How di calculator dey work

Di calculator dey estimate only one tin: di chance say di exposure (wen HIV fit don enter your body) wey you describe don give you HIV. E dey multiply some figures wey dem don publish togeda. E be model, so e dey make tins simple. E no fit see your partner viral load (how much HIV dey di blood), your body, or details wey you no put, and e no fit ever replace test.

1. Risk for each act

Di starting point na how often one act dey pass HIV, if di oda person get HIV and e no dey take treatment wey dey work. Dis na CDC estimate, for every 10,000 exposures.

CDC put oral sex (to use mouth for private part) as “low” (0 to 4 for every 10,000), and dem no give one exact figure; we dey use 1 for every 10,000. For blood wey reach eye, mouth or skin wey get cut, we dey use di estimate from exposure wey health workers get. Wen you pick more than one type, we dey combine dem as separate chances, 1 − (1 − p₁)(1 − p₂)…, instead of to add dem togeda.

We dey count kissing, bite or spit, sharing sex toys, and body fluid wey touch skin wey no get cut as negligible (risk wey small so tey e no matter). CDC put biting, spitting and sharing sex toys as negligible, and HIV no dey pass through saliva (spit) or skin wey no get cut.

Needle jam (needlestick): di details

Di needlestick figure na average. One study for health workers (Cardo et al., 1997) find say dis tins dey raise di odds of infection. If more than one of dem apply, we dey use di biggest one, becos dem dey usually happen togeda.

2. Tins wey dey change di risk for each act

Viral load dey very high for di first weeks afta person get HIV, and studies don find say HIV dey pass several times more for dat period than later; we dey use factor of 10. Undetectable viral load (di HIV for blood don low reach where test no fit see am) dey make di risk from sex zero, as di evidence behind Undetectable = Untransmittable (U=U) show. Treatment still dey lower di risk from sharing needle, but nobody don measure by how much, so we no dey change di risk for needle and blood. Di figures for belle come from one study of couples for seven African countries (Thomson et al., 2018). E find say di risk for each act dey higher for di last part of belle and afta di woman don born. For di early part of belle, di risk dey lower small than di figure wey we dey use.

3. Di chance say di oda person get HIV

Dis one na often di biggest factor for di whole estimate. If you know say di person get HIV, e be 100%. If e test negative and di result still dey up to date, e be 0%. Negative test dey up to date only if e do am afta di window period (di time afta infection wen test neva fit show HIV) wey follow im last possible exposure, and e neva get any risk since dat time. If you no sure, we go treat im status as unknown, becos person wey just get infection fit test negative for di same time wey e fit pass HIV pass.

For partner wey you no know im status, we dey build di figure in four steps.

Where di person come from, and weda na man or woman

We dey start wit di share of adults wey dey 15 to 49 years old wey dey live wit HIV for di country or region wey you choose (UNAIDS estimates via World Bank indicator SH.DYN.AIDS.ZS (adults 15-49), collected 2026-10-29). UNAIDS still dey report how many of di pipo wey get HIV na women, so we dey divide di figure by sex. For South Africa, for example, dat one dey give about 23% for women and 12% for men; for UK, 0.27% for men and 0.13% for women.

UNAIDS no dey publish model estimate for some countries. For China, Japan, Poland, Russia, South Korea, Türkiye, Ukraine, United States we dey use approximate figures from national reports; we mark dem “approx.” for di list.

Weda male partner don do circumcision

Trials for Africa find say circumcision cut men risk to get HIV through vaginal sex by 50 to 60%, so for dat side, men wey don do circumcision no too likely to get HIV. We dey count say di chance for man wey don do circumcision to get HIV na 50% of di chance for man wey neva do am. Becos di country figure don already mix di two togeda, how much dis answer go change tins depend on how common circumcision be for dat country (Morris et al. 2016, corrected 2016). For UK, where about 20.7% of men don do circumcision, di figure for man wey don do circumcision dey 44% lower. For Kenya, where about 91.2% don do am, di figure for man wey neva do circumcision dey 1.8× higher.

Dis one dey change only di chance say e get HIV. If e get am, circumcision no dey lower di risk say e go pass am: one trial for Uganda no see any reduction for how HIV dey pass to female partners (Wawer et al. 2009). We no dey use am for men wey dey sleep wit men, becos di evidence for dem neva clear, or for community figures, becos circumcision rates na for di whole country.

