A dating app pools infection rather than spreading it evenly. Gonorrhoea shows where the pool is. Ebola survivors can carry a reservoir the pool can’t show.
What a dating app does to a network
First, let’s get something out of the way: the author has had a varied sex life: periods of celibacy contrasting periods of high sexual partner count, further contrasting periods of being faithful to one partner with zero cheating. The author has no cock in this fight.
Most writing about dating apps treats them as a matching market. Someone swipes, someone matches, and the interesting question is why the numbers are so lopsided. There’s a second question underneath it: what does the app do to the shape of the sexual network?
It concentrates it.
People in an ordinary city meet through work, through friends, through the pub, through the school gate. Those routes overlap and cross, and the resulting network is messy but relatively diffuse. An app collects a subset of the population that is actively looking and routes their encounters through a much smaller number of contact points. Prevalence of whatever is circulating inside that subset goes up. Prevalence outside it stays where it was. The infection doesn’t spread across a whole generation. It pools in a corner of it.
The survey data bear this out, though not neatly. A 2023 scoping review of STI prevalence among dating-app users found gonorrhoea rates ranging from 1% to 18%, chlamydia from 0.3% to 7.6%, and syphilis from 3.1% to 17.4%. That range is enormous, because the studies draw on different clinics in different countries with different sampling rules. The direction is consistent: a meta-analysis of men who have sex with men found app users were more than twice as likely to have gonorrhoea (OR 2.36) or chlamydia (OR 2.22) as non-users. The app-using population is carrying more infection than the population around it.
Online daters also sort by age more tightly than offline daters do. The matching logic surfaces people of similar age by default, and the result is a set of dense clusters, each internally well connected and externally poorly connected. Apps don’t mix people up. They gather them into piles. The algorithm does the heavy lifting to sort the users into pre-selected social groups before the users have had any opportunity to choose whatsoever.
Gonorrhoea is the map
If you want to know where the piles are, you need something that leaves a mark.
Gonorrhoea is good at this for three reasons. It passes between people readily enough during unprotected sex that most connected partnerships will show up. It clears in about six months untreated, which is short enough that its prevalence reflects the network as it currently exists rather than as it existed years ago. And it’s common enough to measure. A pathogen with very low transmission would sample the network too sparsely to map it. A pathogen with very high transmission and long duration would saturate every edge and stop distinguishing between structures. Gonorrhoea sits in the range where it can still tell you something.
Public health agencies have used it this way for decades, and the numbers that come out are specific. Contact tracing and partner notification give the mixing patterns: who sleeps with whom, by age, by neighbourhood, by social circle. Repeat infection data give acquisition rates. Clinic samples give the core group figures. A 2025 study in a North Carolina county classified 6.7% of patients as belonging to a core group, and that 6.7% accounted for 68% of all gonorrhoea cases. Earlier modelling work put the core at under 2% of the population causing 60% of infections. A third estimate found 0.1% of the population responsible for 22% of cases. The percentages vary by setting. The pattern doesn’t. A small number of people are involved in a large share of transmission.
The bridging figures are equally concrete. About 5.2% of people in one gonorrhoea study reported partners both inside and outside their local sexual network, and those people had more partners and more concurrent partners than people who stayed local. In chlamydia clinic samples, 21% of infected young men in Baltimore and 26% in Denver met the definition of an age-bridger: at least two partners in two months, differing in age by two years or more. These are the people who move an infection between groups that would otherwise stay separate.
Here’s the limit. Gonorrhoea maps the routes. It doesn’t tell you how much traffic a different pathogen will put on them. That’s set by the biology of the other pathogen, not by the shape of the network. Because gonorrhoea clears in months while some pathogens persist for years, the map it produces is a snapshot. It catches the partnerships that exist now. It misses whatever forms in the gaps between surveys. The map is a floor, not a ceiling. The routes it shows are real, and there may be others it didn’t catch.
Gonorrhoea is a bacterium that clears in months after a transmission. The rest of the article is about something that doesn’t: the Bundibugyo Ebola virus.
