"If medical errors were a disease, they would be the sixth leading cause of death in America—just behind accidents and ahead of Alzheimer's."<p>This is wrong, it's <i>at least</i> the third leading cause of death. And realistically if you don't smoke and you exercise once in a while, medical 'oopsy daisies' are probably your #1 risk of death:<p>C.f. The IOM's report To Err is Human: <a href="http://wps.pearsoneducation.nl/wps/media/objects/13902/14236351/H%2007_To%20Err%20Is%20Human.pdf" rel="nofollow">http://wps.pearsoneducation.nl/wps/media/objects/13902/14236...</a><p>And then also read Lucian Leape's commentary in JAMA on why it is probably a huge underestimate: <a href="http://jama.jamanetwork.com/article.aspx?articleid=192842" rel="nofollow">http://jama.jamanetwork.com/article.aspx?articleid=192842</a><p>(Leape is one of the authors of the original IOM report, and the report cites a lot of his own research, including his study estimating that only 5% of medical errors are ever discovered.)<p>C.f. also Barbara Starfield's estimate in JAMA: <a href="http://jama.jamanetwork.com/article.aspx?articleid=192908#REF-JCO00061-1" rel="nofollow">http://jama.jamanetwork.com/article.aspx?articleid=192908#RE...</a><p>Realistically, all of these estimates are at the low end. In fact most of them only count deaths in hospitals (and severely undercount them), when at least half of all medical errors are thought to happen at local doctors offices, plastic surgery clinics, nursing homes, etc.<p>C.f. also the CDC figures for Hospital Acquired Infections, which IIRC the IOM report doesn't count as being medical errors:<p><a href="http://www.cdc.gov/HAI/pdfs/hai/infections_deaths.pdf" rel="nofollow">http://www.cdc.gov/HAI/pdfs/hai/infections_deaths.pdf</a><p>My personal rule of thumb is to avoid taking any drugs or getting any non-trivial medical procedure unless I've read at least three books about it. The problem is that all the papers published in medical journals are basically complete bullshit (except the occasional well-done NIH ones), which means if you want real information about a given drug then the only way to get it is to subpoena the FDA, or read a book written by someone who has.<p>The way drug approval works is that you need 2 tests demonstrating that the drug is better than a placebo, but you're allowed infinite tries to get there. So often a drug will be better than a placebo in only 2 out of 10 trials, but it will still get approved and only those two trials will get published in medical journals. And then those journal articles will have very little in common with the actual data from the FDA trials because the pharma companies completely spin it, which is why the vast majority of the most popularly prescribed drugs are not only no more effective than placebos, but in fact significantly worse when you look at the total quality/length of life.