Screening Principles

Where this shows up in the AKT

• Recalling the Wilson & Jungner criteria for a worthwhile screening programme.
• Spotting screening-specific biases — especially lead-time versus length-time bias.
• Recognising why “improved survival” is not the same as “lower mortality”.

Screening means testing asymptomatic people for a disease. It is deceptively complex: the benefits look obvious, the harms are quiet, and the statistics behind it are exactly where the AKT likes to probe. Two threads matter most — the criteria for when to screen, and the biases that flatter screening results.

Wilson & Jungner criteria (1968)

The classic ten principles for whether a screening programme is worthwhile. You do not need them verbatim — learn the headlines, grouped into four buckets: disease, test, treatment, programme.

About the disease

• The condition should be an important health problem.
• There should be a recognisable latent or early symptomatic stage.
• The natural history of the condition should be adequately understood.

About the test

• There should be a suitable test (sensitive, specific, acceptable).
• The test should be acceptable to the population.

About the treatment

• There should be an accepted treatment for patients with the disease.
• Facilities for diagnosis and treatment should be available.
• There should be an agreed policy on whom to treat.

About the programme

• The cost should be economically balanced against expenditure on care.
• Case-finding should be a continuing process, not a one-off.

The UK National Screening Committee (UK NSC) has since expanded Wilson & Jungner into a fuller set of criteria, grouped under the same four headings, with more rigorous evidence requirements — particularly randomised-trial evidence that screening reduces mortality and avoids overdiagnosis. For the AKT, the original Wilson & Jungner list is what you are most likely to be tested on.

Sensitivity vs specificity — the screening trade-off

A screening test usually has to choose where to set its threshold, and the two errors trade off against each other:

• Push for high sensitivity (catch every case) and you accept more false positives — specificity falls.
• Push for high specificity (avoid false alarms) and you start missing cases — sensitivity falls.

Screening is the first sieve, so it is usually built to be sensitive: better to flag a healthy person for a confirmatory test than to miss real disease. A more specific confirmatory test then follows. That two-step design is deliberate — and it is forced on us by the prevalence problem below.

Why low prevalence wrecks the PPV in screening

This is the bridge back to Chapter 9. Screening targets large, mostly healthy populations, so the prevalence of disease is low — and PPV moves with prevalence:

• When disease is rare, even a tiny false-positive rate produces more false positives than true positives.
• So most positive screening results are false alarms, and the PPV of a single positive is low.
• That is precisely why a positive screen is never the end of the story — it triggers a confirmatory test, and patients should not be alarmed before that result is back.

Screening-specific biases

Lead-time bias

• Screening detects disease earlier. Survival measured from diagnosis looks longer — even if the time of death is unchanged. You started the clock earlier, so it runs for longer, but the patient still dies at the same age.

Length-time bias

• Screening preferentially detects slow-growing, less aggressive disease. Fast, aggressive cancers tend to appear and kill between screening rounds, so the cases caught at screening look more favourable than the underlying disease really is.

Selection bias (the “healthy screenee” effect)

• People who attend screening tend to be healthier overall than those who do not. Outcomes look better partly because of who turns up, not what the test does.

Overdiagnosis

• Detecting disease that would never have caused harm in the patient’s lifetime, leading to overtreatment of indolent conditions — a major issue in PSA screening for prostate cancer, for example.

The traps that cost marks

Lock it in

• Wilson & Jungner = four buckets: disease, test, treatment, programme.
• Screening tests are built to be sensitive (catch cases); a specific confirmatory test follows.
• Low prevalence means low PPV — most positive screens are false alarms, so confirm before you alarm.
• Lead-time = earlier diagnosis, same death; length-time = catching slow-growers; watch for overdiagnosis, and never read “improved survival” as “lower mortality”.