fb tracking

Still Unsafe

Health Letter, March 2014

By Philip Levitt, M.D.

Why the American Medical Establishment Cannot Reduce Medical Errors

The following article first appeared in Skeptic Magazine, volume 18, no. 4. It is reprinted here with permission of the publication.

Every year, there are about 138,000 preventable deaths in hospitals throughout America. At least 30 percent of those deaths are the result of the actions of doctors who by any standards are incompetent.[1],[2] Because it is easier to blame a system than an individual doctor, particularly if he or she is a peer, the medical establishment has chosen to attack the problem of preventable deaths by employing what it calls a systems approach. The theory behind such an approach, which was first widely instituted at the beginning of this century, is that if you standardize the delivery of health care in a hospital, you reduce the number of errors that can be made by fallible humans. Although such an approach has worked successfully in the airline and automobile industries, it has not caused any noticeable drop in the overall number of preventable deaths in American hospitals.

In 2007, I retired from neurosurgery after 32 years. During that time I had been the chief of staff of two hospitals in South Florida over a total period of five years. In those positions I had responsibility for overseeing the professional and ethical conduct of roughly 1,800 doctors, about a tenth of those in the state. It was nearly impossible to discipline any one of them who stepped out of line and endangered patients. A second important consideration was that I knew from daily experience that the systems methods that were put in place in my hospital and in nearly 80 percent of the hospitals in the country during the first decade of this century were unlikely to make a significant dent in the number of deaths due to caretakers’ mistakes. Tactics such as checklists, time outs (where the nurses withhold the scalpel from the surgeon until he or she declares the name of the patient and the name and site of the operation), and labeling the correct extremity for surgery seemed merely to nibble around the edges of the identifiable hard core source of more deaths annually than automobile accidents— the inept doctors. I knew from direct experience and from sitting on numerous quality assurance committees for years and learning of their misadventures that they harmed many more patients than the occasional omission of typical systems tactics such as administering aspirin to a heart attack patient or giving preoperative antibiotics.

While researching the problem, I came upon two documents with overlapping authorship about harmful medical mistakes that contradicted each other. The first, published in the New England Journal of Medicine were the Harvard Medical Practice Studies of 1991, based on data gleaned from scores of hospitals in New York State during the mid- 1980s.[2],[3] They are the gold standard dealing with “adverse events,” a term meaning harm to patients as the result of medical management. “Adverse events” has almost completely replaced the much older term “iatrogenic mishaps” in the medical literature. “Iatrogenic” means doctor caused.

The Harvard Medical Practice Studies contained the results of screening by expert physician examiners of over 30,000 hospital charts for evidence of bad care. Nothing on that scale has been done since. The Harvard Studies analyzed, categorized and put into easily scanned written tables the types of errors that resulted in significant harm to patients. Although follow-up studies that analyzed hospital charts have included enough patients to reach statistical significance—a venerable yardstick of scientific validity—they are still dwarfed in size and detail by the Harvard Studies.

The other document, which appeared in 1999, To Err Is Human, is a report written by a 19-person committee of the Institute of Medicine (IOM).[4] The IOM is a private scientific group that advises Congress and the public on matters of public health. To Err Is Human claimed to have as its scientific and factual basis the Harvard Medical Practice Studies. By projecting to the entire country the data from New York State in the Harvard Studies, it concluded that there were around 98,000 fatal adverse events in America’s hospitals each year. Similar results have been echoed in two more recent studies published in 2010.[5],[6] To Err Is Human did not contain any new facts. Its most significant and far-reaching impact was its recommendations for reducing medical errors. It concluded that the best way to do so was to apply the same general method used to reduce crashes of commercial airplanes, called the “systems approach.” The authors, without scientific data arising from studies of hospitals and medicine, concluded a priori that most of the errors and deaths arose from faulty medical care delivery systems. It promised that if those systems were corrected that the number of deaths would drop by 50 percent in five years. The methods were started in 78 percent of American hospitals.[7] They included a variety of checklists that assured that a heterogeneous group of tasks were always carried out, such as giving aspirin and beta blockers to heart attack patients immediately upon admission to the hospital, polyvalent pneumonia vaccine to elderly patients upon discharge, and antibiotics to surgical patients within the hour before making the surgical incision.

