Archives of Disease in Childhood 1995; 73: 357-363 CURRENT TOPIC Adult learning theory, problem based learning, and paediatrics T J David, Leena Patel 'A medical student faints at the sight of blood. Would you expect his pulse to be fast or slow? What underlying mechanism caused his fainting?' This is the kind of problem that first year medical students in a problem based learning curriculum are asked to investi- gate, while their counterparts in traditional medical curricula are trying to memorise the names of all the holes in the skull. Department of Child Health, University of Manchester T J David L Patel Correspondence to: Professor T J David, University Department of Child Health, Booth Hall Children's Hospital, Charlestown Road, Blackley, Manchester M9 7AA. In conventional medical education, the under- graduate student is given specific learning objectives in a particular discipline (for example biochemistry). Teaching (lectures, tutorials, practicals) is provided, and the students cram as much as possible in order to pass the exami- nation. If successful, the cerebral hard disc is wiped clean, and the student starts afresh with a new discipline. Throughout, the process is in the control of the teacher. Readers of this journal will recognise the conventional model, and may well think that the excellent outcome in their own case means that an ancient, well tried, and tested system should not be tinkered with. A somewhat anarchic alternative, problem based learning, which may disturb the equilibrium of those who are happy with conventional medical education, is explained below. In problem based learning, the learner (undergraduate or postgraduate) decides on his or her own learning objectives as a result of exposure to trigger material (for example a written problem, a video, or a patient). The role of the teacher is to design trigger material that will interest the learner and cause the learner to generate appropriate learning objec- tives, and to provide the appropriate learning resources (for example library materials, com- puterised literature searching, video materials, skills laboratories, clinical exposure). Other than the design and provision of the trigger material, the whole educational process is in the control of the learner. Adults and children learn in different ways Paediatricians are champions of the concept that children are not just small adults. However, the converse is also true. Adults are not just large children. Any discussion about medical education, or an appreciation of how damaging is the conventional system, is diffi- cult without an understanding of how adults learn. Firstly, the word adult has different definitions. Biologically it is the age at which we can reproduce. Legally it is the age at which we can vote, obtain a driving licence, get married without consent, and so on. Socially, we become adult when we start performing adult roles such as full time worker, spouse, parent, and so on. In the context of this article, we are referring to the psychological definition of being responsible for our own lives, and of being self directing. The way that children are taught is the only way of thinking about education that most of us know, for it has dominated school educa- tion, and even adult education, until recently. There are, however, differences in the child leaming and adult leaming models, and these are contrasted below. The jargon word for child leaming is pedagogic, and for adult learn- ing andragogic; we have avoided both terms in this article. THE CONCEPT OF THE LEARNER In the child leaming model, the leamer is a dependent person. The teacher has full responsibility for making all decisions about what should be leamed, how and when it should be leamed, and whether it has been leamed. The leamer submissively carries out the teacher's directions. In the adult leaming model, developed and reviewed by Knowles,' the learner is sel directing. Adults have a need to be perceived as being capable of taking responsibility for themselves; where adults feel that others are imposing their wills, there may be resentment and resistance. However, adults need help in making the transition from being dependent learners to self directed learners, for even though adult learners may be totally self directing in every other aspect of their lives, the minute they walk into a situation labelled 'education' (or any of its synonyms), they tend to hark back to their conditioning at school, assume a role of dependency, and demand to be taught. THE ROLE OF THE LEARNER'S EXPERIENCE In the child learning model of learning, learners enter with little experience that is of much value as a resource in learning. It is the 357 on March 7, 2020 by guest. Protected by copyright. http://adc.bmj.com/ Arch Dis Child: first published as 10.1136/adc.73.4.357 on 1 October 1995. Downloaded from
Transcript
Archives of Disease in Childhood 1995; 73: 357-363
CURRENT TOPIC
Adult learning theory, problem based learning,and paediatrics
T J David, Leena Patel
'A medical student faints at the sight ofblood. Wouldyou expect his pulse to befast orslow? What underlying mechanism caused hisfainting?'This is the kind of problem that first yearmedical students in a problem basedlearning curriculum are asked to investi-gate, while their counterparts in traditionalmedical curricula are trying to memorisethe names of all the holes in the skull.
