Testing should be pretty good. You can look up the "sensitivity" of a test (the rate at which it detects the target) and the "specificity" (in effect how good it is at avoiding false positives). As long as the samples are gathered right the tests are probably good.

Note that false positives do happen. Roxanne Hall had a false HIV+ result years ago, and I think Tiana Lynn had a false hep-c result last year. The tests are intended to be overly sensitive; you can always fix a false positive with a re-test, whereas a false negative might be far more serious.

I have found one estimate at the CDC that suggests condoms are 95% successful at protecting against HIV. I haven't looked for equivalent estimates on other STDs.

It would be nice to see AIM testing of other STDs for surveillance if nothing else - is there a widespread problem with various STDs or not?