Who gets tested changes the result
If testing shifts from broad screening to people with symptoms, positivity can rise even without an equivalent rise in population prevalence. Expanded testing among lower-risk people can move the measure in the other direction.
Access, test availability, and reporting also affect who appears.
Read the unit carefully
| Denominator | Interpretation |
|---|---|
| Tests | Repeated testing of one person may contribute multiple times |
| Unique people | Requires a rule for identifying and counting people |
| Testing episodes | Depends on how an episode is defined |
The numerator must use the same unit and period.
A hypothetical change
If 20 of 100 included tests are positive, positivity is 20%. If the next period adds many negative screening tests, positivity may fall even if the number of positive results stays at 20.
That is a denominator change, not necessarily evidence of a proportional decline in infections.
Use complementary indicators
Testing volume, eligibility, healthcare visits, hospitalizations, and other surveillance sources can help interpret the pattern. Each has its own delays and coverage.
Positivity can be useful when its collection process is understood. The mistake is turning a measure of the tested subset into a population estimate without a design or model that supports that inference.