Despite considerable change in how hospital medicine is practiced and with the vast majority of American hospitals participating, there was no drop in mortality from adverse events in the decade that followed the release of To Err is Human. Two research reports came out at the end of the last decade with very similar results. One of the studies was directed by the Inspector General of the U.S. Department of Health and Human Services, (HHS)[5] and the other, by a group from Harvard and Stanford medical schools.[6] The latter report was published in the November 2010 issue of the New England Journal of Medicine, the same time that the HHS report appeared on the Internet. In To Err is Human, the IOM estimated that 98,000 preventable deaths occurred based on data collected during the 1980s. During the period from the mid-1980s to mid-2000s, no other studies of a similar scope and resulting from an equally rigorous methodology (a careful screening of hospital charts) were reported. In the Health and Human Services (HHS) report based on hospital chart review, 180,000 Medicare patients lost their lives. Its authors considered 44 percent—or 79,200 of the deaths—preventable. How does this compare with the numbers gleaned from the mid-1980s that were reported in To Err is Human? Given that Medicare admissions are somewhat less than half of all hospital admissions, the numbers are, sadly, comparable.

The Harvard-Stanford study looked at the inpatient charts of patients in all age groups in several hospitals in North Carolina. That state was chosen because compared to other states it “had shown a high level of engagement in efforts to improve patient safety,” the authors said in their report. Those “efforts” involved the widespread use of systems tactics as championed by the authors of To Err is Human.[7] A typical example: every patient was given prophylactic antibiotics before surgery to prevent surgical wound infections. However, the use of these antibiotics was not only at the direction of the doctor; should he forget, a nurse was there to remind him. Another example involved the use of medication. A doctor could not prescribe penicillin to a patient without ascertaining whether the patient had any allergies to the medication. Regardless, the prescription was not filled until the pharmacist, checking with his or her pharmaceutical software, also confirmed that the patient, according to the software, had no allergies to penicillin. The system also enabled the pharmacist to double-check the dose and the frequency prescribed by the physician. The nurse, a third party, also checked by asking the patient— even if the doctor already had—and double-checking the patient’s chart. A system of checking and double-checking was in place, no matter how reliable or unreliable, renowned or unknown, the particular physician involved.

At the time of the scientific investigation, Dr. Christopher Landrigan, the senior author of the Harvard- Stanford study, estimated that about 96 percent of the hospitals in North Carolina and 78 percent of hospitals nationwide, were participating in the 100,000 Lives Campaign, an effort meant to save 100,000 lives a year in hospitals by putting in effect various systems approaches.[7] He and the rest of the authors of the study reasoned that if an improvement in patient outcomes were to be found, it would surely have occurred in North Carolina. Instead, the report found a much higher level of mortality from medical mishaps than expected. In doing their research, the authors of the Harvard-Stanford study continuously monitored adverse events—including deaths—and tabulated them by three-month periods over a span of six years. Their hypothesis was that the longer the systems approach was used, and the larger the number of systems methods employed, the more progressive the drop in the number of adverse events. In fact, the number of adverse events remained the same within the six-year period of the study. In 2002—the start of the study—the authors found 15 high severity harms per 100 admissions (with high severity harms defined as temporary harms requiring prolonged hospitalization, permanent harms, lifethreatening harms, and death); in 2007, after the implementation of systems methods, they found exactly the same number.

Using the same method used in To Err Is Human and working with the data accumulated by the authors of the 2010 Harvard-Stanford article, one could project 215,000 deaths of patients in hospitals nationwide, of which 138,000 would be preventable. This number compares unfavorably to the 98,000 preventable deaths nationwide found in To Err is Human, a figure taken from 1984 before systems approaches were in place. If anything, the results were worse.

