Prevention in Focus

Fall 2015 

Are Condoms Over? Research and Reality for Gay Men

By Mark Milano

Reprinted with permission from Achieve, a publication of ACRIA, available at

“Condoms are over. Everyone knows they don’t work for gay sex. If you use them, you’re naïve.”

“Condoms were for all the AIDS guys 30 years ago.”

“The reason there were so many more gay AIDS cases than straight ones in the ’80s and ’90s is that both gays and straights used condoms, but they broke much more for gay men.”

“There’s no point in using condoms because they will break 50% of the time. They’re more trouble than they’re worth.”

These are all direct quotes captured a few months ago by long-time gay activist Rebecca Reinhardt at a West Hollywood gathering of gay college alumni in their 20s and 30s.

I was disappointed but not shocked when she posted these quotes online. Since the steep decrease of in-person HIV education, gay men get almost all their info about safer sex online. It is perhaps due to this that misinformation about HIV transmission is something I regularly encounter in the workshops I give at ACRIA. For example, at a recent workshop on HIV prevention, one participant put oral sex without a condom in the highest risk category, even though it’s actually a very low-risk activity. Many similar statements have convinced me that we need to relaunch in-person or online interactive safer sex workshops, especially for gay men.

But where has the lack of faith in condoms come from? Since I began working as an HIV educator in the late ’80s, I have always taught that correct and consistent condom use was over 98% effective in preventing HIV transmission, based on CDC statements. Had I been wrong all those years? Is there really a big difference in the effectiveness of condoms for vaginal vs. anal sex? I decided to look at some of the claims and the data.

PrEP vs. Condoms

One of the first things I encountered was condom-bashing from PrEP advocates. Now, I think PrEP (taking Truvada to prevent HIV infection) is an important option and should be made available to all who need it. Regardless of how well condoms work, it’s clear that many gay men aren’t using them and that more prevention options are badly needed.

But do we need to bash condoms in order to promote PrEP? One PrEP advocate created a graph showing that PrEP was 99% effective but that, according to the CDC, condoms were only 70% effective. Really? The data I’ve seen on PrEP shows proven effectiveness of around 92% (the 99% number is a mathematical model that needs more research), and that was in studies in which people also used condoms. But since the 99% figure is being used everywhere, I can see why a gay man might abandon condoms for PrEP – who would take a 30% risk with condoms when a pill could lower it to 1%?

I also wonder what effect this could have on HIV-positive gay men. I recently asked my boyfriend, “When I told you 12 years ago that I had HIV, would you have kept seeing me if you thought that condoms were only 70% effective?” He said there was a strong chance he wouldn’t have. My reassurances that condoms would protect him allowed him to be with his first HIV-positive partner.

So if gay men become convinced that condoms offer only partial protection, will those who aren’t taking PrEP become afraid to have sex with poz guys? And if condoms fail so often, how has my boyfriend remained negative all these years without PrEP? We use a condom every time and haven’t had one break. Are we really that far outside the norm? Let’s take a look at the data.

The 70% Study

For years, most of the data around condom effectiveness came from studies of heterosexuals. In my workshops, I often cite a 1994 study by the European Study Group on Heterosexual Transmission of HIV. It found no HIV transmissions among 124 serodiscordant couples (one partner with HIV, one without) who reported using a condom every time. That was powerful evidence of the effectiveness of condoms, long before PrEP was available.

But that was heterosexuals in 1994. What about gay men today? Do condoms offer the same level of protection for anal sex as for vaginal sex? To answer that question, Dawn Smith of the CDC did an analysis of two studies from the late ’90s of HIV-negative MSM (men who have sex with men). (The study includes this disclaimer: “The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the CDC.”)

Smith looked at 4,492 MSM enrolled in an HIV vaccine trial, VAX 004. In that study, 48% reported anal sex with an HIV-positive partner during the study, and 7% acquired HIV infection. She also looked at 3,233 MSM from Project EXPLORE, a behavioral intervention study. In that study, 41% reported anal sex with an HIV-positive partner, and 5% acquired the virus.

Smith estimated the effectiveness of condom use by looking at men who reported they “always used” condoms for anal sex with an HIV-positive partner to those who “never used” them. After comparing the number of HIV infections in the two groups, she concluded that condoms were 70% effective for anal sex.

