Teaching and Learning Forum 97 [ Contents ]
DADA: A diagnosis, assessment and decision aid for the clinical and behavioural sciences
Kim Kirsner, Andrew Page, Danny Boase-Jelinek, Mark Randell
Department of Psychology
The University of Western Australia
School of Speech and Hearing Science
Curtin University of Technology
Our demonstration depicts the first stage of DADA, a Computer-Aided Instruction and Learning package that was designed and developed to facilitate the acquisition of diagnostic expertise by students in clinical and behavioural science. DADA consists of a multimedia database and associated resources which students use to study clinical cases, observe patients being assessed by clinicians, monitor the behavioural cues, criteria and procedures used by experts during diagnosis, view integrated models of cognitive, psycholinguistic, neurological and arterial function, and review behavioural, cognitive and neurological test data. DADA provides a solution to a critical problem in clinical training namely, first hand access to that body of case knowledge and experience which underpins clinical expertise. The first application of DADA involves broad differential diagnosis for acquired neurogenic disorder following stroke.
Considered as a package, DADA meets some of the criteria offered by Grabinger and Dunlop (1996) for a "REAL" system; that is, a system which (1) promotes study and investigation within authentic environments, (2) encourages the growth of student responsibility, initiative, decision making and intentional learning, (3) cultivates collaboration among students and teachers, (4) it utilizes dynamic inter-disciplinary, generative learning that promote higher order thinking processes to give students rich and complex knowledge structures, and (5) it assesses student performance within authentic contexts using realistic displays and problems.
The scientist-practitioner model, that is fundamental to so much of training in psychology reflects the integration that is required between basic science on the one hand, and clinical practice on the other. Although the strong and explicit emphasis upon the scientist-practitioner model of training is probably unique to psychology, common to all the clinical and behavioural sciences is the desire to train students so that the declarative knowledge base of the discipline is integrated with the procedural knowledge acquired through clinical experience. Perhaps nowhere is the need for this integration stronger than in the area of clinical decision making and diagnosis. However, it is this very area where training in all the clinical and behavioural sciences can falter.
The Training Problem
Training in diagnostic skills involves extensive, direct and supervised access to people with disorders from one or more of a set of pre-identified categories. But these objectives are difficult to achieve in practice. Supervised tuition is expensive and time-consuming, even under the typical 'SEE-ONE / DO-ONE' procedure. Notwithstanding these difficulties, the aim in training diagnostic expertise remains to enable students to acquire skills in making judgements about complex and multidimensional presentations.
The critical problem is that there are inadequate numbers of patients with the requisite disorders. In traditional training systems this problem is resolved by providing only very limited access to selected patients, for both training and assessment. But this limits the acquisition of expertise, and forces practitioners to acquire basic skills at the expense of their clients. The major implication of this argument is, then, that effective access to case information is the critical bottleneck where clinical training is concerned. Limited exposure to cases forces students to acquire domain knowledge by reference to general principles and rules alone. This solution is unsatisfactory however; declarative knowledge needs to be supplemented by procedural knowledge, but procedural knowledge can only be acquired by direct contact with cases.
A further problem stems from the fact that, because symptom patterns are consequences of different underlying aetiologies, many symptoms are associated with disorders from several taxonomic categories. For example, voice quality problems may accompany dysarthria, dysphagia or dysphonia. To further complicate training, these taxonomic categories are routinely diagnosed and managed by people from different disciplines.
Another problem concerns the difficulties associated with the provision of equitable supervision across students with different cases or even students across different supervisors. Some cases are harder than others, some supervisors are better than others, and some students are more assertive than others. Combined with a scarcity of patients, these factors compromise training and assessment.
In summary, limited access to cases impedes the acquisition of expertise because it forces students to acquire domain knowledge by reference to general rules and principles. If expertise is to be achieved, rule-based learning needs to be supplemented with the accumulation of diagnostic experience to create a richer learning platform. The foregoing analysis draws attention to the central problem in clinical training. Because the extent of supervised case-based training in any one discipline is limited by practical and ethical problems, adequate preparation of clinical practitioners in the clinical and behavioural sciences is difficult to achieve. These problems define the issue then. Can multimedia tools be refined to solve these perennial and vexing problems of clinical training?
