PrevIQ Docs
Open app

Key concepts

Five ideas that make every number in PrevIQ make sense.

PrevIQ is easy to use once you internalize five ideas. Everything else in this documentation builds on them.

1. The atomic unit

Every prevalence number answers one precise question:

Of all specimens tested for this organism within this panel, how many were detected?

The formula

prevalence_rate = detected / tested      (within one syndrome × panel × organism)

That tuple — (syndrome, panel, organism, tested, detected) — is the smallest unit of data in PrevIQ. Each one is also tagged with the lab (vendor) it came from and the date range it covers.

2. Prevalence is computed within a panel, never across

Important

The same organism can appear in multiple panels. PrevIQ never merges those into one "global" rate, because they mean different things clinically. Streptococcus in a respiratory panel is a different signal than Streptococcus in a vaginosis panel.

This is the single most important rule in the platform. When you see a number, it is always scoped to a specific syndrome and panel context. See Prevalence methodology.

3. The denominator is all tested specimens

PrevIQ's prevalence rate uses every specimen tested for an organism as the denominator — not only the specimens that came back positive on the panel.

Note

Some historical spreadsheets used a "positive-panel-only" denominator, which produces higher apparent rates. PrevIQ's numbers are intentionally more conservative and more defensible. Every report carries a methodology footnote stating this.

4. Not every number is reportable — tiers tell you which

Each rate is classified by how many specimens it's based on:

Tier Sample size (n) Use it for
A n ≥ 100 Clinical & reimbursement decisions
B 30 ≤ n < 100 Decisions, with a note on CI width
C n < 30 Exploration only — not clinically reportable

Tier C rows are always shown muted so the defensible signal stays front and center. See Sample-size tiers.

5. Panel design boils down to three buckets

For panel decisions, PrevIQ sorts each organism into one of three classifications:

  • INCLUDE — Rate ≥ 5% and a reportable sample (Tier A/B). Clinically actionable.
  • WATCH — Rate ≥ 1% and a reportable sample. Worth monitoring.
  • RARE — Everything else — including all Tier C, regardless of apparent rate.

See Panel classifications for the exact rules.

Putting it together

When you read any row in PrevIQ, ask yourself:

  1. What's the context?

    Which syndrome and panel is this organism being measured in?

  2. How solid is it?

    What tier is it, and how wide is the confidence interval?

  3. What does it imply?

    Is it INCLUDE, WATCH, or RARE for panel purposes?