The Nature of Iatrogenic Injury in Medicine and Surgery

The earliest practitioners of medicine recognized and described iatrogenic injury. Iatrogenic (Greek, iatros = doctor, genic = arising from or developing from) literally translates to “disease or illness caused by doctors.” Famous examples exist of likely iatrogenic deaths, such as that of George Washington, who died while being treated for pneumonia with blood-letting. The Royal Medical and Surgical Society, in 1864, documented 123 deaths that “could be positively assigned to the inhalation of chloroform.”² Throughout history, physicians have reviewed unexpected outcomes related to the medical care they provided to learn and improve that care. The “father” of modern neurosurgery, Harvey Cushing, and his contemporary Sir William Osler modeled the practice of learning from error by publishing their errors openly so as to warn others on how to avert future occurrences.^3–5 However, the magnitude of iatrogenic morbidity and mortality was not quantified across the spectrum of health care until the Harvard Practice Study, published in 1991.⁶ This seminal study estimated that iatrogenic failure occurs in approximately 4% of all hospitalizations and is the eighth leading cause of death in America—responsible for up to 100,000 deaths per year in the United States alone.⁷

A subsequent review of over 14,700 hospitalizations in Colorado and Utah identified 402 surgical adverse events, producing an annual incidence rate of 1.9%.⁸ The nature of surgical adverse events were categorized by type of injury and by preventability (Table 1-1).

Table 1-1 Surgical Adverse Events Categorized by Type of Injury and Preventability

These two studies were designed to characterize iatrogenic complications in health care. While not statistically powered to allow surgical subspecialty analysis, it is likely that the types of failures and subsequent injuries that this study identified can be generalized to the neurosurgical patient population. More recent literature supports the findings of these landmark studies.^9–11

The Institute of Medicine used the Harvard Practice Study as the basis for its report, which endorsed the need to discuss and study errors openly with the goal of improving patient safety.⁷ The Institute of Medicine report on medical errors, “To Err Is Human: Building a Safer Health System,” must be considered a landmark publication.¹² It was published in 1999 and focused on medical errors and their prevention. This was followed by the development of other quality improvement initiatives such as the Joint Commission on the Accreditation of Healthcare Organizations (JCAHO) Sentinel Events Program.¹²

One might argue that morbidity and mortality reviews already achieve this aim. The “M&M” conference has a long history of reviewing negative outcomes in medicine. The goal of this traditional conference is to learn how to prevent future patients from suffering similar harm, and thus incrementally improve care. However, frank discussion of error is limited in M&M conferences. Also, the actual review practices fail to support deep learning regarding systemic vulnerabilities¹³; indeed, since M&M conferences do not explicitly require medical errors to be reviewed, errors are rarely addressed. One prospective investigation of four U.S. academic hospitals found that a resident vigilantly attending weekly internal medicine M&M conferences for an entire year would discuss errors only once. The surgical version of the M&M conference was better with error discussion. However, while surgeons discussed adverse events associated with error 77% of the time, individual provider error was the focus of the discussion and cited as causative of the negative outcome in 8 of 10 conference discussions.¹³ Surgical conference discussion rarely identified structural defects, resource constraints, team communication, or other system problems. Further limiting its utility, the M&M conference is reactive by nature and highly subject to hindsight bias. This is the basis for most clinical outcome reviews, focusing solely on medical providers and their decision making.¹⁴ In their report, titled “Nine Steps to Move Forward from Error” in medicine, human factors experts Cook and Woods challenged the medical community to resist the temptation to simplify the complexities that practitioners face when reviewing accidents post hoc. Premature closure by blaming the closest clinician hides the deeper patterns and multiple contributors associated with failure, and ultimately leads to naive “solutions” that are weak or even counterproductive.¹⁵ The Institute of Medicine has also cautioned against blaming an individual and recommending training as the sole outcome of case review.⁷ While the culture within medicine is to learn from failure, the M&M conference does not typically achieve this aim.

