The influence of viewing time on visual diagnostic accuracy: Less is more.

BACKGROUND: Understanding the factors that contribute to diagnostic errors is critical if we are to correct or prevent them. Some scholars influenced by the default interventionist dual-process theory of cognition (dual-process theory) emphasise a narrow focus on individual clinician's faulty reasoning as a significant contributor. In this paper, we examine the validity of claims that dual process theory is a key to error reduction. METHODS: We examined the relationship between a clinical experience (staff and resident physicians) and viewing time on accuracy for categorising chest X-rays (CXRs) and electrocardiograms (ECGs). In two studies, participants categorised images as normal or abnormal, presented at viewing times of 175, 250, 500 and 1000 ms, to encourage System 1 processing. Study 2 extended viewing times to 1, 5, 10 and 20 s to allow time for System 2 processing and a diagnosis. Descriptives and repeated measures analysis of variance were used to analyse the proportion of true and false positive rates (TP and FP) as well as correct diagnoses. RESULTS: In Study 1, physicians were able to detect abnormal CXRs (0.78) and ECGs (0.67) with relatively high accuracy. The effect of experience was found for ECGs only, as staff physicians (0.71, 95% CI = 0.66-0.75) had higher ECG TP than resident physicians (0.63, 95% CI = 0.58-0.68) in Study 1, and staff had lower ECG FP (0.10, 95% CI = 0.03-0.18) than resident physicians (0.27, 95% CI = 0.20-0.33) in Study 2. In other comparisons, experience was equivocal for ECG FPs and CXR TPs and FPs. In Study 2, overall diagnostic accuracy was similar for both ECGs and CXRs, (0.74). There were small interactions between experience and time for TP in ECGs and FP in CXRs, which are discussed further in the discussion and offer insights into the relationship between processing and experience. CONCLUSION: Overall, our findings raise concerns about the practical application of models that link processing type to diagnostic error, or to specific diagnostic error reduction strategies.

Critical thinking, biases and dual processing: The enduring myth of generalisable skills.

CONTEXT: The myth of generalisable thinking skills in medical education is gaining popularity once again. The implications are significant as medical educators decide on how best to use limited resources to prepare trainees for safe medical practice. This myth-busting critical review cautions against the proliferation of curricular interventions based on the acquisition of generalisable skills. STRUCTURE: This paper begins by examining the recent history of general thinking skills, as defined by research in cognitive psychology and medical education. We describe three distinct epochs: (a) the Renaissance, which marked the beginning of cognitive psychology as a discipline in the 1960s and 1970s and was paralleled by educational reforms in medical education focused on problem solving and problem-based learning; (b) the Enlightenment, when an accumulation of evidence in psychology and in medical education cast doubt on the assumption of general reasoning or problem-solving skill and shifted the focus to consideration of the role of knowledge in expert clinical performance; and (c) the Counter-Enlightenment, in the current time, when the notion of general thinking skills has reappeared under different guises, but the fundamental problems related to lack of generality of skills and centrality of knowledge remain. CONCLUSIONS: The myth of general thinking skills persists, despite the lack of evidence. Progress in medical education is more likely to arise from devising strategies to improve the breadth and depth of experiential knowledge.

It's the destination: diagnostic accuracy and reasoning.

While multiple theories exist to explain the diagnostic process, there are few available assessments that reliably determine diagnostic competence in trainees. Most methods focus on aspects of the process of diagnostic reasoning, such as the relation between case features and diagnostic hypotheses. Inevitably, detailed elucidation of aspects of the process requires substantial time per case and limits the number of cases that can be examined given a limited testing time. Shifting assessment to the outcome of diagnostic reasoning, accuracy of the diagnosis, may serve as a reliable measure of diagnostic competence and would allow increased sampling across cases. The present study is a retrospective analysis of 7 large studies, conducted by 3 research teams, that all used a series of brief written cases to examine the outcome of diagnostic reasoning-the diagnosis. The studies involved over 600 clinicians ranging from final year medical students to practicing emergency physicians. For 4 studies with usable reliability data, reliability for a 2 h test ranged from .63 to .94. On average speeded tests were more reliable (.85 vs. .73).To achieve a reliability of .75 required an average test time of 1.11 h for speeded tests and 1.99 for unspeeded tests. The measure was shown to be positively correlated with both written knowledge tests and measures of problem solving derived from OSCE performance tests. This retrospective analysis provides evidence to support the implementation of outcome-based assessments of clinical reasoning.

McMaster at 50: lessons learned from five decades of PBL.

Although educators frequently act as if curricula are as standardized as drug doses (300 mg of PBL t.i.d.), such is not the case. As a case in point, at its inception, Problem Based Learning was hailed as a major curriculum innovation, with the promise of enormous gains in learning outcomes. Very quickly, ecclesiastical debates arose as what was true PBL and what was "modified PBL". Ironically, systematic reviews conducted fairly early in its evolution showed that the gains in learning outcome from PBL were neither large nor uniform (Vernon and Blake in Acad Med 68:550-563, 1993), and the most consistent finding was greater student satisfaction. In this paper, we review five decades of experience with the first PBL curriculum at McMaster. We point out how the curriculum has evolved, both theoretically and practically, in response to external influences, based both on empirical evidence and practical demands. We describe these changes in four broad domains-theoretical rationale, the curriculum, assessment and admissions.

Eyeballing: the use of visual appearance to diagnose 'sick'.

