When medical errors kill


American hospitals have a big problem with unnecessary deaths from medical errors. Estimates of the numbers vary widely, but extrapolating from the best studies, a conservative estimate would be that well over 100,000 people a year die unnecessarily because of errors made by their healthcare teams. And the numbers have remained high despite concerted efforts to bring them down. Why? Because we’ve embraced a so-called solution that doesn’t address the problem.

For the last 14 years, the medical profession has put its faith in a systems approach to the problem. As applied to medicine, systems is a way of standardizing routine aspects of patient care: giving every heart attack patient an aspirin upon entering the hospital, say, or administering antibiotics just before surgery. In hospitals across the country, multiple check lists have been put into place in an attempt to ensure that all members of the treatment team follow accepted protocols.

Reliance on these kinds of systems dates back to 1999, when the Institute of Medicine, a private advisory group to the public and Congress, issued a highly influential report, “To Err Is Human.” It asserted that if the practice of medicine was made more systematic in all hospitals, the number of unnecessary deaths would be cut in half within five years.


The concept was based on the success of such an approach in the field of anesthesia decades ago, but it had been insufficiently tested in medicine overall. And today, despite a widespread embrace of systemized medicine in hospitals across the country, the number of unnecessary deaths hasn’t dropped significantly. Just last week, the New England Journal of Medicine published a study from Ontario, Canada, that compared rates of mortality and complications in Ontario hospitals before and after they began using surgical safety checklists. The conclusion? “Implementation of surgical safety checklists in Ontario, Canada, was not associated with significant reductions in operative mortality or complications.”

There’s a simple reason for that: Most preventable mishaps in hospitals are caused by the acts of individual practitioners, not flawed systems, and there was plenty of evidence of that fact available when the committee wrote “To Err Is Human.”

In 1991, for example, a Harvard Medical Practice Study examined more than 30,000 randomly selected records from 51 hospitals. A table in that study attributed some 61% of harm to patients to either errors of technique (during surgeries and other procedures) or to a failure of doctors to order the correct diagnostic tests. These are both errors of individuals, not systems. The same study found that only 6% of adverse events were due to systems problems.

And studies have continued to draw similar conclusions. A 2008 analysis of 10,000 surgical patients at the University of South Florida found that, of all the complications among those patients, only 4% were attributable to flawed systems. The rest resulted from individual human shortcomings. A 2013 study from Baylor College of Medicine on diagnostic errors found that most failures to diagnose arose from deficient physician performance during doctor-patient interactions, including poor history-taking, inadequate physical examinations or ordering the wrong tests. The study suggested that systems remedies, such as checklists and electronic medical records programs, would not avoid diagnoses missed in these ways because the problem is a cognitive one on the part of doctors.

By 2006, more than 3,000 American hospitals, representing 78% of the acute care beds in the country, had enrolled in a systems-based program called the 100,000 Lives Campaign, an outgrowth of recommendations in “To Err Is Human.” To see what effect that was having, a Harvard-Stanford group examined patient harms in hospitals in North Carolina, the state with the highest rate of enrollment in that campaign, at 96%. As the authors of the 2010 study put it: “We chose North Carolina as a site that was likely to have improvement, since it had shown a high level of engagement in efforts to improve patient safety.” But the authors found no reduction in preventable patient injuries between 2002 and 2007, the period when the systems method was being rolled out. At the same time, a report from Medicare found that the preventable death rate among hospitalized seniors was at least as high as it had been prior to “To Err Is Human.”

The major studies of what causes preventable errors have mostly failed to examine whether some doctors had a disproportionate number of bad outcomes. That failure has caused researchers to miss the most significant source of patient harm: incompetent doctors. But a few studies have made the problem clear. Dr. Robert Oshel, who was chief statistician for the National Practitioner Data Bank, examined medical malpractice suits nationwide and found that 2% of American doctors were responsible for 50% of the payouts over a 20-year period. And an Australian group examined 19,000 complaints against doctors over a 10-year period and found that 3% of Australia’s doctors accounted for 49% of complaints.


The institutions meant to protect patients from inept physicians are not doing an adequate job. Not convinced? Consider these facts:

• The average American hospital revokes the privileges of one doctor every 20 years.

• Only 250, or 0.04%, of the nation’s 650,000 physicians lose their licenses annually. At that rate, it would take 50 years to remove the most dangerous 2% of doctors.

• Only 1% of patient harms are reported by hospitals to state health departments in the 26 states that require them to report all of them. Since 61% of such events are caused by the acts of individual doctors, an essential conduit for reporting erring doctors is all but blocked.

These things all need to be addressed if we are to prevent even some of the 1 million patient deaths likely to occur in the next decade.

Philip Levitt is a retired neurosurgeon who served as chief of staff of two hospitals during a 31-year career.