Denver HIV Risk Score Calculator
An 8-question screening aid based on the published clinical prediction rule of Haukoos et al. (American Journal of Epidemiology, 2012). Answer the questions to get a score, a risk band, and the calibrated probability of undiagnosed HIV from the original study.
The 8 questions
The wording follows the original research items. All questions are optional in the sense that you may stop at any time — but every unanswered question blocks a score. Nothing you enter leaves your browser: this page works 100% offline with no data sent anywhere.
Your result
Answer all 8 questions above and your score will appear here automatically.
How the score works
Points below are exactly as published (Supplemental Table 1 of the authors' 2015 validation paper, identical to the 2012 derivation paper). The total ranges from -14 to 81.
| Item | Answer | Points |
|---|---|---|
| Age | <22 or >60 years | 0 |
| 22–25 or 55–60 years | +4 | |
| 26–32 or 47–54 years | +10 | |
| 33–46 years | +12 | |
| Gender | Female | 0 |
| Male | +21 | |
| Race/ethnicity | White or Other* | 0 |
| Black | +9 | |
| Hispanic | +3 | |
| Sex with a male | Yes | +22 |
| Vaginal intercourse | Yes | −10 |
| Receptive anal intercourse | Yes | +8 |
| Injection drug use | Yes | +9 |
| Past HIV test | Yes | −4 |
* “Other” = American Indian/Alaska Native, Native Hawaiian, or Pacific Islander (non-Hawaiian), per the paper. “No” answers score 0.
Risk bands and calibrated probabilities
| Band (paper's labels) | Score | Newly-diagnosed HIV in the study | 95% CI | n/N |
|---|---|---|---|---|
| Very low | <20 | 0.31% | 0.20–0.45% | 27/8,782 |
| Low | 20–29 | 0.41% | 0.29–0.57% | 36/8,677 |
| Moderate | 30–39 | 0.99% | 0.63–1.47% | 24/2,431 |
| High | 40–49 | 1.59% | 1.02–2.36% | 24/1,505 |
| Very high | ≥50 | 3.59% | 2.73–4.63% | 57/1,588 |
The authors describe scores of 30 or higher as indicating increased risk (worth prioritizing for testing offers) and 50 or higher as the highest risk.
Limitations — read before trusting the number
- Specific study population. The score was derived from patients of a Denver STD clinic (1996–2008) and validated in U.S. urban emergency departments. It may not generalize to other countries, to the general population, or to today's landscape of testing, treatment, and PrEP.
- Not a diagnosis. A high score does not mean you have HIV; a low score does not mean you don't. Only an HIV test can determine status.
- Probabilities are population rates. “3.59%” means about 36 in 1,000 people in that band in the original studies had newly diagnosed HIV — most did not.
- A refined 6-variable version exists (Haukoos et al., 2015) that drops the two anal/vaginal intercourse items; this page implements the original 8-variable score from the 2012 derivation paper.
Frequently asked questions
What is the Denver HIV Risk Score?
The Denver HIV Risk Score is a clinical prediction rule developed by Haukoos and colleagues (American Journal of Epidemiology, 2012) to estimate the probability of undiagnosed HIV infection. It combines 8 items — age, gender, race/ethnicity, sexual behaviors, injection drug use, and past HIV testing — into a score from -14 to 81. It is a screening aid for deciding who should be offered HIV testing, not a diagnostic tool: it cannot tell anyone whether they have HIV.
How is the score calculated? Can you show a worked example?
Each answer carries points from the published paper. For example: a 38-year-old man (+21 male, +12 age 33–46) who is Black (+9), has sex with men (+22), and reports receptive anal intercourse (+8) scores 72 points (+21 (male) +12 (age 33–46) +9 (Black) +22 (sex with a male) +8 (receptive anal intercourse)), which falls in the Very high band — an estimated 3.59% probability of undiagnosed HIV (95% CI 2.73–4.63%) in the study population. A second example: a 28-year-old woman (+10 age 26–32) who reports vaginal intercourse (−10) and a past HIV test (−4) scores -4, in the Very low band (0.31%). Both examples are computed by this page's engine from the paper's point table.
