DENOMINATORS & SOURCE QUALITY
How to interpret the evidence
Do not combine different endpoints
| Number | What it means - and does not mean |
|---|---|
Prevalence | The fraction with a defined finding at a time/place. It is not the chance one sex act transmits infection. |
Per-act transmission | Conditional on a relevant infected source and a defined act; not the risk from every unknown-status partner. |
Partnership positivity | May reflect many acts, both directions, shared external partners or prior infection; not a per-act rate. |
Vaccine / condom effectiveness | Relative comparison between studied conditions. A relative reduction is not the residual absolute risk. |
Antibody prevalence | May describe historical infection or immunity, not current infectiousness. The interpretation is pathogen-specific. |
Confidence interval | Sampling/statistical uncertainty for one estimate, not protection against applying it to the wrong population. |
Mgen inputs: observed, not invented
NHANES 2017-2018 tested urine from 3,945 US participants ages 14-59; weighted prevalence was 1.7% (95% CI 1.1-2.7). The Swedish study recruited 235 club attendees: 9/118 women and 5/117 men were genital-positive, giving 14/235 = 5.96%, rounded to 6.0% for scenarios. [04][06]
MyGeniUS used 1,743 remnant urogenital specimens from selected clinic attendees in four US regions in 2020, with sampling quotas. Overall prevalence was 16.6% (95% CI 14.9-18.5); clinic-specific estimates ranged from 9.9% to 23.5%. The highest site is an observed high scenario, not a representative estimate for all multipartner networks. [05]
Other input limitations
Ureaplasma 40% and 50% are SSI clinical carriage benchmarks, not a contemporary probability sample of party attendees. Trichomonas inputs come from distinct population/clinic estimates cited by CDC. HSV values are historical serology; HPV values are historical genital-DNA estimates. These endpoints are deliberately kept separate. [09][24][36][40]
No blanket "asymptomatic multiplier" is applied: studies testing participants can already detect symptom-free infections. Annual diagnosis counts are not prevalence. Community-wide saturation is not assumed. The main pages identify evidence gaps instead of filling them with arbitrary inputs.
REPRODUCIBLE CALCULATIONS
Model methods and audit
The independent-partner model
For n unique partners, each with probability p of the specified finding at the modeled contact time:
P(at least one) = 1 - (1 - p)^n
For unequal independent probabilities: P = 1 - product(1 - p_i). A six-month partner count is the observation window; it does not convert a point-prevalence estimate into six-month incidence. A repeat encounter with the same partner is not a new independent partner draw.
What protection does to this model
A condom does not change whether the partner already has an infection. It changes the conditional chance of transmission during the covered act. A possible acquisition model is 1 - product(1 - p_i q_i), where q_i summarizes transmission across relevant contacts with partner i. It cannot be calculated responsibly when the required q_i evidence is missing.
Network dependence and boundaries
Independence is a simplifying assumption, not a best-case guarantee. At equal marginal prevalence p, general bounds for at least one positive partner are p to min(1, n p). Perfectly linked infection states yield p; disjoint positive states can approach the upper bound. Clustering can change the applicable prevalence as well as dependence. Neither direction should be assumed automatically.
Numerical checks
| Calculation | Result |
|---|---|
p=0.166; n=5 | 59.65% |
p=0.235; n=10 | 93.14% |
p=0.15; n=10 | 80.31% |
Same p=0.166; n=5, dependence bounds | 16.6% to 83.0%; not a confidence interval |
Minimal reproduction
from math import expm1, log1p
def encounter_probability(p, n):
if not 0 <= p <= 1 or n < 0 or int(n) != n:
raise ValueError("Invalid prevalence or partner count")
if n == 0: return 0.0
if p == 1: return 1.0
return -expm1(n * log1p(-p))
Tables show bucket endpoints, not confidence intervals. Values above 99.9% remain labeled >99.9%, not 100%. Zero partners in this interval means no modeled new partner encounter; it does not establish infection-free status. No combined "any STI" score is calculated because coinfection dependence is not specified.