The network has two axes, and the connected dyad is the unit
Every partnership can be typed twice: on the age axis (same-age or age-gap) and on the graph axis (inside a local social network or between two of them). The two are independent. A same-age edge can cross graphs; an age-gap edge can stay inside one. Four cells result, each with a different reach. Under the illustrative central case, same-age intra-graph is about 42% of partnerships, same-age inter-graph about 28%, age-gap intra-graph about 18%, and age-gap inter-graph about 12%. The cell that crosses both boundaries is the rarest and has the widest reach. The age-gap logic closely follows the logic of general out-group high-partnered adults with the caveat that only a part of the age-gap dyads follow this logic. This is a statistical inference: It’s common but not universal that the older partner is a high-partnered adult.
The edge is not the unit of transmission. A closed dyad, two people with no other partners, produces zero transmissions regardless of its age gap, because the pathogen has no way in. The unit is the connected dyad: a partnership in which at least one partner has an outside partner. The age-gap share of connected dyads is higher than the age-gap share of all dyads, because age-gap partnerships are disproportionately casual, and casual partnerships are where concurrency lives. Under illustrative connectivity assumptions, the age-gap share rises from roughly 30% of all dyads to roughly 39% of connected dyads at the marginal level, and from 20% to 23% for app-mediated edges.
The connected dyad is a chain position, not a single event. One transmission comes in, and the newly infected node now has outgoing edges. The count of possible transmissions is the count of directed edges from an infectious node to a susceptible partner, and each confirmed transmission generates the receiver’s outgoing edges.
Who actually passes it on
The following comes under the assumption that everyone washes their hands after entering their home or a shared space, avoids hugging and kissing strangers, avoids consuming food from the same plate with strangers and keeps fresh produce behind a screen, ordering it from a sales staff. Everyone knows that this won’t consistently happen, hence publishing this paragraph becomes necessary.
The percentages are the polite version. The profiles behind them are more recognisable than a tier model suggests, and each one calls for a different intervention.
Repeat infection is the strongest single signal. The people who keep coming back to the clinic are the people whose partners are drawn from a network where the bacterium is already circulating. Gonorrhoea doesn’t produce lasting immunity, so the same person can be reinfected within months, and reinfection is both a marker of exposure and a mechanism that keeps prevalence up. A repeat positive is the most informative test result available. The intervention that fits is expedited partner therapy, treating a diagnosed person’s partners without requiring them to attend a clinic first, because the bottleneck isn’t willingness, it’s the second appointment. The second is repeat screening at short intervals: three-monthly rather than annual for anyone with a positive in the last year. In case of Ebola transmission the only thing that works is automatic quarantine and aggressive contact tracing. In casual encounters, condom and manual stimulation are very much recommended.
Young men should know that if a woman you’re seeing repeatedly visits a healthcare professional for screening, they very likely are sexually active with multiple partners regardless of what they say. Young women should know the same about men, statistically more often about charming, older men. Act accordingly and require them to get tested before you have any unprotected sex with them. A condom with a high-partner woman is a non-negotiable. Manual stimulation is an effective way to prevent transmission.
In both genders, if they avoid replying or deflect from questions about their recent or concurrent partners, it usually means they have them. Therefore, they count as high-partner adults.
Have honest discussion about exclusivity with a high-partner individual regardless of their gender and sexual orientation, involving a professional counselor if possible.
High-partner men are a category defined by acquisition, not by orientation. The structural position is the same whether the partners are men or women: partner counts in the tens per year, overlapping partnerships, a circuit of venues and platforms that keeps the same people in contact, and a prevalence that is high because the network is dense. The shape of the network differs, and the shape determines which mitigants bite.