Im community

For some countries, official surveillance dey show say HIV rate dey much higher or lower for some communities. Na social and economic tins and sexual networks (who dey sleep wit who) dey cause am, no be di body. Where dis figures dey, one optional question go let you use dem instead of di country figure:

Di US figures na pipo wey dem don diagnose wit HIV, and we scale am up for di share wey neva know dia status. National figures for oda US communities, like Hispanic/Latino and Asian American pipo, neva dey inside yet. Di UK figures na di latest wey dem publish by community, but dem don old reach about ten years.

Groups wey get higher HIV rate

If you be man and you sleep wit man, e don already be man wey dey sleep wit men, so dat figure dey apply even if you no tick anytin. We dey raise di odds instead of to multiply di percentage, so for countries wey HIV common well-well, di result go still be correct probability. Di UNAIDS figures dey measure new infections, no be di ones wey don already dey, so here dem be approximation. Wen many of dis groups apply, or you still choose community figure, we dey use di one wey give di highest result instead of to pile dem on top each oda, becos di groups dey overlap.

Treatment

Plenti pipo wey get HIV dey take treatment, and pipo wey dia viral load don undetectable no fit pass HIV through sex. So for exposure through sex, we dey count only pipo wey get HIV wey dia viral load neva suppress (neva low well wit treatment). We dey estimate dat share from di treatment coverage for each country (UNAIDS, through World Bank), multiply by di 93% of pipo wey dey take treatment wey dia viral load don suppress for di whole world (UNAIDS 2024). For needle and blood exposures we dey count everybody wey get HIV, becos U=U neva establish for dose ones. Where treatment data no dey, we no dey adjust anytin, and dis one dey lean to di side of higher estimate. Key populations (groups wey get higher HIV rates) often get lower treatment coverage than di national average, so for dem dis step fit lower di estimate small more than e suppose.

4. More than one exposure

Wit di same partner, im status no dey change between di acts, so we dey apply di prevalence once: prevalence × (1 − (1 − p)ⁿ). Wit different partners wey you no know dia status, each act na separate chance: 1 − (1 − prevalence × p)ⁿ. Partners wey dia viral load don suppress dey inside only for blood exposures. Risk dey add up as e dey repeat, but e no go ever reach di point wey e go dey sure.

5. PrEP and PEP

PrEP (pre-exposure prophylaxis: HIV medicine wey you dey take before any exposure to protect yourself) na before exposure, and PEP (post-exposure prophylaxis: HIV medicine wey you go take afta exposure to stop di virus) na afta. Di PrEP figures for consistent use (to take am every time as e suppose be) base on CDC estimate. CDC estimate say daily PrEP dey lower risk by about 99% for men wey dey sleep wit men, and we dey apply dis one to anal sex and oral sex. For heterosexual men and women (pipo wey dey sleep wit opposite sex), CDC give 88 to 90%, based on recent use and no be consistent use; we dey use 90% for vaginal sex. For pipo wey dey inject drug, CDC give 74 to 84%; we dey use 74% for all blood exposures.

Di lower levels of adherence (wen person no dey take di medicine every time) and how PEP dey work less as time dey pass na our own assumptions: PEP dey work beta di sooner you start am, but no exact figures dey for each day. If you miss PEP doses, di risk wey remain go double. We no dey combine di two: if you tok say you dey take PrEP, di PEP option go off.

6. Negative test result

Negative test dey lower di estimate wit Bayes’ rule. Di question na how likely e be say di test miss infection for di day wey dem do am. Before di window of di test open, e no dey tell us anytin. Inside di window, we dey assume say di chance to miss dey fall steady, on logarithmic scale, from 90% go 0.5%. Afta di window close, we dey take negative result as final, wit only very small allowance for mistake. If you put many negative tests, we dey use di one wey give di most information.

Di 90% starting point dey careful on purpose, so early negative result no go lower di estimate as much as e fit lower am. PEP fit delay di time wey test go detect HIV, so we no dey use tests wey you do inside di 28 days wey you dey take PEP.

7. Risk levels

Di label wey dey near your number dey follow fixed levels:

• Low well well: under 0.01% (less than 1 out of 10,000)
• Low: 0.01% to 0.1%
• Medium: 0.1% to 1%
• High: 1% or more (1 out of 100 or higher)