Go back to IDI's Home Page...    Home SearchSite Index
 

Journal of Computer-Based Instruction


An Interactive Videodisc Drama:
The Case of Frank Hall
William G. Harless
National Library of Medicine

A videodisc-based, voice-driven, microprocessor-controlled case model designed by the author is described. The model supports experiential, contextual, and discovery learning in a variety of educational settings. General design considerations for this technique are reviewed.

The advent of videodisc technology has generated renewed interest and enthusiasm among medical educators for creative applications of computer-based instruction. In fact, this new technology has transformed the concept of computer-aided learning and provided the educator with unprecedented capabilities for the presentation of educational material. The opportunity for new educational designs and new theoretical approaches to the educational process has never been greater. To explore this opportunity, the author developed a voice-driven, videodisc-based, microprocessor-controlled model which supports experiential, contextual, and discovery learning strategies in a variety of educational settings. The model utilizes techniques of drama and contemporary cinematography to pique the interest of the student, and voice recognition technology to enhance involvement in the dramatic situation through the illusion of natural language interaction. The result is a new educational format for classroom and small group teaching: interactive drama.

The critical events and outcomes within an interactive drama are determined through probability analysis and decision tables. There is no predetermined event or outcome as a result of action taken by the user. An instructor may prescribe factors within the drama, which influence the events and outcomes, but may not dictate in advance any event or outcome. This important aspect of interactive drama creates the opportunity for discovery learning in the classroom, since teacher and student alike are actively participating in, and spontaneously responding to, evolving situations relevant to the learning goals. In this way interactive drama has the potential to improve the quality of student-student exchanges as well as student-teacher relationships in the classroom.

Interactive drama provides: 1) an engaging and relevant context for teaching cogent points; 2) the opportunity to participate actively in making decisions about important situations relevant to educational goals, observing the results of interventions, and discussing the implications with instructors and peers; and 3) the opportunity for both students and instructors to engage in true discovery learning since the course of events and outcomes of a drama are not predictable – even by the instructor.

Technology Configuration

The hardware used to develop and utilize an interactive drama currently consists of an IBM PC/XT microprocessor; a Votan VPC-2000 voice system; a Pioneer LD-V6000 videodisc player, and a Sony color monitor. The effective integration of the capabilities of these technologies is essential to the realization of the educational strategies embodied in the interactive drama format. Specifically, the interactive videodisc allows believable dramatization of predetermined situations through immediate and random access to a variety of dynamic video and audio material. Voice recognition makes the technology inconspicuous to the knowledgeable user and allows uncued, natural language interaction with the characters in the drama. The microprocessor controls the voice-driven, random access video/audio system, supports an intelligent interface between the user and the system, allows for probability-based decision making concerning critical occurrences and outcomes of the case, and monitors the tracking necessary to evaluate student performance.

An Example of Interactive Drama

A prototype interactive drama has been developed in the field of medicine, The drama portrays a situation in the life of a fictitious patient named Frank Hall, who is played by a professional actor. Frank Hall is a 46-year-old man who comes to an emergency room, complaining of weakness and abdominal pain, following an episode of vomiting blood two days earlier. The simulated patient illustrates the medical content and sociological issues related to a variety of gastroenterology problems confounded by alcoholism. The Frank Hall case contains a variety of scenes depicting the patient’s current experience in the hospital and significant past experiences (through flashbacks) related to his present illness. Other scenes depict elements of the physical examination, relevant diagnostic tests and procedures, visuals of x-rays, endoscopic views and so forth, whenever appropriate. The user (the instructor, in concert with students) assumes the role of the physician in this interactive drama. Through vocal inquiries and commands, the physician interviews the simulated patient and directs the course of action. The system responds instantly to each inquiry and command.

Prior to the simulation, the instructor trains the system to recognize 135 control words consisting mainly of medical history concepts (presenting complaint, pain, smoking habits, drinking habits, etc.) and terminology (hematocrit, urinalysis, endoscopy, etc.) to diagnose and manage a patient. This training requires 25 - 30 minutes. The computer records the instructor’s voice print in the form of a digitized voice pattern. These patterns are stored and subsequently used by the computer to recognize the instructor’s inquiries and commands during the presentation. The system’s recognition rate following training is greater than 95 percent.

Once the system has been trained, the instructor can act as the teacher, guide, facilitator, and spokesperson for the class. The drama can unfold in ways that are never totally predictable. Figure 1 is a flowchart of the various decision points of the Frank Hall case. Each triangular figure on the chart represents a point at which the various outcomes are determined through a probability model within the system. Each time a decision point is reached, a random number is generated by the system and matched to the probability ranges of each of the outcome possibilities. The match determines the course of action taken at that time. An example of a typical interaction with Frank Hall follows:

The case begins with a full color action scene, including background music, of Frank Hall driving a large, battered automobile into the hospital parking lot. The scene changes to a disheveled Frank Hall at the emergency room reception desk.

Frank Hall (to the ER nurse)

"I’m sick; I need to see a doctor."

