Audit surveillance denominator definitions for MSM HIV/PrEP rates

[smjenness]

Context

Our model calibration targets (HIV incidence/prevalence among MSM, PrEP coverage, etc.) carry an implicit MSM denominator definition. If those denominators are built on multi-year recall (Grey/Purcell-style) populations, then setting N from a 12-month-recall (Islek/NHANES-style) source creates a denominator mismatch between our model and its calibration data. Steve Goodreau flagged this as a critical alignment question.

This issue scopes the audit needed to verify which definition our calibration targets actually use.

Investigation tasks

• CDC HIV Surveillance reports: identify the recall window and behavioral definition of MSM used in MSM-specific incidence and prevalence rates.
• Georgia Department of Public Health HIV epidemiology reports: same.
• WA DOH and PHSKC reports: same (PHSKC's BRFSS basis is partly documented in the companion popsize2026 issue on PHSKC).
• AIDSVu PrEP coverage estimates: identify the denominator basis.
• CDC Estimated Number Eligible for PrEP: identify the denominator basis.
• NHBS-MSM (RDS-based behavioral surveillance): identify recruitment criteria and how that maps to a population-level denominator.
• Summarize: do current calibration targets implicitly assume a 5-year/meta-analytic denominator, a 12-month/NHANES denominator, an identity-based denominator, or something else?

Priority

High. Directly informs the decision on which N to use.

Source

Plan: inst/popsize/msm_population_size_discussion.md, §Q2.