Description

	• Although unavoidable human fallibility plays a part in causation of most errors, medical errors tend to occur in error-permissive or error-promoting environments.
	• Medical errors can be analyzed according to their classification, eg, diagnostic, therapeutic, medication errors, etc.

Etiology of medical errors. The traditional “person approach” to medical errors places blame on an individual or small group for having committed “human error,” after a cursory evaluation to rule out technical or mechanical failure. In recent decades, research into errors in multiple fields, including medicine, has shown that although unavoidable human fallibility plays a part in almost all errors, these errors tend to occur in error-permissive or error-promoting environments. This type of error is often referred to as a “latent error” because it exposes a latent shortcoming in the system. Thus, identifying the cause as human error is not satisfactory in understanding the pathogenesis of the error, nor is it helpful in preventing future errors. Factors such as suboptimal staffing or work schedules, inadequate nurse-patient ratios, poorly designed equipment, poor illumination, excessive distractions, excessive reliance on memory, and violation of the principles of ergonomics (human factors engineering) in the work environment can predispose individuals to commit errors.

The basic causal factors underlying an adverse or sentinel event are referred to as root causes, and root-cause analysis is a systematic method of analysis used to identify root causes. When a root-cause analysis uncovers a medical error as a contributing factor to an event, the event is, at least theoretically, preventable. From 2005 to 2016, 12,946 sentinel events have been received and a full analysis is available on the Joint Commission website (17). Because the reporting of most sentinel events to the Joint Commission is voluntary, it represents only a small proportion of actual events. Therefore, these data are not an epidemiologic data set, and no conclusions can be drawn about the actual relative frequency of events or trends in events over time.

Classification of medical errors. Categories of medical errors are listed in Table 1.

Table 1. Classification of Medical Errors

Errors in diagnosis
	• failure to order appropriate diagnostic • misinterpretation of diagnostic results • overdiagnosis
Errors in treatment
	• failure or undue delay in starting treatment • lack of judgement in selecting appropriate treatment
Errors in medications/prescription
	• pharmacy errors: wrong medications or wrong patient • multiple medications with adverse interactions
Nosocomial infections
	• transmission of infections to patients due to negligence • inappropriate use of antibiotics
Errors relevant to surgical procedures
	• surgery at incorrect site or wrong side of body • leaving nontherapeutic foreign bodies in the wound • technical errors due to lack of competence of the surgeon • unanticipated permanent disability resulting from a technical error during surgery • anesthetic errors, eg, wrong medications • mismatched blood transfusion
Medical malpractice
	• delivering substandard medical care • professional negligence in patient care • use of unapproved methods of treatment known to have harmful effects

Epidemiology

	• Epidemiological studies show that incidence of medical errors is high, particularly in hospitalized patients.
	• Incidence varies according to the type of medical error.

According to a September 2019 World Health Organization report on patient safety, medical errors harm some 40% of patients in primary and outpatient care worldwide (www.who.int). Approximately 400,000 persons die every year in the United States due to medical errors (22). A systematic review pooled the results of 8 studies and showed that 9.2% of hospitalized patients experience adverse events, of which 43.5% are preventable, implying they are due to medical errors (08). The exact incidence of medical errors (including those that do not lead to patient harm, ie, near-misses) is unknown but is likely to be even higher.

Errors in diagnosis continue to occur with an estimated incidence of 10% to 20% and account for up to 80,000 deaths per year and substantially more patient harm (13).

Nosocomial infections are the most common adverse events. According to World Health Organization data from several countries, prevalence in mixed patient populations of high-income countries is about 7.6% and about 10% in low-and middle-income countries. Patients with methicillin-resistant Staphylococcus aureus are estimated to be 64% more likely to die than those with a nonresistant form of the infection.

Unsafe surgical care or procedures cause complications in up to 25% of patients. Approximately 7 million surgical patients annually suffer serious complications, and 1 million of these die during or immediately after surgery.

Clinical applications

	• Errors related to the use of medications are common.
	• Diagnostic errors are frequent as well as significant as they are an important basic of healthcare and may be due to physician- as well as patient-related factors.
	• Medical errors are mostly preventable, eg, nosocomial infections; various methods of prevention are described.
	• Various aspects of medical errors are discussed, eg, legal, ethical, and economic.

