The problem with medical tests.

Here's why you shouldn't put much stock into medical tests.

Example 1:

Let's say that you are an average couple and you want find out if your child-to-be will have Down syndrome. The doctor tells you that there is a test with 95% accuracy. You take the test, and it's positive for Down syndrome. Now, how worried should you be?

To begin with, the chances were approximately 1 and 500 that your child has Down syndrome. This means that out of every 1000 people born, 2 will have Down syndrome.

Now let's give these thousand people the test mentioned above, which is 95% accurate. It turns out that 50 will be tested positive for Down syndrome. But we've just seen that only 2 actually have Down syndrome. This means that they're still only a one out of 25 chance that your child has Down syndrome, which is only 4%.

In other words, if you are a typical pregnant couple, and you take a medical test for Down syndrome which is 95% accurate, and you test positive for Down syndrome, there is a 96% chance your kid does NOT have Down syndrome. Does that make sense?

Example 2:

Let's say that you already have a child with Down syndrome, and therefore the chance is 1 in 100 that your next child will have Down syndrome. You take the test mentioned above and it comes out positive.

(This means that for every 100 people tested, there will be five false positives, and one true positive. This means your chances of having a child with Down syndrome would be one out of five, or 20%.)

In other words, if you are a couple in a high-risk category, and you take a test for Down syndrome which is 95% accurate, and it comes out positive, there is an 80% chance your child will NOT have Down syndrome.

Now how valuable do these medical tests sound?