A Human Factors Approach to Improving Patient Safety

Murphy’s law—that whatever can go wrong will—is the common-sense explanation for medical mishaps. The science of safety (and how to create it), however, is not common sense. The field of human factors engineering grew out of a focus on human interaction with physical devices, especially in military or industrial settings. This initial focus on how to improve human performance addressed the problem of workers that are at high risk for injury while using a tool or machine in high-hazard industries. In the past several decades, the scope of this science has broadened. Human factors engineering is now credited with advancing safety and reliability in aviation, nuclear power, and other high hazard work settings. Membership in the Human Factors and Ergonomics Society in North America alone has grown to over 15,000 members. Human factors engineering and related disciplines are deeply interested in modeling and understanding mechanisms of complex system failure. Furthermore, these applied sciences have developed strategies for designing error prevention and building error tolerance into systems to increase reliability and safety, and these strategies are now being applied to the healthcare industry.^16–21 The specialty of anesthesiology has employed this science to reduce the anesthesia-related mortality rate from approximately 1 in 10,000 in the 1970s to over 1 in 250,000 three decades later.²² Critical incident analysis was used by a bioengineer (Jeffrey Cooper) to identify preventable anesthesia mishaps in 1978.²³ Cooper’s seminal work was supplemented by the “closed-claim” liability studies, which delineated the most common and severe modes of failure and factors that contributed to those failures. The specialty of anesthesiology and its leaders endorsed the precepts that safety stems more from improved system design than from increasing vigilance of individual practitioners. As a direct result, anesthesiology was the first specialty to adopt minimal standards for care and monitoring, preanesthesia equipment checklists similar to those used in commercial aviation, standardized medication labels, interlocking hardware to prevent gas mix-ups, international anesthesia machine standards, and the development of high-fidelity human simulation to support crisis team training in the management of rare events. Lucien Leape, a former surgeon, one of the lead authors of the Harvard Practice Study, and a national advocate for patient safety, has stated, “Anesthesia is the only system in healthcare that begins to approach the vaunted ‘six sigma’ (a defect rate of 1 in a million) level of clinical safety perfection that other industries strive for. This outstanding achievement is attributable not to any single practice or development of new anesthetic agents or even any type of improvement (such as technological advances) but to application of a broad array of changes in process, equipment, organization, supervision, training, and teamwork. However, no single one of these changes has ever been proven to have a clear-cut impact on mortality. Rather, anesthesia safety was achieved by applying a whole host of changes that made sense, were based on an understanding of human factors principles, and had been demonstrated to be effective in other settings.”²⁴ The Anesthesia Patient Safety Foundation, which has become the clearinghouse for patient safety successes in anesthesiology, was used as a model by the American Medical Association to form the National Patient Safety Foundation in 1996.²⁵ Over the subsequent decade, the science of safety has begun to permeate health care.

The human factors psychologist James Reason has characterized accidents as evolving over time and as virtually never being the consequence of a single cause.^26,27 Rather, he describes accidents as the net result of local triggers that initiate and then propagate an incident through a hole in one layer of defense after another until irreversible injury occurs (Fig. 1-1). This model has been referred to as the “Swiss cheese” model of accident causation. Surgical care consists of thousands of tasks and subtasks. Errors in the execution of these tasks need to be prevented, detected, and managed, or tolerated. The layers of Swiss cheese represent the system of defenses against such error. Latent conditions is the term used to describe “accidents waiting to happen” that are the holes in each layer that will allow an error to propagate until it ultimately causes injury or death. The goal in human factors system engineering is to know all the layers of Swiss cheese and create the best defenses possible (i.e., make the holes as small as possible). This very approach has been the centerpiece of incremental improvements in anesthesia safety.

FIGURE 1-1 Reason’s model of accident causation. Accidents (adverse outcomes) require a combination of latent failures, psychological precursors, event triggers, and failures in several layers of the system’s “defense in depth.”

(Copyright Dr. Reason.)

One structured approach designed to identify all holes in the major layers of cheese in medical systems has been described by Vincent.^28,29 He classifies the major categories of factors that contribute to error as follows:

If this schema is used to structure a review of a morbidity or mortality, that review will be extended beyond myopic attention to the singular practitioner. Furthermore, the array of identified factors that undermine safety can then be countered systematically by tightening each layer of defense, one hole at a time. I have adapted active error management as described by Reason and others into a set of steps for making incremental systemic improvements to increase safety and reliability. In this adaptation, a cycle of active error management consists of (1) surveillance to identify potential threats, (2) investigation of all contributory factors, (3) prioritization of failure modes, (4) development of countermeasures to eliminate or mitigate individual threats, and (5) broad implementation of validated countermeasures (Fig. 1-2).

FIGURE 1-2 Sequence of steps for identifying vulnerabilities and then implementing corrective measures.

(Copyright Blike 2002.)

The goal is to move from a reactive approach based on review of actual injuries toward a proactive approach that anticipates threats based on a deep understanding of human capabilities and system design that aids human performance rather than undermines it.

A comprehensive review of the science of human factors and patient safety is beyond the scope of this chapter; neurosurgical patient safety has been reviewed, including ethical issues and the impact of legal liability.³⁰ Safety in aviation and nuclear power has taken over four decades to achieve the cultural shift that supports a robust system of countermeasures and defenses against human error. However, it is practical to use an example to illustrate some of the human factors principles introduced. Consider this case example as a window into the future of managing the most common preventable adverse events associated with surgery (see Table 1-1).

Contributory Factor Analysis