High-partner men in dense male networks are the group the app data describe best, because the apps were adopted early and heavily there and the concentration effect is strongest. The network is dense, the same people recur, and prevalence is high because of it. The mitigants that work shorten the time between infection and treatment. Three-monthly screening including extragenital sites (throat and rectum, self-collected) is the core intervention, because a large share of gonorrhoea in this population sits at sites a urine test misses. Pre-exposure prophylaxis for HIV has been transformative. Its arrival was followed in several settings by rising bacterial STI rates, because condom use declined once the HIV risk was chemically managed, and the guidelines now pair PrEP with frequent bacterial screening for that reason. Doxycycline post-exposure prophylaxis reduces chlamydia and syphilis substantially; its effect on gonorrhoea is weaker and inconsistent, and there are real concerns about resistance, so it isn’t a substitute for screening. The meningococcal B vaccine shows roughly 30–40% cross-protection against gonorrhoea in observational studies, modest but the first vaccine signal against the disease. This group is also where regular screening and PrEP have done the most good: the core is not a fixed caste, and interventions aimed at it work. In the emergence of the Ebola, condoms should immediately be reintroduced into this community. In case of Ebola transmission the only thing that works is automatic quarantine and aggressive contact tracing. The density of the community does most of the heavy lifting for the contact tracing. You really don’t want to be the super-spreader in a close-knit gay community. You don’t have to believe the author, but please don’t be that guy. Condom is the best way to prevent infection in this context, closely followed by manual stimulation. If you choose to have unprotected sex with a high-partner man, testing the high partner man before the sexual encounter prevents the vast majority of the possible transmissions. In casual encounters, condom and manual stimulation are very much recommended. If you have to do it, a petite passade, a little fling, is a far better arrangement than full-on casual: the emotion becomes your condom.
High-partner heterosexual men sit in a different network shape. Their edges fan out into many separate female networks rather than circulating within one dense male network. Each new partner is a new entry point into a pool that may not otherwise have been connected to the last one. The prevalence inside any single one of those female pools is lower than in a dense male core, but the number of pools is larger, and the reach is wider. Heterosexual high-partner men are therefore less likely to be part of a self-sustaining core and more likely to be the route by which infection arrives in populations that would otherwise stay clear.
For women: He may transmit Ebola to you, and you may transmit Ebola to someone else, or he may infect you and your friends. Either way, he transmits Ebola from another sexual network. This man should be subjected to precautions: Condom, manual stimulation and testing before unprotected sex.
These men cannot be public about their pattern of womanizing, because other men would condemn them if they were. For a heterosexual man, being exposed as a womanizer in a community of women increases his appeal because he signals pre-selection, but other men are incentivized to be jealous and guarded around him and look for a way to catch him in an act of perceived or real improper behavior. This is called chapernoning, or more colloquially white-knighting. It’s more prevalent in environments where lots of men have been left out of the mating market. It’s simply something we have to live with, because that’s how the human psyche works. The phenomenon is so strong that a man doesn’t even have to be a womanizer or even heterosexual to trigger this instinct in other men: he just has to be handsome and an unknown quantity.
Men who are a known quantity in the community don’t trigger the chaperoning instinct, because they’re the high status men who often are married and have affairs. They’re the most prolific super-spreaders of all other STI’s between communities due to their disproportionately high mobility compared with all of the other segments of men. It would be prudent to conclude that this is the case regarding Ebola as well.
The mitigants differ accordingly. Clinical screening matters less, because the site distribution is different. What matters more is that the bridger himself gets tested at all, since he has no dense network applying social pressure to do it, and his partners are unlikely to be in a clinic that would contact him. Self-sampling kits, workplace testing, and any route that doesn’t require him to present at a sexual health clinic do most of the work. Rapid dispersal of test kits is paramount, but on a positive result the patient ought to immediately be quarantined. Aggressive contact tracing and exhaustive cataloguing of Ebola survivors matters more here than in the dense-network case, because the partner notification chain is the only mechanism by which his other partners learn anything. Cataloguing the potential transmitters is paramount, because it mitigates against social pressure to introduce draconian measures. In case of Ebola transmission the only thing that works is automatic quarantine and aggressive contact tracing. The bridging function carries the risk, rather than the count on its own. Condom is the best way to prevent infection in this context, closely followed by manual stimulation. If you choose to have unprotected sex with a high-partner man, testing the high partner man before the unprotected sexual encounter prevents the vast majority of the possible transmissions. In casual encounters, condom and manual stimulation are very much recommended. If you have to do it, a petite passade, a little fling, is a far better arrangement than full-on casual: the emotion becomes your condom.
