Auswahl der wissenschaftlichen Literatur zum Thema „Biais cognitifs – Diagnostic“

Abstract Increasing evidence suggests that vision impairment may be an important modifiable risk factor for dementia, particularly in low- and middle-income settings where the prevalence of uncorrected vision impairment is high. Although prior studies in various settings, including India, have found strong associations between vision impairment and cognition, this work has not considered potential bias in cognitive testing due to vision impairment. We evaluated and adjusted for measurement differences by vision impairment status using data from the Longitudinal Aging Study in India–Diagnostic Assessment of Dementia (LASI-DAD) study (N=3780). We used Multiple Indicators Multiple Causes models to assess differential item functioning (DIF) (e.g. bias) in cognitive testing by objective near and distance vision impairment. We estimated associations between vision impairment and cognition adjusting for hypothesized confounders before and after DIF adjustment. Although there was statistical evidence of DIF (near vision: 3/10 items, distance vision: 4/10 items), differences between DIF-unadjusted and -adjusted scores were small compared to the standard error of measurement, indicating no evidence of clinically meaningful measurement differences. Both near and distance vision impairment were associated with cognition before and after DIF-adjustment; after DIF-adjustment, severe near and distance vision impairment were associated with -0.43 [95% CI -0.53--0.33] and -0.60 [-0.76--0.43] standard deviation units lower cognitive scores compared to those with normal vision, respectively. In well-conducted large-scale surveys, bias in cognitive testing due to vision impairment is likely minimal, even in low- and middle-income settings. Findings strengthen the evidence base on vision impairment as a risk factor for dementia.

This cross-sectional study aimed to clarify how cognitive biases and situational factors related to diagnostic errors among physicians. A self-reflection questionnaire survey on physicians’ most memorable diagnostic error cases was conducted at seven conferences: one each in Okayama, Hiroshima, Matsue, Izumo City, and Osaka, and two in Tokyo. Among the 147 recruited participants, 130 completed and returned the questionnaires. We recruited primary care physicians working in various specialty areas and settings (e.g., clinics and hospitals). Results indicated that the emergency department was the most common setting (47.7%), and the highest frequency of errors occurred during night-time work. An average of 3.08 cognitive biases was attributed to each error. The participants reported anchoring bias (60.0%), premature closure (58.5%), availability bias (46.2%), and hassle bias (33.1%), with the first three being most frequent. Further, multivariate logistic regression analysis for cognitive bias showed that emergency room care can easily induce cognitive bias (adjusted odds ratio 3.96, 95% CI 1.16−13.6, p-value = 0.028). Although limited to a certain extent by its sample collection, due to the sensitive nature of information regarding physicians’ diagnostic errors, this study nonetheless shows correlations with environmental factors (emergency room care situations) that induce cognitive biases which, in turn, cause diagnostic errors.

Objective:Population studies have shown that Black individuals are at higher risk for MCI and dementia than White individuals but are more likely to be underdiagnosed or misdiagnosed. Although multiple contributory factors have been identified in relation to neurocognitive diagnostic disparities among persons of color, few studies have investigated race-associated differences in MCI and dementia classification across diagnostic methods. The current study examined the agreement of cognitive classification made via semi-structured interview and neuropsychological assessment.Participants and Methods:Only participants assigned normal cognitive status or cognitive impairment with presumed Alzheimer’s etiology were included in the study. Baseline visit data in the National Alzheimer’s Coordinating Center (NACC) dataset was collected to compare correspondence of cognitive classification (normal cognition, MCI, dementia) via semi-structured interview (Clinical Dementia Rating; CDR) with formal NACC diagnostic determination. NACC diagnostic determination was further separated by single clinician and consensus diagnostic methods. Inter-rater agreement was evaluated using chi-squared tests, and respective analyses were stratified for race (Black vs White), ethnicity (Hispanic vs Non-Hispanic), and education (<12 years vs. >12 years).Results:The sample size included 4,739 Black and 26,393 White participants across 43 Alzheimer’s Disease Research Centers (ADRCs). Inter-rater analyses between CDR (semi-structured interview) versus single-clinician and formal consensus NACC diagnostic methods showed strong (all (pc>.70) consistency in cognitive diagnoses overall, irrespective of race, ethnicity, and education. The percentage of agreement between diagnostic methods was nearly 100% for those categorized as cognitively normal or with dementia. However, the agreement for MCI was considerably lower (ranging from 28-74%) and revealed a disparity in diagnostic method between Black and White individuals. White individuals diagnosed with MCI via CDR (CDR total =0.5) were more likely to be labeled as having dementia regardless of NACC diagnostic method. However, Black individuals diagnosed with MCI via CDR were equally likely to be diagnosed as cognitively normal or with dementia via the formal consensus method.Conclusions:Irrespective of race and other demographic variables, diagnostic methods had high agreement for groups labeled with normal cognition and dementia. Agreement was consistently lower for the group labeled with MCI, with Black individuals having greater variability in diagnostic differentials when diagnosed via formal consensus method. The results of the study suggest that neuropsychological assessment continues to be an integral component of diagnosing individuals with MCI, reducing possible sources of bias.

