DOCSLIB.ORG

Cognitive Debiasing 1: Origins of Bias and Theory of Debiasing

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Cognitive Debiasing 1: Origins of Bias and Theory of Debiasing NARRATIVE REVIEW BMJ Qual Saf: first published as 10.1136/bmjqs-2012-001712 on 23 July 2013. Downloaded from Cognitive debiasing 1: origins of bias and theory of debiasing Pat Croskerry,1 Geeta Singhal,2 Sílvia Mamede3 1Department of Pediatrics, ABSTRACT used. They are fast, usually effective, but Division of Medical Education, Numerous studies have shown that diagnostic also more likely to fail. As they are largely Dalhousie University, Halifax, — Nova Scotia, Canada failure depends upon a variety of factors. unconscious, mistakes when they occur 1–3 2Baylor College of Medicine Psychological factors are fundamental in —are seldom corrected. In contrast, Texas Children’s Hospital, influencing the cognitive performance of the Type 2 processes are fairly reliable, safe Houston, Texas, USA decision maker. In this first of two papers, we and effective, but slow and resource inten- 3Institute of Medical Education Research, Rotterdam, Erasmus discuss the basics of reasoning and the Dual sive. The intuitive mode of decision University Medical Center, Process Theory (DPT) of decision making. The making is characterised by heuristics— Rotterdam general properties of the DPT model, as it applies short-cuts, abbreviated ways of thinking, ‘ ’ Correspondence to to diagnostic reasoning, are reviewed. A variety maxims, seen this many times before , Professor Pat Croskerry, Division of cognitive and affective biases are known to ways of thinking. Heuristics represent an of Medical Education, Dalhousie compromise the decision-making process. They adaptive mechanism that saves us time University, Clinical Research mostly appear to originate in the fast intuitive and effort while making daily decisions. Centre, 5849 University Avenue, PO Box 15000, Halifax, NS, processes of Type 1 that dominate (or drive) Indeed, it is a rule of thumb among cogni- Canada B3H 4R2; decision making. Type 1 processes work well tive psychologists that we spend about [email protected] most of the time but they may open the door for 95% of our time in the intuitive mode.4 biases. Removing or at least mitigating these We perform many of our daily activities Received 23 November 2012 Revised 1 May 2013 biases would appear to be an important goal. through serial associations—one event Accepted 9 June 2013 We will also review the origins of biases. The automatically triggers the next with few Published Online First consensus is that there are two major sources: events of deliberate, focused, analytical 23 July 2013 innate, hard-wired biases that developed in our thinking. We have a prevailing disposition evolutionary past, and acquired biases to use heuristics, and while they work established in the course of development and well most of the time, they are vulnerable http://qualitysafety.bmj.com/ within our working environments. Both are to error. Our systematic errors are termed associated with abbreviated decision making in biases,3 and there are many of them— the form of heuristics. Other work suggests that over a hundred cognitive biases5 and ambient and contextual factors may create high approximately one dozen or so affective risk situations that dispose decision makers to biases (ways in which our feelings influ- particular biases. Fatigue, sleep deprivation and ence our judgment).6 Bias is inherent in cognitive overload appear to be important human judgment, and physicians are, of determinants. The theoretical basis of several course, also subject to them. approaches towards debiasing is then discussed. Indeed, one of the principal factors on October 2, 2021 by guest. Protected copyright. – All share a common feature that involves a underlying diagnostic error is bias.7 9 deliberate decoupling from Type 1 intuitive Post hoc analyses of diagnostic errors10 11 Open Access Scan to access more processing and moving to Type 2 analytical have in fact suggested that flaws in clin- free content processing so that eventually unexamined ical reasoning rather than lack of knowl- intuitive judgments can be submitted to edge underlie cognitive diagnostic errors, verification. This decoupling step appears to be and there is some experimental evidence the critical feature of cognitive and affective that, at least when problems are complex, debiasing. errors were associated with intuitive judg- ments and could be repaired by analytical – INTRODUCTION reasoning.12 14 Moreover, a few experi- ▸ http://dx.doi.org/10.1136/ bmjqs-2013-002387 Clinical decision making is a complex mental studies have supported the claim process. Clinical decisions about patient’s that bias may misdirect diagnostic reason- diagnoses are made in one of two modes: ing, thus leading to errors.15 16 While the To cite: Croskerry P, either intuitive or analytical, also referred evidence for the role of bias in medical Singhal G, Mamede S. BMJ to, respectively