Department of ChildHealth, University ofManchesterT J DavidL Patel
Correspondence to:Professor T J David,University Department ofChild Health, Booth HallChildren's Hospital,Charlestown Road, Blackley,Manchester M9 7AA.
In conventional medical education, the under-graduate student is given specific learningobjectives in a particular discipline (for examplebiochemistry). Teaching (lectures, tutorials,practicals) is provided, and the students cramas much as possible in order to pass the exami-nation. If successful, the cerebral hard disc iswiped clean, and the student starts afresh witha new discipline. Throughout, the process is inthe control of the teacher.
Readers of this journal will recognise theconventional model, and may well think thatthe excellent outcome in their own case meansthat an ancient, well tried, and tested systemshould not be tinkered with. A somewhatanarchic alternative, problem based learning,which may disturb the equilibrium of thosewho are happy with conventional medicaleducation, is explained below.
In problem based learning, the learner(undergraduate or postgraduate) decides onhis or her own learning objectives as a result ofexposure to trigger material (for example awritten problem, a video, or a patient). Therole of the teacher is to design trigger materialthat will interest the learner and cause thelearner to generate appropriate learning objec-tives, and to provide the appropriate learningresources (for example library materials, com-puterised literature searching, video materials,skills laboratories, clinical exposure). Otherthan the design and provision of the triggermaterial, the whole educational process is inthe control of the learner.
Adults and children learn in differentwaysPaediatricians are champions of the conceptthat children are not just small adults.However, the converse is also true. Adults arenot just large children. Any discussion aboutmedical education, or an appreciation of how
damaging is the conventional system, is diffi-cult without an understanding of how adultslearn. Firstly, the word adult has differentdefinitions. Biologically it is the age at whichwe can reproduce. Legally it is the age at whichwe can vote, obtain a driving licence, getmarried without consent, and so on. Socially,we become adult when we start performingadult roles such as full time worker, spouse,parent, and so on. In the context of this article,we are referring to the psychological definition ofbeing responsible for our own lives, and ofbeing self directing.The way that children are taught is the only
way of thinking about education that most ofus know, for it has dominated school educa-tion, and even adult education, until recently.There are, however, differences in the childleaming and adult leaming models, and theseare contrasted below. The jargon word forchild leaming is pedagogic, and for adult learn-ing andragogic; we have avoided both terms inthis article.
THE CONCEPT OF THE LEARNERIn the child leaming model, the leamer is adependent person. The teacher has fullresponsibility for making all decisions aboutwhat should be leamed, how and when itshould be leamed, and whether it has beenleamed. The leamer submissively carries outthe teacher's directions. In the adult leamingmodel, developed and reviewed by Knowles,'the learner is sel directing. Adults have a needto be perceived as being capable of takingresponsibility for themselves; where adults feelthat others are imposing their wills, there maybe resentment and resistance. However, adultsneed help in making the transition from beingdependent learners to self directed learners, foreven though adult learners may be totally selfdirecting in every other aspect of their lives, theminute they walk into a situation labelled'education' (or any of its synonyms), they tendto hark back to their conditioning at school,assume a role of dependency, and demand tobe taught.
THE ROLE OF THE LEARNER'S EXPERIENCEIn the child learning model of learning,learners enter with little experience that is ofmuch value as a resource in learning. It is the
357
on March 7, 2020 by guest. P
rotected by copyright.http://adc.bm
j.com/
Arch D
is Child: first published as 10.1136/adc.73.4.357 on 1 O
experience of the teacher, the textbook writer,and the audiovisual aids' producer that counts.The backbone is transmission technique -lectures, assigned reading, and audiovisualpresentations. By contrast, adults enter into aneducational activity with greater volume anddifferent quality of experience from youth;adults are themselves a rich resource for oneanother. Another consequence of an adult'sexperience is that it becomes increasingly thesource of an adult's self identity. If asked 'whoare you?', your 8 year old son might give hisname, address, name of school, and theoccupation of his father; his self identity isderived almost entirely from external sources.However, adults derive their self identity fromtheir experience. So, if in an educational situa-tion an adult's experience is not valued,ignored, or not made use of, it is not just theexperience that is being rejected; it is theperson. Hence the importance of using theexperience of adults as a resource for learning.There is also a negative consequence of adult-hood. Because of their experience, adults oftenhave developed habitual ways of thinking andacting, preconceptions about reality, pre-judices, and defensiveness about former waysofthinking and doing. Adult learners need helpto become more open minded.