AbstractObjectives:The aim of this study was to determine the presence and severity of potential cultural and language bias in widely used cognitive and other assessment instruments, using structural MRI measures of neurodegeneration as biomarkers of disease stage and severity.Methods:Hispanic (n=75) and White non-Hispanic (WNH) (n=90) subjects were classified as cognitively normal (CN), amnestic mild cognitive impairment (aMCI) and mild dementia. Performance on the culture-fair and educationally fair Fuld Object Memory Evaluation (FOME) and Clinical Dementia Rating Scale (CDR) between Hispanics and WNHs was equivalent, in each diagnostic group. Volumetric and visually rated measures of the hippocampus entorhinal cortex, and inferior lateral ventricles (ILV) were measured on structural MRI scans for all subjects. A series of analyses of covariance, controlling for age, depression, and education, were conducted to compare the level of neurodegeneration on these MRI measures between Hispanics and WNHs in each diagnostic group.Results:Among both Hispanics and WNH groups there was a progressive decrease in volume of the hippocampus and entorhinal cortex, and an increase in volume of the ILV (indicating increasing atrophy in the regions surrounding the ILV) from CN to aMCI to mild dementia. For equivalent levels of performance on the FOME and CDR, WNHs had greater levels of neurodegeneration than did Hispanic subjects.Conclusions:Atrophy in medial temporal regions was found to be greater among WNH than Hispanic diagnostic groups, despite the lack of statistical differences in cognitive performance between these two ethnic groups. Presumably, unmeasured factors result in better cognitive performance among WNH than Hispanics for a given level of neurodegeneration. (JINS, 2018,24, 176–187)

Abstract Cognitive bias is an unavoidable aspect of human decision-making. In breast radiology, these biases contribute to missed or erroneous diagnoses and mistaken judgments. This article introduces breast radiologists to eight cognitive biases commonly encountered in breast radiology: anchoring, availability, commission, confirmation, gambler’s fallacy, omission, satisfaction of search, and outcome. In addition to illustrative cases, this article offers suggestions for radiologists to better recognize and counteract these biases at the individual level and at the organizational level.

Objective:Two prospective studies in patient with brain tumours were performed comparing the Mini Mental State Examination (MMSE) and the Montreal Cognitive Assessment (MoCA). The first assessed their feasibility and the second compared their diagnostic accuracy against a four-hour neuropsychological assessment (NPA). The introduction of the NPA decreased accrual and retention rates. We were therefore concerned regarding potential selection bias.Methods:Ninety-two patients were prospectively accrued and subsequently divided into three categories: a) no NPA required b) withdrew consent to NPA c) completed NPA. In order to quantify any potential bias introduced by the NPA, patient demographics and cognitive test scores were compared between the three groups.Results:There were significant differences in age (p<0.001), education (p=0.034), dexamethasone use (p=0.002), MMSE (p=0.005), and MoCA scores (p<0.001) across the different study groups. Furthermore, with increasing involvement of the NPA, patients' cognitive scores and educational status increased, while their age, dexamethasone use, and opioid use all decreased. Individuals who completed the NPA had higher MoCA scores than individuals who were not asked to complete the NPA (24.7 vs. 20.5; p < 0.001). In addition, this relationship held when restricting the analyses to individuals with brain metastases (p < 0.001).Conclusions:In this study, the lengthy NPA chosen introduced a statistically and clinically significant source of selection bias. These results highlight the importance of selecting brief and well tolerated assessments when possible. However, researchers are challenged by weighing the improved selection bias associated with brief assessments at the cost of reduced diagnostic accuracy.

