I took a metaphorical stroll the other day through a funhouse of mirrors, or maybe a better analogy would be watching while Alice grew and shrank as she sampled different "Drink Me" bottles. It had to do, first, with attempts to put a number on surgical mix-ups—operating on the wrong patient, wrong side of the patient, or wrong part of the patient, not to mention performing the wrong procedure—and second, whether the number even matters.
My attention was caught by a letter in the November issue of the Archives of Surgery from Richard Croteau, a physician (and former NASA rocket scientist) who is in charge of patient safety programs at the Joint Commission International Center for Patient Safety. His letter, which had a rather discouraged tone, was a response to a study about those OR nightmares that had concluded there may be at least twice as many as are reported, and maybe 20 times as many—probably at least 1,300 to 2,700 annually, or roughly 1 for every 30,000 to 60,000 surgeries. The question posed by the study's authors was whether such problems can be prevented, their point being that it's harder to do if there are twice as many (or more) than anybody thought.
Croteau called the errors "devastating." They "are completely preventable and should never happen," he wrote. And yet, he noted, the number of such incidents reported to the Joint Commission has actually been rising since mid-2004, when a new rule took effect requiring surgeons to run through a careful checklist before they make the first incision. In a written response, the researchers replied that many hospitals clearly were not following the rule, called the "universal protocol." Moreover, they said, there may be as much as 98 percent underreporting of these incidents; the commission has no enforcement authority except for its power to withhold accreditation, and it can't do that if it doesn't find out about a problem—which it won't if an event is never reported.
The final twist is that not many months before the study appeared, the very same journal had published another analysis of wrong-site surgeries. That one showed that such incidents occur about once in every 115,000 operations. Which do you believe? And how much difference does any particular figure make if the sticking point is how to get hospitals to change?