Dr. Landrigan, the lead author of the Harvard- Stanford paper and the director of Harvard’s patient safety program, said, “We found that harms remain common, with little evidence of widespread improvement.” Rather than looking for some other cause, Dr. Landrigan simply noted that the systems approach needed to be used more effectively. “Further efforts are needed to translate effective safety interventions into routine practice and to monitor health care safety over time.”

A striking exception to these dismal results was the prevention of blood infections from central intravenous lines which the Centers for Disease Control and Prevention declared to be saving 3,000 to 6,000 lives a year during the same period as the Inspector General’s and the Harvard-Stanford studies.[8] This occurred largely as the result of the efforts of Dr. Peter Pronovost of Johns Hopkins Medical School. However, those savings got swamped by all the other preventable deaths.

The failures of the systems approach were foreshadowed in the data of Harvard Medical Practice Studies. Table 7 of the second installment, which listed the types and numbers of adverse events shows that at least 61 percent of all adverse events could be laid at the feet of individual physicians and, according to the authors of the Harvard Medical Practice Studies, only six percent were attributable to systems problems.[2] The 61 percent comprised technical mishaps during surgery and other procedures and failure to order the correct diagnostic test. For verification in 2013, I asked one of the authors of To Err is Human, Dr. Joseph Scherger, about technical and diagnostic errors and he readily admitted that these were not amenable to systems fixes.[9] He could not reconcile for me his and his co-authors’ great familiarity with the results of the Harvard Medical Practice Studies with the solution they proposed for saving lives, the systems approach. Neither could Dr. Lucian Leape of Harvard’s School of Public Health, an author of both documents.[10]

The medical profession and its chief watchdogs— the state boards of licensure—had no trouble believing what was said in To Err is Human. In a joint statement in 2008 the medical boards of all 50 states and the provinces of Canada wrote that most medical errors were the result of faulty care delivery systems and not the fault of the individual physicians about whom they received complaints on a regular basis.[11]

The deadly 2 percent

Based on information from the National Practitioner Data Bank, a federal data base which stores records of malpractice judgments, loss of licensure and hospital privilege revocations for all the doctors in the country, it is clear that a small number of doctors—about two percent—are responsible for half the cases in which a patient is seriously and unnecessarily harmed in the process of being treated. Dr. Robert Oshel, formerly the associate director for research and disputes at the Data Bank—he has since retired—confirms that the misdeeds of two percent of the physicians in practice during the last twenty years, from about 1990-2010, resulted in half of the money paid out in malpractice cases.[1] In other words, a very small number of doctors are responsible for a disproportionately large number of errors.

The distribution of error rates is not unique to hospitals in this country. Marie M. Bismark, a senior research fellow at the University of Melbourne, who is both a physician and a lawyer, found in a national sample of nearly 19,000 formal healthcare complaints lodged against doctors in Australia between 2000 and 2011, that “three percent of Australia’s medical workforce accounted for 49 percent of complaints and one percent accounted for a quarter of complaints.”[12]

Given such a small number of grossly incompetent doctors, it should be easy to identify and prevent them from practicing. Unfortunately, in most cases, they continue to practice. The average American hospital drops only one doctor from its staff every twenty years.[1] About 250 doctors lose their licenses each year, or 0.04 percent of the total number of practicing physicians, which is about 650,000.[1] At that rate it would take 50 years to remove the most incompetent doctors—the two percent— from practice. Why do the vast majority of these doctors keep practicing? One reason is the pervasive leniency of the hospital peer review committees and state medical boards, the main institutions set up to deal with the problem.

This small number of incompetent physicians not only causes serious errors in terms of patient health; they also cost society a huge amount of money. Dr. Donald Berwick, the chief of Medicare and Medicaid from July 2010 to December 2011— he is currently a senior fellow at the Center for American Progress—estimated that $300 billion a year are spent on the waste that results from poor execution of care and on over treatment that subjects patients to care that is unsupported by science and that cannot possibly help them.[13] Giving unnecessary care is the favorite method of many less scrupulous physicians for padding their earnings.