These are the exact same measures, as the older generation of readers can, hopefully, remember from the HIV era.
Commercial sex cuts both ways. The worker sits at a hub with many clients, and where condom use is low and testing scarce, the worker’s infection rate can be high. The client is the one who carries it home. His own partner count may be low, yet the edge from client to worker joins two networks that would otherwise never touch. Truck drivers, seasonal and migrant workers, people who travel for work and return to a regular partner: this is the classic bridging population, and it moves infection across distance. For the worker, the evidence supports decriminalisation and occupational health support, because both change the bargaining position rather than the worker’s intentions. Where sex work is criminalised, workers can’t report violence, can’t organise, and can’t insist on condoms without risking arrest. Peer-led outreach outperforms clinic-based services because it reaches people who won’t walk into a clinic. For the client, the mitigant is condom availability at the point of transaction and testing that reaches him where he is (self-sampling kits, workplace testing) rather than a clinic he won’t attend. A mandatory condom, an organized brothel with mandatory testing for everyone entering it staffed with bouncers and nurses, and prostitute labor unions with minimum salary thresholds, workplace healthcare, pension insurance and welfare contributions would do the work, but is the society ready for that conversation yet? Until we get there: In casual encounters, condom and manual stimulation are very much recommended. If you have to do it, a petite passade, a little fling, is a far better arrangement than full-on casual: the emotion becomes your condom.
Interacting with high-partner adults should be assumed needing precautions against an Ebola transmission, non-negotiable. Condom, manual stimulation or testing in advance. The only exception is someone who demonstrates high-fidelity exclusivity in a long-term relationship and wants one out of their own initiative.
Okay, the article has obviously skipped voluntary self-quarantine for 21 days before a sexual encounter, but how do you police that?
As we hopefully can clearly remember, the public completely ignored the Covid social distancing mandates when it came to George Floyd protests. Ebola transmission doesn’t care about your moral panics either. It will transmit whether someone were racist or not. This is a recurring feature of human history. During the Black Death, religious flagellant processions were permitted to march from town to town, spreading the plague while seeking divine intervention. The moral urgency of the ritual always overrules the clinical logic of containment. If the human psyche requires a cleansing ritual in order to follow orders, at least this process could be one that doesn’t actually transmit a virus that is guaranteed to kill half of the young and health infected and far more than half of the old and frail. Vice trade ran amok during the both pandemics, from Tinder and drugs via courier delivery during Covid to underground taverns and archway prostitution during the Black Death.
As a result, a discerning reader would likely conclude that it would be prudent to avoid high-partner adults as sexual partners altogether. Yes, this is precisely the most prudent thing to do. The precautions are for the cases where necessary prudence or verification are lacking. The same basic hygiene that your dad applied when trying to avoid catcing HIV, except that the current virus spreads faster and stays in latent reservoir, whereas HIV spread consistently and slowly.
If the reader is the high-partner adult, then obviously, try to avoid getting infected. Allegedly, it may be bad for your health and you may risk others. Therefore, the precautions are absolutely necessary.
Age-disparate partnerships do a specific job in the network, but only under a condition.
Commonly under socio-economic conditions that concentrate wealth in the older cohorts, a younger adult’s partner pool is small and local. An older partner’s pool is larger and wider. The edge between them joins a narrow network to a wide one, and it moves infection efficiently in one direction. Commonly, the rule of thumb is that the older partner gets tested first, but the transmission doesn’t care about statistics. The question is, can transmission occur? yes or no. The Ebola infection transmits in the both directions, therefore the both should get tested, if they have any other partner, current, recent, or concurrent.
