2 ms·

The problem is that misdiagnosis is really easy. Lets take HIV for example. First symptoms of infection are usually a high fever and very sick -- like a reall

The problem is that misdiagnosis is really easy. Lets take HIV for example. First symptoms of infection are usually a high fever and very sick -- like a really bad flu (after this initial spike, you can go asymptomatic for years). So you go to the doctor with flu-like symptoms, and they give you an HIV test. Now if you haven't had unprotected sex or blood transfusion with an at risk partner the odds of you having contracted HIV is extremely low. So the general recommendation is not to test you for HIV. But you say, "I want to rule it out. Give me the HIV test." Modern medical testing is rarely 100%. The HIV test is 99% accurate, but 1% of the time you won't have HIV, but it will say you do. So imagine a population of 1000 zero risk people who take the test. On average 10 are going to come back as HIV positive. And maybe start an HIV regimine, rather than just realizing that its just the flu, which would have gone away in another day. The point is you try to focus on what is common first. You rule those things out first, before diving into the more obscure. This not only reduces healthcare costs, but also reduces the burden on the healthcare system, and likely improves the health of virtually everybody.

duncanj 15y ago
I understand what you're trying to say, but in your specific example, it's more like 15/1000 would be sent on for confirmatory testing (western blot), of which only about 1 in 10,000 would be false positives. In your scenario, this screening of those who request the test but otherwise do not need it would result in statistically very few false positives, on the order of 1 in 175,000 of those who make the request. Perhaps a more likely error would be among those who request testing for Lyme disease (currently a popular self diagnosis) and thus get put on an unnecessary regimen of doxycycline. I'm racking my brain trying to think of an example that would be much, much worse. Perhaps testing for antinuclear antibodies and getting diagnosed with lupus?
- kenjackson 15y ago
  Yes, I was trying to think of a "chronic" condition where a test result is sufficient to believe you have the illness (that is, the lack of symptoms is not sufficient). But I think we both agree that HIV is a generous case in that the testing is fairly accurate. There are many illnesses with no great tests. The more obscure, the more likely there won't be a test for it. So you end up in a situation where you go to the doctor with a rash and your diagnosis is "an allergy to strawberries or leprosy".
swampthing 15y ago
Ideally, if your HIV test comes back positive, you'd also be given information such as how likely false positives are and what the economic / health consequences of starting the HIV regimine (versus getting a second test) are. I understand that the way things are currently structured, it might be detrimental to the system as a whole to dive into the more obscure stuff. However, it's safe to say in this case, this boy suffered as a result. Ideally, there wouldn't be this tradeoff between the efficiency of the system and any one individual's health. I definitely don't think this ideal will be easy to achieve, or know what needs to happen to make it possible, but I think we should strive for it. I realize it's pretty lame to critique the system without offering a solution - but I felt compelled to post because the article struck me as an example of how the medical system can fail people, as much as it's an example of the power of FB, etc. (yet it didn't seem to strike people as such).
- kenjackson 15y ago
  Ideally, if your HIV test comes back positive, you'd also be given information such as how likely false positives Unfortunately, try telling someone that the ELISA test is 99% accurate to resasure them that they probably aren't HIV positive. :-) Ideally, there wouldn't be this tradeoff between the efficiency of the system and any one individual's health. If we could diagnose w/ 100% accuracy then it would be easy. If our medications had no side effects and resistance to antibiotics didn't then it would be easy. But once you put those two things together, even throwing economics aside, you have to be careful that you're not harming more people than helping, net. While FB saved the day here, I know from my doctor friends that online diagnosis in the hands of most people hasn't been net beneficial. With that said, in the hands of some (and I suspect that those of us on HN are in this minority) we can provide real value to doctors.
  - swampthing 15y ago
    If we could diagnose w/ 100% accuracy then it would be easy. If our medications had no side effects and resistance to antibiotics didn't then it would be easy. But once you put those two things together, even throwing economics aside, you have to be careful that you're not harming more people than helping, net. Well, in my last paragraph, I said it probably wouldn't be easy. Doesn't mean we shouldn't try. I admit, I have a bias towards personal choice - I suppose it's possible society would be worse off with more freedom in this case.

duncanj 15y ago

I understand what you're trying to say, but in your specific example, it's more like 15/1000 would be sent on for confirmatory testing (western blot), of which only about 1 in 10,000 would be false positives. In your scenario, this screening of those who request the test but otherwise do not need it would result in statistically very few false positives, on the order of 1 in 175,000 of those who make the request. Perhaps a more likely error would be among those who request testing for Lyme disease (currently a popular self diagnosis) and thus get put on an unnecessary regimen of doxycycline. I'm racking my brain trying to think of an example that would be much, much worse. Perhaps testing for antinuclear antibodies and getting diagnosed with lupus?

kenjackson 15y ago
Yes, I was trying to think of a "chronic" condition where a test result is sufficient to believe you have the illness (that is, the lack of symptoms is not sufficient). But I think we both agree that HIV is a generous case in that the testing is fairly accurate. There are many illnesses with no great tests. The more obscure, the more likely there won't be a test for it. So you end up in a situation where you go to the doctor with a rash and your diagnosis is "an allergy to strawberries or leprosy".

swampthing 15y ago

Ideally, if your HIV test comes back positive, you'd also be given information such as how likely false positives are and what the economic / health consequences of starting the HIV regimine (versus getting a second test) are. I understand that the way things are currently structured, it might be detrimental to the system as a whole to dive into the more obscure stuff. However, it's safe to say in this case, this boy suffered as a result. Ideally, there wouldn't be this tradeoff between the efficiency of the system and any one individual's health. I definitely don't think this ideal will be easy to achieve, or know what needs to happen to make it possible, but I think we should strive for it. I realize it's pretty lame to critique the system without offering a solution - but I felt compelled to post because the article struck me as an example of how the medical system can fail people, as much as it's an example of the power of FB, etc. (yet it didn't seem to strike people as such).

kenjackson 15y ago
Ideally, if your HIV test comes back positive, you'd also be given information such as how likely false positives Unfortunately, try telling someone that the ELISA test is 99% accurate to resasure them that they probably aren't HIV positive. :-) Ideally, there wouldn't be this tradeoff between the efficiency of the system and any one individual's health. If we could diagnose w/ 100% accuracy then it would be easy. If our medications had no side effects and resistance to antibiotics didn't then it would be easy. But once you put those two things together, even throwing economics aside, you have to be careful that you're not harming more people than helping, net. While FB saved the day here, I know from my doctor friends that online diagnosis in the hands of most people hasn't been net beneficial. With that said, in the hands of some (and I suspect that those of us on HN are in this minority) we can provide real value to doctors.
- swampthing 15y ago
  If we could diagnose w/ 100% accuracy then it would be easy. If our medications had no side effects and resistance to antibiotics didn't then it would be easy. But once you put those two things together, even throwing economics aside, you have to be careful that you're not harming more people than helping, net. Well, in my last paragraph, I said it probably wouldn't be easy. Doesn't mean we shouldn't try. I admit, I have a bias towards personal choice - I suppose it's possible society would be worse off with more freedom in this case.