The multi-media solution
One possible solution to many of the problems described can be found within Computer-Aided Instruction and Learning (CAIL). Whereas most CAIL programmes are designed to support the development of a selected skill or skills from a broad domain, we have developed a training system that occupies an unusual role in student instruction. The system that we have developed, called Diagnosis, Assessment and Decision Aid (DADA), has the principal objective to be essentially the same as that of an undergraduate degree programme in the clinical and behavioural sciences. The first version of the package is designed to facilitate the acquisition of broad diagnostic skills by clinicians exposed to stroke patients. It involves training about acquired neurogenic communication disorders such as dysphasia and dysarthria, which will have a particular attraction for speech pathologists. However the general principles developed in this project are equally applicable to any domain where human behaviour is a central element in diagnosis, involving neuropsychology, depression and anxiety for example.
The design of DADA was shaped by the following considerations:
- A Case-Based Instruction system should support access to audiovisual and test data for cases from the target domain and one or more adjacent domains.
One possible criterion for the range of cases for a given system involves the concept of confusion, and the possibility that the Data-Base of cases should include cases from all categories that could be confused with the target disorder, although this would need to be qualified by reference to the expertise of the referent.
- A Case-Based Instruction system should include and represent the decision-making procedures used by domain experts.
Although a Data-Base of cases is a necessary condition for Case-Based Instruction, it is not a sufficient condition. The system should also include a procedure to assist students to move from the declarative knowledge that they have acquired during formal instruction to the procedural expertise involved in practice.
- A Case-Based Instruction should provide access to resources from all knowledge domains that are critical to diagnosis at a time and in a form that can be used to facilitate decision-making.
In training programmes for Speech Pathologists, for example, students will be exposed to specialised material involving Cognition, Linguistics, Neurology, Communication Disorders, Anatomy and Physiology, Medical Disorders and Clinical Practice. Unfortunately, however, this material is usually provided by staff without clinical expertise, and the information is not therefore taught as an integral element in the diagnostic process.
The aim of the programme was to train skills in making broad differential diagnosis, which involves making distinctions among acquired neurogenic communication disorders including dysphasia, dysarthria, dyspraxia, dysphagia and right hemisphere syndrome as distinct from narrow differential diagnosis involving finer distinctions between, for example, lexical and syntactic difficulties in dysphasia. The detailed design of the programme was influenced by the instruction objectives:
- Students should understand the broad diagnostic categories for acquired neurogenic communication disorders following cerebro-vascular accident.
- Students should acquire pattern recognition skills involving behavioural and other cues sufficient to support broad differential diagnosis.
- Students should master the decision-making processes required for broad differential diagnosis.
- Students should acquire cognitive, linguistic, neurological, motor and cerebral blood flow knowledge in a form that will enable them to use this information for broad differential diagnosis.
Overview of system architecture
DADA achieves its objectives by providing:
DADA provides knowledge to students in three general forms. The first of these involves knowledge about specific cases. The second involves information that is designed to facilitate acquisition of diagnostic skills. The third form involves background knowledge about language, cognition, cerebral blood flow and other topics. The second and third of these forms correspond to procedural and declarative knowledge as these terms are used in Cognitive Psychology. The first form provides the instances or episodes on which expertise is based.
- access to an interactive multi-media database comprising sets of cases selected and organized to support the acquisition of knowledge and diagnostic skills about acquired neurogenic disorders such as dysphasia, dyarthria and dysphagia
- a procedural resource consisting of a scaffold specifying critical decision steps for medical diagnosis and symptom classification
- a portfolio of declarative resources including:
- neurological models depicting blood supply and the impact of stroke on blood supply
- cognitive and psycholinguistic models depicting some of the processes involved in language reception and production and the possible impact of strokes on those processes
- conventional case data including material normally available to clinicians
- a case analysis interface designed to facilitate integration of the procedural and declarative resources
Overview of system interface
The interface for DADA supports two forms of access to the system, structured access, where novices follow a fixed path involving the video vignettes for a particular case, and unstructured, where experienced students are free to move through the resources and vignettes in any order. Under structured access for example, students are expected to select an answer about a particular vignette, after which they may be given feedback about their decision and access to an expert commentary associated with their response. The question for each vignette has to be answered before the novice can proceed to the next vignette. In order to answer each question, the students may observe the vignette again, or access the resource portfolio. The students can also indicate the contribution of specific resources to their decisions.