If the older partner has no other partner, and the younger partner has no other partner, the couple is a closed dyad and the age gap carries no transmission route. The edge is a dead end by construction. So the border condition is simple: if the older partner has any other partner, current, recent, or concurrent, the older partner is a candidate for a full screen. If the older partner has none, the older partner is not a possible transmitter on account of the age gap. This works in reverse: if the younger partner has any other partners, screen the younger partner.
The question on the screening form should be about the partner’s other partners and recent changes in the relationship’s shape, not about the partner’s age. Believe what the partner does, not only what they say. If they’re a high partner adult, they should use a condom, switch to manual stimulation or get tested. If they aren’t sexually active with lots of people, they’re as safe as a same-age partner in a similar arrangement. In casual encounters, condom and manual stimulation are very much recommended. If you have to do it, a petite passade, a little fling, is a far better arrangement than full-on casual: the emotion becomes your condom.
Undiagnosed and asymptomatic infection is its own category because the intervention is different. Gonorrhoea is frequently silent, particularly at rectal and pharyngeal sites, and an infection that produces no symptoms produces no reason to seek care. The person carries it for months and every partner in that window is exposed. Repeat infection is a problem of access to treatment. Silent carriage is a problem of screening people who have no reason to think they need it. The two things that raise screening uptake are self-sampling and non-clinical settings. A kit posted to a home address, returned by post, gets tested by people who will not attend a clinic. Point-of-care testing matters more than it sounds, because the week-long gap between test and result is where a large share of patients are lost. Symptom-based testing is structurally incapable of finding this category.
This is a sound formula also for Ebola testing: Ebola is often silent for up to 21 days before symptoms appear, so rapid dispersal of test kits is paramount, but on a positive result the patient ought to immediately be quarantined.
Again, one more time: In casual encounters, condom and manual stimulation are very much recommended. If you have to do it, a petite passade, a little fling, is a far better arrangement than full-on casual: the emotion becomes your condom.
Concentration settings do the rest. Prisons and detention, drug treatment, residential care, worker camps, any place where a population is gathered, has sexual contact. Testing should be mandatory before entering these facilities, otherwise they become lethal public health hazards continuously transmitting Ebola into their surrounding population. These are concentrated populations. The mitigant is condom provision and access to healthcare inside the setting, available rather than on request and not routed through a clinician. Abstinence policy has been tried repeatedly and the transmission data do not support it. Again, one more time: In casual encounters, condom and manual stimulation are very much recommended. If you have to do it, a petite passade, a little fling, is a far better arrangement than full-on casual: the emotion becomes your condom.
Importantly, Ebola doesn’t ask whether you’ve consented to the sexual act or not: a transmission is a transmission. This means that you should take precautions and avoid situations where you may be coerced or forced to perform sexual acts in the first place.
Ebola is indifferent to blame, and it will kill you in roughly half the cases, plus half the people in your immediate family and peer group, whereas arguably an average street encounter with a volatile individual doesn’t result in equally serious ramifications. Or maybe it does. Maybe the author just hasn’t been to rough enough neighborhoods yet.
Core membership is a position, not an identity. People move in and out as circumstances change. At population level the structure is stable; at individual level it isn’t, and any policy that treats the current core as a permanent list will be wrong within a year. Poverty, housing instability, addiction, incarceration, stigma, and poor access to care all push people into denser networks with higher prevalence and less ability to leave. The behavioural patterns matter for transmission. They sit downstream of conditions most people didn’t choose, and the measures that work are the ones that change access.
The survivors nobody is following
Ebola virus persists in parts of the body that the immune system doesn’t police properly. The testes are one of them. Viral RNA has been found in the semen of male survivors for up to 965 days after symptoms began, more than two and a half years. The 2021 Guinea outbreak was genetically traced back to a survivor of the 2013–2016 epidemic, roughly five years after that person was originally infected. A separate cluster was linked to a survivor who probably transmitted the virus about 470 days after his own illness.
This isn’t a steady drip. It looks more like long silences interrupted by bursts. Most survivors never transmit. A few do, years later, and the timing is the problem. An outbreak gets declared over, surveillance stands down, and then a case appears in a place with no obvious chain of infection. That is the shape of a reservoir reactivating.