as Type 1 and Type 2 pro- diagnostic error is still scarce, research Qual Saf 2013;22:ii58–ii64. cesses. The former are more commonly findings in other domains13is sufficient ii58 Croskerry P, et al. BMJ Qual Saf 2013;22:ii58–ii64. doi:10.1136/bmjqs-2012-001712 Narrative review BMJ Qual Saf: first published as 10.1136/bmjqs-2012-001712 on 23 July 2013. Downloaded from to justify concerns with the potential adverse influ- such as functional MRI (fMRI) can be experimentally ence of bias on diagnostic reasoning. Two clinical employed. examples of biased decision making leading to diag- In this first paper, we discuss how biases are gener- nostic failure are given in Case 1 and Case 2, dis- ated, and situations that make physicians more vulner- played, respectively, in boxes 1 and 2. From the case able to bias. Building upon dual process theories descriptions, the mode of decision making the phys- (DPTs) of reasoning,13we discuss the origins of ician was relying upon cannot be determined. This is indeed a limitation of studies on diagnostic reasoning, which only have indirect evidence or post hoc infer- ence of reasoning processes, at least until other tools Box 2 Example of biased decision making leading to diagnostic failure: case 2 A mildly obese, 19-year-old woman is admitted to a psy- chiatric hospital for stabilisation and investigation. She Box 1 Example of biased decision making leading has suffered depressive symptoms accompanied by to diagnostic failure: case 1 marked anxiety. Over the last week she has had bouts of rapid breathing which have been attributed to her A 55-year-old man presents to a walk-in clinic towards anxiety. However, she has also exhibited mild symptoms the end of the evening. It has been a busy day for the of a respiratory infection and the psychiatry resident clinic and they are about to close. His chief complaint is transfers her to a nearby emergency department (ED) of constipation. He has not had a bowel movement in a tertiary care hospital to ‘rule out pneumonia’. She is on 4 days, which is unusual for him. He complains of pain no medications other than birth control pills. At triage, in his lower back and lower abdomen and also some tin- she is noted to have an elevated heart rate and respira- gling in his legs. He thinks that he will feel better with a tory rate. She is uncomfortable, anxious and impatient, laxative because what he has tried so far has not and does not want to be in the ED. She is noted to be worked. He was briefly examined by the physician who ‘difficult’ with the nurses. did not find anything remarkable on his abdominal After several hours she is seen by an emergency medi- examination. There was some mild suprapubic tender- cine resident who finds her very irritable but notes ness, which was attributed to the patient’s need to nothing remarkable on her chest or cardiac examination. urinate. Bowel sounds were good, his abdomen was soft However, to ensure that pneumonia is ruled out he and there were no masses or organomegaly. The phys- orders a chest X-ray. He reviews the patient with his ician prescribed a stronger laxative and advised the attending noting that the patient is anxious to return to http://qualitysafety.bmj.com/ patient to contact his family doctor for further follow-up. the psychiatric facility and is only at the ED ‘because she During the night, the patient was unable to urinate was told to come’. He expresses his view that her symp- and went to the emergency department. On examination, toms are attributable to her anxiety and that she does his lower abdomen appeared distended and he was not have pneumonia. Nevertheless, he asks the attending found to have a residual volume of 1200cc on catheter- to review the chest X-ray to ensure he has not missing isation. His rectum was markedly distended with soft something. The attending confirmed that there was no stool that required disimpaction. He recalled straining his evidence of pneumonia and agreed with the resident lower back lifting about 4 days earlier. The emergency that the patient could be returned to the psychiatric physician suspected cauda equina syndrome and this was hospital. on October 2, 2021 by guest. Protected copyright. confirmed on MRI. He was taken immediately to the While awaiting transfer back to the psychiatric hospital operating room for surgical decompression. He did well the patient requests permission on several occasions to postoperatively and regained full bladder control. go outside for a cigarette and is allowed to do so. Later, Comment: The patient was initially misdiagnosed and on the sidewalk outside the ED, the patient has a cardiac might have suffered permanent loss of bladder function arrest. She is immediately brought back into the ED but requiring lifelong catheterisation. The principle biases for could not be resuscitated. At autopsy, massive pulmonary the physician who saw him in the clinic were framing, saddle emboli are found as well as multiple small emboli search satisficing and premature diagnostic closure.