READINESS TO LEARNIn the child learning model, learners maturefrom grade to grade, becoming ready to learnwhat they need in order to advance to the nextgrade level. Readiness is largely a function ofage. In contrast, adults become ready to learnwhen they experience a need to know or dosomething. Adults do not need to wait for readi-ness to develop naturally. Things can be done toinduce it, such as exposing learners to moreeffective role models, and to diagnostic experi-ences in which they can assess gaps betweenwhere they are and where they want to be.
ORIENTATION TO LEARNINGIn the child learning model, learners are subjectorientated. They see learning as a process ofacquiring prescribed matter content. The cur-riculum is organised according to subjects.Because adults are motivated after they experi-ence a need in their life, they enter with a lifecentred, task centred, or problem centred orienta-tion to learning. For adult learners, learningexperiences need to be orientated to life ratherthan to subject matter.
MOTIVATION TO LEARNIn the child learning model, learners aremotivated primarily by external pressures fromparents and teachers, competition for grades,and the consequences of failure. Althoughadults also respond to these external motivatorsand others such as a better job and salaryincrease, the adult model predicts that morepotent motivators will be internal: self esteem,recognition, better quality of life, greater selfconfidence.
CHOOSING THE MODELAlthough for centuries we only had one model,as will be evident from the above, we now havetwo sets of assumptions about learners.Despite the phrases child learning and adultlearning, these models are not entirely agedependent. In some situations, such as whenlearners of any age are entering a totallystrange territory (for example, confronting acomplex machine they have never seen before),they may be entirely dependent on didacticinstruction before they can take much initia-tive. In many more instances, however, partic-ularly with adults, the adult learning modelwould be more applicable, particularly if thelearners have had some orientation to selfdirected learning. The adult learning modelmay apply to children more than we think. Forexample, children are very self directed in theirlearning outside school and could also be moreself directed in school. Children bring someexperience with them, and this experiencecould be used as a resource for some kinds oflearning. Children are also more ready to learnwhen they experience a need to know thanwhen they are told they have to learn.
Briefhistory ofproblem based learningin medicineIn 1965, the McMaster University in Ontarioset up an innovative medical school whichemployed problem based learning rather thanconventional medical education. Others havefollowed, most notably the University ofLimburg at Maastricht, where a medicalschool modelled on McMaster was establishedin 1974. Much of the further research anddevelopment of problem based learning inmedicine has taken place in Maastricht, withthe result that McMaster is now incorporatinga number ofinnovations from the Netherlands.The 'disease' has spread to other parts ofNorth America, Australia (most notably atNewcastle, New South Wales), and the ThirdWorld. Europe has been slightly resistant toinfection, some of the notable exceptions beingLinkoping in Sweden and Manchester. Thelatter is the first UK medical school to havetotally converted the preclinical course into afully problem based course, although thereare signs that others are likely to implementproblem based learning.
What happens in problem basedlearning?The process follows identifiable steps. Thestudents are provided with so-called stimulusor trigger material. An example might be thefainting medical student episode, or it might bea more detailed problem, a video or a patientwith a problem. The nature of the stimulusmaterial is critical; the trick is to producematerial that is stimulating, contextual, andwill lead the students to formulate most or allof the learning goals set by the course manage-ment group - this is discussed later. Studentswork cooperatively in a small group (eight to10). They identify the problems, they use prior
358
on March 7, 2020 by guest. P
rotected by copyright.http://adc.bm
j.com/
Arch D
is Child: first published as 10.1136/adc.73.4.357 on 1 O
Adult learning theory, problem based learning, and paediatrics
The 'seven jump': the tutorial process
1. Clarify terms and concepts in the problem which areunknown to you.
2. Define the problem, that is list the phenomena to beexplained.
3. Explain the problem; try to produce as many differentexplanations for the phenomena as you can think of Useprior knowledge and common sense.