For Bundibugyo virus specifically, there are no BDBV-specific data on how long the virus persists or how long a survivor remains infectious. A 2026 review in the International Journal of Infectious Diseases states this directly and notes that current guidance is extrapolated from other Ebola viruses. A 2012 genomic analysis of a Bundibugyo outbreak noted that the virus evolves slowly and that genetically similar strains have appeared years apart, which “raise[s] the possibility of persistent, asymptomatic infection with later reintroduction into humans,” but the authors present that as a hypothesis rather than a finding. The mechanism is plausible. The rate is unknown.
What we do know is the scale of the exposure. The 2026 outbreak in eastern Democratic Republic of the Congo had produced 515 laboratory-confirmed cases and 91 confirmed deaths as of June 2026, with no authorised vaccine and no specific antiviral. Every one of those survivors is a potential carrier. The men among them are the ones with the anatomical site that harbours it.
Put that next to the network map and the shape of the risk becomes clear, even though the size of it doesn’t. If a survivor reactivates and forms new partnerships, gonorrhoea data tell you which of those partnerships lead into a wider network and which lead nowhere. A reactivation that lands in a closed dyad infects one person and stops. A reactivation that lands in someone with several partners and connections into another social circle has somewhere to go. The map doesn’t predict the event. It tells you what the event would reach.
The tier model sharpens the placement. A female survivor cannot exceed the lowest tier on the sexual pathway, because she has no testicular reservoir. A male survivor with no partners during persistence is also in the lowest tier: the reservoir exists, but it has no outlet. A male survivor with any partners during persistence is in the top tier by reservoir status alone, regardless of his partner acquisition rate, because the persistence amplifier dominates the acquisition term. The reservoir doesn’t need to sustain a sexual chain. It needs to generate sparks.
The cures, and what they approximate
Here is everything the piece has argued, compressed into a list.
Barriers. Condoms, used consistently and correctly, cut gonorrhoea transmission by roughly 90%. For Ebola the per-act probability is already low, and a barrier pushes it lower still.
Route elimination. Non-penetrative sex removes the reservoir’s outlet entirely, because Ebola persists in semen and has not been found in saliva, sweat, or on skin. Again: In casual encounters, condom and manual stimulation are very much recommended.
Treatment speed. Expedited partner therapy treats a diagnosed person’s partners without a second clinic visit. Point-of-care testing closes the gap between test and result where patients are lost. Three-monthly screening for anyone with a positive in the last year catches reinfection inside its window.
Screening coverage. Self-sampling kits reach people who will not attend a clinic. Extragenital sites (throat and rectum) catch the infections a urine test misses. Opt-out screening in concentration settings reaches people who were not already worried.
Supply inside the setting. Condoms available rather than on request. Testing that comes to where people are. Continuity of treatment on release or discharge, because the exit is where people disappear from care.
Structural conditions. Decriminalisation of sex work, because a worker who can report violence can also insist on a condom. Housing and healthcare access, because poverty pushes people into denser networks and keeps them there. PrEP paired with frequent bacterial screening, because managing one risk chemically does not manage the other.
The border condition for age-disparate partnerships. Screen the younger partner if the older partner has any other partner. Do not screen on age alone, because a closed dyad is a dead end whatever the age gap.
And for the survivor. Register him. Follow him. Counsel him. Test his partners when a reactivation is suspected. Keep the line open for years rather than months, because the outbreak being declared over does not mean the reservoir is over.
The effective measures change what is available, where it is available, and how few steps it takes to get it. They shorten the distance between an infection and its treatment, or between a person and a barrier. The ones that rely on a person making a better choice at the moment of desire work least well. That is a comment on what interventions are up against.
And then there is the measure that is not an approximation.