AT LEAST ONE PARTNER WITH THE SPECIFIED FINDING
Full tables / Mgen and Ureaplasma
Mgen: calculated percentages
| Partners | 1.7% national | 6.0% club | 16.6% clinic | 23.5% high clinic |
|---|---|---|---|---|
0 | 0.0 | 0.0 | 0.0 | 0.0 |
1-2 | 1.7–3.4 | 6.0–11.6 | 16.6–30.4 | 23.5–41.5 |
3-5 | 5.0–8.2 | 16.9–26.6 | 42.0–59.7 | 55.2–73.8 |
6-10 | 9.8–15.8 | 31.0–46.1 | 66.3–83.7 | 80.0–93.1 |
11-20 | 17.2–29.0 | 49.4–71.0 | 86.4–97.3 | 94.7–99.5 |
21 | 30.2 | 72.7 | 97.8 | 99.6 |
30 | 40.2 | 84.4 | 99.6 | >99.9 |
50 | 57.6 | 95.5 | >99.9 | >99.9 |
Inputs: US national urine testing (2017-2018), a Swedish genital club sample, and selected US clinics (2020). All scenarios assume independent partners and stable, applicable contact-time prevalence. Neither condoms nor number of sex acts is an input. [04][05][06]
Ureaplasma carriage: calculated percentages
| Partners | 40% carriage scenario | 50% carriage scenario |
|---|---|---|
0 | 0.0 | 0.0 |
1-2 | 40.0–64.0 | 50.0–75.0 |
3-5 | 78.4–92.2 | 87.5–96.9 |
6-10 | 95.3–99.4 | 98.4–>99.9 |
11-20 | 99.6–>99.9 | >99.9 |
21 | >99.9 | >99.9 |
30 | >99.9 | >99.9 |
50 | >99.9 | >99.9 |
These are carriage probabilities, not disease, symptom, transmission or treatment-need probabilities. SSI benchmarks are approximate and not stratified to a specific party network. [09]
21+ is an open category. Exact-count examples at 21, 30 and 50 partners are shown instead of treating everyone above 20 as the same person. A large encounter probability does not establish when, where or whether acquisition occurred.
HISTORICAL SCENARIOS, NOT INDIVIDUAL FORECASTS
Full tables / other endpoints
Chlamydia and trichomoniasis: percentages
| Partners | CT 1.7% | Trich 0.5% | Trich 2.1% | Trich 14.6% |
|---|---|---|---|---|
0 | 0.0 | 0.0 | 0.0 | 0.0 |
1-2 | 1.7–3.4 | 0.5–1.0 | 2.1–4.2 | 14.6–27.1 |
3-5 | 5.0–8.2 | 1.5–2.5 | 6.2–10.1 | 37.7–54.6 |
6-10 | 9.8–15.8 | 3.0–4.9 | 12.0–19.1 | 61.2–79.4 |
11-20 | 17.2–29.0 | 5.4–9.5 | 20.8–34.6 | 82.4–95.7 |
21 | 30.2 | 10.0 | 36.0 | 96.4 |
30 | 40.2 | 14.0 | 47.1 | 99.1 |
50 | 57.6 | 22.2 | 65.4 | >99.9 |
CT: one Swedish club sample. Trichomonas: CDC-cited US male, US female and selected clinic-female estimates respectively. Do not assign a scenario solely from gender or identity; matching population, anatomy and time matters. [04][24]
HSV antibodies and HPV DNA: percentages
| Partners | HSV-1 47.8% | HSV-2 11.9% | Any HPV 42.5% | High-risk HPV 22.7% |
|---|---|---|---|---|
0 | 0.0 | 0.0 | 0.0 | 0.0 |
1-2 | 47.8–72.8 | 11.9–22.4 | 42.5–66.9 | 22.7–40.2 |
3-5 | 85.8–96.1 | 31.6–46.9 | 81.0–93.7 | 53.8–72.4 |
6-10 | 98.0–99.8 | 53.2–71.8 | 96.4–99.6 | 78.7–92.4 |
11-20 | >99.9 | 75.2–92.1 | 99.8–>99.9 | 94.1–99.4 |
21 | >99.9 | 93.0 | >99.9 | 99.6 |
30 | >99.9 | 97.8 | >99.9 | >99.9 |
50 | >99.9 | 99.8 | >99.9 | >99.9 |
HSV: US ages 14-49, 2015-2016 serology. HPV: US ages 18-59, 2013-2014 genital DNA. These do not measure shedding, cancer risk or current 2026 network prevalence. HPV columns overlap. [36][40]
No generic partner-count table is assigned to gonorrhea, infectious syphilis, transmissible HIV, active HBV/HCV, HAV, mpox or Shigella without an appropriately matched input. Their clinical pages instead show valid risk, state or prevention benchmarks.