4. Arrange the explanations proposed; try to produce acoherent description of the processes that you thinkunderlie the phenomena.
5. Formulate learning goals.6. Attempt to fill the gaps in your knowledge through
individual study.7. Share your findings with your group and try to integrate
the knowledge acquired into a comprehensive explanationfor the phenomena. Check whether you know enough.
knowledge to try and explain them by brain-storming, and then they formulate learninggoals. The students then attempt to fill gaps inknowledge through individual study andinquiry. The students share their findings andtry to integrate the knowledge acquired into a
comprehensive explanation for the phenomenaor problem. In Maastricht, the process hasbeen divided into seven discrete steps, knownas the 'seven jump' (the term comes from a
Dutch children's song); this is shown in thetable.The process is best illustrated with an
example taken from our own experiments withproblem based learning in paediatrics. Two'firms' of four students are combined to form a
group of eight; they work together throughoutthe eight week fourth year paediatric clerkship.The group meets twice a week, on a Mondayand a Thursday. At each meeting the group
elects a chairperson and a secretary; the latteris a scribe who records items on a flip chart.An example of a problem would be as
follows:
Rapidly fatal illnessJohn Smythe, aged 4 years, was noted tobe unwell and feverish one morning, andhis mother called the GP who arrived at1 1 10. He found that the axillary tempera-ture was 38-6°C and when he examinedJohn's chest he noted some purpuric spotson both arms. He telephoned the hospital,and agreed admission with the paediatricsenior house officer. An ambulance was
called, and John arrived in the accidentand emergency department at 1155. Bythe time he had arrived his condition haddeteriorated, and the purpuric rash was
noted to have extended on to the trunkand legs. Benzylpenicillin 1 million unitswas given intravenously at 1205, butdespite supportive treatment he thenrapidly deteriorated and had a cardio-respiratory arrest from which he could notbe revived at 1240.
Explain the skin lesions and rapiddeath in terms of the underlying process.
Having ensured that there were no termsthat the group could not understand (step 1),they would then proceed to step 2 and brieflylist the phenomena to be explained. A typicallist would include fever, purpuric rash, rapiddeterioration, and death. (This is an exception-ally short list; many problems would generate a
larger number of phenomena.) Step 3 is anattempt to produce as many explanations aspossible for each of the phenomena. There isno need to weed out unlikely or doubtfulexplanations at this stage. In step 4, there is anattempt to produce a coherent description ofthe processes. This is relatively difficult for thisparticular problem; our students have a poorunderstanding of dermatological terms suchas purpura, never mind the pathogenesis ofpurpura. Interestingly, not all groups appreci-ate that the child probably had meningococcalsepticaemia. Steps 3 and 4 are the most diffi-cult, and may take 20 to 30 minutes. Thenfollows step 5, in which the group makes a listof what it feels should be the learning goals -
topics to be explored before the next tutorial.Step 6 comprises whatever happens betweenthis and the next tutorial. It will include a vari-able mixture of textbook reading, literaturesearching, and the use of other learningresources, for example, video tapes, going totalk to a dermatologist or pathologist orhaematologist or whoever the individual stu-dent thinks can help. Each student addressesevery learning goal; the learning goals are notdivided between different students. At thesecond tutorial, the students share their find-ings, indicating both what they have found andtheir source. The information they find is oftenconflicting, and an important aim is tocritically appraise the evidence for or against aparticular statement: 'the textbook says ..., butwhat is the evidence?'
There are a number of practical issues. Therole of the tutor is difficult, especially fordoctors who are most accustomed to providinginformation rather than helping students tothink. Remaining silent when students saysomething wrong (which they frequently do) isespecially difficult, but absolutely essential. Ifstudents appear to be going completely off therails, for example during steps 4 or 7, then thetutor needs to try to provoke some lateralthinking. To do this without being directivecan be hard. One of a number of approaches isto ask the proponent of a particular incorrectidea how he or she can reconcile this with other(conflicting) facts which have been presented.It is desirable that the role of chairperson andscribe should rotate among the group. Thechairperson has the important role of control-ling over eager or domineering contributorsand encouraging the more reluctant membersof the group; it is important that all memberscontribute - the only spectator should be thetutor.