CONTEXT: Prior studies suggest that clinicians can categorise patients in an emergency room as 'sick' or 'not sick' using rapid visual assessment. The rapid nature of these decisions suggests clinicians are relying on pattern recognition or System 1 processing; however, this has not been studied experimentally. In this study, we explore the accuracy of these decisions using patient disposition (discharge, admission to ward or admission to critical care) as an objective outcome, and collect evidence to argue for the use of System 1 processing in the 'sick' or 'not sick' decision process. METHODS: Fourteen practising emergency physicians reviewed 25 videos of patients presenting to the emergency room. They were asked to predict patient disposition (discharge, admission to ward or admission to critical care) and estimate whether they were 'sick' or 'not sick' using a continuous slider on a 'sick' scale from 'not sick' (0) to 'sick' (100). We collected decision time and asked physicians to identify how they came to the decision using a continuous slider on a 'system processing' scale from 'knew immediately' (0) to 'deliberated intently' (1). RESULTS: Inter-rater reliability judging 'sick' was computed as an intraclass correlation coefficient (ICC) of 0.54. Agreement among physicians in predicting disposition was 68% with ICC of 0.44, and accuracy at predicting disposition was 55%. Physicians made their decision in an average of 10 - 11 seconds and rated 70% of their decisions as < 0.5 on the scale from 'knew immediately' (0) to 'deliberated intently' (1). CONCLUSIONS: Experienced emergency physicians are able to visually assess patients rapidly and predict disposition in a very short time, albeit with fair reliability and lower accuracy than reported previously. Subjectively, they reported that the majority of decisions were on the side of 'knew immediately', consistent with the application of System 1 processing.

Examining the Influence of Context and Professional Culture on Clinical Reasoning Through Rhetorical-Narrative Analysis.

According to the dual process model of reasoning, physicians make diagnostic decisions using two mental systems: System 1, which is rapid, unconscious, and intuitive, and System 2, which is slow, rational, and analytical. Currently, little is known about physicians' use of System 1 or intuitive reasoning in practice. In a qualitative study of clinical reasoning, physicians were asked to tell stories about times when they used intuitive reasoning while working up an acutely unwell patient, and we combine socio-narratology and rhetorical theory to analyze physicians' stories. Our analysis reveals that in describing their work, physicians draw on two competing narrative structures: one that is aligned with an evidence-based medicine approach valuing System 2 and one that is aligned with cooperative decision making involving others in the clinical environment valuing System 1. Our findings support an understanding of clinical reasoning as distributed, contextual, and influenced by professional culture.

Assessing the quality of published genetic association studies in meta-analyses: the quality of genetic studies (Q-Genie) tool.

BACKGROUND: Advances in genomics technology have led to a dramatic increase in the number of published genetic association studies. Systematic reviews and meta-analyses are a common method of synthesizing findings and providing reliable estimates of the effect of a genetic variant on a trait of interest. However, summary estimates are subject to bias due to the varying methodological quality of individual studies. We embarked on an effort to develop and evaluate a tool that assesses the quality of published genetic association studies. Performance characteristics (i.e. validity, reliability, and item discrimination) were evaluated using a sample of thirty studies randomly selected from a previously conducted systematic review. RESULTS: The tool demonstrates excellent psychometric properties and generates a quality score for each study with corresponding ratings of 'low', 'moderate', or 'high' quality. We applied our tool to a published systematic review to exclude studies of low quality, and found a decrease in heterogeneity and an increase in precision of summary estimates. CONCLUSION: This tool can be used in systematic reviews to inform the selection of studies for inclusion, to conduct sensitivity analyses, and to perform meta-regressions.

Reexamining our bias against heuristics.

Using heuristics offers several cognitive advantages, such as increased speed and reduced effort when making decisions, in addition to allowing us to make decision in situations where missing data do not allow for formal reasoning. But the traditional view of heuristics is that they trade accuracy for efficiency. Here the authors discuss sources of bias in the literature implicating the use of heuristics in diagnostic error and highlight the fact that there are also data suggesting that under certain circumstances using heuristics may lead to better decisions that formal analysis. They suggest that diagnostic error is frequently misattributed to the use of heuristics and propose an alternative view whereby content knowledge is the root cause of diagnostic performance and heuristics lie on the causal pathway between knowledge and diagnostic error or success.

Dual process models of clinical reasoning: The central role of knowledge in diagnostic expertise.

RATIONALE: Research on diagnostic reasoning has been conducted for fifty years or more. There is growing consensus that there are two distinct processes involved in human diagnostic reasoning: System 1, a rapid retrieval of possible diagnostic hypotheses, largely automatic and based to a large part on experiential knowledge, and System 2, a slower, analytical, conscious application of formal knowledge to arrive at a diagnostic conclusion. However, within this broad framework, controversy and disagreement abound. In particular, many authors have suggested that the root cause of diagnostic errors is cognitive biases originating in System 1 and propose that educating learners about the types of cognitive biases and their impact on diagnosis would have a major influence on error reduction. AIMS AND OBJECTIVES: In the present paper, we take issue with these claims. METHOD: We reviewed the literature to examine the extent to which this theoretical model is supported by the evidence. RESULTS: We show that evidence derived from fundamental research in human cognition and studies in clinical medicine challenges the basic assumptions of this theory-that errors arise in System 1 processing as a consequence of cognitive biases, and are corrected by slow, deliberative analytical processing. We claim that, to the contrary, errors derive from both System 1 and System 2 reasoning, that they arise from lack of access to the appropriate knowledge, not from errors of processing, and that the two processes are not essential to the process of diagnostic reasoning. CONCLUSIONS: The two processing modes are better understood as a consequence of the nature of the knowledge retrieved, not as independent processes.