Dans cette thèse, nous nous sommes intéressés à l'utilisation des méthodes formelles pour guider le diagnostic des erreurs humaines dans des situations d'accidents. L'application des méthodes formelles dans un tel contexte pose plusieurs difficultés. La première est de pouvoir expliquer à l'aide de la logique mathématique des situations incohérentes, donc en contradiction avec cette logique. La deuxième est de pouvoir comparer les différents diagnostics. En effet, une décision incorrecte n'est jamais le fruit du hasard mais se base sur les croyances, les désirs et les intentions de l'opérateur. Ainsi, toute erreur ne se vaut pas et il est nécessaire de formaliser et de définir ce qui fait un bon diagnostic. La première partie de la thèse présente un état de l'art des travaux en sciences humaines et sociales (SHS) sur l'erreur humaine. Nous montrons qu'il est nécessaire de distinguer deux aspects : la détermination des causes d'une prise de décision erronée et la compréhension de ces causes par la recherche de biais cognitifs. Nous présentons ensuite les principaux modèles informatiques pour la modélisation du raisonnement et l'étude de l'erreur humaine. Nous montrons que le diagnostic fondé sur la cohérence (consistency-based diagnosis) et l'opérateur de révision de croyance AGM constitue une bonne piste pour l'explication d'erreurs humaines. La deuxième partie de la thèse s'intéresse à la modélisation d'une situation d'accident et au diagnostic des décisions humaines erronées dans cette situation. Nous nous sommes basés pour cela sur une logique de croyances inspirée de la logique BDI pour la modélisation des situations d'accidents. Nous avons développé un algorithme de diagnostic itératif basé sur un opérateur de révision de croyance minimale respectant l'axiomatique AGM. Cet algorithme de diagnostic itératif à l'avantage de faciliter la distinction des erreurs de nature différentes. De plus, celui-ci est correct et complet par rapport à un algorithme de diagnostic minimal. La troisième contribution de la thèse réside dans notre travail pour définir formellement la plausibilité d'un diagnostic. Nous nous sommes basés pour cela sur la littérature des sciences humaines et plus précisément des biais cognitifs. Pour cela, nous avons développé une première taxonomie formelle des biais qui permet de définir des caractéristiques logiques communes entre les biais. À partir de cette taxonomie, nous avons pu définir huit biais cognitifs rattachés aux biais présent dans la littérature. Nous avons ensuite considéré que plus un diagnostic peut être expliqué par les biais, plus le diagnostic est plausible. Nous avons alors étudié la validité de ce modèle informatique sur deux cas d'étude d'accident de l'aviation civile. Nous montrons que nous retrouvons les explications proposées par le Bureau d'Enquêtes et d'Analyses ainsi que des explications non envisagées par les enquêteurs. Nous proposons enfin plusieurs perspectives pour améliorer notre approche. Nous pensons notamment prendre en compte les émotions et les interactions sociales dans la modélisation de la situation d'accident afin d'augmenter la variété de diagnostic possible. Enfin, nous souhaitons étendre l'évaluation des diagnostics par une méta-évaluation des biais cognitifs ainsi que par la prise en compte de l'intention d'action<br>In this thesis, we are interested in the use of formal methods to guide the diagnosis of human errors in accident situations. The application of formal methods in such a context raises several difficulties. The first one is to be able to explain with the help of mathematical logic situations that are incoherent and therefore in contradiction with this logic. The second is to be able to compare the different diagnoses. Indeed, an incorrect decision is never the work of a hazard but is based on the beliefs, desires and intentions of the operator. Thus, not all errors are equal and it is necessary to formalize and define what makes a good diagnosis. The first part of the thesis presents a state of the art of human and social sciences (HSS) work on human error. We show that it is necessary to distinguish two aspects: the determination of the causes of erroneous decision making and the understanding of these causes through the search for cognitive biases. We then present the main computer models for modeling reasoning and studying human error. We show that consistency-based diagnosis and the belief revision operator AGM is a good way to explain human errors. The second part of the thesis deals with the modeling of an accident situation and the diagnosis of human errors in this situation. We have based our work on a belief logic inspired by the BDI logic for the modeling of accident situations. We have developed an iterative diagnosis algorithm based on a minimal belief revision operator respecting the AGM axiomatic. This iterative diagnosis algorithm has the advantage of facilitating the distinction of errors of different nature. Moreover, it is correct and complete compared to a minimal diagnosis algorithm. The third contribution of the thesis lies in our work to formally define the plausibility of a diagnosis. We based our work on the literature of human sciences and more precisely on cognitive biases. For this purpose, we have developed a first formal taxonomy of biases that allows us to define common logical characteristics between biases. From this taxonomy, we were able to define eight cognitive biases related to the biases present in the literature. We then considered that the more a diagnosis can be explained by the biases, the more plausible the diagnosis is. We then studied the validity of this computer model on two cases of civil aviation accidents. We show that we find the explanations proposed by the Bureau d'Enquêtes et d'Analyses as well as explanations not considered by the investigators. Finally, we propose several perspectives to improve our approach. In particular, we intend to take into account emotions and social interactions in the modeling of the accident situation in order to increase the variety of possible diagnoses. Finally, we wish to extend the evaluation of the diagnoses by a meta-evaluation of the cognitive biases as well as by taking into account the intention of action