Errors related to the use of medication. These are a leading cause of injury and avoidable harm in health care systems. The global cost associated with medication errors has been estimated at $42 billion annually. The top 5 medications associated with errors include anticoagulants, insulin, opiates, intravenous potassium, and hypertonic saline. Drugs that sound alike are or have spelling similarities may lead to mixups in prescription and administration. The medications “epinephrine” and “ephedrine” have led to much patient harm, as each has a different purpose and can have serious adverse effects if administered incorrectly.

Errors in diagnosis. Diagnostic errors have received less attention than other types of medical errors including medication errors, surgical errors, and hospital-acquired infections. Major causes of errors in diagnosis are listed in Table 2.

Table 2. Major Causes of Diagnostic Errors

Physician factors
	• Inadequate clinical examination of patient and ordering inappropriate tests • Overconfidence in one’s diagnostic ability and reluctance in consulting colleagues • Poor diagnostic reasoning skills • Impaired physical or emotional well-being due to fatigue, stress, illness, medications, etc.
Patient factors
	• Deliberate misinformation • Malingering
Miscellaneous factors
	• Inadequate data gathering due to time constraints and overwork • Missing significant patient information about previous tests • Distractions during patient encounters such as phone calls and staff

In 2015, the Committee on Diagnostic Error in Health Care of the National Academies of Sciences, Engineering, and Medicine published “Improving Diagnosis in Health Care” as a follow-up to the Institute of Medicine’s report “To Err is Human,” and this report has 3 major themes (07):

	(1) Diagnostic errors have been underappreciated, even though the correct diagnosis is a critical aspect of health care. This report highlights the importance of the issue and what should be done about this challenge.
	(2) The report defines diagnostic error, which is central to the issue, from the patient’s viewpoint as “the failure to (a) establish an accurate and timely explanation of the patient’s health problem(s) or (b) communicate that explanation to the patient”.
	(3) Diagnosis is a collaborative effort, rather than a single physician effort. A diagnostic error is not always due to human error; often, it occurs because of errors in the health care system.

In 2000, the Institute of Medicine did not think that the time was ripe to ask for mandatory reporting but recommended that diagnostic error should be treated as a key component of quality improvement efforts throughout all settings of healthcare.

Measures for reducing diagnostic errors are based on improvement, correction, or eliminations of some of the factors causing these errors. Suggestions include re-assessment of a physician’s overconfidence in diagnostic skills and a deeper understanding of the reasoning processes leading to diagnosis as well as the pathogenesis of a diagnostic error (28). Development of validated metrics for diagnostic errors has been proposed for determining the frequency of occurrence of various errors to elucidate their downstream health and economic implications (18). Identification of health system factors associated with diagnostic errors could be used to reduce the rates of misdiagnosis.

Since the National Academies of Sciences, Engineering, and Medicine’s report in 2015, literature on diagnostic safety has grown rapidly. An update was presented at the annual international meeting of the Society to Improve Diagnosis in Medicine, which focused on search of articles published between 2016 and 2018 and 3 key topics emerged from this review with debates around these topics in a pro/con format (05):

	(1) Definitions of diagnostic error and related concepts are evolving since the 2015 report. One criticism of definition of diagnostic errors by the National Academies is that it does not include “overdiagnosis”, which may have undesirable effects.
	(2) Medical educators are grappling with new approaches to teaching clinical reasoning and diagnosis.
	(3) Potential of artificial intelligence to advance diagnostic excellence is bearing results but there is some concern that reliance on artificial intelligence could have other undesirable effects in diagnostic safety. Variables used in current machine learning algorithms are based on numerical data or discreet text inputs found in electronic data sets. Thus, algorithms can handle imaging studies, which can be distilled into pixels, voxels, and other discreet quanta, but may not process nuances or irregularities. The response to this criticism is that the computational power of artificial intelligence will enhance diagnostic safety as the connectivity as well as updatability of computer networks give them an advantage over human decision-makers who have difficulty in keeping up-to-date with new knowledge. Artificial intelligence can rapidly integrate new scientific discoveries in diagnostic algorithms and can also improve our understanding of diagnostic errors.