Set new partner acquisition to zero and every term in the model goes to zero with it. Expected onward transmission is the product of per-act probability, partner acquisition, and infectious duration. If acquisition is zero, the product is zero, regardless of the pathogen, regardless of concurrency, regardless of whether a reservoir is sitting in a survivor’s body years after the outbreak ended. A closed dyad has no route in. That is the arithmetic, and it is why every other measure on the list is best understood as a way of getting partway to something that one arrangement gets all the way to.
If you want the single most effective thing a person can do to avoid an STI, it is to form a dyad with one other person who has been screened, and to stop having sex with anyone else. Everything else here is for the people who will not do that, or cannot, or did once and then their circumstances changed, which is most people, which is why the rest of the list exists. If you want to sample someone from outside of the dyad, a three-some with condom or manual stimulation prevents most of the risks of infection. Cheating is more likely to sustain transmission. Most couples don’t need to do that so this isn’t a recommendation. If you have to do it, a petite passade, a little fling, is a far better arrangement than full-on casual: the emotion becomes your condom.
Two things are worth adding, because that sentence is true and also incomplete. A dyad is only closed if both people keep it closed, so the measure depends on a promise being kept, and promises are not a public health instrument. And the measure is not available to everyone: people leave relationships, people are widowed, people are in relationships that were never monogamous to begin with. A message that offers only the perfect option reaches the people who already took it. The partial measures are measured against the behaviour people actually have rather than against a standard they do not meet.
The population that is already at zero
One group is already past the whole list.
The measure is binary per period: partnered sex or none. Partnership status is a different variable, since a person can have a partner and no sex, or no partner and sex. Self-directed activity does not count, because it has no route to another person. Zero means zero events with anyone in the window.
For that group, expected onward transmission is zero. Per-act probability times partner acquisition times duration equals zero when acquisition is zero. No pathogen, no reservoir, no concurrency changes it, because there is no edge for anything to travel along.
The figures:
| Group | Zero partnered sex in period |
|---|---|
| G7 women | 52% (IFOP) |
| G7 men | 37% (derived; band 35–40%) |
| Britain, 16–74, past year | 15.9% men / 22.2% women (Natsal-3) |
| US men, past year | 24.5% |
| US men 18–24, virgin | 42% (from 21%) |
| German men 18–30, past 12 months | 20.3% (from 7.5%) |
| Adults in high-income countries | one-third to one-half |
The G7 male figure is derived rather than measured. IFOP published the G7 result for women only, and 37% applies the Natsal-3 sex ratio of 0.72 to the IFOP women’s figure. The national male series are measured directly, use different instruments, and run lower.
The share is rising on every series that measures it. US male sexlessness in the past year went from 15.5% to 24.5%. Male virginity among 18–24-year-olds went from 21% to 42%. German men aged 18–30 went from 7.5% to 20.3% between 2005 and 2016.
Men’s sexlessness is lower than women’s because gay men tend to have more sex with more partners, while lesbian and bisexual women tend to have long-term relationships disguised as friendships with one woman at a time with long dry periods between partners.
None of the interventions above reach this group. Every item on the list acts on an edge. There is no edge. The list is not addressed to them because there is nothing to address.
No surveillance system records them either. They generate no cases. A prevention apparatus that counts cases will never see the population that produced none, which is the population that has already arrived where the rest of the list is trying to get.
Not having sex at all is the best prevention, but it does come with the side-effect of not having a partner.
The catalogue is the countermeasure
For the Ebola survivor, the arithmetic has one addition. No new partners during the persistence window collapses the sexual route to zero, which is why counselling survivors about partner numbers is a direct application of the model rather than a moral suggestion. Condoms reduce a per-act probability that is already low. Non-penetrative sex removes the route entirely. The intervention that matters most is the boring one.
The most effective way to stop the spread of Ebola in non-sexual transmission is to wash your hands. Even infected semen is relatively safe on skin, if you wash your hands before you put them in your mouth. For most of the time, an Ebola survivor won’t transmit the infection, but they have a possible future active transmission window.
If the male survivors of the 2026 Bundibugyo outbreak aren’t systematically registered and followed through medical services, their reactivations won’t be linked back to the outbreak that produced them. They’ll look like sporadic cases. They might look like small clusters with no obvious source. They might not be detected at all. A reservoir that nobody is looking for doesn’t stay quiet forever. It finds a way out eventually.