Load more

Recommended publications
  • Avoiding the Conjunction Fallacy: Who Can Take a Hint?
    Avoiding the conjunction fallacy: Who can take a hint? Simon Klein Spring 2017 Master’s thesis, 30 ECTS Master’s programme in Cognitive Science, 120 ECTS Supervisor: Linnea Karlsson Wirebring Acknowledgments: The author would like to thank the participants for enduring the test session with challenging questions and thereby making the study possible, his supervisor Linnea Karlsson Wirebring for invaluable guidance and good questions during the thesis work, and his fiancée Amanda Arnö for much needed mental support during the entire process. 2 AVOIDING THE CONJUNCTION FALLACY: WHO CAN TAKE A HINT? Simon Klein Humans repeatedly commit the so called “conjunction fallacy”, erroneously judging the probability of two events occurring together as higher than the probability of one of the events. Certain hints have been shown to mitigate this tendency. The present thesis investigated the relations between three psychological factors and performance on conjunction tasks after reading such a hint. The factors represent the understanding of probability and statistics (statistical numeracy), the ability to resist intuitive but incorrect conclusions (cognitive reflection), and the willingness to engage in, and enjoyment of, analytical thinking (need-for-cognition). Participants (n = 50) answered 30 short conjunction tasks and three psychological scales. A bimodal response distribution motivated dichotomization of performance scores. Need-for-cognition was significantly, positively correlated with performance, while numeracy and cognitive reflection were not. The results suggest that the willingness to engage in, and enjoyment of, analytical thinking plays an important role for the capacity to avoid the conjunction fallacy after taking a hint. The hint further seems to neutralize differences in performance otherwise predicted by statistical numeracy and cognitive reflection.
    [Show full text]
  • Cognitive Bias Mitigation: How to Make Decision-Making More Rational?
    Cognitive Bias Mitigation: How to make decision-making more rational? Abstract Cognitive biases distort judgement and adversely impact decision-making, which results in economic inefficiencies. Initial attempts to mitigate these biases met with little success. However, recent studies which used computer games and educational videos to train people to avoid biases (Clegg et al., 2014; Morewedge et al., 2015) showed that this form of training reduced selected cognitive biases by 30 %. In this work I report results of an experiment which investigated the debiasing effects of training on confirmation bias. The debiasing training took the form of a short video which contained information about confirmation bias, its impact on judgement, and mitigation strategies. The results show that participants exhibited confirmation bias both in the selection and processing of information, and that debiasing training effectively decreased the level of confirmation bias by 33 % at the 5% significance level. Key words: Behavioural economics, cognitive bias, confirmation bias, cognitive bias mitigation, confirmation bias mitigation, debiasing JEL classification: D03, D81, Y80 1 Introduction Empirical research has documented a panoply of cognitive biases which impair human judgement and make people depart systematically from models of rational behaviour (Gilovich et al., 2002; Kahneman, 2011; Kahneman & Tversky, 1979; Pohl, 2004). Besides distorted decision-making and judgement in the areas of medicine, law, and military (Nickerson, 1998), cognitive biases can also lead to economic inefficiencies. Slovic et al. (1977) point out how they distort insurance purchases, Hyman Minsky (1982) partly blames psychological factors for economic cycles. Shefrin (2010) argues that confirmation bias and some other cognitive biases were among the significant factors leading to the global financial crisis which broke out in 2008.
