METHODS IN MEDICINE METHODS IN MEDICINE A Descriptive Study of Physicians' Behaviour J. RIDDERIKHOFF Department of Family Medicine, Erasmus University, Rotterdam, the Netherlands KLUWER ACADEMIC PUBLISHERS DORDRECHTI BOSTON I LONDON Library of Congress Cataloging in Publication Data Rldderlkhoff, J., 1932- Methods In medIcIne: a descrlpt1ve study of physlc1ans' behavIour I J. Ridderlkhoff, p, cm. Bibliography: p. Includes index, ISBN-13: 978-94-010-6984-7 e-ISBN-13: 978-94-009-1097-3 DOl: 10.1007/978-94-009-1097-3 1. Physlciar.s--"'sychology, 2. Problem-solvIng. 3. DecisIon -.aking. 4. Medicine--Practice--Psychologlcal aspects. I. TItle. [DNLM: 1. Behavl0r. 2. Physlcians--psyChology. 3. Professional Practice. W 87 R543ml R690.R53 1989 610.69'52'OI9--dc19 DNLM/DLC for Library of Congress 88-12961 CIP Published by Kluwer Academic Publishers, P.O. Box 17, 3300 AA Dordrecht, The Netherlands Kluwer Academic Publishers incorporates the publishing programmes of D. Reidel, Martinus Nijhoff, Dr W. Junk and MTP Press. Sold and distributed in the U.S.A. and Canada by Kluwer Academic Publishers, 101 Philip Drive, Norwell, MA 02061, U.S.A. In all other countries, sold and distributed by Kluwer Academic Publishers Group, P.0. Box 322, 3300 AH Dordrecht, The Netherlands. All Rights Reserved © 1989 by Kluwer Academic Publishers Softcover reprint of the hardcover I st edition 1989 No part of the material protected by this copyright notice may be reproduced or utilized in any form or by any means, electronic or mechanical, including photocopying, recording or by any information storage and retrieval system, without written permission from the copyright owner. Dedicated to my wife, Diny, my two sons, Hans and Paul, and particularly to my parents, without whose creation this book would never have been born. Contents Preface XI Introduction XIII Methods in medicine XIV Planning and investigation XVI Medical science XX Organization of this book XXI Chapter I: Medical knowledge 1 Introduction 1 A historical view 2 The concept of disease 4 Taxonomy of diseases 7 Medical data 11 Reliability of data 13 Data recording 15 Diagnosis 17 Prediction 19 Treatment 22 Summary 24 Chapter II: Ways of reasoning 27 Introduction 27 Science and belief 28 Methods in science 31 Deductive reasoning 33 Hypothetico-deductive reasoning 36 Inductive reasoning 38 Probabilistic reasoning 42 Conditions for the valid application of the probability theory 44 Logical implications of probability theory 47 Summary 50 Chapter III: Human thinking and problem-solving and their relationship to medicine 52 Introduction 52 Human cognition and the development of thought 53 Problem-solving 54 Thought processes 56 VIII CONTENTS The process of memory storage and recall 58 The process of forgetting 60 Problem-solving in complex tasks 63 I. Restructure of the problem 63 II. Limiting the information 63 Studies of physicians' thought processes 63 The validity of inference 66 What is a problem? 67 What is a medical problem? 69 Problem space and task environment 72 Problem-solving processes 74 The approach to problem-solving 75 Studies of physicians' problem-solving 77 The validity of clinical judgements 79 Summary 81 Chapter IV: Clinical decision-making 83 Introduction 83 Historical background 84 Towards a decision theory 85 Some thoughts on utility theory 85 Decision theory 89 Bayes' theorem 91 Constraints on the rational decision maker 96 Meeting the rational decision maker 100 Artificial intelligence 106 Alternative decision-making models 107 Satisficing theory 109 Incrementalism 109 Garbage can model 110 Conflict model 111 Summary 116 Chapter V: Models and methods 118 Introduction 118 Models in medical reasoning 119 Styles of problem-solving and decision-making 122 The models 124 Deductive model 124 Inductive model 125 Hypotheses 127 Subtypes of inductive reasoning 129 Landmarks in the models 131 Instruments 135 The paper patient 135 Case histories for the investigation 140 CONTENTS IX Hypotheses and symptoms 141 Hypotheses levels 143 Observational methods 144 Simulation procedure 148 Summary 151 Chapter VI: Results of the study 153 Introduction 153 The criteria for the strategies 154 Deductive reasoning 154 Inductive reasoning 155 Pattern recognition 155 Inductive-heuristic reasoning 156 Inductive-algorithmic reasoning 156 Relevancy and redundancy in information processing 160 On hypotheses 162 On symptoms and diagnoses 166 Symptom clustering 166 Decomposition procedure 168 On questions, answers, and time 171 Ratio effectiveness 178 Content effectiveness 178 Retrieval rate 178 On probabilities and (un)certainties 179 On patient management and therapy 184 On diagnostic gradation 186 On experience and consequences 191 Summary 196 Chapter VII: Reflections, conclusions, consequences 200 Introduction 200 The inductive method in medicine 202 On the circular process 204 On hypotheses 205 On diagnoses 208 On prognosis and treatment 210 On learning and experience 212 On medical research 215 Closing remarks 217 References 219 Author index 227 Preface Clinicians spend their working lives making decisions. such decisions are usually made in interlocking streams rather than in the discrete circumscribed contexts so beloved of scientists. When the clinician encounters a patient a complex interactive process is initiated in which the clinician searches his memory to match the symptoms and signs indicated by the patient with the complex disease models which he carries in his head. He then makes choices about further questions or tests in order to clarify his understanding of the patient's problem and to formulate a management or treatment plan. In recent years there has been increasing interest in how clinicians make such decisions and a realization that decision-making in clinical medicine is virtually the same as that in many other professional contexts. The scientific study and formal teaching of clinical decision-making is a relatively young discipline. Less than 20 books have so far appeared which take explicit account of the theoretical and experimental decision-making literature in medicine and other related disciplines. This book is a distinctive and important contribution to this growing field. It combines a comprehensive critical analysis of a wide range of relevant philo