What, then, is the solution to the large number of preventable deaths that happen each year specifically because of incompetent doctors? Combining the estimates of both Dr. Oshel and the Harvard Studies this number equals about 42,090—more than the average number of people who die from auto accidents each year.

The advocates of the systems approach try to avoid the issue of the incompetent doctor. They say we can reduce the number of preventable deaths by streamlining and routinizing the practice of medicine. However, that approach has not been successful in either reducing the number of preventable deaths or improving the practice of medicine.

Currently, hospitals do a poor job of disciplining incompetent doctors in spite of laws that exist to ferret them out. Alan Levine and Dr. Sidney Wolfe of Public Citizen, a nonprofit, lobbyist group based in Washington, D.C. and Austin, TX, in examining the National Practitioner Data Bank, have found that although all hospitals are required by law to report serious disciplinary measures taken against a physician on their staff, 47 percent of American hospitals have never reported a single doctor.[14] Any practicing doctor of integrity will agree: it is highly doubtful that there are no incompetent doctors in 47 percent of American hospitals.

In addition, according to a separate report by the Inspector General of the Department of Health and Human Services (HHS), 26 states have laws requiring hospital administrations to report harms inflicted on patients to the state departments of health and state medical boards, but only one percent of adverse events are actually reported even in these states.[15] This data is critical to finding poorly performing doctors as, conservatively, 61 percent of adverse events are caused by the errors of individual physicians, not systems failures. Why don’t hospitals do the required reporting? Stuart Wright, the deputy inspector general for HHS (for evaluations and inspections) wrote in his Memorandum Report to the Acting administrator of the centers of Medicare and Medicaid Services that this low rate of reporting is more likely the result of a hospital’s failure to identify events rather than from its neglect to report known events.[15]

In fact, the opposite is true. Most hospital administrators know exactly who the incompetent doctors are and very rapidly learn of their mistakes. Elaborate mechanisms are set in place for reporting to the CEO on a daily basis so he or she does not get blindsided by doctors, nurses or family members calling to complain. Once these poor outcomes are detected, however, hospitals are loath to act because both the hospital and the doctor become subject to fines, bad publicity and the loss of licenses. Until hospitals have stronger incentives to report adverse events, most will do their utmost to avoid conveying the details to state authorities.

Two modest proposals

One obvious solution is to impose sufficiently high penalties—a fine of at least $250,000—on hospitals that fail to report to the state board doctors who commit disabling or fatal errors. If hospitals don’t report errors, then how will any agency know to levy such a fine? Even when hospitals fail to report adverse events involving negligence, the state boards can and do find out about them through other sources: the patient, his or her family, attorneys, judges, or malpractice insurance companies. If the complaints made are deemed to be legitimate then the doctor in question may lose his or her license. Therefore, the loss of a physician’s license is the one crucial and indelible marker of medical negligence and adverse events. If it is found that a hospital has failed to report the incident that caused the physician to lose his license, the fine will be imposed. No CEO of any hospital could withstand the publicity associated with such a case, let alone the fine. The existence of such a law would affect all hospitals in a state, not just the ones investigated or fined. Because no one can predict what adverse event might lead to the revocation of a doctor’s license, it is likely that many more events will be reported to the state. The second proposal concerns the two percent who cause half the damages to medical malpractice plaintiffs in America, a special group of repeat offenders. The identities of the two percent are a closely held secret of the National Practitioner Data Bank. It would take congressional action to disclose them. Publicity about the money, not to mention the lives lost, because of this small hard-core group of inept doctors, could sway public opinion. And to prevent these doctors from practicing, Congress could take away their participation in Medicare and Medicaid. Currently, few doctors other than those involved in Medicare or Medicaid fraud ever lose their right to participate.