When the students has completed the vignettes for a particular case, they should have formed an impression of the pattern of behavioural symptoms that are critical for that decision, and the way in which the resources can be used to refine their decision. The objective is to demonstrate to the students how information from a variety of resources can be integrated to support and inform a complex decision, and to lead him or her through that process.
The knowledge resources were formulated on the basis of "decomposed" expert knowledge based on knowledge elicitation. Interviews with experts were used to create a Question-Answer-Commentary loop for each vignette for each case in the DataBase. The material was collected and prepared according to a technique based on the recognition primed decision procedure developed by Klein (1990).
Relative to Klein, the major modification was that the time line involved critical behaviours rather than critical decision points. Thus, decomposition of the decision path involves critical behavioural observations rather than critical decisions, although each behavioural observation makes a contribution to the final diagnosis or classification.
Is DADA a "Rich Environment for Active Learning"?
Considered as a multimedia package, DADA meets some of the criteria offered by Grabinger and Dunlop (1996) for a Rich Environment for Active Learning. The first criterion identified by Grabinger and Dunlop is that the package should promote study and investigation within authentic environments. By providing students with access to cases DADA provides a simulation of an authentic environment. DADA falls short of the ideal in two respects though; students are not given access to the full patient interview, and there is no provision for direct interaction with the patient, characteristics that require the presence of a virtual patient. The second criterion is that the package encourage the growth of student responsibility, initiative, decision making and intentional learning. Operation of DADA is predicated on student initiative and decision-making, however, where procedural skill is concerned, the presence of incidental learning is probably to be desired. The third criterion is that the package should to cultivate collaboration among students and teachers. DADA does not, as it stands, meet this criterion except in so far as it provides students with practice in the type of questions that they could meaningfully ask an expert clinician. The fourth criterion is that the package should to utilize dynamic inter-disciplinary, generative learning that promotes higher order thinking processes to help students rich and complex knowledge structures. DADA is rich in inter-disciplinary resources and, because students must master and integrate these resources to solve diagnostic decisions, the package provides ideal conditions for integrating complex knowledge structures while engaging in authentic, problem-solving tasks. The fifth criterion is that the package should assess student performance within authentic contexts using realistic displays and problems. An assessment component based on these considerations is under development.
Grabinger, R. S. & Dunlop, J. C. (1996). Rich environments for active learning. Association for Learning Technology Journal, 3(2), 5-34.
Klein, G. (1990). Recognition primed decisions. In W. Rouse (Ed.), Advances in Man-Machine Systems Research, Vol. 5, pp. 47-92. JAI Press.
|Please cite as: Kirsner, K., Page, A., Boase-Jelinek, D., Randell, M. and Hird, K. (1997). DADA: A diagnosis, assessment and decision aid for the clinical and behavioural sciences. In Pospisil, R. and Willcoxson, L. (Eds), Learning Through Teaching, p177-181. Proceedings of the 6th Annual Teaching Learning Forum, Murdoch University, February 1997. Perth: Murdoch University. http://lsn.curtin.edu.au/tlf/tlf1997/kirsner.html|
[ TL Forum 1997 Proceedings Contents ]
[ TL Forums Index ]
HTML: Roger Atkinson, Teaching and Learning Centre, Murdoch University [firstname.lastname@example.org]
This URL: http://lsn.curtin.edu.au/tlf/tlf1997/.html
Last revision: 2 Apr 2002. © Murdoch University
Previous URL 12 Jan 1997 to 2 Apr 2002 http://cleo.murdoch.edu.au/asu/pubs/tlf/tlf97/kirs177.html