Problem designThe challenge is to design a problem that isboth interesting and will lead most students tothe intended learning goals. The design ofwritten problems is complex; the learninggoals have to be appropriate for the level ofknowledge of the students, and must fit inwith the general aims and objectives of thecourse. A problem may need to be amendedseveral times before it is satisfactory. In theabove example the intended learning goals
359
on March 7, 2020 by guest. P
rotected by copyright.http://adc.bm
j.com/
Arch D
is Child: first published as 10.1136/adc.73.4.357 on 1 O
were (i) the pathogenesis of meningococcalsepticaemia, (ii) the mechanisms of shock,and (iii) toxins and inflammatory mediators.Pieces of information may inadvertently leadstudents off at a tangent. For example, pro-viding the exact temperature in the problemdescribed above often caused students todiscuss how high this was, and whether ornot the height of the fever was a contributoryfactor. (In the less focused McMasterapproach, such diversions would be regardedpositively.) Giving the dose of penicillincreated a diversion for all groups of students,who embarked on a discussion as to what wasor was not the correct dose of penicillin, andwhether over or under dosage could have con-tributed to the outcome. The amended prob-lem stated that the dose was appropriate forthe weight and condition of the child. It is ofcourse neither possible nor appropriate toeliminate all possible causes for diversion.Some students offer a possible explanation forthe further deterioration as an anaphylacticreaction to penicillin, and clearly this is anappropriate topic to consider.
Practical difficultiesIn our own pilot studies, attendance attutorials has been voluntary. However, at thefirst tutorial, the process is briefly outlinedto the students, and the need for regularattendance is self evident. The number of poorattenders has been negligible. A greaterchallenge are the silent minority; some are shy,fearful of 'getting it wrong' (common), or idle.One role for the tutor is to ensure that thechairperson is trying in a kindly way to involvethese members of the-gro-up. By far the greatestchallenge, however, is that our students arequite unaccustomed to controlling their leam-ing in this way. They regularly look to the tutorfor non-verbal clues as to whether or not theyare on the right track; a deadpan response isimportant. Even though we explain the aimsand objectives of problem based leaming,students remain very concemed to knowwhether they 'got it right' or not. The greatestdifficulties are the students' low capacity forself direction; many find it hard to find thingsout for themselves, and despite being in thefourth year of an undergraduate medicaldegree they often have poor ability to use thefull resources of the library. Experience else-where is that students may take many monthsto become accustomed to self directedproblem based learning. Many of the problemsour students have experienced will disappearin two years' time, when students reachingthe fourth year will have encounteredManchester's fully problem based first twoyears.
How does problem based learning work?Problem based learning is an approach that isbased on research into how adults learn mosteffectively. The comments below are extractedfrom more detailed discussions of the basis ofproblem based learning.24
ACTIVATION OF PRIOR KNOWLEDGEActivation of prior knowledge facilitates thesubsequent processing of new information.New knowledge has to 'adhere' to prior knowl-edge, and for integration to take place the priorknowledge has first to be retrieved. Studieshave shown that recall of new information ispoor when subjects fail to reactivate relevantknowledge. Small group discussion of a prob-lem, as described above, does ensure that priorknowledge is activated, and used to try toexplain the phenomena observed.
ELABORATION OF KNOWLEDGEElaboration of knowledge at the time of learn-ing, as happens in problem based learning,enhances subsequent retrieval. Memory ofsubject matter and ability to use knowledge isenhanced when students have the opportunityto elaborate on the knowledge at the time oflearning. Elaboration can take several formssuch as discussion, answering questions, orusing the knowledge to understand a problem.
CONTEXTUAL LEARNINGThe difficulty in remembering the name of acolleague when encountered at the shops is anillustration of the influence of context. Anotherexample may be the difficulty that some peoplehave with passing examinations, where recallof information is in a different context to thatin which it was acquired. The context meansall the features of the environment at the timeof learning, including items which are quiteirrelevant. So, for example, when materialabout a particular patient is recalled, irrelevantfeatures (for example, the child was in the endcubicle on ward 4 where the plaster has comeoff the wall) will be recalled in addition toclinically important features. Learning a disci-pline (for example, anatomy, biochemistry, orpharmacology) outside of a clinical context isan effective way to ensure that little informa-tion is recalled beyond the date of the exami-nation. Problem based learning addresses thisissue, because all relevant concepts are learnedin the context of a human or clinical problem.