These are not by any means the only solutions, but they at least seek to address the real source of the problem—the incompetent physician. The problem of the incompetent doctor is a constant, weaving throughout the history of modern medicine. Back in 1958, David Allman, president of the American Medical Association, exhorted the association’s component bodies to root out bad doctors: “Any reluctance to reprimand an erring colleague does irreparable harm to our profession. Any use of a whitewash brush to sweep dirt under the rug imperils our disciplinary system. Any compromise with personal moral convictions damages the very character which makes a man or a woman a good doctor.”[16] His words, however powerful, did not result in any notable action.

The irony is that the continuing tolerance for the incompetent physician is harmful not only to the public, but also to the majority of hard-working, competent physicians. The inept physicians drain huge amounts of money out of a vast system that also needs to reimburse the competent ones, tarnish the reputation of the profession as a whole, and create malpractice pitfalls for those doctors who work on the same patients with them. Lawyers are often obliged by circumstances and the law to sue doctors merely because their names appear in the hospital chart before the process of legal discovery clears the field of inappropriate defendants. This often takes years.

In the meantime, the inept doctors of America are still in place, sprinkled among the 5,700 hospitals of our country. They remain almost untouched by a so-called scientific systems approach that statistically has not succeeded in lowering or even maintaining the number of preventable deaths. So the silent casualties continue. More than a million patients will die because of medical error in the next decade. Until we confront the real problem, these unnecessary deaths will continue.


[1] Oshel, R., 2012. Personal communication.

[2] Leape L.L., T. A. Brennan, et al. 1991. The nature of adverse events in hospitalized patients. Results of the Harvard Medical Practice Study II. New England Journal of Medicine, 324:377- 384.

[3] Brennan, T.A., L.L. Leape, et al. 1991. Incidence of adverse events and negligence in hospitalized patients. Results of the Harvard Medical Practice Study I. New England Journal of Medicine, 324:370-376.

[4] Kohn L.T., J.M. Corrigan, M. S. Donaldson (Eds.). 1999. To Err is Human: Building a Safer Health System. National Academies Press: Washington, DC.

[5] Department Of Health And Human Services Office of Inspector General. Adverse events in hospitals: national incidence among Medicare beneficiaries. November 2010.

[6] Landrigan C.P., G.J. Perry, et al. 2010. Trends in rates of patient harm resulting from medical care. New England Journal of Medicine, 363:2124-2134.

[7] Ibid., References 19 and 20: 19. North Carolina Center for Hospital Quality and Patient Safety. About us. (http://www. nc.qualitycenter.org/about.lasso. 20. Institute for Healthcare Improvement. A network that works! The 100,000 LivesCampaign nodes. Cambridge, MA: IHI, 2006. (http://www.ihi.org/IHI /Topics/Improvement/Spread ingChanges/ImprovementStories/ ANetworkThatWorks1000 00LivesCampaignNodes.htm.)

[8] Centers for Disease Control and Prevention. 2010.

[9] Scherger, J. 2013. Personal communication.

[10] Leape, L.L. 2011. Personal communication.

[11] Federation of Medical Regulatory Authorities of Canada, Federation of State Medical Boards and Milbank Memorial Fund. 2008. Medical Regulatory Authorities and the Quality of Medical Services in Canada and the United States, 4.

[12] Bismark, M.M., M. J. Spittal, et al. 2013. Identification of doctors at risk of recurrent complaints: A national study of healthcare complaints in Australia. Quality and Safety in Health Care, 1-9.

[13] Berwick, D.M. and A. D. Hackbarth. 2012. Eliminating waste in U.S. health care, JAMA, 307 No. 14, 1513-1516.

[14] Levine, A. and S. Wolfe. 2009. Hospitals drop the ball on physician oversight. Public Citizen, May 27: www.citizen.org/hrg.

[15] Wright, S. 2012. Memorandum report. Few Adverse Events in Hospitals Were Reported to State Adverse Events Reporting Systems, OEI-06-09-00092, July 19.

[16] Ameringer, C.F. 1999. State Medical Boards and the Politics of Public Protection. Johns Hopkins University Press, 35.