Prevention of medical errors. Several medical errors are preventable. For example, a significant proportion of nosocomial infections are preventable. In-hospital patient falls due to negligent care are a common and largely preventable group of adverse events. The summary data of sentinel events reviewed by The Joint Commission introduced several patient safety goals to assist institutions and healthcare practitioners from 1995 to 2020 in creating a safer practice environment for patients and providers; more information can be found on the following website:www.jointcommission.org. The Joint Commission goals include:

	• Identify patient safety dangers and risks
	• Prevent infection by hand-cleaning, post-op infection antibiotics, catheter changes, and central line precautions
	• Prevent errors in surgery by making sure the correct surgery is done on the correct body part; pause before surgery to double-check
	• Use device alarms and make sure that alarms on medical equipment are heard and checked quickly

Organizational, team, and individual approaches. Error prevention can be approached at the organizational, team, and individual level. At the organization level, there are cultural, systems, and budgetary considerations in establishing a “safety culture.” Organizations can prevent errors by ensuring that there is fundamental leadership and institutional awareness of the potential for errors and their causes and by seeing that patient safety takes priority over budgetary, scheduling, and political considerations (33). Organizations have found success at reducing medical errors by adopting an open and blame-free approach to errors and adverse events.

Healthcare providers who focus on effective teamwork alter the strict hierarchical team structure (referred to as "flattening the authority gradient"), encourage any member of the team to raise concerns about error-prone situations, and communicate across medical disciplines, which may be successful in reducing errors.

At the individual level, the importance of work schedules on reducing fatigue has been acknowledged, and the Accreditation Council for Graduate Medical Education (ACGME) began regulating works hours in 2003 and overhauled these regulations in 2011. The work environment can be modified to build in constraints (cues or barriers that prevent errors) and affordances (cues or aides that facilitate proper actions by making them easier than incorrect actions) to reduce the potential for errors. Constraints can be physical or procedural, with examples including connectors that only fit with proper intravenous tubing (physical) or electronic order systems that require an extra step for nonstandard dosing (procedural). Examples of affordances include safety needles that make capping used needles easy (physical) and the use of timeouts prior to procedures (procedural). Designing a workplace with these ideas in mind is a major focus of ergonomics, or human factors engineering. Decreasing environmental potential for distractions, interruptions, and excessive noise may be beneficial, as is the provision of adequate information support for clinical tasks to reduce reliance on rote memory as memory is fallible. Technology and automation provide innovations as well as constraints, and therefore there is a tendency to view them as universal solutions to prevent errors, but they do not eliminate all errors and can even facilitate new types of errors. Thus, technology should be introduced thoughtfully, and its effects monitored carefully. Finally, education and training of medical students, residents, and fellows should include exposure to the principles and practical application of patient safety concepts.

A cross-sectional study using data from the Human Diagnosis Project (Human Dx), a multicountry data set of ranked differential diagnoses by physicians, concluded that a collective intelligence approach was associated with higher diagnostic accuracy compared with that of individuals, which included specialists (02).

As the lead federal agency supporting patient safety for the country, Agency for Healthcare Research and Quality (AHRQ) research activities covering the topic has a broad portfolio (24). Peer Review of a Report on Strategies to Improve Patient Safety, a major research dissemination vehicle has been the Making Healthcare Safer (MHS) series of reports, the most recent of which was released in 2020 (Hall et al 2020). These reports represent highly curated and prioritized summaries of important safety research and publications, and the AHRQ draft report includes an excellent overview of the MHS reports and their important contributions. In addition to its review of the AHRQ draft report, the committee offers several technical suggestions for enhancing the usefulness of the MHS report series.

Fixing error-prone spots in healthcare system. Error-prone loci in the healthcare system can be identified and eliminated before they cause an error. In a study of risk factors for medication errors in elderly, hospitalized patients, factors significantly associated with increased risk for medication errors included Charlson Comorbidity index score greater than 8, number of medications greater than 9, and length of hospital stay greater than 13 days (04). More studies of this type might help hospitals identify situations that carry high risk for medical error and target specific factors for remedy.

Computerized provider order entry (CPOE) systems. CPOE systems involve electronic prescription of patient services, which eliminates the need for handwritten paper orders and saves costs in hospitals. A study found that introduction of CPOE systems in hospitals reduced medical errors and adverse drug events, particularly when they are combined with measures for alerting physicians and other healthcare providers of pending lab or medical errors (06). Major barriers to adoption of CPOE systems in hospitals are high initial costs and low physician acceptance. In a large national study in the United States, physicians' satisfaction with their electronic health records and CPOE was generally low because the amount of time spent on clerical tasks put them at higher risk for professional burnout (29). Another concern is that introduction of computers into the examination room may shift physicians’ focus away from the patient and the human interaction.

Machine learning for preventing medical errors. It is generally recognized that if every medical decision made by a clinician were to be instantly reviewed by a team of relevant experts who can provide guidance if the decision is wrong, the likelihood of medical errors would be reduced. However, there are not enough medical experts for reviews, and it would take too long for available experts to read through a patient’s history. If all medically relevant data used by millions of clinicians to make decisions in caring for billions of patients are analyzed by machine-learning models, it may be possible to deliver the best possible error-free care to all patients (26).