If you register every Ebola survivor, you can offer them medical services in anticipation of the virus becoming active, and through counseling and medical services prevent them from spreading it during a future active transmission window.
Against that, the gonorrhoea map is genuinely useful. It shows which networks are dense, which people connect them, and where a reactivation would travel if it happened. That’s enough to prioritise. You can identify the survivors who are embedded in the highest-risk networks and make sure they’re supported first, rather than working through a list alphabetically.
None of this is a prediction with a date on it. The reactivation rate for Bundibugyo is unknown. The per-act transmission probability is unknown. What we do know is that the persistence window is measured in years, that the reservoir exists, and that app-mediated networks are unusually dense. Taken together, those three facts describe a real possibility, whether or not it happens in any given year.
The pathogen doesn’t care who built the network or what the people in it call themselves. It follows the links. Gonorrhoea shows where the links are. Ebola waits in the reservoir. The precautions cut the links or make them visible, and following the survivors makes the reservoir visible too.
Knowing where the traffic could go isn’t the same as knowing how much of it there will be. But if nobody has drawn the map, nobody can see where it’s heading.
Taking necessary precautions, arguably, may be better than risking a persistent, global Ebola pandemic.
Ultimately, banging different people is fun, but the price for the leisure is eternal vigilance mainly towards infections. Avoiding and reporting toxic dates may be a valuable side-quest, but never the main game because from the perspective of a new STI you’re just a new, mal-adapted host. Never forget where you came from: the animal kingdom.
To be honest, this essay describes a hypothetical but probable scenario in about a decade from the time of its writing, October 2026 contignent on an unsuccessful or partially successful containment of the Bundibugyo Ebola virus. If the epidemic is successfully contained and the transmission chains are mapped well enough to geographically isolate transmission in the Eastern Democratic Republic of the Congo, it will never become a globally persistent latent specter. If the epidemic is poorly contained, the scenario becomes all but certain. In today’s connected world, the choice is almost binary.
References
Bongiovanni, A., et al. (2026). Bundibugyo virus disease: Transmission dynamics, infectiousness and viral persistence. International Journal of Infectious Diseases.
Chan, P. A., Towery, C., Poceta, J., et al. (2018). A network analysis of sexually transmitted diseases and online hookup sites among men who have sex with men. Sexually Transmitted Diseases, 45(7), 462–468.
Hethcote, H. W., & Yorke, J. A. (1984). Gonorrhea transmission dynamics and control. Springer.
Hulseberg, C. E., et al. (2021). Molecular analysis of the 2012 Bundibugyo virus disease outbreak. Cell Reports Medicine, 2(8), 100351.
IFOP. (2024). Observatoire du désir féminin (Volet 1): Les femmes face à la baisse du désir au sein du couple. IFOP Group.
Irigoin, M. E., Elliot, J., & Guerrero Antequera, M. (2023). Sexually transmitted infections and prevention strategies in dating application users: A scoping review. Revista Chilena de Salud Pública.
Jennings, J. M., et al. (2007). Age-bridging among young, urban, heterosexual males with asymptomatic Chlamydia trachomatis. Sexually Transmitted Infections, 83(2), 136–141.
Keita, A. K., et al. (2021). Resurgence of Ebola virus in 2021 in Guinea suggests a new paradigm for outbreaks. Nature.
Mercer, C. H., et al. (2013). Changes in sexual attitudes and lifestyles in Britain through the life course and over time: findings from the National Surveys of Sexual Attitudes and Lifestyles (Natsal). The Lancet, 382, 1781–1794.
Okiring, J., et al. (2022). Sexual partnership concurrency and age disparities associated with sexually transmitted infection and risk behavior in rural communities in Kenya and Uganda. International Journal of Infectious Diseases.
Sexually Transmitted Diseases. (n.d.). Spatial bridges for the importation of gonorrhea and other sexually transmitted infections.