Using Illness Scripts to Teach Clinical Reasoning Skills to ... - STFM

Vol. 42, No. 4255Medical Student EducationUsing Illness Scripts to Teach Clinical ReasoningSkills to Medical StudentsAnna Lee, PhD; Gavin M. Joynt, MB, BCh; Alex K.T. Lee, MPhil;Anthony M.H. Ho, MD; Michele Groves, PhD; Alexander C. Vlantis, MB, BCh;Ronald C.W. Ma, MB, BChir; Colman S.C. Fung, MB,BS; Cindy S.T. Aun, MDBackground and Objectives: Most medical students learn clinical reasoning skills informally duringclinical rotations that have varying quality of supervision. We conducted a randomized controlledtrial to determine if a workshop that uses “illness scripts” could improve students’ clinical reasoningskills when making diagnoses of patients portrayed in written scenarios. Methods: In 2007–2008,53 fourth-year medical students were randomly assigned to either a family medicine (intervention)or psychiatry (control) clerkship at The Chinese University of Hong Kong. Students in the interventiongroup participated in a 3-hour workshop on clinical reasoning that used illness scripts. Theworkshop was conducted with small-group teaching using a Web-based set of clinical reasoningproblems, individualized feedback, and demonstration of tutors’ reasoning aloud. The effectivenessof the intervention was assessed using the Diagnostic Thinking Inventory (DTI) and the measurementof individual students’ performance in solving clinical reasoning problems (CRP). Results: The postinterventionoverall DTI scores between groups were similar (mean difference 0, 95% confidenceinterval [CI]= -7.4 to 7.4). However, the total scores on the CRP assessment were 14% (95% CI=8%to 21%) higher in the intervention group than in controls. Conclusion: A workshop on illness scriptsmay have some benefit for improving diagnostic performance in clinical reasoning problems.(Fam Med 2010;42(4):255-61.)The ability to diagnose effectively and accuratelyrequires an appropriate clinical knowledge and soundclinical reasoning skills. Medical educators agree thatclinical reasoning is a central component of physiciancompetence. 1 Currently, most students learn clinicalreasoning skills informally in clinical rotations withvarying quality of supervision.When making a diagnosis, students often reason byan iterative process of hypothesis generation and testingof one symptom at a time. 2 Since they may be unable tosynthesize clinical features into meaningful clusters orsyndromes, their clinical reasoning styles are typicallycharacterized by poorly organized knowledge structure.In comparison, expert clinicians reason by organizingFrom the Department of Anaesthesia and Intensive Care (Dr Lee, Dr Joynt,Mr Lee, Dr Ho, Dr Aun), Department of Otorhinolaryngology, Head andNeck Surgery (Dr Vlantis), Department of Medicine and Therapeutics (DrMa), School of Public Health and Primary Care (Dr Fung), The ChineseUniversity of Hong Kong, Prince of Wales Hospital, Shatin, NT, Hong Kong,China; and Faculty of Health Sciences, University of Queensland, Brisbane,Queensland, Australia (Dr Groves).and prioritizing syndrome recognition through comparingand contrasting key clinical features in makinga diagnosis (“illness scripts”). 2 The illness scriptconcept 3 provides a theoretical framework to explainhow medical diagnostic knowledge can be organizedfor diagnostic problem solving.Illness scripts are a product of extensive experiencewith patients superimposed on a formal knowledgestructure. 4 Expert clinicians use illness scripts mostof the time in their clinical reasoning since it involvesa knowledge-driven model of pattern recognition thatmay be more efficient than hypothetico-deductivereasoning used by students. 5The Chinese University of Hong Kong undergraduatemedical curriculum is a body-system-based modeltaught over 5 years. In response to faculty members’perception that our students had poor clinical reasoningskills when starting clinical rotations during their fifthand final year of medical school, we constructed a brief3-hour educational intervention to improve those skills.Our hypothesis was that an educational interventionwould promote and improve students’ competency in

