Diagnostic and screening tests - negative predictive value

Definition

Negative predictive value is the proportion of people with a negative test result who truly do not have the disease - like positive predictive value, NPV is directly dependent on disease prevalence in the tested population, meaning the same test's NPV varies depending on the pre-test probability of disease
The clinically most directly relevant statistic to a patient asking "given my negative result, how reassured should I be that I don't have this disease" - complementing positive predictive value as the other prevalence-dependent predictive measure

Because NPV depends on prevalence, the same test applied to a low-prevalence population produces a higher NPV (a negative result is more reassuring when disease is already unlikely), while applied to a high-prevalence population, NPV falls (a negative result is less reassuring when disease is common, since even a low false-negative rate generates a meaningful number of missed cases in a population with many true positives) - this inverse relationship to PPV's prevalence pattern is a frequently tested point

Calculated as true negatives divided by (true negatives plus false negatives) in the specific tested population - this calculation is sample/population-specific, unlike sensitivity/specificity
Estimating NPV for an individual patient requires accounting for their specific pre-test probability, not just general population prevalence

A negative result in a high pre-test-probability patient (strong clinical suspicion) should be interpreted more cautiously than the same negative result in a low pre-test-probability patient, given NPV's prevalence-dependence
Understanding NPV's prevalence-dependence is essential to correctly counselling a patient about how reassuring their specific negative result actually is
Prevalence-dependence of NPV (opposite direction to PPV's pattern), high pre-test probability and reduced reassurance value of a negative result, NPV vs sensitivity distinction (a frequently confused pair)

A negative test result is less reassuring in a patient with high clinical suspicion (high pre-test probability) than in a patient with low suspicion, even using the identical test - this is because NPV falls as disease prevalence/pre-test probability rises, meaning a negative result in a strongly clinically suspicious patient should prompt further investigation rather than being treated as definitively excluding the diagnosis

4 more sections, plus exam facts

Premium unlocks every note across every specialty, and the full exam fact library behind it.