Apesar de muitos estudos terem sido realizados para avaliar a acurácia dos métodos de detecção de cárie, poucos avaliam possíveis fatores capazes de influenciar o desempenho dos examinadores, como por exemplo, a experiência de cárie da criança. Dessa forma, este estudo teve como objetivo avaliar se a experiência de cárie da criança influencia o desempenho do exame visual e radiográfico na detecção de lesões de cárie proximais em molares decíduos. A amostra foi composta por 80 crianças de 3 a 6 anos de idade, as quais foram classificadas de acordo com a experiência passada de cárie, considerando lesões cavitadas. Dois examinadores calibrados avaliaram 526 superfícies proximais. Inicialmente foi realizado o exame visual, e em seguida, o exame radiográfico. Como padrão de referência, outros dois examinadores realizaram o exame visual direto das superfícies após separação temporária com elástico ortodôntico. A sensibilidade, especificidade e acurácia dos métodos visual e radiográfico, isolados e em associação, foram calculadas e comparadas entre si considerando os limiares de lesões não cavitadas e de lesões cavitadas. Análises de regressão de multinível de Poisson foram utilizadas para avaliar a influência da experiência de cárie e de outras variáveis na performance das estratégias de diagnóstico. Análises estratificadas pela experiência de cárie da criança também foram conduzidas. A associação do exame radiográfico ao exame visual não melhorou a acurácia na detecção de lesões proximais de cárie em dentes decíduos nos dois limiares de um modo geral. No entanto, foi observada uma influência da experiência de cárie apenas no exame visual. A detecção de lesões não cavitadas em crianças com maior experiência de cárie foi superestimada, provavelmente devido a um viés de confirmação. Já para lesões cavitadas, a detecção pelo exame visual foi subestimada, indicando um possível viés de representatividade. Já o exame radiográfico não sofreu influência de vieses cognitivos, e a performance desse método isolado ou associado ao visual de forma simultânea foi melhor nas crianças com maior experiência de cárie. Portanto, conclui-se que a experiência de cárie da criança influencia o desempenho dos examinadores ao utilizar o exame visual na detecção de lesões proximais de cárie em dentes decíduos.<br>Despite many studies have already been conducted to investigate the accuracy of caries detection methods, few investigations have evaluated the influence of some factors on the examiners\' performance, such as child\'s caries experience. Thus, this study aimed to evaluate if the child\'s caries experience exerts some influence on the performance of visual and radiographic methods for the detection of approximal caries lesions in primary molars. Eighty children aged from 3 to 6 years were selected and classified according to the past caries experience considering cavitated lesions. Two calibrated examiners evaluated 526 approximal surfaces for the presence of caries lesions using visual inspection and radiographic methods. As reference standard, two other examiners checked the surfaces by direct visual inspection after the temporary separation with orthodontic rubbers. Sensitivity, specificity, and accuracy obtained with visual inspection and radiographic method, alone or associated, were calculated and compared considering non-cavitated and cavitated lesions thresholds. Poisson multilevel regression analyses were conducted to evaluate the influence of the caries experience on the performance of diagnostic strategies. Radiographic examination and visual inspection performed associated did not improve the accuracy in detecting approximal caries lesions in both thresholds. However, an influence of child\'s caries experience was observe only on the visual inspection. The detection of non-cavitated caries lesions in children with higher caries experience was overestimated, probably due to confirmation bias. On the other hand, considering cavitated caries lesions, the performance of visual inspection was underestimated, indicating the occurrence of representativeness bias. Nevertheless, the radiographic method did not suffer influence of any type of cognitive bias, and the performance of this method, alone or simultaneously associated with visual inspection, was better in children with higher caries experience. In conclusion, the child\'s caries experience exerts influence on visual inspection in detecting approximal caries lesions in primary teeth.

