How this instrument works
This calculator implements the Denver HIV Risk Score, published by Haukoos and colleagues in the American Journal of Epidemiology in 2012 and derived and validated from patients screened for HIV at a metropolitan sexually transmitted disease clinic and in emergency departments. Eight factors — age, sex, race/ethnicity, whether the patient has had sex with a male, receptive anal intercourse, vaginal intercourse, injection drug use, and prior HIV testing — combine into a single point total. Age contributes points in a U-shaped pattern that peaks at 12 points for ages 33 through 46 and falls to zero at the youngest and oldest ends of the range, reflecting the age distribution of undiagnosed HIV observed in the derivation cohort rather than a simple straight-line relationship with age.
It's worth being direct about a genuine design choice here. The score includes race and ethnicity as a scoring variable — Black race adds 9 points and Hispanic ethnicity adds 3 points, while White, other, and American Indian/Alaska Native categories add none — because the derivation cohort showed real differences in HIV prevalence across these groups, and including them measurably improved how well the score identified undiagnosed infections. Whether risk-scoring tools should incorporate race and ethnicity at all is a genuinely debated question in clinical risk-prediction more broadly: doing so can improve a tool's statistical performance on the population it was built from, but it also means two people with identical behavioral risk factors receive different point totals based on race alone, which is worth stating plainly rather than glossing over.
This is a screening-prioritization tool, built for clinicians and public health testing programs deciding how to allocate limited HIV testing resources — not a diagnostic test and not a substitute for one. A score in the very-low category does not rule out HIV risk for a specific individual; the score was derived and validated on population-level screening data, and anyone who requests testing or reports behavioral risk factors should be able to get tested regardless of where their point total lands.
- Enter Age in years — the score assigns U-shaped points that peak at ages 33 to 46.
- Set Sex is male to Yes or No.
- Select Race/ethnicity: White/other/American Indian or Alaska Native, Black, or Hispanic.
- Set Has had sex with a male and Receptive anal intercourse to Yes or No based on sexual history.
- Set Vaginal intercourse, Injection drug use, and Prior HIV testing to Yes or No.
- Read the age points, total Denver HIV Risk Score, and the resulting risk category.
Worked example — three risk profiles
A 28-year-old Black man who has had sex with a male, including receptive anal intercourse, with no vaginal intercourse, no injection drug use, and no prior HIV testing: age 28 falls in the 26-32 band worth 10 points, and the total score is 10 (age) + 21 (male) + 9 (Black) + 22 (sex with a male) + 8 (receptive anal intercourse) = 70, landing in the very-high-risk category (≥50).
A 19-year-old woman with only vaginal intercourse and a history of prior HIV testing scores 0 (age, below the 22-year threshold) + 0 (race/ethnicity) − 10 (vaginal intercourse, a protective factor in this model) − 4 (prior testing, also protective) = −14, well into the very-low-risk category (<20) — both vaginal intercourse and prior testing carry negative point weights.
A 40-year-old Hispanic man with none of the specific behavioral risk factors listed scores 12 (age, in the 33-46 peak band) + 21 (male) + 3 (Hispanic) = 36, in the moderate-risk category (30-39) from age, sex, and race/ethnicity alone, without any behavioral risk factor contributing points.
Questions
Is the Denver HIV Risk Score a diagnostic test?
No. It's a screening-prioritization tool designed to help clinicians and public health testing programs decide who to prioritize for HIV testing, not to diagnose HIV infection. Only a laboratory HIV test can diagnose infection — this score estimates how closely someone's profile resembles patients with undiagnosed HIV in the derivation cohort, and a low score doesn't rule out infection for any individual.
Why does the score include race and ethnicity as an input?
The derivation cohort — patients screened for HIV at a Denver clinic and in emergency departments — showed real differences in HIV prevalence by race and ethnicity, and including them as scoring variables measurably improved how well the score identified undiagnosed infections in that population. This is a genuinely debated design choice in clinical risk-scoring more broadly: it can improve statistical performance on the population a tool was built from, while also meaning two people with identical behavioral risk factors get different point totals based on race alone.
What does a 'very high' risk category mean in practice?
A score of 50 or higher places someone in the very-high-risk category, the top of five bands (very low, low, moderate, high, very high) the Haukoos et al. instrument uses to stratify HIV testing priority. It signals that, based on the factors entered, this person's profile resembles those with a comparatively higher probability of undiagnosed HIV in the derivation cohort — it is not a probability of infection itself, and it should prompt testing, not assumption.
Should someone with a low score skip HIV testing?
No. A low or very-low score reflects population-level patterns in the derivation cohort, not a guarantee for any individual. Anyone who wants an HIV test, or who has risk factors not fully captured by this eight-factor model, should be able to get tested regardless of their calculated score — this tool is meant to help allocate limited screening resources, not to gatekeep who is offered a test.
What do the U-shaped age points represent?
Age contributes 0 points below 22, rises to 4, then 10, peaks at 12 points for ages 33 through 46, then falls back through 10 and 4 points before returning to 0 above age 60. This U-shaped pattern reflects the actual age distribution of undiagnosed HIV observed in the Haukoos et al. derivation cohort, rather than assuming risk rises or falls in a straight line with age.
What inputs does the calculator need?
Age in years, sex, race/ethnicity (White/other/American Indian or Alaska Native, Black, or Hispanic), whether the person has had sex with a male, receptive anal intercourse, vaginal intercourse, injection drug use, and prior HIV testing history. All eight combine into the total Denver HIV Risk Score and its risk category.
References
Read this first: This instrument computes a screening figure from population formulas — it is not a diagnosis, and it cannot see the whole picture a clinician can. Use it to inform a conversation, not to replace one.