Study shows Mexican migrants have high rates of HIV

[question everything]

Tuesday, June 14, 2005

Study shows Mexican migrants have high rates of HIV

By: LAURA WIDES - Associated Press

LOS ANGELES ---- Mexican migrants are infected with HIV at a higher rate than the general population in California and Mexico, according to two studies released Monday.

The findings marked a significant change from past research that found migrants engaged in high risk behavior but were unlikely to contract the disease.

A study of Mexican migrants in California found that 0.6 percent were infected with HIV. That compares to numbers released Monday by the CDC, which put the prevalence rate as of 2003 at around 0.4 percent in the general U.S. population.

More..
http://www.nctimes.com/articles/2005/06/14/news/state/61305204907.txt

Universitywide AIDS Research Program: http://uarp.ucop.edu/

[mbperrin]

1. For more slurs about various immigrant groups,

visit

http://gyral.blackshell.com/names.html

Lessee, diseased, retarded, unhygienic, criminal...and that's just the IRISH!

For a nation of immigrants, we have very short memories...

[Igel]

2. Your post confuses and concerns me.

Does it mean that the Mexican migrants can't possibly have higher rates of HIV infection?

Or does it mean that we shouldn't do the research to find out?

Or, if they do, we shouldn't bother to say it in polite company? In that case, do we allow it to be mentioned at conferences and in journals that the MSM might possibly peruse? Do we allow the research community to discuss it? Or do we expect the researchers that produce the number try to deal with it?

Of greater concern is not reporting the confidence interval: I assume 95%, but the number of positives was low, and the methodology wasn't described.

And of possibly even greater concern is this paragraph:
"A separate study of migrants in their home states in Mexico found the rate of infection was even higher, at 1.1 percent, compared to 0.3 percent among the general Mexican population."

[HeeBGBz]

3. So does that mean they became infected while they were here?

Then taking it home to Mexico.

[mbperrin]

4. Their post is the one that's confusing...

as you point out, no methodology, no confidence interval, nothing, just the blanket assertion of the higher rate, and the idea of studying them in their hometown is even weirder...are Mexicans in their hometown migrants?

I was just pointing out that historically, this country has vilified the most recent immigrant groups as a way to keep their wages low and expectations down, and then when we need another wave from somewhere else, they become the new "dirt".

My German/Scot/Dutch ancestors came to this country, but only 1 of my 4 grandparents' families was documented, so I'm not sure that we'd have been welcomed at that time either...this is what I mean by a short memory...but blaming Mexican migrants for the HIV pandemic is just typical, typical, typical...

Not against studies, but give me real ones with real numbers, methods, etc...

[DFWdem]

5. Well, you could click the link in the article

The article has a link to the Research Program where they posted the testing methodology and results. Generally speaking, a newspaper produced for mass consumption by an entire city is not going to detail the testing methods used by the statisticians. Rather, they will summarize the results into layman's terms.

http://uarp.ucop.edu/ca_collaborations/modules/dissemination.html

Try that link

[KurtNYC]

6. Sample size was 791

so the margin of error is high. It says only that "five people tested HIV positive" so if, in this limited sample, they had had two fewer HIV+ persons they would have gotten a number equal to the average for the USA.

5 / 791 = .006 or .6%
3 / 791 = .0039 or .4%

Heck if they had only 2 HIV+ they could have written an article explaining how mexican migrants have a lower rate of HIV than Americans. Of course they wouldn't write THAT article.

[redqueen]

7. Of course they didn't...

That would be responsible reporting. Can't have that...

[DFWdem]

9. M.O.E. not necessarily high

A sample size of 791 isn't too small, statistically speaking. You can usually get pretty accurate results (90% - 95% C.I.) with a sample size of 40 or more. Do you honestly think that professional statisticians from both California and Mexico would produce a study like this based on a sample size that wouldn't yield at least a 90% C.I.?

[KurtNYC]

10. A 3% margin of error is considered reasonable for most surveys

And a sample size of ~800 should produce results within 3% MOE but this study looks for the incidence of something which is found in 0.2 to 0.3% of the population. As I showed in the math above, the inclusion or exclusion of even 1 person with HIV skews the number greatly on a relative basis.

In other words, if you accept their numbers then the incidence of HIV antibodies is 50% higher in migrant populations compared to the USA in general (0.6% versus 0.4%) but a 3% margin or error means the real number within the migrant population could be as high as 3.6% or as low as 0.0001%. The sample size is not large enough to be statistically valid since what we are looking for has an incidence smaller than the margin of error.

Would statisticians produce such a study? Well, it wasn't one study. Whoever composed the article compares numbers from several studies and sources which used differing methods to acquire their samples. It is likely that George Lemp knows the sample size is small but he addressing issues in a population that doesn't get much attention -- he will make do with what he has. I blame the reporter for skewing this into something of significance. I see people make the unjustified and totally erroneous conclusions from data all the time. Sometimes this is intentional as it was when the food lobbyists said last month that being overweight reduces your "chances of death" (? - whatever that means). They were actually looking at the weights (BMI) of dead people and many people lose significant amounts of weight in the end stages of terminal conditions.

And the headline -- "Study shows Mexican migrants have high rates of HIV" is unsupported by the data. 4 in 1000 in said to a "high rate"?! How then would they describe the incidence among pregnant women in Uganda -- 300 in 1000?

[DFWdem]

11. 3% of which number?

3% of 6 = .18, so the number of migrants with HIV will fall into the range, 5.82 - 6.18, 97% of the time

[Laughing Mirror]

8. Mexican migrants are desperately poor people

and poor people have much worse health than the general population.

It should be noted however that the rate of HIV in the US general population is 0.4, while in Mexico, it's lower: 0.3. As categories go, the HIV rate among blacks in US may be as high as 2 percent, not much better than the rate attributed to the Mexican migrants.

http://www.cbsnews.com/stories/2005/02/25/health/main676628.shtml

The HIV infection rate has doubled among blacks in the United States over a decade while holding steady among whites — stark evidence of a widening racial gap in the epidemic, government scientists said Friday.

Other troubling statistics indicate that almost half of all infected people in the United States who should be receiving HIV drugs are not getting them.

...

The surveys look only at young and middle-aged adults who live in households, excluding such groups as soldiers, prisoners and homeless. Thus, health officials believe the numbers probably underestimate true HIV rates in this country.

Still, they show a striking rise in the prevalence of the AIDS virus from 1 percent to 2 percent of blacks. White rates held steady at 0.2 percent. Largely because of the increase among blacks, the overall U.S. rate rose slightly from 0.3 percent to 0.4 percent.

[Chauga]

12. Delightfully racist freeper thread on this story.

http://www.freerepublic.com/focus/f-news/1422793/posts