Improving communications to prevent medical errors. Miscommunications are a leading cause of serious medical errors. An approach to hospital handoffs, called I-PASS bundle, has been shown to improve communications and reduce medical errors by 23% without hindering workflow (30). It uses a mnemonic to alert providers to all the issues that need to be covered during a handoff, but also includes a written handoff tool, communication training, a sustainability campaign, and a process for feedback. Several medical centers are now implementing the approach hospital-wide and additional studies into the approach are planned. Developers advise hospitals interested in the approach to first gather data and survey providers to make a case for the intervention.

Response to medical errors. Ideally, in the case of harm to a patient, the situation should be stabilized medically, and the patient and medical decision makers should be informed of the incident. Disclosure to the patient should be empathetic, honest, and full. Most sources recommend that the senior physician involved in the patient’s care be involved in this disclosure. The discussion should also include an outline of the medical team’s plan to monitor the situation and mitigate harm from the adverse event. Finally, the patient should be informed that the incident will be formally reviewed to prevent similar events from occurring in the future. The AMA suggests that physicians offer apologies if they are at fault for an adverse outcome of treatment, though the fear of litigation often dissuades physicians from apologizing.

Medical errors, particularly in diagnosis, may occur in the emergency department as quick decisions have to be made, sometimes with less than optimal information available, and prevention may not be possible. Transparency about errors is necessary and patients or their families should be informed about all errors when discovered. A systematic approach to apologizing creates an environment of mutual trust and shared decision making that is beneficial to all aspects of patient care (09).

Physician attitudes to medical errors. Physicians may be reluctant to report errors or take ownership for them for fear of being sued, and although this is a legitimate concern, mechanisms are in place to permit the use of medical errors to improve the quality of care and reduce future errors without repercussion.

Analysis of responses of over 3000 United States and Canadian physicians found that their attitudes about discussing errors with patients, colleagues, and institutions are related to several predictable factors that are potentially modifiable by educational and institutional strategies (03). Factors supporting transparent communications about medical errors were: (1) female sex; (2) United States versus Canadian physicians; (3) academic versus private practice; (4) belief that disclosure decreased likelihood of litigation; (5) belief that system changes occur after error reporting; and (6) previous experience in disclosing a serious error.

Physicians at all levels of expertise and specialty should be encouraged to work with their legal department, risk-management services, and hospital administration to recognize medical errors, analyze them for their root causes, and take the necessary steps to prevent future error. The point of action may take place on a personal, team, service, department, or system level.

Resident education, patient safety, and medical errors. Education about medical errors and adverse events has increased in residency programs. Since 1983, the Accrediting Council of Graduate Medical Education (ACGME) has required surgical residency programs to hold regular morbidity and mortality conferences. The earlier conferences were not focused on medical errors, but during the past decade, the ACGME has required that residents demonstrate proficiency in systems-based practice and participate in identifying system errors and implementing potential systems solutions. Neurology residency requirements include the following: “All residents must receive training in how to disclose adverse events to patients and families. Residents should have the opportunity to participate in the disclosure of patient safety events, real or simulated (01).

The New York-Presbyterian Hospital, Weill Cornell Medical Center, has reported on successes in improving quality of care and patient safety by establishing a house staff quality council and creating a resident quality and patient safety position (12).

To determine the effect of increased attending physician supervision of residents on patient safety, a randomized clinical trial was performed on an inpatient general medical service of a large academic medical center using a crossover design and rate of medical errors as outcome (11). Increase in direct attending physician supervision did not significantly reduce the medical error rate. Other factors should be assessed. Residency programs should reconsider their balance of patient safety, learning needs, and resident autonomy.

Healthcare providers’ attitudes to medical errors. Although healthcare providers usually believe in the benefits of error disclosure, surveys of patients reveal that full disclosure is uncommon. A study has defined a 6-part error disclosure as: (1) an admission of the error, (2) a description of the event or error, (3) a description of the proximate, or first, effect of the error, (4) a description of the link between the error and the effect, (5) a description of the final harm to the patient, and (6) a description of the link between the error and the harm (10). Using this definition to analyze descriptions of error disclosure by healthcare providers in focus groups, the authors found that only 27% of error disclosures by healthcare providers met criteria for full disclosure; an additional 52% described partial disclosure. Because patients may register a partial disclosure as insufficient, this might account for the discrepancy between healthcare provider belief and the patient experience.

