Surgeons have an encouraging term for their mistakes: "never events," because they should never happen. Yet they do, with startling frequency.

At least 4,000 times a year in the United States, surgeons leave a sponge or instrument inside a patient, perform an incorrect procedure, or operate on a wrong body part or even a wrong patient, according to a recent study from the Johns Hopkins University School of Medicine.

That estimate is probably low, based as it is on malpractice claims, because many errors don't turn into legal actions. Some go undetected altogether.

In one in 15 cases, the mistake leads to a patient's death.

Why can't preventable mistakes be prevented? Studies have long established the value of operating-room checklists and other simple strategies such as marking the surgery site with indelible ink. Evidently, they aren't universally employed. Some surgeons may think that they take too much time, or that it's enough to rely on their own caution and competence.

To change that, what's needed is an accurate, public accounting of surgical errors. As Martin Makary, the leader of the Johns Hopkins study and the author of the new book "Unaccountable," said, "To fix a problem, you need to measure it."

Hospitals' own voluntary reporting systems miss most errors - in fact, almost all of them, according to a study of "adverse events," a term that also includes unanticipated but less easily preventable problems such as infections. Software from the federal Agency for Healthcare Research and Quality can screen patient discharge records for "patient safety indicators" (complications), and states and Medicare and Medicaid use it to gauge hospitals' performance. This approach isn't much better, though: It fails to account for nine out of 10 adverse events.

Far more effective, the study found, is third-party review of medical charts. The idea is to find "triggers" that indicate something went wrong - a change in surgical procedure, say, or a postsurgical patient's being returned to the operating room or sent to the intensive-care unit. When a trigger is spotted, the case is investigated. In the adverse-events study, the Institute for Healthcare Improvement's Global Trigger Tool detected more than 90 percent of hospitals' adverse events.

This approach can require time and effort, but in hospitals with digital patient records, it can be programmed to work automatically. Several large hospitals have already begun using the trigger tool this way. It's fast and accurate, according to David Classen of Pascal Metrics, a patient safety organization based in Washington, who led the adverse-events study.

Some states require hospitals to report their mistakes. That's good, but a national reporting system would be better. All hospitals should report using the same standards of data collection and the same definitions of procedures and errors.

With public accounting, hospitals would be answerable for patterns of recurring error. Consider: Of the 9,500 doctors in Makary's study who had been involved in one surgical mistake, 12 percent had at least one more. Hospitals would have a new incentive to improve by, for example, mandating checklists, instituting safety training and making sure the workplace culture supports members of surgical teams who speak up when they see something amiss. Lack of good communication is a root cause of surgical never events, the Joint Commission, a nonprofit health-care accreditation organization, has found.

Finally, accurate reporting could bolster the Centers for Medicare and Medicaid Services' strategy of denying reimbursement for costs due to errors. In 2008, it stopped paying for incidents in which a foreign body is left in a surgical patient. Almost immediately, reported cases dropped by half, according to an article in the Jan. 23 Journal of the American Medical Association. Unfortunately, under the current arrangements fewer reports needn't mean fewer incidents.

Considering that medical errors of all sorts - inside and outside the operating room - cost Medicare alone an estimated $17.1 billion a year, making them rarer would not only prevent injury and save lives. It would also put a significant dent in health-care costs.