TRANSFER OF PRINCIPLES AND CONCEPTSIn conventional medical education, studentsin their preclinical years are taught basicprinciples or concepts of biomedical science,often with examples of their application, in thehope that these concepts will be available andused in the solution of patients' problems. Thisis referred to as transfer, the extent to which aconcept or principle learned in one context canbe transferred or applied to another problemthat requires the same principles for resolution.Unfortunately, research has shown that trans-fer is unlikely to occur unless the context orsurface similarity of two problems is matchedvery closely. Studies of problem solving haveshown that any change in the surface featuresof a problem impede transfer, so that a prob-lem solver does not recognise the similarity ofthe underlying concept.
360
on March 7, 2020 by guest. P
rotected by copyright.http://adc.bm
j.com/
Arch D
is Child: first published as 10.1136/adc.73.4.357 on 1 O
Adult learning theory, problem based learning, and paediatrics
Experiments contrasted a group (the 'mem-orise' group) who were asked to read andremember a problem and then had the solutionexplained, with another group (the 'problemsolvers') who were engaged in solving aproblem and who were then given the solutionalso. Then both groups were given sets of testproblems that illustrated the same problem invery different contexts from those in which ithad been presented originally. The resultsshowed that those who were asked to solvethe problem and received feedback aboutthe problem transferred the concept to the newproblem nearly 90% of the time; this compareswith about 60% for those who were simplyasked to memorise the problem. Other experi-ments showed that (a) simply engaging inproblem solving, without feedback, gave noadvantage over the memorise group; and (b) inthe memorise group, the performance wasabout 60%, whether or not feedback had beengiven. Learning concepts while trying to solve aclinical problem in problem based learning isanalogous to the problem solving conditions ofthe above experiments, while teaching aconcept and then having students apply itto a problem is analogous to the memorisecondition.
INTRINSIC INTERESTA student's intrinsic interest in the subject mat-ter, with a consequent impact on the motivationto learn, is enhanced by problem based learn-ing.5 The power of interest is easily overlooked.One of us recalls the experience of puttingdown a gripping paperback thriller in favourof a remarkably entertaining and intrinsicallyinteresting book about the very dullest of allsubjects (bar Latin that is), statistics.6
LIFELONG LEARNINGThe end point of conventional medical educa-tion is passing the final examinations andqualifying as a doctor. There is little focus onthe capacity to cope with the changes inmedical practice that will occur during the next30 years of clinical practice. There is increasingappreciation that one aim of undergraduateeducation should be to equip future doctors fora lifetime of self motivated and self directedlearning, a goal that is not achieved by passivelearning in a succession of lectures. There isevidence that students in a problem basedcurriculum acquire more self learning skillsthan students in a conventional curriculum,and that the difference is sustained well beyondthe duration of the curriculum.
Problem based learning is easilyunderminedIn a sense, problem based learning is not veryrobust. It is easy to undermine and devalue orshort circuit the whole process.
It is critical that students formulate theirown learning objectives. This process is shortcircuited if students are given specific learningobjectives. In specific problems, it is essential
to avoid 'advance organisers', that is labelswhich focus the students. The 'rapidly fatal ill-ness' problem described above would be oflittle use if labelled 'endotoxin and its harmfuleffects'. Tutor behaviour is important; forexample, providing direct answers will shortcircuit the whole process. The topic of thetutor role, the problems of small groups, andthe degree of subject matter expertise of thetutor are subjects in their own right, and arediscussed elsewhere.7-9 Lectures devalueproblem based learning and have no place,other than to explain particularly difficulttopics which are not readily accessible in otherways.The single biggest challenge is the appropri-
ate design of pass-fail examinations*. In con-ventional medical education, students arehighly exam driven.10 They study to pass 2ndMB, then 3rd MB, and then final MB, oftensubdivided into a whole series of individualpass-fail examinations in separate subjects.Their activities are very heavily influenced bythe need to pass exams. Their attendance at,and interest in, various learning opportunitiesis governed by the relevance to examinations.Problem based learning is not compatible withconventional examination systems; studentssimply find out what they need to pass theexams, and concentrate on this objective. Thesimplest solution is to avoid all pass-failexaminations. This was until recent years thesituation at McMaster, where the first suchexamination faced by students was theCanadian State Licensing Examination. (Thereader can see why we used the word anarchic!The reader's instant response might be thatthis could not possibly work for one's ownstudents - without examinations they woulddo no work at all. The answer is that in thepresent totally exam driven system, this isundoubtedly true, but it is a function of thesystem and not the students. If one wants tochange student behaviour, then first of all thesystem must change.)