sophical, statistical, psychological and medical literature with an interesting set of experimental observations of primary care physicians. Dr. Ridderikhoff shows great erudition and wide command of a large reference literature. Dr. Ridderikhoff takes a firmly descriptive rather than prescriptive viewpoint on understanding clinical decision-making. The key issues addressed in the book are 'How scientific is clinical medicine as a discipline?' and 'Do primary care physicians use inductive or deductive reasoning in their clinical practice?' In recent years following the seminal work of elstein et al. (1978) the emphasis in clinical decision-making studies has been on the hypothetico-deductive processes as the principal method of clinical reasoning. By contrast Dr. Ridderikhoff's experimental findings lend support to his assertion that induction is the method of choice for primary care physicians. This book has as it main strength (1) its form and realistic grasp of the philosophical foundations of clinical practice and (2) its perspective of the working clinician both in its critical analysis of the literature and in the design of the studies. Dr. Ridderikhoff has been influenced in his perspective by his many years in private practice before entering academic. The book consists essentially of two parts. In the first part after a broad introduction touching on the philosophical and scientific complexities of the clinical method he focuses not only on utility and probability theories but devotes considerable space to a critical analysis of the literature on human thinking and problem-solving and on the nature of medical knowledge. He XII PREFACE tackles these complex areas with a clinician's eye but also with a firm founda tion in philosophical analysis. The description of the Bayesian viewpoint is highly critical with very effective use of the original texts especially those of Von Neumann and Morgenstern. In the second part of the book Dr. Ridderikhoff sets out to describe clinical decision-making in primary care settings using simulated cases. He had one major aim in these studies, namely, to try to establish by careful observation of working clinicians in clinically realistic settings whether the predominant method of clinical reasoning was, inducative or deductive. This was an ambi tious agenda and the design and methodology are very explicitly described in the book. He has used a carefully selected set of cases which reflect the case-mix of primary care physicians. The form of simulated methodology which he makes use of reflects the wide range of possible questions and tests available to the primary care physician. The studies were conducted in the physician's own office and incorporated the time pressures found in real practice settings. The verbal mode of interaction is used t9 reflect clinical reality. In order to be able to decide whether induction or deduction is the method of choice he sets up cognitive 'landmarks' in the experiments so as to be able to detect the key differences reflected6y the two methods under study. The results of these studies strongly favour induction (in one or other of its strategic forms) as the primary inferential approach used by primary care physicians. Another important finding is that clinical experience seems to have no influence on any of the measures of performance which he has developed for this stu5fy! This is a fascinating and provocative book which challenges our assumptions about clinical reasoning. It will be of particular value to working clinicians, teachers of clinical medicine as well as cognitive and educational psychologists with an interest in medical education. Thomas R. Taylor University of Washington, 1988 Introduction But in fact, we know nothing from having seen it; for the truth is hidden in the deep Democritus Do patients benefit from their doctor's advice? Do physicians know what they are doing? How do we know what the best advice is? Are the methods doctors use in medicine the most appropriate, valid, reliable, and efficient ones? Do our methods reach the standards of present -day conceptions of science? Or do we have to admit that medicine, as is often claimed, is an art, not a science? Must we allude to Kuhn's opinion that medicine is a craft, like calendering and metallurgy, or do we have to share Braithwaite's opinion that medicine is a science because of its 'natural domain'? If we concede the former viewpoint then we are unable to answer the first questions. Art, in contrast with science, is based on experience, which is subjective, incapable of precise analysis, irreproducible, and impossible to measure. So we are facing the old problem of demarcation between the empirical sciences (theories, hypotheses, explanatious) and metaphysical con cepts (art, prescience, pseudoscience). But when medicine is not accepted as a science how can physicians possibly trace what they are doing? How can they trace the quality of their problem solving, their decision-making? It will cast us in a dilemma as many people prefer the more personal approach of an art-like practice to the (assumed) more impersonal scientific approach. We take the standpoint that in the interest of patients, physicians, as well as health care provisions, we have to make an effort to clarify the question of science versus art. As is often the case with these big questions, a conclusive answer in the sense of a binary statement cannot be given. We can try to highlight a number of elements and features of both cultural domains. We can try to specify a number of characteristics of science as opposed to art. However, a clear definition, a world-wide consensus about 'science' cannot be found. So we have to take a personal standpoint in order to deal with the 'big question'. According to my philosophical guide, Popper (1959, 1983), the aim of science is to find satisfactory explanations of whatever strikes us as being in need of explanation. A satisfactory explanation can be found in a theory which allows for genuine prediction. Whenever a prediction falls short of explanation from a theory, the theory has to be refuted. A theory or a conjecture belongs to the empirical sciences if and only if it is falsifiable. In other words, the theory must be able to specify under what conditions the theory can be falsified.