Calculated Risks: How to know when numbers deceive you...

[Q3JR4]

I finished this book by Gerd Gigerenzer a month ago, and I thought I'd share one of the pieces of information presented in the book.

On pages 231-232, the book has two paragraphs that I found interesting...

‌In April 1995, a previously healthy 36-year-old American construction worker suffering from fatigue was tested for HIV ^{Reimer et al (1997)}. The ELISA, a test for HIV approved by the U.S. Food and Drug Administration, returned a negative result. Two months later, the man, who had lost 27 pounds, was admitted to a hospital with shortness of breath, diarrhea, and other symptoms. A second ELISA test indicated that he was HIV-negative, and routine laboratory tests turned up no other illnesses. The patient was discharged without a diagnosis. In August of the same year, the man was hospitalized again. An ELISA and a Western blot performed by the Utah Department of Health laboratory came back negative. At that point, the physicians decided to interview his wife, from whom he had separated two years previously. She reported having had sexual contact with an HIV-infected partner before their marriage and told the doctors that this earlier partner had recently died of AIDS. In 1994, she developed pneumonia and tested positive for HIV, a fact of which the construction worker was unaware. He reported that during their marriage, he had had sex with his wife without using a condom, but he had had no sexual contact with her since their separation. Because of his history of exposure and the strong clinical evidence that his immune system was compromised, the doctors performed a series of additional laboratory tests, which eventually revealed that he had the same strand of HIV as his ex-wife.

During his examination, the construction worker made it known that he had--in good faith--donated blood more than 30 times in the previous four years. In each case, the routine ELISA--which is used to screen all blood donations for the HIV virus--returned a negative result. The consequences for those who received the construction worker's blood are unknown. This HIV-infected man had tested negative for HIV no fewer than 35 times over a four year period.

Q3JR4 says:
I have been, so far, unsuccessful in finding corroborating evidence for the existence of this man. So to my fellow DUers I issue a challenge. The passage I've given you contains all of the relevant citations placed at the same place they were in the book. If you can find the original article or other corroborating evidence, I will personally do SOMETHING OF EXTREME IMPORTANCE in your name. Also major kudos will go your way. :)

Q3JR4
Disclaimer:
This case took place in 1995. Refinements to both the Western Blot test and the ELISA have undoubtedly occurred since then to decrease the risk of false positives or false negatives. Unfortunately we cannot eliminate the risk of both, so the possibility does, however, exist that someone out there will be tested and receive one or the other. Though we cannot guard against either case, what we can do is work to minimize our risk. That means that you all should do whatever it is you want to, but use the correct kind of protection that is relevant to your particular situation.

Also, nothing says lovin' like, "Honey I love you but please wrap it up first."
Also also, "Entry by invitation only, proper attire required."


Edited to add: Damn spelling errors. Good thing most of you are asleep.

[fasttense]

1. Here is a study describing him.

In October 1995, the Utah Department of Health referred to CDC blood samples obtained from a man who had had onset of persistent fatigue and malaise during January 1995. During January-June 1995, he had sought medical care at several clinics. When he was admitted to a hospital in June because of respiratory illness and recent weight loss of 27 lbs, HIV-EIA was negative. In August, he was admitted with lung-biopsy-confirmed Pneumocystis carinii pneumonia (PCP) and a CD4+ count of 129 cells/uL; an HIV-EIA again was negative. The patient reported frequently donating plasma at a plasmapheresis center from August 1990 through April 1994. Review of records at the plasmapheresis center identified 33 donations by the patient. At the time of each donation, testing on an aliquot of the donated plasma was negative by HIV-EIA (Table_1).

http://www.cdc.gov/mmwr/preview/mmwrhtml/00040569.htm

I hope this helps.

[Q3JR4]

3. Sweet! You rock.

Changed my sig line to reflect this. :)

Q3JR4.

[fasttense]

5. To tell you the truth, it really was simple.

I knew the CDC would have a write up of this case, if it was true. So I simple went to the CDC website and did a search.

But the facts of the case are very disturbing. Getting blood is a much more risky proposition than I thought. And to think people may be suffering from aides and not know it.

Another reason to jettison a cheating partner. The diseases they bring home may kill you.

[Odin2005]

2. Yikes, that's scary!

:scared:

[Q3JR4]

4. Yep.

But as the book likes to point out again and again, nothing is certain.

Q3JR4.
Except death, and lottery...assuming you never play.

[Q3JR4]

6. As with anything, you can never be 100% certain

about what you're getting. You walk outside your house there's a risk that a falling bullet, or a meteor will kill you. You drive your car, walk, ride a bike there's a risk you won't come home again. Around 267 people die each year due to grizzly attacks, and around 38,000 die from bathtub related injuries.

We shouldn't throw the baby out with the bathwater because HIV tests are still the current best way we can know about the health of those we plan to be with (and also our own health). Just have to be aware that even if the test comes out negative (or positive) there is a small change that the test could be wrong. That's why they recommend that people get tested every 6 months, and they check a positive oral test with a blood test and then another blood test to check THOSE results. If you test positive with a single blood test then they conduct a blood test to confirm the results. Each of the tests in a chain have to come from a different manufacturer.

Finally, this is the ONLY case I can cite where this has happened. That's one case in literally millions of HIV tests. I think the vast majority of Americans who get tested for HIV have nothing to worry about. There is always the chance that something's off, but that chance is low enough that I'm not going to think twice about getting tested in the future (provided I ever find another willing, and single partner).

Q3JR4.