In order to avoid the steering effect ofexaminations, it is necessary to break the linkbetween education and testing. To give justone of a number of possible examples (forother examples see Feletti et alI 1) of how thiscan be done, educators at Maastrichtachieved this by designing an examination,the 'progress test', that was so comprehensiveas to be virtually impossible to study for, andby separating the assessment of progressthrough the curriculum from performance onany single examination. This has recentlybeen taken up at McMaster, where it isknown as the personal progress index.12The principle of the progress test is that itconsists of a set of multiple choice questions(chosen because of cost effectiveness andhigh predictive validity) that is administeredto all medical students in every year of thecourse four times a year. The questions aretaken from the whole domain of medicine,
*Known in educational jargon as summative assessments, asdistinct from formative assessments, examinations the solepurpose of which is to provide feedback to the student.
361
on March 7, 2020 by guest. P
rotected by copyright.http://adc.bm
j.com/
Arch D
is Child: first published as 10.1136/adc.73.4.357 on 1 O
and all students, regardless of which year ofstudy, take the same examination. The resultsare graded so that up to one half a standarddeviation below the mean is taken to be satis-factory (S). Between 0 5 and 1 0 standarddeviations below the mean is classed as doubt-ful (D). More than 1 0 standard deviationsbelow the mean is classified as unsatisfactory(U). After each test, the student receivesa test performance report describing per-formance in different domains, and a perfor-mance profile report in which performance isrelated to previous performance and to theperformance of peers from the same year.The progress test system has a number ofunusual advantages:
* Research has shown that it is impossibleto study for the test. Thus the progress testeliminates activities solely aimed at passingexams.
* The test allows an individual student tofollow progress throughout the course, and ithas been demonstrated that students doindeed make steady progress throughout thecourse.
* Unlike the conventional system where astudent studies a subject, is examined in it,and then forgets about it, the progress testencourages continuing interest in subjectspreviously studied, and research has shown,for example, improved knowledge in basicsciences occurring during the final years of thecourse.
* It makes it much easier to make structuralchanges to the curriculum. For example, onecould switch a clinical attachment block intothe preclinical years, which would be muchharder to implement with an 'end of year'examination structure.
* There are no 're-sit' examinations, butsimply further progress tests every threemonths.
* Students get detailed feedback after eachtest, so that they can see in which areas they areweak.
* The system minimises 'preparationanxiety', partly because the students becomevery familiar with the routine of taking the test,and partly because preparation is impossible,and known to be impossible.A theoretical drawback to the progress test is
regarded as being the fact that one could passthe exam without any knowledge at all of, forexample, anatomy. Research has shown, how-ever, that in practice this does not happen. InMaastricht, the students also receive pass-failskills tests. In McMaster, there has beenemphasis on subjective tutorial evaluation bytutors and peers. The test is of course primar-ily a test of knowledge and understanding,rather than one that, for example, calls forreasoned decision making in the face of anunfamiliar situation. Also, the progress test,like all multiple choice examinations, presentschoices, whereas in real life doctors have togenerate the choices themselves.
