisable latent or early symptomatic stage; and its natural history, including the progression from latent to declared disease, should be adequately understood
• About the test — there should be a suitable test or examination, which is acceptable to the population, and safe, simple, valid (sensitive and specific), reliable and cheap
• About the treatment — there must be an accepted treatment for patients with recognised disease; facilities for diagnosis and treatment must be available; there should be an agreed policy on whom to treat as patients; the cost should be economically balanced against expenditure on medical care as a whole; and case-finding should be a continuing process, not a once-and-for-all project
• The most important criterion IN practice IS that early treatment must improve the outcome. If treating early makes no difference to the eventual result, screening merely lengthens the time the patient knows about the disease without lengthening life, and is a harm rather than a benefit
• The biases that make screening appear effective when it is not are the subtlest and most examinable part of the topic:
• Lead-time bias — screening advances the moment of diagnosis without postponing death, so survival from diagnosis appears longer while the date of death is unchanged. The patient has gained more years of knowing, not more years of living
• Length-time bias — screening preferentially detects slow-growing, less aggressive disease, simply because such cases spend longer in the detectable preclinical phase; rapidly fatal cases arise and declare themselves between screening rounds (interval cancers). Screened cases therefore appear to do better because they were a milder selection to begin with
• Overdiagnosis — the detection of disease that would never have caused symptoms or death in the person’s lifetime. It is the most serious harm of screening, because every overdiagnosed person becomes a patient, is treated, and can only be harmed — and it inflates survival statistics spectacularly
• Selection (volunteer, healthy screenee) bias — those who attend screening are healthier, wealthier and more health-conscious than those who do not
• The consequence for evaluation — survival rates and stage distribution cannot be used to evaluate a screening programme, because all four biases inflate them. Only a randomised trial with disease-specific mortality in the whole invited population as the outcome can establish benefit
• The harms of screening must be stated — false positives, causing anxiety and unnecessary invasive investigation; false negatives, giving false reassurance and delaying presentation; overdiagnosis and overtreatment; complications of the test itself; the opportunity cost; and the medicalisation of healthy people
• Screening in INDIA — NCD screening for hypertension, diabetes and the common cancers under the national programme at Health and Wellness Centres; visual inspection with acetic acid (VIA) for cervical cancer, chosen because it is cheap, needs no laboratory and gives an immediate result, allowing screen-and-treat in a single visit; clinical breast examination rather than mammography, on grounds of cost and feasibility; oral visual examination; antenatal screening; and neonatal screening
• Survival rates cannot evaluate screening; only disease-specific mortality in a randomised trial can
• Lead-time gives more years of knowing, not more years of living, which is why survival lengthens with no benefit
• Length-time bias selects the indolent cases, since aggressive disease declares itself between screening rounds
• Overdiagnosis is the most serious harm, because those detected can only be harmed by the treatment that follows
• A screening test is not diagnostic, and every positive must be followed by a confirmatory test
• Early treatment must improve the outcome, or screening merely lengthens the period of knowing
• Screening must be a continuing process, not a one-off campaign, which is an explicit Wilson and Jungner criterion
• Facilities for diagnosis and treatment must exist before screening begins, or detected cases cannot be helped
• VIA allows screen-and-treat in one visit, which is why it suits Indian settings better than cytology
• Clinical breast examination is used in India rather than mammography, on grounds of cost and feasibility
• Volunteers for screening are healthier than non-attenders, which biases any uncontrolled comparison
• False negatives give false reassurance and may delay presentation, which is a real harm rather than merely a missed benefit
• High-risk screening is usually more efficient than mass screening, for reasons that follow from predictive value