Screening

Screening

Screening offers a test or assessment to people without recognised symptoms of a target condition to identify who may need further assessment or preventive action. It is a pathway, not merely a test: invitation, uptake, result, confirmation, treatment and follow-up all influence its benefits and harms. This page explains that pathway, the arithmetic of test results, and how to judge whether a programme improves health at an acceptable use of resources.

Screening finds risk or possible disease, not a final diagnosis

The intended population and target condition must be defined before selecting a test. A positive screen generally means that further assessment is indicated; it does not automatically establish disease. A negative screen does not guarantee absence of disease, and guidance for someone with symptoms may follow a different diagnostic pathway.

Stage	Question for a programme	Possible failure
Eligibility and invitation	Who is offered screening, at what age or risk and interval?	People who could benefit may not be reached.
Test and interpretation	Which threshold, specimen and quality controls apply?	False results or indeterminate tests may occur.
Confirmation	Can a positive screen receive timely diagnostic assessment?	More positive screens can overload follow-up capacity.
Care after diagnosis	Is effective treatment or prevention available?	Detection may not translate into better outcomes.
Repeat and review	How are later rounds, harms and programme quality monitored?	Missed disease and cumulative false positives may be obscured.

Screening programmes need more than a detectable early stage. The condition, test, follow-up and intervention must jointly offer a favourable balance of benefits and harms for the population offered screening. The correct comparison is a plausible no-screening or alternative-screening pathway, not simply the count of cases found.

Why prevalence changes a positive result's meaning

Sensitivity is the proportion of people with the target condition who test positive, and specificity is the proportion without it who test negative. In an apparently healthy population where the condition is uncommon, false positives can outnumber true positives even for a reasonably accurate test. Positive predictive value depends on prevalence as well as test performance.

Imagine 1,000 eligible asymptomatic people, 20 of whom have the target condition at the time of screening. A fictional test with 90% sensitivity detects 18 and misses 2; with 95% specificity, it correctly returns a negative result for 931 of the 980 without the condition and gives 49 false positives. Of 67 positive results, $18/67\approx26.9%$ are true positives under these assumptions.

Actual condition status	Screen positive	Screen negative	Total
Condition present	18 true positives	2 false negatives	20
Condition absent	49 false positives	931 true negatives	980
Total	67	933	1,000

This arithmetic is a teaching example, not the performance of a real screening test. Verification of condition status in practice may be incomplete or affected by the diagnostic reference used. Changing the test threshold often trades sensitivity against specificity; a higher detection count alone cannot establish net benefit.

Work out the complete programme cost

A budget should cover invitation and outreach, test administration, confirmatory procedures, treatment, adverse events, quality assurance and repeat rounds where relevant. Count the eligible population, participation and positive results explicitly. A screening test that is inexpensive per person may create a large diagnostic workload or downstream treatment costs.

At an illustrative £20 per screen, testing all 1,000 people costs $1{,}000\times £20=£20{,}000$. If all 67 positive results undergo confirmation at £100 each, confirmation costs $67\times £100=£6{,}700$; test plus confirmation is £26,700, or £26.70 per person screened. These figures exclude outreach, the cost or harm of complications, treatment, future rounds and the value of any health gains.

Spreadsheet item	Illustrative formula	Result
Condition present	=1000*0.02	20.
True positives	=20*0.90	18.
False positives	=980*(1-0.95)	49.
Positive predictive value	=18/(18+49)	About 26.9%.
Screening and confirmation	=1000*20+67*100	£26,700.
Cost per person screened	=26700/1000	£26.70 for these two components.

Cost per true case detected would be another intermediate measure, but it would not value earlier treatment or account for overdiagnosis. An economic evaluation needs patient outcomes and consequences over an appropriate horizon. It should distinguish a budget impact estimate for a specific payer from a cost-effectiveness analysis of health gained and resources used.

Weigh benefit against several kinds of harm

Earlier detection can improve outcomes only when it leads to effective action early enough to matter. Screening can also cause false-positive anxiety and procedures, false reassurance, incidental findings and overdiagnosis of conditions that would not have become clinically important. Overdiagnosis differs from a false positive: the condition may truly be present, but detecting and treating it may not benefit that person.

Survival measured from the date of diagnosis can appear longer simply because screening moved the diagnosis earlier, even if death occurs at the same time; this is lead-time bias. Prefer patient-relevant outcomes such as disease-specific mortality, overall mortality where informative, morbidity and quality of life, together with harms. Compare screened and unscreened groups appropriately rather than treating more early-stage diagnoses as proof of lives saved.

Set up and monitor a fair programme

Population benefit depends on access to the full pathway. Invitation methods, language, transport, test availability, follow-up capacity and treatment access can vary across groups. Programme evaluation should report invitation, uptake, timely confirmation, treatment, benefits and harms by relevant population groups, not only the number of tests performed.

• Specify the target: State the condition, asymptomatic population, interval and action after each result.
• Confirm positives: A screen is an entry point to diagnosis and care, not a final disease label.
• Measure missed disease: Negative results and interval cases matter alongside detections.
• Account for overdiagnosis: A genuine diagnosis can still be clinically unhelpful when disease would never cause harm.
• Include the pathway: Costs and outcomes of confirmation, treatment and repeat rounds belong in the comparison.
• Review delivery: Test quality, follow-up completion, capacity and unequal access can reverse the intended benefit.

Sources and further reading

The WHO screening-programmes guide describes the programme pathway and the need to maximise benefit while minimising harm. The WHO cancer-screening guide distinguishes early detection and precursor treatment, while the US Preventive Services Task Force discussion of breast-screening harms illustrates false positives and downstream harms in a particular condition. The numbers above are original fictional inputs, not screening advice or estimates for a real test.