Evaluation ofproblem based learningLocal pilot studies have been evaluated only by
student feedback, which has been very posi-tive. We are currently developing a randomisedtrial of problem based learning in paediatrics,because most of the studies thus far have beenof problem based learning during the preclini-cal years. Thus, both at Maastricht andMcMaster, problem based learning is mainlyapplied to the preclinical years (four years atMaastricht and two years at McMaster), theclinical part of the curriculum (two years atMaastricht and one year at McMaster) beingconventional clinical clerkships. There aresome reports from elsewhere about introduc-ing problem based learning into the clinicalpart of the curriculum.13The literature on the outcome of problem
based learning has been reviewed in detailelsewhere. 14 15 Sceptics can take comfort inthe fact that there is no objective evidencethat, for example, medical graduates from aproblem based course are better doctors(whatever that term might mean) than thosewho experienced a conventional education.Enthusiasts will point to studies demonstrat-ing better long term retention of knowledge,better study behaviours, greater student satis-faction, better performance of graduates,better lifelong learning, greater enthusiasm ofteachers, and other positive outcomes inschools that have studied the effects ofproblem based learning.
ConclusionsMedical education has a responsibility toensure that students are helped to developgenerally applicable competences such as life-long education, critical reasoning, dealingwith unfamiliar situations, collaborating interms, and various levels of communication,from obtaining information to counselling,adaptation to change, and participation inchange. None of these are well served by aconventional passive learning environ-ment.
Reading about French cooking is a poorsubstitute for the real thing, and may be diffi-cult to understand for someone accustomed tothe plain English fare ofbacon and eggs or fishand chips. Thus it is difficult to fully conveyour enthusiasm for problem based learning inthis short review. For the interested reader, wewould strongly recommend a visit to theUniversity of Limburg at Maastricht or theMcMaster University at Hamilton, Ontario,both of which run short (one to three week)workshops on various aspects of problembased learning as applied to medicine. In addi-tion, the University ofLimburg runs a distancelearning Master's Programme in HealthProfessions Education, in which one ofus (LP)is currently participating. For further informa-tion, write to us, or direct to: Office forInternational Relations, Faculty of Medicine,University of Limburg, POB 616, 6200 MDMaastricht, Netherlands or Programme forFaculty Development, Faculty of HealthSciences - Room 3N51g, McMasterUniversity, 1200 Main Street West, Hamilton,Ontario, Canada L8N 3Z5.
David, Patel
on March 7, 2020 by guest. P
rotected by copyright.http://adc.bm
j.com/
Arch D
is Child: first published as 10.1136/adc.73.4.357 on 1 O
6 Norman GR, Streiner DL. Biostatistics. The bare essentials. StLouis: Mosby, 1994.
7 Tiberius RG. Small group teaching: a trouble-shooting guide.Monograph series 22. Toronto: OISE Press (The OntarioInstitute for Studies in Education), 1990.
8 Schmidt HG, Van der Arend A, Moust JHC, Kokx I, BoonL. Influence of tutors' subject-matter expertise on studenteffort and achievement in problem-based learning. AcadMed 1993; 68: 784-91.
9 Schmidt HG. Resolving inconsistencies in tutor expertiseresearch: does lack of structure cause students to seektutor guidance? AcadMed 1994; 69: 656-62.
10 Newble DI, Jaeger K. The effect of assessments and exami-nations on the learning of medical students. Med Educ1983; 17: 165-71.
11 Feletti GI, Saunders NA, Smith AJ. Comprehensive assess-ment of final-year medical student performance based onundergraduate programme objectives. Lancet 1983; ii:34-7.
12 Blake J, Johnson A, Mueller CB, et al. Progress reporton the personal progress index. Pedagogue 1994;5(2): 1-6. (Pedagogue is published triannually by theProgram for Educational Development, Room 3N51Health Sciences Centre, McMaster University, 1200Main Street West, Hamilton, Ontario, CanadaL8N 3Z5.)
13 Chamberland M, Des Marchais JE, Charlin B. CarryingPBL into the clerkship: a second reform in the Sherbrookecurriculum. Annals of Community-Oriented Education1992; 5: 235-47.
14 Albanese MA, Mitchell S. Problem-based learning: a reviewof literature on its outcomes and implementation issues.AcadMed 1993; 68: 52-81.
15 Berkson L. Problem-based learning: have the expectationsbeen met? AcadMed 1993; 68 (suppl 10): S79-88.
on March 7, 2020 by guest. P
rotected by copyright.http://adc.bm
j.com/
Arch D
is Child: first published as 10.1136/adc.73.4.357 on 1 O
