HIV statistics and do people actually understand them.

[broncofan]

Without knowing much about statistics but as someone who has tried to look at these numbers, I prefer looking at statistics that reflect the risk that inheres in the act itself. So I will look at numbers for serodiscordant couples (where one partner has hiv and the other does not), and know that it is providing a high water mark. Therefore, if I get a number for bareback or condom sex, I am getting a number that is based on the assumption that the other person has hiv.

Even with these statistics, there is obviously variation within the samples. Some people have rough sex, some people don't. Some people who use condoms use them well, some people don't put them on well. But it makes it less likely I'm deluding myself about my partner's status and just have a number that pertains to the relative risks of each act.

The ultimate question though is, how risky is the act and how likely is your partner to have hiv. The latter question really does depend a lot and you have to use your judgment. But it helps to get numbers overall to be able to anchor some sort of estimate. If you are a man sleeping with a man, what are the percentages for men who sleep with men. If you are sleeping with a ts, then what are those overall. Then you make adjustments from there for your own use and edification. But with more specific cohorts you can get more relevant data but eventually you get so specific you run out of data.

[Ts RedVeX]

One thing that might be worth mentioning here is that statistics are only numbers and it does not matter whether your probability is 0.99 or 1^-100, you may still be unlucky and get HIV during your first encounter, or have bareback sex with an HIV positive person 99999 and still remain healthy. So keep in mind that whether you are taking the statistic for an "average homosexual man" or "average heterosexual woman" you are talking about an abstract entity and neither of those "average persons" exist in reality.

[broncofan]

And I do appreciate that the above method doesn't have the rigor of the calculations on the first page. But for someone who knows his status and is concerned about a single act, I look at these numbers afterwards and this is the sort of thing I think about. I personally like to keep separate the risk in the act and the risk that the partner is infected, but what you two are discussing on the first page is obviously the way to go about it rigorously.

[trish]

Without knowing much about statistics but as someone who has tried to look at these numbers, I prefer looking at statistics that reflect the risk that inheres in the act itself. So I will look at numbers for serodiscordant couples (where one partner has hiv and the other does not), and know that it is providing a high water mark. Therefore, if I get a number for bareback or condom sex, I am getting a number that is based on the assumption that the other person has hiv.

I think what RedVex refers to as g (the probability of acquiring the HIV virus upon having a single bareback receptive encounter with an HIV infected partner) is an example of the sort of numbers you’re talking about. These are probabilities of HIV transmission per exposure with the type of exposure specified (e.g. insertive anal with condom, or receptive oral without condom etc.). The chart she linked, I presume, gives a list of such stats. These probabilities tend to look small to the inexperienced eye (usually below a tenth of a percent). Which is why some people like to look at what happens to the probability of transmission as the number of exposures increases. That’s all we’re doing on the first page. The formula we’re using is practically as old as probability theory itself. Nothing fancy. Everything depends on the inputs - the transmission rate per exposure (which is the number I think you were referring to) and the probability that your partner carries the virus (which is the number RedVex denoted by the letter x).

An aside: I’m not sure the value that RedVex uses for x is the correct one, but let’s go ahead and use it. Suppose Mr.K sees an escort, say MissMystery, a few times every year. Suppose MissMystery has had receptive anal once daily since she started her escorting business. Given the initial inputs for g and x suggested in Post#1, RedVex and I both calculate that the probability that MissMystery has acquired the HIV by the end of her first year in business is 0.02 (rounding up). If Mr.K started seeing MissMystery just after her first year in business, then this should be the value he gives to the number x; i.e. it’s his probability of having sex with an HIV infected partner. Notice it’s considerably larger than the value of 0.00676 that MissMystery assigned to x. The moral here is that all these values (of x and g) change with time, location, and from person to person.

[Ts RedVeX]

Yeah. They depend on a load of factors. However, your x should only have x=0.02 as Trish suggested if MissMystery has not gotten tested recently and proven to be HIV-negative, in which case you would be able to safely use the original statistics provided those have not changed over her 1 year of work.

