Lucian Leape’s legacy for patient safety Authors: Atul A Gawande Published: 2026 “Errors must be accepted as evidence of systems flaws, not character flaws.”—Lucian Leape1 In losing Lucian Leape to heart failure last 30 June, at the age of 94 years, we lost a rare physician whose insights fundamentally changed how clinicians and the public understand the practice of medicine. His contributions helped transform medicine’s moral self-conception from one that prized perfection and regarded error as a character failure to one that prizes humility and understanding failure. Medical mistakes were his laboratory. They were both his subject and his inspiration. And his career was a study in audacity and late blooming. He was born on 7 November 1930, in Bellevue, Pennsylvania. His father worked for a steel company. His mother was a schoolteacher. He studied chemistry at Cornell, where he met his wife, Marty. He served in the US Navy, then went to Harvard Medical School, graduating in 1959. He completed general and thoracic surgery residency at Massachusetts General Hospital, then a paediatric surgery fellowship at Boston Children’s Hospital. His first appointment was at the University of Kansas. He helped found the American Pediatric Surgery Association in 1970. He returned to Boston in 1973 as Chief of Pediatric Surgery at Tufts Medical Center. He had a steady research output, writing just under half of his 200 academic papers and book contributions during his surgical career. He published findings on subjects from child burn treatment to nutritional deficiencies like rickets to paediatric reflux disease. Paediatric surgery was a new speciality, and he was among the pioneers inventing the field as they practised. Several of his papers also show interest in safety from the start. One of his first papers, in 1967, on 148 cases of testicular torsion, was entitled, ‘Invitation to Error’.2 He reported an adverse event in a child who developed calcification of the leg from calcium infusions. He detailed a technique for snaring broken intravenous catheters from the right heart of infants and wrote a series of searing reports on ingestion of lye drain cleaners due to inadequate child safety measures. Preventable injury was already an animating cause. In 1986, he branched out to do a mid-career fellowship at RAND in epidemiology and health policy. He was hooked, and at the age of 56 years, he made the risky move of departing his chairmanship and full-time practice to become an Adjunct Professor at the Harvard School of Public Health (HSPH). He had been given an opportunity to join an ambitious project to investigate the medical malpractice system, the Harvard Medical Practice Study, which was the brainchild of Howard Hiatt, the former dean of HSPH, and Paul Weiler, a Harvard Law School professor. The Harvard Medical Practice Study Hiatt recruited Troy Brennan, an internist, lawyer and HSPH faculty member, to lead the study. The effort involved a detailed review of the medical records of over 30 000 patients admitted to 51 hospitals in New York, USA, in 1984. This was well before the era of digital records. The project required fielding teams of nurses and doctors to hand review reams of paper records. Leape’s role was humble. He joined as record review supervisor after the departure of the original one. The study produced three key papers, all published in the New England Journal of Medicine in 1991. The first, with Brennan as lead author, documented that adverse events occurred in 1 in 27 patients, and 15% resulted in death.3 This rate of complications was far higher than had been recognised, and a quarter of them were judged due to negligence, meaning substandard care. The result sent shockwaves through the medical profession. A second paper, with Russell Localio as lead author, reviewed the records of the 47 patients in the study cohort who filed malpractice claims.4 Just eight of them had negligent adverse events identified on either the initial record review or a subsequent review. In other words, 98% of the 280 people identified as injured by substandard care had no opportunity for compensation for their losses. The record of success of claims filed in such cases was also known to be mixed at best. The results indicated, therefore, that the medical malpractice system failed virtually every patient injured through negligence. The authors would argue that the findings favoured a no-fault compensation system, which could produce faster, more reliable and less costly compensation for injured patients, as demonstrated by the no-fault vaccine injury compensation system created in 1986 in the USA.5 Leape, however, proposed an additional, very different analysis centred on a profound question: Why do medical personnel fail at what they set out to do for patients? At the simplest level of analysis, just two reasons are possible. Failure could be due to ignorance—we lack adequate knowledge of how to avoid failure; or it could be due to execution failure—the knowledge exists but an individual, or a group of individuals, fails to apply it correctly. In other words, failure could be due to error. To perform the