This book provides, without the use of specialist language, a description of diagnostic reasoning and error and a discussion of steps that could improve diagnostic accuracy. Drawing on work in cognitive psychology, it presents the key characteristics of human reasoning. It notes that complex cognitive tasks such as medical diagnosis require a synergy of intuition and analytical thinking and introduces the concept of bias. The book considers the value of current classifications of disease, the meaning of diagnostic thresholds, and the potential for overdiagnosis. It examines the role of the patient-centred approach in this context. It develops a description of the diagnostic process, provides illustrative examples and metaphors, and refers to the dual-process model. It suggests that medical training does not consistently provide a coherent account of diagnostic thinking and the associated risks of error. It considers the role of probability in diagnostic reasoning, noting the contribution and the limitations of both informal and mathematical estimates. It refers to clear evidence that error in medical diagnosis is a prevalent and potent cause of harm and may result from systems factors or cognitive glitches such as bias and logical fallacy. It presents cases with commentaries, highlighting the cognitive processes in diagnostic successes, near misses, and disasters. It concludes with proposals for change, notably in institutional culture; in professional culture, education, and training; and in the structure of medical records. The book advocates the development and deployment of computerized diagnostic decision support. It argues that these changes could significantly enhance patient safety.

Chapter 10 addresses the salient alternative effects on knowledge ascriptions by developing the epistemic focal bias account. According to this account, denials of knowledge in the face of a salient alternative often amount to false negatives. But while this is argued to be central to a comprehensive diagnosis, it is recognized that other psychological factors may also influence this class of judgments, and some of these are discussed. Furthermore, the epistemic focal bias account is integrated with a number of assumptions drawn from cognitive pragmatics. In this manner, Chapter 10 provides an empirical account and philosophical diagnosis of the puzzling pattern of knowledge ascriptions constituted by salient alternative effects.

Anxiety disorders are highly common in children and adolescents and are associated with significant impairment. This group of disorders includes a broad range of specific diagnoses that often co-occur. Well-established assessment measures exist to facilitate accurate differential diagnosis and characterization of anxiety disorders. Evidence-based treatments also are available. Cognitive behavior therapy has a uniquely broad and robust evidence base, although newer treatments such as attention bias modification training and parent accommodation interventions are the source of growing attention. Current research in the field includes attempts at understanding the basic nature of anxiety disorders, the development of new treatments, and innovative approaches to addressing the key challenge of limited access to treatment.