This is can be explained with the "gambler's fallacy" where the gambler thinks that if he has a "lucky strike" then he is more likely to lose in the following go or vice versa. In this case, if an escort gambling with her health for 1 year gets tested and proven HIV-negative, then your x=0.00681 again.

As to the g, you should probably just look up some UK speciffic stats and determine whether you actually want to get fucked bareback, as that way her using the condom does not make much of a difference in your sensations, and you might use a lower probability factor here.

[trish]

Yeah. They depend on a load of factors. However, your x should only have x=0.02 as Trish suggested if MissMystery has not gotten tested recently and proven to be HIV-negative, in which case you would be able to safely use the original statistics provided those have not changed over her 1 year of work.

Very true. Indeed, if MissMystery is the only person MrK's having sex with then the value of his x is zero. At least it starts out being equal to zero. Note too, that if MissMystery gets checked once per year, then by the end of the year MrK's value of x is back to 0.02 again. Actually MrK's x increases with each of MissMysteries exposures until she has her next 'all clear' whenever that is.

[filghy2]

Since I work in the sex industry and I often hear people saying bad things about barebacking, yand then I also see many other people not being concerned about contracting HIV too much, I find it interesting what makes all those people think so differently. This thread is by any means not meant to encourage anyone to bareback or conclude whether barebacking is good or bad. It is supposed to be more about understanding information provided by media correctly as well as how much the way of looking at the info changes one's personal the feeling about it.

Lemme know what you think, especially if you reckon that this approach is wrong.

But why are you trying to reinvent the wheel when it has already been invented? The probability you are trying to calculate is for unprotected receptive anal sex, partner unknown status, which is given in the same table as 0.27% (1/370). The probability of remaining uninfected after N sexual encounters is therefore (1-0.0027)^N.

The point your calculation misses is the one I already explained in the other thread, which is that if a person is willing to have unprotected sex then that tells you something important about them - namely, that they are more likely to have engaged in risky behaviour in the past. So it is not like a random draw from the gay population - the risk will be significantly higher. That means it is not possible to derive the probability in the way you have tried to do. In technical terms, this is a conditional probability, , where the probability of having unprotected sex with someone and the probability of them being infected are not independent. https://en.wikipedia.org/wiki/Conditional_probability Multiplying probabilities is only valid when the events are independent.

I haven't closely examined the studies underpinning the published probability estimates, but I understand they are based on the sexual experiences of a sample of the population that is designed to be representative. That means they should take account of the more risky behaviour of people having unprotected sex. There are lots of challenges in getting this right, but it is still likely to be more accurate than any 'back of the envelope' calculation. Also, the published probabilities are derived from a range of studies, so they don't depend on one particular study.

[davejones65]

Anything from the government, wikisleezia and the cdc cant be trusted and is a lie.

[buttslinger]

I don't do statistics, and I've said it here before, that I recall a report that in Washington DC, something like 98% of AIDs cases were amongst the class of people making less than 10,000 dollars a year.
One time I was completely smashed and following a prostitute down the hall of an abandoned hotel where all the nice old wooden doors had been removed, so that each room I gazed into was like a level towards Dante's inferno, people engaging in every conceivable variation of IV drug or sexual risky behavior. It was like a movie. The height of the crack epidemic.
I remember she stole my wallet.
I missed AIDs but I did run into some Hep-C that my body fought off.
My crack dealer had AIDs, her name was Dianne and her only talent was knowing where the best rock in town was and getting me to pay for it.
She got murdered before she was scheduled to die in jail.
Down and Out people have fantastic sex lives if they're young.

[Post Op Preferred]

The CDC recently published a summary of a meta study based on 9 million HIV tests that it funded, and analyzed. 2.7% of self identified MtF's were HIV positive (one tenth of previously published estimates, as noted in the briefiing). 0.09% of men were HIV+. https://www.cdc.gov/hiv/pdf/policies...nder-brief.pdf. the underlying meta study, https://www.cdc.gov/mmwr/volumes/66/wr/mm6633a3.htm, takes note of its sample size, bias and geographic limitations, but it provides a useful data point for the missing "x."