Throughout history, the practice of medicine has presented practitioners with challenging and difficult decisions. They have been asked to provide healing using every possible resource, while often being forced to place their patients at risk due to flawed treatments, imperfect science, limited resources, financial constraints, and more.
The frequently quoted study by the United States (US) Institute of Medicine (IOM), To Err is Human, noted that errors cause between 44,000 and 98,000 deaths every year and over one million injuries.1 More recently, error was reported as the third leading cause of death in the US with 250,000 deaths per year accounting for 9.5% of all deaths.2 In 2005, medical mistakes, medication errors, or test errors were reported in as many as 34% of all patients in the US—the highest rate of any nation.3 A follow-up study in 2016, found little improvement.4 Although these early studies classified adverse events as 100% preventable, subsequent reviewers estimate 3%–5% of deaths were probably preventable.5
The ethics of error in medicine can only be understood after defining its moral value and its effects. Furthermore, defining error remains elusive, and its incidence, for example, in emergency medicine, is extraordinarily difficult to quantify. In theory, the rate of error in emergency medicine should be extraordinarily high, as the gatekeepers of a multifaceted complex health care system are often the first stop in evaluation and treatment.6 Short evaluation time, a hectic environment, minimal or often disjointed and difficult-to-access health care records, and little time for patient–physician rapport-building and shared decision-making may all lead to error in emergency medicine. Yet, one study using automated select case reviews reported an error rate of 9.5%, representing an error rate per year of only 0.13% if the analysis was extended to all emergency department patients.7 This discrepancy in error rates is likely multifactorial and affects all medical specialties; physicians and caregivers may be reluctant to report error due to feelings of guilt, fear of retribution from patients, and apprehension related to possible loss of employment or damage to reputation.7
In truth, we believe error is subjective. The Merriam–Webster dictionary defines “error” as an act or condition of ignorant or imprudent deviation from a code of behavior.8 In the case of medical error, the relevant code of behavior is the standard of care; however, this standard may vary from specialty to specialty and even from institution to institution, compromising our ability to identify deviation from the standard of care. On the other hand, relying on individual judgment without rules or defined standards reinforce subjectivity and high variability in care. Prior data have suggested a lack of inter-rater reliability between case reviewers in assigning error, which may reflect the difficulty distinguishing between judgment calls and errors, as well as varying individual reviewers’ ability to overcome outcome bias in assessing whether an error has been made as well as who made the error.7 Furthermore, providers may have distinct thresholds for assigning errors when the desire to protect a colleague conflicts with the duty to improve the system. Studies that have assessed error have often failed to use providers with actual expertise in the standard of care of the particular discipline.9 Reviews by providers unfamiliar with a given specialty or local conditions may add to the subjectivity of error assignment. The lack of scientific methodology in error assignment has been identified as a concern as early as 1996 but has rarely been considered even in the studies that supported the opinions in the IOM report.10
Even if error is accepted within the ethical practice of medicine, the cost of practicing medicine in a system replete with error may be prohibitive. With health expenses at a critical level, as a society, we cannot afford to allow errors to go undetected. In 1999 the IOM estimated that the annual cost of medical errors ranged from US$17 billion to US$29 billion.1 This staggering amount may underestimate the true cost of errors. Yet, errors such as omissions and delays are hard to identify unless they cause serious adverse reportable events. Error may create additional indirect costs, which arise from defensive medical practice and administrative overheads to prevent lawsuits resulting from errors and preventable adverse events. Thus, a health care system that acknowledges error as a consequence of normative ethical practice must create systems to minimize error.
Error reduction, in turn, should attempt to decrease patient harm and improve the entire health care system by shifting resources away from unnecessary care needed to ameliorate the effects of the error, back to helping the patient and caregivers more directly. Such an approach would serve as a model to increase the cost-effectiveness of the health care system and the value of the care delivered.
Despite the complexities in quantifying and defining error and the continuing needs to better refine what constitutes error, error in medicine exists. In reality, we practice medicine based on Bayesian theorem, relying on probabilities to make diagnoses, based on prior knowledge or history of conditions that might be related to the diagnosis and then ordering tests to help confirm or rule out that diagnosis. One would hope that every time we evaluate a patient and test a diagnosis, the post-test probability of that patient having a certain disease will be higher than the pretest probability of their having that disease. Yet, even if we always ordered the correct test, testing is confounded by false positive and false negative results, and additionally, we make diagnoses based on likelihood ratios of a patient having a certain disease, rather than certainty, setting practitioners up for error.
We would like to discuss a number of ethical and moral considerations that arise from practicing medicine in the shadow of error. First, when a practitioner does err, at what point does that practitioner no longer have a right to continue to practice? Second, if the error is egregious such that its outcomes result in loss of life or limb, should this further preclude the practitioner from practicing? Third, should a student, trainee, or junior physician be allowed to train with the assumption that their care will likely be more error-prone then that of the more seasoned physician? Lastly, does an apology ameliorate error, or is it futile?
Our response to each of these four ethical concerns is discussed below.