    [Show full text]
  • Drug Exposure Definition and Healthy User Bias Impacts on the Evaluation of Oral Anti Hyperglycemic Therapies
    Drug Exposure Definition and Healthy User Bias Impacts on the Evaluation of Oral Anti Hyperglycemic Therapies by Maxim Eskin A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Health Policy Research School of Public Health University of Alberta © Maxim Eskin, 2017 ABSTRACT Accurate estimation of medication use is an essential component of any pharmacoepidemiological research as exposure misclassification will threaten study validity and lead to spurious associations. Many pharmacoepidemiological studies use simple definitions, such as the categorical “any versus no use” to classify oral antihyperglycemic medications exposure, which has potentially serious drawbacks. This approach has led to numerous highly publicized observational studies of metformin effect on health outcomes reporting exaggerated relationships that were later contradicted by randomized controlled trials. Although selection bias, unmeasured confounding, and many other factors contribute to the discrepancies, one critical element, which is often overlooked, is the method used to define exposure. Another factor that may provide additional explanation for the discrepancy between randomized controlled trials and observational study results of the association between metformin use and various health outcomes is the healthy user effect. Aspects of a healthy lifestyle such as following a balanced diet, exercising regularly, avoiding tobacco use, etc., are not recorded in typical administrative databases and failure to account for these factors in observational studies may introduce bias and produce spurious associations. The influence of a healthy user bias has not been fully examined when evaluating the effect of oral antihyperglycemic therapies in observational studies. It is possible that some, if not all, of the benefit, observed with metformin in observational studies may be related to analytic design and exposure definitions.
    [Show full text]
  • Graphical Techniques in Debiasing: an Exploratory Study
    GRAPHICAL TECHNIQUES IN DEBIASING: AN EXPLORATORY STUDY by S. Bhasker Information Systems Department Leonard N. Stern School of Business New York University New York, New York 10006 and A. Kumaraswamy Management Department Leonard N. Stern School of Business New York University New York, NY 10006 October, 1990 Center for Research on Information Systems Information Systems Department Leonard N. Stern School of Business New York University Working Paper Series STERN IS-90-19 Forthcoming in the Proceedings of the 1991 Hawaii International Conference on System Sciences Center for Digital Economy Research Stem School of Business IVorking Paper IS-90-19 Center for Digital Economy Research Stem School of Business IVorking Paper IS-90-19 2 Abstract Base rate and conjunction fallacies are consistent biases that influence decision making involving probability judgments. We develop simple graphical techniques and test their eflcacy in correcting for these biases. Preliminary results suggest that graphical techniques help to overcome these biases and improve decision making. We examine the implications of incorporating these simple techniques in Executive Information Systems. Introduction Today, senior executives operate in highly uncertain environments. They have to collect, process and analyze a deluge of information - most of it ambiguous. But, their limited information acquiring and processing capabilities constrain them in this task [25]. Increasingly, executives rely on executive information/support systems for various purposes like strategic scanning of their business environments, internal monitoring of their businesses, analysis of data available from various internal and external sources, and communications [5,19,32]. However, executive information systems are, at best, support tools. Executives still rely on their mental or cognitive models of their businesses and environments and develop heuristics to simplify decision problems [10,16,25].
    [Show full text]
  • Cognitive Biases in Economic Decisions – Three Essays on the Impact of Debiasing
    TECHNISCHE UNIVERSITÄT MÜNCHEN Lehrstuhl für Betriebswirtschaftslehre – Strategie und Organisation Univ.-Prof. Dr. Isabell M. Welpe Cognitive biases in economic decisions – three essays on the impact of debiasing Christoph Martin Gerald Döbrich Abdruck der von der Fakultät für Wirtschaftswissenschaften der Technischen Universität München zur Erlangung des akademischen Grades eines Doktors der Wirtschaftswissenschaften (Dr. rer. pol.) genehmigten Dissertation. Vorsitzender: Univ.-Prof. Dr. Gunther Friedl Prüfer der Dissertation: 1. Univ.-Prof. Dr. Isabell M. Welpe 2. Univ.-Prof. Dr. Dr. Holger Patzelt Die Dissertation wurde am 28.11.2012 bei der Technischen Universität München eingereicht und durch die Fakultät für Wirtschaftswissenschaften am 15.12.2012 angenommen. Acknowledgments II Acknowledgments Numerous people have contributed to the development and successful completion of this dissertation. First of all, I would like to thank my supervisor Prof. Dr. Isabell M. Welpe for her continuous support, all the constructive discussions, and her enthusiasm concerning my dissertation project. Her challenging questions and new ideas always helped me to improve my work. My sincere thanks also go to Prof. Dr. Matthias Spörrle for his continuous support of my work and his valuable feedback for the articles building this dissertation. Moreover, I am grateful to Prof. Dr. Dr. Holger Patzelt for acting as the second advisor for this thesis and Professor Dr. Gunther Friedl for leading the examination board. This dissertation would not have been possible without the financial support of the Elite Network of Bavaria. I am very thankful for the financial support over two years which allowed me to pursue my studies in a focused and efficient manner. Many colleagues at the Chair for Strategy and Organization of Technische Universität München have supported me during the completion of this thesis.