256 April 2010 Family Medicineclinical reasoning in a systematic manner. The objectiveof our study was to determine if a brief workshopto teach students to refine their knowledge organizationthrough the use of illness scripts could improvestudents’ clinical reasoning skills for diagnosis.MethodsThe workshop started with a 20-minute lecture to allstudents. It introduced the key elements of the clinicaldiagnostic process. 6 This was followed by a 1-1/4 hoursmall-group tutorial on problem representation and developingan illness script and a final tutorial, also 1-1/4hours long, in the computing laboratory on developingand selecting an appropriate illness script. A manualfor teachers wishing to conduct a similar workshop isavailable from the authors by request.Problem Representation and Developingan Illness ScriptThe objective in the workshop was for students toacquire the skill of articulating the patient’s problemand to develop an illness script based on two clinicalscenario articles from the New England Journalof Medicine’s Clinical Problem-solving series. 7,8 Thepurpose was to help students develop a succinct andcoherent case presentation from the clinical problempresented in the journal article and to organize pertinentclinical data into a structured template (illness script) toformulate the basis of reaching a diagnosis. We chosethe articles based on the complexity of the scenario andthe perceived interest to students.A PowerPoint presentation of the clinical scenarioswas prepared by one faculty member before the workshop.Three other physicians from the Faculty of Medicineacted as tutors for the interactive sessions.The students were given the details of the clinicalpresentation, history, and examination results duringthe workshop. The tutor’s task was to guide the studentsto identify the important findings and to help themdevelop a presentation of the patients’ problems usingthe teaching strategies described by Bowen. 6 Studentswere expected to characterize the clinical problempresented in the two articles as “an elderly man with apersistent cough” 7 and “a middle-aged woman with anacute swollen and painful left leg.” 8The tutor then reasoned aloud to compare and contrasthis expert problem representation with those ofthe students. Students then developed an illness scriptby listing the enabling (predisposing conditions or riskfactors), fault (pathophysiological insult), and clinicalconsequences (key signs, symptoms, complaints) withfeedback from the tutor. A sample problem illnessscript for pertussis in “an elderly man with a persistentcough” 7 is shown in Appendix 1.Developing and Selecting an AppropriateIllness ScriptThe objective of this session was for students toacquire the skill of developing and selecting an appropriateillness script in formulating the most probablediagnosis for a given clinical problem. The purposewas to help students prioritize multiple diagnoses byidentifying discriminating features for each diagnosticconsideration. We developed a Web-based program thatwas comprised of a set of 20 clinical reasoning problems(CRPs) for teaching, combined with a scoring systemfor assessment. Our Web-based program was based onpreviously validated paper-based CRPs developed byGroves et al in conjunction with 22 family medicinephysicians at the University of Queensland. 9 Details ofthe scoring system for CRP are described elsewhere. 9The reliability of the paper-based CRP for medicalstudents ranged from 0.61 to 0.83. 9Prior to the implementation of the workshop, wetested the Web-based program on eight final-year studentswho were not involved in the study, to identifycomputing errors in the program and scoring system.The Web-based CRPs for post-workshop assessmenthad acceptable reliability (Cronbach’s alpha=0.80).For each CRP, students had to nominate the two mostlikely diagnoses, list the clinical features that supportedor opposed each diagnosis, and weigh each of the clinicalfeatures (slightly relevant, somewhat relevant, veryrelevant) from a series of drop-down menus. For thefirst CRP, two tutors interacted with students by askingthem to list all important findings from the case, createa problem representation based on those findings,generate and prioritize diagnostic considerations thatidentify discriminating features for each consideration,identify findings from the case to support the diagnosis,and lastly to identify and