What do the risk categories mean?
The study authors grouped scores into five bands with these observed rates of newly diagnosed HIV: Very low (score below 20): 0.31%; Low (20–29): 0.41%; Moderate (30–39): 0.99%; High (40–49): 1.59%; Very high (50 or more): 3.59%. The authors describe scores of 30 or higher as indicating increased risk and 50 or higher as the highest risk. These percentages come from the original study populations and will differ in other settings.
What score means I should get an HIV test?
No score can answer that for you — only a clinician who knows your situation can. The researchers considered scores of 30 and above as marking increased risk worth prioritizing for testing offers in emergency departments. Regardless of your score here, the responsible next step is to discuss HIV testing with a clinician, clinic, or local health department. If you may have had a recent exposure, ask about testing timelines and emergency options promptly rather than waiting.
How accurate is the Denver HIV Risk Score?
In a large external validation using a national HIV testing cohort (4.8 million testing events), the score discriminated between people with and without newly diagnosed HIV with an area under the ROC curve of 0.77, and its predicted probabilities matched observed rates closely (calibration slope 1.09). That is good for a brief questionnaire, but far from perfect: most people even in the highest band do not have HIV, and some infections occur in lower bands. Treat the result as a rough risk estimate, not a verdict.
Who was this score developed on? Does it apply to me?
It was derived from patients of a Denver, Colorado sexually transmitted disease clinic (1996–2008) and validated in urban emergency departments in the United States. It may not generalize well to other countries, to the general population outside clinics and emergency departments, or to today's HIV landscape — testing, treatment, and prevention (including PrEP) have changed substantially since the data were collected. Use the result as an educational estimate, not a personal medical assessment.
Why do some answers subtract points, like vaginal intercourse (−10) or a past HIV test (−4)?
The points are statistical weights from the original regression model, not moral judgments. In the derivation population, people reporting vaginal intercourse (without the other risk factors) or a previous HIV test were less likely to have newly diagnosed HIV, so the model assigns negative points. The score describes patterns in that specific study population — it says nothing about any individual's worth or character.
I got a low score. Does that mean I definitely do not have HIV?
No. A low score means people with similar answers had a low rate of newly diagnosed HIV in the original studies — about 3 in 1,000 in the very-low band — but it cannot rule out infection for any individual. Only an actual HIV test can determine someone's status. If you have had a potential exposure or symptoms that worry you, seek testing and medical advice regardless of this score.
Is this calculator a medical diagnosis?
No. This page is an educational implementation of a published research score. It is not a medical device, not a diagnosis, and not a substitute for professional medical advice, diagnosis, or treatment. Never disregard professional advice or delay seeking it because of a result shown here.
What should I do with my result?
Use it as a starting point for a conversation, not a conclusion. Consider discussing HIV testing with a clinician, sexual health clinic, or community testing service — they can advise on the right test and timing for your situation. If cost or privacy is a concern, many health departments offer free, confidential testing. If you may have been exposed very recently, contact a clinician or emergency service promptly to ask about time-sensitive options.
Sources
- Haukoos JS, Lyons MS, Lindsell CJ, et al. Derivation and Validation of the Denver Human Immunodeficiency Virus (HIV) Risk Score for Targeted HIV Screening. Am J Epidemiol. 2012;175(8):838–846. PMID: 22431561.
- Haukoos JS, Hopkins E, Bucossi MM, et al. Validation of a Quantitative HIV Risk Prediction Tool Using a National HIV Testing Cohort. J Acquir Immune Defic Syndr. 2015;68(5):599–603. Supplemental Appendix, Supplemental Table 1 (point table and risk-band labels).
- Haukoos JS et al. Targeted Rapid HIV Screening Using the Denver HIV Risk Score (poster, SAEM 2012) — author-published scoring table.