    [Show full text]
  • Working Memory, Cognitive Miserliness and Logic As Predictors of Performance on the Cognitive Reflection Test
    Working Memory, Cognitive Miserliness and Logic as Predictors of Performance on the Cognitive Reflection Test Edward J. N. Stupple ([email protected]) Centre for Psychological Research, University of Derby Kedleston Road, Derby. DE22 1GB Maggie Gale ([email protected]) Centre for Psychological Research, University of Derby Kedleston Road, Derby. DE22 1GB Christopher R. Richmond ([email protected]) Centre for Psychological Research, University of Derby Kedleston Road, Derby. DE22 1GB Abstract Most participants respond that the answer is 10 cents; however, a slower and more analytic approach to the The Cognitive Reflection Test (CRT) was devised to measure problem reveals the correct answer to be 5 cents. the inhibition of heuristic responses to favour analytic ones. The CRT has been a spectacular success, attracting more Toplak, West and Stanovich (2011) demonstrated that the than 100 citations in 2012 alone (Scopus). This may be in CRT was a powerful predictor of heuristics and biases task part due to the ease of administration; with only three items performance - proposing it as a metric of the cognitive miserliness central to dual process theories of thinking. This and no requirement for expensive equipment, the practical thesis was examined using reasoning response-times, advantages are considerable. There have, moreover, been normative responses from two reasoning tasks and working numerous correlates of the CRT demonstrated, from a wide memory capacity (WMC) to predict individual differences in range of tasks in the heuristics and biases literature (Toplak performance on the CRT. These data offered limited support et al., 2011) to risk aversion and SAT scores (Frederick, for the view of miserliness as the primary factor in the CRT.
    [Show full text]
  • 10. Sample Bias, Bias of Selection and Double-Blind
    SEC 4 Page 1 of 8 10. SAMPLE BIAS, BIAS OF SELECTION AND DOUBLE-BLIND 10.1 SAMPLE BIAS: In statistics, sampling bias is a bias in which a sample is collected in such a way that some members of the intended population are less likely to be included than others. It results in abiased sample, a non-random sample[1] of a population (or non-human factors) in which all individuals, or instances, were not equally likely to have been selected.[2] If this is not accounted for, results can be erroneously attributed to the phenomenon under study rather than to the method of sampling. Medical sources sometimes refer to sampling bias as ascertainment bias.[3][4] Ascertainment bias has basically the same definition,[5][6] but is still sometimes classified as a separate type of bias Types of sampling bias Selection from a specific real area. For example, a survey of high school students to measure teenage use of illegal drugs will be a biased sample because it does not include home-schooled students or dropouts. A sample is also biased if certain members are underrepresented or overrepresented relative to others in the population. For example, a "man on the street" interview which selects people who walk by a certain location is going to have an overrepresentation of healthy individuals who are more likely to be out of the home than individuals with a chronic illness. This may be an extreme form of biased sampling, because certain members of the population are totally excluded from the sample (that is, they have zero probability of being selected).