analysis, he and Brennan each carried out an independent re-review of all the adverse event records with preventability as the standard. Leape would be the lead author on this paper about the nature of these events.6 It would report that they found only a minority were due to ignorance. The majority of the complications—at least 58%, and as many as 69% on further analysis—were preventable with existing knowledge.7 These were mostly not due to negligence—that is, the degree of error did not exceed accepted norms. For example, half of the adverse events resulted from surgical procedures and most were technical in nature. Injuries to a vital nerve, organ or major blood vessel can be common risks, yet often be preventable—such as with more experience—but typically do not involve negligence. The findings would drive a fierce debate about error and blame in medicine. They heralded a new phase in the history of medicine. For millennia, human health was governed primarily by ignorance. We lacked the knowledge to understand most diseases and conditions human beings face, let alone what could be done about them. But now knowledge had progressed to the extraordinary point that most failures were a matter of human agency. So what do we do about that? Do we require more teaching and certification? Do we impose more regulation and punishment? Or do we simply accept that risk is an inherent, normal part of the complexity of medicine and unavoidable? A groundbreaking analysis on error in medicine In December 1994, Leape would publish a blockbuster JAMA paper, simply titled, ‘Error in Medicine’.1 It argued that errors have become a major cause of death and disability, that they are human and that they are nonetheless largely avoidable. He detailed the shocking magnitude of the numbers found in the Harvard Medical Practice Study and described them as ‘the equivalent of three jumbo-jet crashes every 2 days’. But he argued that addressing these injuries would require abandoning the traditional focus in medicine (and malpractice systems) on individual accountability—away from what he called the ‘perfectibility model’. This was the belief that if clinicians were sufficiently trained and motivated, they would make no mistakes. The model’s tools were training and punishment. Its result was silence. As studies from psychology and human factors engineering had amply shown, he wrote, “All humans err frequently. Systems that rely on error-free performance are doomed to fail.” Leape detailed the known mechanisms of error and the many approaches in other industries—which made processes easier to do right and harder to do wrong—that had produced far better reliability than we achieved in medicine. Instead of punishing accidents, we needed to embrace the concept of systemic error and the science of its prevention, which were widely accepted in such fields as aviation and nuclear power but unfamiliar to most people in medicine. This was a profound moral reframing. “Hospitals,” he wrote, “rely on individuals not to make errors rather than assuming they will.”1 The remedy was not shame but design: reduce reliance on memory, simplify tasks, standardise processes, build redundancy and provide feedback loops. In this way, ‘Error in Medicine’ gave medicine its language for change—systems, forcing functions, standardisation, checkpoints, buffers, redundancy, closed loops—translating scattered disciplines into a shared grammar of improvement. The article’s most controversial prescription was a cultural change: “Errors must be accepted as evidence of systems flaws not character flaws. Until and unless that happens, it is unlikely that any substantial progress will be made in reducing medical errors.” The JAMA article provoked intense controversy. Editor George Lundberg faced calls for his removal for printing the piece. And that same month, a terrible event occurred that demonstrated how far medicine and the public were from embracing Leape’s prescription. The long road of patient safety On 3 December 1994, a 37-year-old Boston Globe health reporter named Betsy Lehman died at the Dana Farber Cancer Institute from a fourfold overdose of chemotherapy for her breast cancer. An oncology fellow made a dose miscalculation for her and another patient, who survived but sustained serious organ damage. The blunder was missed by at least six other doctors and nurses who countersigned the order, plus pharmacists and others involved in the care. Heavy media coverage focused on the practitioners involved.8 The oncology fellow lost his licence for 3 years. The senior oncologist lost her privileges and her appointment was not renewed. Three pharmacists were reprimanded. 