Test Performance in Three Hypothetical Clinical Scenarios*
Panel a: Scenario 1 – Low Pretest Probability (LFA) | ||||
---|---|---|---|---|
Dx | No Dx | PPV | NPV | |
Prevalence 0.1%, Sn 90, Sp 95 | ||||
(+) | 9 | 500 | 1.8% | 99.9% |
(-) | 1 | 9490 | ||
Prevalence 0.2%, Sn 90, Sp 95 | ||||
(+) | 18 | 499 | 3.5% | 100.0% |
(-) | 2 | 9481 |
Panel b: Scenario 2 – Intermediate Pretest Probability (RT-PCR Initially Negative) | ||||
---|---|---|---|---|
Dx | No Dx | PPV | NPV | |
Prevalence 40%, Sn 85%, Sp 95% | ||||
(+) | 34 | 3 | 91.9% | 90.5% |
(-) | 6 | 57 | ||
Prevalence 60%, Sn 85%, Sp 95% | ||||
(+) | 51 | 2 | 96.2% | 80.9% |
(-) | 9 | 38 |
Panel c: Scenario 3 – High Pretest Probability (RT-PCR) | ||||
---|---|---|---|---|
Dx | No Dx | PPV | NPV | |
Prevalence 80%, Sn 90%, Sp 95% | ||||
(+) | 72 | 1 | 98.6% | 70.4% |
(-) | 8 | 19 | ||
Prevalence 80%, Sn 80%, Sp 95% | ||||
(+) | 64 | 1 | 98.4% | 54.3% |
(-) | 16 | 19 |
LFA, lateral flow assay; RT-PCR, real-time polymerase chain reaction.
*These tables describe the distribution of 10,000 (panel a) or 100 patients (panels b and c) with and without the disease (COVID-19) compared to with positive and negative test results.
†Sensitivity (Sn), the proportion of infected individuals who test positive.
‡specificity (Sp), the proportion of uninfected individuals who test negative.
§Dx, patient has disease; no Dx, patient does not have disease.
||(+), test result is positive; (-), test result is negative.
¶Positive predictive value (PPV), the probability that a positive test result actually means one has COVID-19.
¶¶Negative predictive value (NPV), the probability that a negative test result actually means one does not have COVID-19.