I had heard such low numbers quoted before – in the 1980s by conservatives like Jesse Helms, who wanted to show how ineffective condoms were. But those studies included people who were given condoms and never used them. Smith was saying that even when people said they used condoms every time, they offered only 70% protection for anal sex. Since her study was well done and carried great weight, I arranged an interview to ask her to address how the results were being misinterpreted. Here are some quotes from our talk.

Dr. Smith, your study is being cited all over the web as “The CDC reports that using a condom every time for anal sex is only 70% effective.” Is that what you meant to say?

It is correct that in our analysis, when condoms were reported to be used 100% of the time for anal sex, they were 70% effective.

But when people report that they used a condom every time, is that always accurate or is there over-reporting?

There may be over-reporting or underreporting. We don’t have a gold standard for condom use, so the best thing we can do is to ask people repeatedly over time about their condom use.

Some people are saying that your study says condoms fail 30% of the time.

That’s not what we said. This was not a study of condom failure. What we said is that if you compare people who report that they always use condoms for anal sex to people who report that they never use condoms, there are 70% fewer HIV infections among those who report they always used a condom. So our best estimate of condom effectiveness for anal sex is 70%.

But why do other studies of condom failure have rates that are so much lower?

Our study could include cases where the condom slipped or broke and the users were not aware of it. Nothing in life is 100%. It’s not surprising to me that we don’t find 100% protection from this or any other intervention.

I’m just trying to respond to people who say your study shows condom failure rates of 30%.

Then you’re talking to the wrong person. We never used the number 30%. If somebody out there is using that number, you need to talk to them about that. We said it was a 70% reduction in HIV incidence. You shouldn’t come up with another spin on what we said – you should use our language.

Condom Failure

I had hoped Dr. Smith would help me explain why her study was so different from what I had taught for years, but that didn’t happen. So I decided to look at the research again, this time focusing on studies of condom failure during anal sex. (“Condom failure” refers to how often the condom breaks or slips off during sex.) Could it be true that they fail much more during anal sex than during vaginal sex? Here’s what I found.

A 2011 online survey of 944 men conducted by the Rollins School of Public Health found a condom failure rate for anal sex of 4.1%. A Dutch study of 671 gay men found a failure rate of 3.7%, but when people used them correctly (with a water-based lubricant), that number dropped to 1.7%.

A study from London’s City University looked at 283 gay male couples who had been randomly assigned to use either standard or thicker condoms for anal sex. Each couple was provided with nine condoms and completed a questionnaire after each sexual act. But the study found no real difference between the two types of condoms. When they were used correctly, the standard condoms failed 2.5% of the time and the thicker condoms 2.3 %. It was the way they were used that determined whether they failed. The researchers recommended that gay men:

  • Unroll the condom on the penis, not before putting it on.
  • Use additional lubricant.
  • Apply lubricant to the outside of the condom only.
  • Apply lubricant in and around the anus.

The authors also noted:

Previous studies have consistently reported that some men report frequent incidents of condom failure whereas others report never having experienced failure at all…. Behaviors associated with condom failure include a history of failure, the infrequent use of condoms, engaging in anal rather than vaginal intercourse, not using lubricant consistently, the use of inappropriate (not water-based) lubricant, and the age of the condom used.

In fact, Dr. Smith’s analysis found a large difference between the two studies it included. While the behavioral intervention study showed that condoms were 86% effective, the vaccine study found they were only 61% effective. Why the large difference? Is it possible that men in the behavioral study learned better condom skills over time?

Whatever the reason, the take-away message is that condoms seldom fail during anal sex if they are used correctly. Avoiding air bubbles, unrolling the condom on the penis (not before putting it on), and using a lot of the right (water-based) lubricant are especially critical if condoms are to protect during anal sex. Once you learn how, it’s simple to use them correctly.


The studies on condom failure I found confirmed that the number I’d been using for years was right: when used correctly, condoms work close to 98% of the time, even for anal sex. So why was Smith’s study so different?