    [Show full text]
  • The Challenge of Debiasing Personnel Decisions: Avoiding Both Under- and Overcorrection
    University of Pennsylvania ScholarlyCommons Management Papers Wharton Faculty Research 12-2008 The Challenge of Debiasing Personnel Decisions: Avoiding Both Under- and Overcorrection Philip E. Tetlock University of Pennsylvania Gregory Mitchell Terry L. Murray Follow this and additional works at: https://repository.upenn.edu/mgmt_papers Part of the Business Administration, Management, and Operations Commons Recommended Citation Tetlock, P. E., Mitchell, G., & Murray, T. L. (2008). The Challenge of Debiasing Personnel Decisions: Avoiding Both Under- and Overcorrection. Industrial and Organizational Psychology, 1 (4), 439-443. http://dx.doi.org/10.1111/j.1754-9434.2008.00084.x This paper is posted at ScholarlyCommons. https://repository.upenn.edu/mgmt_papers/32 For more information, please contact [email protected]. The Challenge of Debiasing Personnel Decisions: Avoiding Both Under- and Overcorrection Disciplines Business Administration, Management, and Operations This journal article is available at ScholarlyCommons: https://repository.upenn.edu/mgmt_papers/32 1 The Challenge of Debiasing Personnel Decisions: Avoiding Both Under- and Over-Correction Philip E. Tetlock, * Gregory Mitchell, ** and Terry L. Murray *** Introduction This commentary advances two interrelated scientific arguments. First, we endorse Landy's (2008) concerns about the insufficient emphasis placed on individuating information by scholars eager to import social-cognition work on stereotyping into employment law. Building on Landy’s analysis, we emphasize that greater attention needs to be given to the power of accountability and teamwork incentives to motivate personnel decision-makers to seek and utilize individuating information that is predictive of job-relevant behavior. Second, we note how easy it is for exchanges between proponents and skeptics of unconscious stereotyping to lead to ideological stalemates.
    [Show full text]
  • Mitigating the Attraction Effect with Visualizations Evanthia Dimara, Gilles Bailly, Anastasia Bezerianos, Steven Franconeri
    Mitigating the Attraction Effect with Visualizations Evanthia Dimara, Gilles Bailly, Anastasia Bezerianos, Steven Franconeri To cite this version: Evanthia Dimara, Gilles Bailly, Anastasia Bezerianos, Steven Franconeri. Mitigating the Attrac- tion Effect with Visualizations. IEEE Transactions on Visualization and Computer Graphics, Insti- tute of Electrical and Electronics Engineers, 2019, TVCG 2019 (InfoVis 2018), 25 (1), pp.850 - 860. 10.1109/TVCG.2018.2865233. hal-01845004v2 HAL Id: hal-01845004 https://hal.inria.fr/hal-01845004v2 Submitted on 22 Aug 2018 HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in France or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés. This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TVCG.2018.2865233, IEEE Transactions on Visualization and Computer Graphics Mitigating the Attraction Effect with Visualizations Evanthia Dimara, Gilles Bailly, Anastasia Bezerianos, and Steven Franconeri Abstract—Human decisions are prone to biases, and this is no less true for decisions made within data visualizations. Bias mitigation strategies often focus on the person, by educating people about their biases, typically with little success. We focus instead on the system, presenting the first evidence that altering the design of an interactive visualization tool can mitigate a strong bias – the attraction effect.
    [Show full text]
  • How a Machine Learns and Fails 2019
    Repositorium für die Medienwissenschaft Matteo Pasquinelli How a Machine Learns and Fails 2019 https://doi.org/10.25969/mediarep/13490 Veröffentlichungsversion / published version Zeitschriftenartikel / journal article Empfohlene Zitierung / Suggested Citation: Pasquinelli, Matteo: How a Machine Learns and Fails. In: spheres: Journal for Digital Cultures. Spectres of AI (2019), Nr. 5, S. 1–17. DOI: https://doi.org/10.25969/mediarep/13490. Erstmalig hier erschienen / Initial publication here: https://spheres-journal.org/wp-content/uploads/spheres-5_Pasquinelli.pdf Nutzungsbedingungen: Terms of use: Dieser Text wird unter einer Creative Commons - This document is made available under a creative commons - Namensnennung - Nicht kommerziell - Keine Bearbeitungen 4.0/ Attribution - Non Commercial - No Derivatives 4.0/ License. For Lizenz zur Verfügung gestellt. Nähere Auskünfte zu dieser Lizenz more information see: finden Sie hier: http://creativecommons.org/licenses/by-nc-nd/4.0/ http://creativecommons.org/licenses/by-nc-nd/4.0/ © the author(s) 2019 www.spheres-journal.org ISSN 2363-8621 #5 Spectres of AI sadfasdf MATTEO PASQUINELLI HOW A MACHINE LEARNS AND FAILS – A GRAMMAR OF ERROR FOR ARTIFICIAL INTELLIGENCE “Once the characteristic numbers are established for most concepts, mankind will then possess a new instrument which will enhance the capabilities of the mind to a far greater extent than optical instruments strengthen the eyes, and will supersede the microscope and telescope to the same extent that reason is superior to eyesight.”1 — Gottfried Wilhelm Leibniz. “The Enlightenment was […] not about consensus, it was not about systematic unity, and it was not about the deployment of instrumental reason: what was developed in the Enlightenment was a modern idea of truth defined by error, a modern idea of knowledge defined by failure, conflict, and risk, but also hope.”2 — David Bates.
