SPSC Screening Test flashcards

32 cards. Read the term and the definition here. The deck at the top of this page is optional.

Sensitivity

Proportion of people who truly have the condition and are correctly flagged by the test: TP / (TP + FN). High sensitivity means few false negatives, so a negative result is reassuring.

Specificity

Proportion of people without the condition who are correctly classified as negative: TN / (TN + FP). High specificity means few false positives, so a positive result is more trustworthy.

Positive predictive value

Probability that a person with a positive result actually has the condition: TP / (TP + FP). It rises with prevalence and falls when the same test is applied to a lower-risk population.

Negative predictive value

Probability that a person with a negative result truly lacks the condition: TN / (TN + FN). It is high when prevalence is low, which is why negative screens reassure more in rare disease.

Likelihood ratio positive

Sensitivity divided by (1 - specificity). It states how much a positive result multiplies the pre-test odds; values above 10 usually shift a diagnosis substantially.

Likelihood ratio negative

(1 - sensitivity) divided by specificity. It states how much a negative result multiplies the pre-test odds; values below 0.1 usually exclude the condition in practice.

Receiver operating characteristic curve

A plot of sensitivity against one minus specificity at every possible cut-off. It shows the trade-off available from a test and the thresholds that maximize either measure.

Area under the curve

Summary of ROC performance ranging from 0.5 (no discrimination, equivalent to chance) to 1.0 (perfect separation). It equals the probability a random case scores above a random non-case.

Cut-off threshold

The fixed value separating screen-positive from screen-negative. Lowering it raises sensitivity and false positives; raising it raises specificity and false negatives.

Prevalence

Proportion of the screened population that has the condition at the time of testing. It drives predictive values but does not change sensitivity or specificity of the assay.

Number needed to screen

The reciprocal of the absolute risk reduction attributable to screening: how many people must be screened to prevent one adverse outcome. It grows large when events are rare.

Cohen's kappa

Agreement beyond chance between two raters or two tests: (observed agreement minus chance agreement) divided by (1 minus chance agreement). Values near zero indicate chance-level agreement.

Reliability versus validity

Reliability is consistency of repeated measurements; validity is agreement with the true condition or reference standard. A test can repeat the same wrong answer reliably.

Reference standard

The best available method for establishing whether the condition is truly present, such as biopsy or definitive imaging. Screening results are judged against it.

Lead-time bias

Earlier diagnosis lengthens the interval from diagnosis to death without postponing death. Survival measured from diagnosis therefore looks better even when screening saves no lives.

Length bias

Slow-growing disease remains detectable for longer, so screening disproportionately finds indolent cases. This inflates apparent survival benefit relative to aggressive disease.

Selection bias in screening

People who volunteer for screening often differ in health and risk from those who do not. Comparing volunteers with non-volunteers overstates benefit unless groups are randomized.

Verification bias

Occurs when patients with positive screens receive more thorough confirmatory testing than those with negative screens, so missed disease among negatives goes uncounted.

Spectrum bias

Test accuracy measured in severely ill, referred patients overstates performance in the mild, early, asymptomatic population for which screening is actually intended.

Overdiagnosis

Detection of a condition that would never have caused symptoms or death during the person's lifetime. It produces treatment without benefit and is a core harm of screening.

False positive result

A screen-positive result in a person who does not have the condition. It triggers anxiety, further testing, cost, and occasionally procedural complications.

False negative result

A screen-negative result in a person who does have the condition. It delays diagnosis and is the main harm when sensitivity is low.

Intention-to-screen analysis

Analyzing participants by randomized allocation regardless of whether they attended screening. It preserves randomization and avoids healthy-adherer bias.

Randomized screening trial

The reference design: participants are randomly assigned to invitation or usual care, and condition-specific mortality is compared. It controls selection and lead-time bias.

Wilson-Jungner criteria

Classic conditions for a screening program: important health problem, accepted treatment, detectable latent stage, suitable test, agreed treatment policy, and cost-effectiveness.

Absolute risk reduction

Difference in event rates between screened and unscreened groups. It conveys real benefit size, unlike relative reduction, which can appear large on a tiny baseline.

Cost-effectiveness threshold

The maximum acceptable cost per quality-adjusted life-year a health system will pay. Screening programs exceeding it are usually not adopted despite clinical benefit.

Informed consent for screening

Disclosure that a screen may yield false positives, false negatives, and overdiagnosis, plus the follow-up pathway, before the person agrees to be tested.

Confirmatory diagnostic test

The test performed after a positive screen to establish diagnosis. It is usually more invasive or costly, so a low positive predictive value multiplies work-up burden.

Blinding in test accuracy

Readers of the screening test and of the reference standard must be unaware of each other's results. Blinding prevents review bias from inflating measured accuracy.

Pre-test probability

Estimated likelihood of disease before the screen, from prevalence or clinical features. Combined with likelihood ratios it yields the post-test probability.

Screening versus diagnostic testing

Screening tests asymptomatic people to estimate risk in a population; diagnostic tests confirm or exclude disease in individuals. They face different accuracy demands.