I decided to look at self-reporting. As Dr. Smith reported, the studies she analyzed asked people if they used condoms “always”, “sometimes”, or “never”. Was it possible that people weren’t being entirely honest, or misremembering how often they used condoms? As someone who’s taken part in several studies that asked specific details about my sex life, I know how hard it is to be honest when an interviewer is asking very intimate questions. Shame, embarrassment, and wishful thinking can definitely influence your answers.

And in fact, studies have found this to be the case. A 1995 study found that self-reported condom use was not associated with lower rates of STDs. A 2003 microbicide study stated:

Although most women believed they were accurate in their answers about condom use, a number of women admitted to exaggerating their condom use or believed other women did, citing concerns related to interviewers’ expectations of them as the primary reason.

A 2008 study of 186 young women who reported consistent condom use found that 34% of them actually had sperm in their vaginal fluid. A 2010 study looked at 11 surveys of female sex workers and MSM from five Central American countries and the Dominican Republic. People who reported always using condoms were asked to think again about the specified time period, and to confirm that there was no instance in that time period when a condom had not been used. The researchers found that in all the surveys, the number of people reporting consistent condom use decreased after follow-up questions were asked. Among MSM, the difference in self-reported condom use dropped as much as 37%. In one of the surveys, the number of MSM who reported using a condom every time dropped from 89% to 52% after re-questioning. The authors concluded:

Given the amount of recall bias and social desirability bias usually associated with condom use survey items, a measure that identifies a lower proportion of condom users than initially found is taken to be more accurate. The two follow-up questions examined in this study appear to substantially reduce the proportion of people claiming consistent condom use. As most behavioral surveys rely on self-reported measures, the addition of such questions could significantly improve estimates of consistent condom use.

Planned Parenthood addressed condom failure rates in a 2011 fact sheet:

High failure rates in some studies occur because many people over-report contraceptive use to shift the responsibility for an unintended pregnancy to a “faulty” contraceptive. Such over-reporting artificially inflates failure rates (Trussell, 1998). Condom failure rates are also inflated because some young people have been shown to inaccurately report condom uses, use condoms incorrectly, and respond to survey questions with what they perceive to be socially desirable answers (Rose, 2009). In fact, most people who use condoms do not experience breakage or slippage. Most condom failures occur among a minority of users because they are less experienced and/or less careful about using condoms than more successful users (Steiner, 1994).

In a study of [heterosexual] couples in which one partner was HIV positive, only one case of infection occurred among those who remained sexually active and used condoms consistently and correctly. In contrast, the incidence of HIV infection was 14% with inconsistent use (Deschamps, 1996). Another study found that among a group of couples who used condoms consistently, 2% of the uninfected partners contracted HIV over the course of the two-year study. This contrasts with 12% of partners who became infected in couples that reported inconsistent or no condom use (Saracco, 1993).

Other Concerns

Another important caveat to Smith’s study is that she only included men who reported sex with an HIV-positive partner. So a man who used a condom one time with an HIV-positive partner but never used a condom with 100 presumed HIV-negative partners would still have been included in the “always used” group. And in fact, the men who reported they always used a condom with HIV-positive partners said they were less likely to use a condom with partners they thought were negative. If those partners actually had HIV, that could account for some of the infections in the “always use” group, making the 70% number artificially low.

Finally, I looked at that 50% figure some gay men mention. That comes from a 2014 study by Robert Remis, which concluded:

Among Ontario MSM in 2009, an estimated 92,963 HIV-negative men had 1,184,343 episodes of anal sex with a condom and 117,133 anal sex acts without a condom with an HIV-positive partner. Of the 693 new HIV infections, 51% were through anal sex with a condom….

Really? That was far lower than even Smith’s study. Why such a dramatic difference? Well, it turns out that Remis’ study is simply a mathematical model. No one was actually enrolled or followed over time – it was all based on estimates. First, the authors estimated the number of MSM in Ontario, using the last census – but how reliable is that? Then, they estimated that the “annual number of anal sex acts per person” was 100 – seriously? Then they estimated that 91% of them used a condom for anal sex – once again, seriously? They then looked at the number of MSM with HIV in Ontario, and estimated that 51% of HIV infections occurred during anal sex with a condom. They also said that estimate could range from 7% to 77%.

The many different things estimated in this analysis make it highly questionable, and also account for the broad range of the estimate. For those reasons, I won’t be using it in any of my workshops.