Objective. The systematic review assessed evidence on the diagnosis, treatment, and monitoring of attention deficit hyperactivity disorder (ADHD) in children and adolescents to inform a planned update of the American Academy of Pediatrics (AAP) guidelines. Data sources. We searched PubMed®, Embase®, PsycINFO®, ERIC, clinicaltrials.gov, and prior reviews for primary studies published since 1980. The report includes studies published to June 15, 2023. Review methods. The review followed a detailed protocol and was supported by a Technical Expert Panel. Citation screening was facilitated by machine learning; two independent reviewers screened full text citations for eligibility. We abstracted data using software designed for systematic reviews. Risk of bias assessments focused on key sources of bias for diagnostic and intervention studies. We conducted strength of evidence (SoE) and applicability assessments for key outcomes. The protocol for the review has been registered in PROSPERO (CRD42022312656). Results. Searches identified 23,139 citations, and 7,534 were obtained as full text. We included 550 studies reported in 1,097 publications (231 studies addressed diagnosis, 312 studies addressed treatment, and 10 studies addressed monitoring). Diagnostic studies reported on the diagnostic performance of numerous parental ratings, teacher rating scales, teen/child self-reports, clinician tools, neuropsychological tests, EEG approaches, imaging, and biomarkers. Multiple approaches showed promising diagnostic performance (e.g., using parental rating scales), although estimates of performance varied considerably across studies and the SoE was generally low. Few studies reported estimates for children under the age of 7. Treatment studies evaluated combined pharmacological and behavior approaches, medication approved by the Food and Drug Administration, other pharmacologic treatment, psychological/behavioral approaches, cognitive training, neurofeedback, neurostimulation, physical exercise, nutrition and supplements, integrative medicine, parent support, school interventions, and provider or model-of-care interventions. Medication treatment was associated with improved broadband scale scores and ADHD symptoms (high SoE) as well as function (moderate SoE), but also appetite suppression and adverse events (high SoE). Psychosocial interventions also showed improvement in ADHD symptoms based on moderate SoE. Few studies have evaluated combinations of pharmacological and youth-directed psychosocial interventions, and we did not find combinations that were systematically superior to monotherapy (low SoE). Published monitoring approaches for ADHD were limited and the SoE is insufficient. Conclusion. Many diagnostic tools are available to aid the diagnosis of ADHD, but few monitoring strategies have been studied. Medication therapies remain important treatment options, although with a risk of side effects, as the evidence base for psychosocial therapies strengthens and other nondrug treatment approaches emerge.