    [Show full text]
  • Fixing Biases: Principles of Cognitive De-Biasing
    Fixing biases: Principles of cognitive de-biasing Pat Croskerry MD, PhD Clinical Reasoning in Medical Education National Science Learning Centre, University of York Workshop, November 26, 2019 Case q A 65 year old female presents to the ED with a complaint of shoulder sprain. She said she was gardening this morning and injured her shoulder pushing her lawn mower. q At triage she has normal vital signs and in no distress. The triage nurse notes her complaint and triages her to the fast track area. q She is seen by an emergency physician who notes her complaint and examines her shoulder. He orders an X-ray. q The shoulder X ray shows narrowing of the joint and signs of osteoarthrtritis q He discharges her with a sling and Rx for Arthrotec q She is brought to the ED 4 hours later following an episode of syncope, sweating, and weakness. She is diagnosed with an inferior MI. Biases q A 65 year old female presents to the ED with a complaint of ‘shoulder sprain’. She said she was gardening this morning and sprained her shoulder pushing her lawn mower (Framing). q At triage she has normal vital signs and in no distress. The triage nurse notes her complaint and triages her to the fast track area (Triage cueing). q She is seen by an emergency physician who notes her complaint and examines her shoulder. He orders an X-ray (Ascertainment bias). q The shoulder X ray shows narrowing of the joint and signs of osteoarthrtritis. He explains to the patient the cause of her pain (Confirmation bias).
    [Show full text]
  • Croskerry MD, Phd, FRCP(Edin)
    Clinical Decision Making + Strategies for Cognitive Debiasing Pat Croskerry MD, PhD, FRCP(Edin) International Association of Endodontists Scottsdale, Arizona June 2019 Financial Disclosures or other Conflicts of Interest None It is estimated that an American adult makes 35,000 decisions a day i.e. about 2200 each waking hour Sollisch J: The cure for decision fatigue. Wall Street Journal, 2016 Decision making ‘The most important decision we need to make in Life is how we are going to make decisions’ Professor Gigerenzer Is there a problem with the way we think and make decisions? 3 domains of decision making Patients Healthcare leadership Healthcare providers Patients Leading Medical Causes of Death in the US and their Preventability in 2000 Cause Total Preventability (%) Heart disease 710,760 46 Malignant neoplasms 553,091 66 Cerebrovascular 167,661 43 Chronic respiratory 122,009 76 Accidents 97,900 44 Diabetes mellitus 69,301 33 Acute respiratory 65,313 23 Suicide 29,350 100 Chronic Liver disease 26,552 75 Hypertension/renal 12,228 68 Assault (homicide) 16,765 100 All other 391,904 14 Keeney (2008) Healthcare leadership Campbell et al, 2017 Healthcare providers US deaths in 2013 • 611,105 Heart disease • 584,881 Cancer • 251,454 Medical error Medical error is the 3rd leading cause of death Estimated number of preventable hospital deaths due to diagnostic failure annually in the US 40,000 – 80,000 Leape, Berwick and Bates JAMA 2002 Diagnostic failure is the biggest problem in patient safety Newman-Toker, 2017 Sources of Diagnostic Failure The System 25% The Individual 75% Graber M, Gordon R, Franklin N.
    [Show full text]