Fear of litigation is the biggest barrier to disclosure. Another fear is the risk of losing professional liability insurance coverage if a hospital or a physician accepts being at fault for the medical error without prior consent of the insurance company. A Swiss study found that the practice of some liability insurance companies of inhibiting communication with harmed patients after an error is incompatible with the ethic of transparency regarding medical errors and suggested that hospitals move away from total reliance on liability insurance with an additional measure for legal protection to prevent the loss of liability insurance coverage for apologizing to harmed patients (23).

Initiatives for patient safety. The term “patient safety” applies to initiatives designed to prevent adverse or sentinel events resulting from medical error.

Since 2005, the Joint Commission has required that all sentinel events and their root-cause analyses be reported. Thus, concurrent with error disclosure to patients, formal reports should be issued via existing hospital reporting mechanisms. A committee then performs a detailed root-cause analysis to identify system factors leading to an error and establish a plan to prevent similar errors in the future. Formal training in root-cause analysis might increase its frequency of use as well as its effectiveness (14).

Economic impact of medical errors. Medical errors have huge economic consequences for the United States healthcare system, with some studies indicating a cost of up to several billion dollars a year. Nosocomial infections alone increase healthcare costs that exceed 1 billion dollars a year.

Medical errors in the practice of neurology and neurosurgery. This article deals with errors in the practice of medicine in general. The practice of neurology involves challenges in diagnosis and therapy. There are many pitfalls and an error could have harmful consequences. Neurologists may be involved in harmful errors and should familiarize themselves with tools to prevent, report, and examine errors. Lessons learned from review of errors would be helpful in preventing them and improving future management of disorders involved. There is a trend towards greater transparency following medical errors, but substantial progress is still needed. Neurologists can contribute to this movement by promoting open reporting and disclosure to patients (34). Examples of medical errors in practice of neurology are:

Headache. Among headache patients, most frequent errors occur in the management of trigeminal autonomic cephalalgias and hemicrania continua, which are relatively rare and include diagnostic delay, misdiagnosis, and the use of inappropriate treatments (32). The result is the unnecessary prescription of ineffective and sometimes invasive treatments, which may have harmful consequences for patients.

Nosocomial infections. Following subarachnoid hemorrhage, nosocomial infections have been reported in 37% of patients and mostly occur in the first week after ictus (19). The authors of this study suggest the need for further investigations to determine if general hygienic measures, infection awareness, minimizing the duration of mechanical ventilation and use of catheters/drains can reduce these infections and improve functional outcome.

Neurosurgical procedures. Errors involving high risk neurosurgical procedures can have serious consequences including death and severe neurologic disability in survivors. Several studies have started to analyze the frequency and types of errors that neurosurgical patients experience, along with their effects and causes (27). Data from these studies are guiding new innovations such as checklists and computerized order entry to reduce and prevent errors. Filming of operative procedures has been done for teaching purposes, but operative reports and recall of the surgeon and operating room staff is currently the basis for review in case of mishaps, which is not reliable. There is movement in the United States to install video cameras in operating rooms to document possible adverse events and, thereby, prevent similar ones from occurring in the future, but critics are concerned that recording devices may compromise patient and physician. This may also lead to malpractice lawsuits.

Multidisciplinary time-outs have become standard of care in neurosurgery. Despite proximity and overlapping personnel, there is considerable variability between hospitals in the practice of time-outs. This lack of uniformity, allowed for by flexible World Health Organization guidelines, may reflect the origins of surgical time-outs in general surgery, rather than neurosurgery, underscoring the potential for time-out optimization with neurosurgery-specific considerations (21).

Stroke. Rates of error in stroke diagnosis range from 2% to 26% for under-diagnosis or false-negative cases and 30% to 43% for over-diagnosis or false-positive cases (20). Whereas harm associated with stroke over-diagnosis is low, failure to diagnose stroke promptly, particularly in cases with atypical presentation, is associated with poor outcomes, as it deprives patients of treatments that need to be carried out within a short time window following the onset of stroke.

Approximately 9% of cerebrovascular events are missed at initial presentation in the emergency department, and risk of misdiagnosis is high when presenting neurologic complaints are mild, nonspecific, or transient, indicating that they are related to symptom-specific factors (31). There is a need for further research in studying causes and designing diagnostic error-reduction strategies in symptom-specific subgroups.