compare alternative diagnoses.6 Students were assisted if necessary at each stepto reach reasonable conclusions and, when appropriate,positive feedback for correct responses was provided.This was then supplemented by a tutor reasoning aloudto summarize the process and illustrate key componentssuch as the relevance of supporting clinical featuresassociated with the correct diagnoses.Students were encouraged to gradually work onfour other CRPs individually during the session, withindividual feedback from the tutors as they reasonedaloud. Thus, students were developing their own illnessscripts for each CRP attempted. At the end of eachcompleted CRP, the student was given an automatedscore by the program that gave immediate quantitativefeedback on their clinical reasoning skills. Toward theend of the session, the tutors reasoned aloud for theselected CRPs to illustrate the gap between actual anddesired expert performance. A sample CRP is shownin Appendix 2. The overall score for CRPs from theresponses shown in the tables was 28/37, as visceral

260 April 2010 Family MedicineConclusionsThe key elements in the intervention involved theuse of integrated small-group teaching, individual feedbackto students when developing appropriate illnessscripts, highly motivated tutors who reasoned aloud tostudents to highlight the desired performance, and thedevelopment of a Web-based set of clinical reasoningproblems for teaching and assessments. Our initialresults are encouraging. They can serve as a rationalefor implementation of a one-morning clinical reasoningmodule to improving clinical diagnostic skills ofmedical students.Acknowledgments: The work described in this paper is supported by a grantfrom the University Grants Committee Teaching Development TrienniumGrant of the Hong Kong Special Administrative Region, China (ProjectNo. 4170255).This work was presented at The Chinese University of Hong Kong,Faculty of Medicine Curriculum Day meeting in September 2008 in HongKong.Corresponding Author: Address correspondence to Dr Lee, The ChineseUniversity of Hong Kong, Prince of Wales Hospital, Department of Anaesthesiaand Intensive Care, Shatin, NT, Hong Kong, China. +852-2632-2735.Fax: +852-2637-2422. annalee@cuhk.edu.hk.Re f e r e n c e s1. Norman G. Research in clinical reasoning: past history and currenttrends. Med Educ 2005;39(4):418-27.2. Schmidt HG, Norman GR, Boshuizen HP. A cognitive perspectiveon medical expertise: theory and implication. Acad Med1990;65(10):611-21.3. Barrows HS, Feltovich PJ. The clinical reasoning process. Med Educ1987;21(2):86-91.4. Norman G. Building on experience—the development of clinical reasoning.N Engl J Med 2006;355(21):2251-2.5. Charlin B, Tardif J, Boshuizen HP. Scripts and medical diagnosticknowledge: theory and applications for clinical reasoning instructionand research. Acad Med 2000;75(2):182-90.6. Bowen JL. Educational strategies to promote clinical diagnostic reasoning.N Engl J Med 2006;355(21):2217-25.7. Cornia PB, Lipsky BA, Saint S, Gonzales R. Clinical problem-solving.Nothing to cough at—a 73-year-old man presented to the emergencydepartment with a 4-day history of nonproductive cough that worsenedat night. N Engl J Med 2007;357(14):1432-7.8. Fazel R, Froehlich JB, Williams DM, Saint S, Nallamothu BK. Clinicalproblem-solving. A sinister development—a 35-year-old woman presentedto the emergency department with a 2-day history of progressiveswelling and pain in her left leg, without antecedent trauma. N Engl JMed 2007;357(1):53-9.9. Groves M, Scott I, Alexander H. Assessing clinical reasoning: amethod to monitor its development in a PBL curriculum. Med Teach2002;24(5):507-15.10. Bordage G, Grant J, Marsden P. Quantitative assessment of diagnosticability. Med Educ 1990;24(5):413-25.11. Beullens J, Struyf E, Van DB. Diagnostic ability in relation to clinicalseminars and extended matching questions examinations. Med Educ2006;40(12):1173-9.Appendix 1Sample Problem Illness Script for Pertussis in “An Elderly Man With a Persistent Cough” 7 ** The illness script forms the logical construct underlying the symptoms and signs making up the recognizable patterns for making a diagnosis.

Medical Student EducationVol. 42, No. 4261Appendix 2Sample Clinical Reasoning Problem From the Web-based Program