Objectives. Diagnostic errors are a known patient safety concern across all clinical settings, including the emergency department (ED). We conducted a systematic review to determine the most frequent diseases and clinical presentations associated with diagnostic errors (and resulting harms) in the ED, measure error and harm frequency, as well as assess causal factors. Methods. We searched PubMed®, Cumulative Index to Nursing and Allied Health Literature (CINAHL®), and Embase® from January 2000 through September 2021. We included research studies and targeted grey literature reporting diagnostic errors or misdiagnosis-related harms in EDs in the United States or other developed countries with ED care deemed comparable by a technical expert panel. We applied standard definitions for diagnostic errors, misdiagnosis-related harms (adverse events), and serious harms (permanent disability or death). Preventability was determined by original study authors or differences in harms across groups. Two reviewers independently screened search results for eligibility; serially extracted data regarding common diseases, error/harm rates, and causes/risk factors; and independently assessed risk of bias of included studies. We synthesized results for each question and extrapolated U.S. estimates. We present 95 percent confidence intervals (CIs) or plausible range (PR) bounds, as appropriate. Results. We identified 19,127 citations and included 279 studies. The top 15 clinical conditions associated with serious misdiagnosis-related harms (accounting for 68% [95% CI 66 to 71] of serious harms) were (1) stroke, (2) myocardial infarction, (3) aortic aneurysm and dissection, (4) spinal cord compression and injury, (5) venous thromboembolism, (6/7 – tie) meningitis and encephalitis, (6/7 – tie) sepsis, (8) lung cancer, (9) traumatic brain injury and traumatic intracranial hemorrhage, (10) arterial thromboembolism, (11) spinal and intracranial abscess, (12) cardiac arrhythmia, (13) pneumonia, (14) gastrointestinal perforation and rupture, and (15) intestinal obstruction. Average disease-specific error rates ranged from 1.5 percent (myocardial infarction) to 56 percent (spinal abscess), with additional variation by clinical presentation (e.g., missed stroke average 17%, but 4% for weakness and 40% for dizziness/vertigo). There was also wide, superimposed variation by hospital (e.g., missed myocardial infarction 0% to 29% across hospitals within a single study). An estimated 5.7 percent (95% CI 4.4 to 7.1) of all ED visits had at least one diagnostic error. Estimated preventable adverse event rates were as follows: any harm severity (2.0%, 95% CI 1.0 to 3.6), any serious harms (0.3%, PR 0.1 to 0.7), and deaths (0.2%, PR 0.1 to 0.4). While most disease-specific error rates derived from mainly U.S.-based studies, overall error and harm rates were derived from three prospective studies conducted outside the United States (in Canada, Spain, and Switzerland, with combined n=1,758). If overall rates are generalizable to all U.S. ED visits (130 million, 95% CI 116 to 144), this would translate to 7.4 million (PR 5.1 to 10.2) ED diagnostic errors annually; 2.6 million (PR 1.1 to 5.2) diagnostic adverse events with preventable harms; and 371,000 (PR 142,000 to 909,000) serious misdiagnosis-related harms, including more than 100,000 permanent, high-severity disabilities and 250,000 deaths. Although errors were often multifactorial, 89 percent (95% CI 88 to 90) of diagnostic error malpractice claims involved failures of clinical decision-making or judgment, regardless of the underlying disease present. Key process failures were errors in diagnostic assessment, test ordering, and test interpretation. Most often these were attributed to inadequate knowledge, skills, or reasoning, particularly in “atypical” or otherwise subtle case presentations. Limitations included use of malpractice claims and incident reports for distribution of diseases leading to serious harms, reliance on a small number of non-U.S. studies for overall (disease-agnostic) diagnostic error and harm rates, and methodologic variability across studies in measuring disease-specific rates, determining preventability, and assessing causal factors. Conclusions. Although estimated ED error rates are low (and comparable to those found in other clinical settings), the number of patients potentially impacted is large. Not all diagnostic errors or harms are preventable, but wide variability in diagnostic error rates across diseases, symptoms, and hospitals suggests improvement is possible. With 130 million U.S. ED visits, estimated rates for diagnostic error (5.7%), misdiagnosis-related harms (2.0%), and serious misdiagnosis-related harms (0.3%) could translate to more than 7 million errors, 2.5 million harms, and 350,000 patients suffering potentially preventable permanent disability or death. Over two-thirds of serious harms are attributable to just 15 diseases and linked to cognitive errors, particularly in cases with “atypical” manifestations. Scalable solutions to enhance bedside diagnostic processes are needed, and these should target the most commonly misdiagnosed clinical presentations of key diseases causing serious harms. New studies should confirm overall rates are representative of current U.S.-based ED practice and focus on identified evidence gaps (errors among common diseases with lower-severity harms, pediatric ED errors and harms, dynamic systems factors such as overcrowding, and false positives). Policy changes to consider based on this review include: (1) standardizing measurement and research results reporting to maximize comparability of measures of diagnostic error and misdiagnosis-related harms; (2) creating a National Diagnostic Performance Dashboard to track performance; and (3) using multiple policy levers (e.g., research funding, public accountability, payment reforms) to facilitate the rapid development and deployment of solutions to address this critically important patient safety concern.