Using the Damn Things

Of course, the real question is: will people use condoms? When it comes to gay men, the answer seems to be “no”. Smith’s study found that “only 16% of MSM reported consistent condom use during anal sex with male partners of any HIV status over the entire observation period.” Other studies have also found low rates of consistent use – an internet survey conducted by George Mason University asked over 14,000 MSM if they had used a condom the last time they had anal sex. Over 56% reported that they knew their partner had had sex with other men in the prior six months. In spite of that, condom use ranged from 26% to 56% – clearly not high enough to make a dent in the epidemic.

Even so, we need to get the facts out to young gay men about condom effectiveness, and we must respond to statements like this from a recent Facebook debate:

With high local prevalence and less-than-pristine condom use, condoms are very nearly worthless…. [Y]ou guys lied about that and obfuscated it for decades. It’s STILL the easiest fact to troll condom-nazis with, drives them NUTS. That remains proof that activists’ omissions, lies and distortions about condoms lasted far longer and embedded more deeply than any lies or distortions about PrEP.

Do younger gay men really think there was some kind of conspiracy of silence coordinated by safer-sex advocates to overestimate the efficacy of condoms? What would motivate gay men to lie to other gay men about something as critical as condom effectiveness? That’s the kind of malevolence we used to ascribe only to people like Jesse Helms. Should we now accuse each other of it?

We must educate young gay men about proper condom use. Perhaps one part of the problem is that gay men no longer get in-person condom demonstrations that were common in safer sex workshops in the ’80s and ’90s, and so never learn the steps so important to their correct use.

AIDS activists fought long and hard to get condom demonstrations in high schools. But do they still happen? The teens I’ve talked to say HIV education in high schools is sorely lacking – particularly when it comes to gay sex.

Another thing that has always bothered me is that I’ve never seen a porn actor actually put on a condom. There’s lots of bad dialogue and foreplay and then – cut – the condom is magically on!

Would it be that much of a turn-off to show the guy actually put it on? Think how many thousands of gay men could have learned proper condom technique in a fun and sexy way if that had been done for the last 30 years.

But if many gay men don’t use them, does it matter how well they work? Yes – it matters because we need more, not fewer, prevention options. Adding PrEP to the prevention arsenal is an important step that could dramatically change the course of the epidemic. But gay men need to know that even if they’re not on PrEP, it’s possible to protect themselves. Many men have done that for years. If you use a condom correctly every time you have anal sex, your chances of getting HIV are extremely low. And, unlike PrEP, condoms offer protection against other STDs, like hepatitis C, not to mention being far cheaper than Truvada, which can cost over $10,000 a year.


Here’s what I learned while researching this article:

  • The CDC has not said that condoms are only 70% effective for anal sex.
  • A CDC researcher reported that an analysis of two 1999 studies found 70% fewer infections when comparing men who said they used a condom every time for anal sex to men who said they never used them.
  • Studies have found that people often tell researchers they used a condom more often than they actually did.
  • Studies of condom failure rates have found that when used correctly, they break or slip about 2% of the time during anal sex.
  • The majority of gay men do not use a condom every time they have anal sex.

The reality is that the message “Use A Condom Every Time” has not worked, and is not going to work well enough to stem the epidemic. Fortunately, we have a new tool, PrEP, that (once again, when used correctly) may match condoms in effectiveness. Studies of microbicides are ongoing, and hopefully will show positive results soon. And intermittent PrEP and long-lasting injectable PrEP will soon be important additional options.

We should also address the aversion many men have toward condoms. For many gay men, skin-on-skin contact is an important part of sexual intimacy. But others just don’t like the way a condom feels. For them, finding the right lube is essential, as is choosing the right condom.

Polyurethane condoms, available since the ’90s, offer the advantage of allowing users to choose any lube they like – including oil-based products like hand lotion – removing one of the major reasons for condom failure.

So, are condoms over? No. The availability of new prevention options does not require us to abandon a tool that has been proven effective for decades. Let’s move forward in the search for new prevention options without bashing the ones we have – or those who promote them.

About the author(s)

Mark Milano is the Editor of Achieve, a publication of the AIDS Community Research Initiative of America (ACRIA).