18 nurses faced disciplinary hearings, of whom 13 were reprimanded and had to recertify for their licences and two accepted probation. Leape spoke out loudly against these actions. In a Boston Globe opinion, he noted that the clinicians managed over 100 protocols with widely varying drug dosages.9 The mistaken doses were not higher than seen in some patients. This was not a case of bad clinicians, but of good clinicians and a terrible accident waiting to happen. “Punishment is indicated for willful misconduct, reckless behavior, and unjustified deliberate violation of rules,” he wrote. “But not for errors… This concept of error as sin and the use of punishment for its control was abandoned long ago by industries outside of health care.” Such punishment would, if anything, discourage clinicians from admitting and examining errors. Instead, attention should focus on the system flaws that failed the patients and clinicians. His work fostered the creation of the field of patient safety and a generation of practitioners who would join it. He collaborated with the Institute of Medicine (IOM) and a group of eminent leaders to co-author a landmark 1999 report, ‘To Err is Human’. Based on new work in Utah and Colorado confirming the Harvard Medical Practice Study, it estimated that 44 000–98 000 Americans died each year from medical errors.10 11 And it named specific, immediate actions that could be taken—from computerised order entry systems proven to eliminate most medication errors, to standardised protocols for reducing hospital infections. The report galvanised action across the world. Its emphasis on systems innovation would also inspire advances in many fields, including the development of WHO’s Safe Surgery Checklist, a team communication routine that would reduce surgical complications and deaths by over one-third.12 An entire institutional ecosystem would emerge, much of it with Leape’s involvement. The Agency for Healthcare Research and Quality was founded later that year. The Institute for Healthcare Improvement advanced the creation of Quality and Patient Safety Officer positions in hospital executive teams. The National Patient Safety Foundation, the Lucian Leape Institute, the Betsy Lehman Center for Patient Safety in Massachusetts and other institutions were founded to develop leadership and strategy, update standards and track progress. Centres for health system research and innovation, such as Ariadne Labs, would develop new solutions and new talent. These fostered widespread change in the culture and practice of medicine. Transformations in the culture of medicine The work of advancing patient safety remains far from complete. But 15 years after the IOM medical error report, another terrible hospital event unfolded that revealed how radically approaches to error had evolved. In September 2014, a man aged 41 years presented to a Dallas hospital emergency department with dizziness, abdominal pain, nausea, headache, runny nose and nasal congestion.13 He was evaluated, given a series of scans and tests, diagnosed with sinusitis and sent home with antibiotics. The diagnosis was incorrect. It was a missed case of Ebola virus disease in a traveller from Liberia where an outbreak raged. The patient, Eric Thomas Duncan, grew sicker over the next 3 days at his apartment. He was returned by ambulance, and the disease was now promptly diagnosed. By then, he had exposed more than 100 people while symptomatic. Panic ran through his Dallas community and well beyond. Duncan would die after ten days in intensive care. Two nurses were also infected and survived. An investigation uncovered multiple failures.13 The emergency nurse who had first seen the man, following electronic medical record (EMR) prompts, screened for travel history and recorded that he had just returned from West Africa. The EMR flagged that as critical information, but the nurse did not notice the alert. The physician missed the information because it required looking beyond the physician view to the nursing assessment. The patient spiked a fever of 103°F during admission, which was also recorded in the EMR, but not verbally communicated to the physician. His Systemic Inflammatory Response Score, displayed on the ED monitors, rose from 0 to 3 (out of 4), but that too was not noticed and the fever subsided before discharge. The incident drew saturation-level media attention and Congressional hearings. It was notable, however, what did not happen this time. The clinicians involved were not suspended or fired. No one lost their licence. Public outrage was directed at why the systems had failed. Those systems had, in fact, advanced remarkably. But the investigation revealed that harried staff were now deluged with flags and alerts, creating overload and overreliance on the EMR for communication of critical information. Since then, critical event notification systems have been developed with escalation protocols—such as pager alerts—for the most urgent concerns. The episode traumatised the community and the hospital, and one person died and two people were infected because of the missteps. But instead of attacking the staff, who had simply been human, they fixed the systems. Leape’s call for culture change had reached an entire nation and beyond. That happened not only because of the research he had done, but also because of the movement he had helped build. Leape’s wife, Marty, died earlier this year. They are survived by their sons James, Jonathan and Gerald, and seven grandchildren. We in medicine and public health were lucky to have him for nearly a century of time. His legacy will live on even longer.