A dialog between Frank and the nurse takes place for approximately 30 seconds at which point the action freezes. This freeze in the action is referred to as a "wait state." serves as a prompt to the student to take charge and respond. There are no menus or cues in the simulation. The user (instructor and students) must determine the course of action. They may take the patient’s history, order diagnostic tests, admit him to the hospital, or release him. The system will remain in a wait state until a verbal command is given.

Taking the Patient’s History

User (speaking into the microphone)

"Begin."

To interview the patient, the instructor requests orally "Mr. Hall." This request returns the patient to the screen looking directly at the class, waiting to respond to question Or, the instructor may say, "Start the workup," which elicits a scene of Frank Hall in an examining room having his vital signs taken by the ER nurse. The scene changes t the patient in an examining gown, again looking directly the group, and waiting to answer questions. These, too, are wait states where the instructor and students must determine the course of action, In this example they decide to interview the patient.

User "What’s your problem?"
Frank Hall (looking directly out at the student) "I’m sick. I’ve been throwing up blood ... I’m afraid I might have cancer ..." (Wait state.)
User "When did it begin?"
Frank Hall "Well, it started last week. I was in this restaurant ..."

The scene becomes a "flashback" to the restaurant showing Frank Hall talking to the waitress and beginning to feel sick. The patient continues to describe the situation as a flashback scene to his bathroom and the episode of hematemesis is portrayed. Upon completion of the flashback scene, the patient reappears in the examining room in a wait state. This user-controlled dialog with the simulated patient can continue through a complete social and medical history, or the user may digress to request test procedures or other information.

User "Vital signs."

Values for Frank Hall’s temperature, blood pressure, pulse, and respiration appear as superimposed text over a scene of the nurse taking his vital signs. To continue talking with the patient, the user simply says the patient’s name and he reappears on the screen.

User "Mr. Hall."
(Frank Hall reappears.)
Frank Hall "Yes."
User "You mentioned your wife."
Frank Hall "Rita’s my wife...er, she was my wife. We’re separated now... She’s filing for divorce. We’ve been married for 15 years."

The interactive drama is designed to allow the class to probe as deemed appropriate to elicit information from the patient. The user may do this by asking open-ended questions such as: "Is there more?" "Tell me more." "Is that all?" the patient continues to respond until he has nothing more to say about that concept. Probing questions concerning Frank Hall’s marriage reveal, through flashback, a wife-beating incident and other pertinent information about the patient and his relationships.

Managing the Patient

When enough information has been gathered, the class may decide to admit the patient to the hospital for diagnosis and treatment. If so, the user is prompted by a prerecorded voice (monitor) to provide the differential diagnoses the group is considering. The system confirms each of the user’s verbal entries with on-screen text.

User "Admit the patient."
Monitor voice "Please state your differential

diagnoses, one at a time

User: "Hemorrhagic gastritis."
Monitor voice "Is there another?"
User "Mallory-Weiss Syndrome."
Monitor voice "Another?"
User "Cancer."
Monitor voice "Is there more?"
User "That’s all."

 

The system then asks the user to provide, in the same manner, a plan for diagnostic workup and for immediate management. With this information, the computer com- poses Frank Hall’s hospital experience by selecting scenes for the diagnostic procedures ordered. The scenes are then organized, in proper chronological sequence, with other scenes depicting the patient’s activities, boredom, and relationships in the hospital. The instructor and students may simply observe these scenes, by watching the patient as he goes through each day, or stop the action to revise the diagnostic workup and/or therapeutic plan.

At any time during any scene, the user may use the command word "thoughts" to hear what the patient is thinking. The purpose of this feature is to remind students to be aware of the stresses with which patients must cope, which often are never verbalized and remain unknown to the physician. This unique capability of the technology makes it possible to raise consciousness concerning the patient’s emotional state and broaden their perspective of the patient’s clinical experience.

The instructor also may command the system to move ahead in time. For example, by stating, "Second afternoon," the simulation advances and shows scenes for the requested day and time. This enables the instructor to condense time and obtain results of diagnostic procedures ordered earlier. The results of the tests are superimposed as text over appropriate background scenes. The command "interpretation" elicits vocal interpretations of test results by a medical expert, which often include visuals of the procedures.

During this entire time the responsibility for managing the patient remains in the hands of the instructor and students. The system provides no cues, prompts, or menus which might by themselves elicit a response.

Probability-Based Decision Making

The TIME interactive drama model utilizes a probability algorithm to make decisions concerning certain occurrences, and to provide autonomous responses depicting outcomes during the course of the classroom interaction. There are eight decision points in the Frank Hall prototype. Each decision point relies on a specific table of probability values to determine the outcome of the situation in question. The probability values are dynamic and may be user-influenced during the course of the case by: 1) relevant inquiry; 2) intelligent intervention; and 3) perseverance. An expert medical committee determines the specific content items which may influence the probability values.

This dynamic internal decision making provides a lifelike mystery to the TIME interactive case study, and offers the opportunity for discovery learning to unfold in the classroom. Even the instructor cannot always predict the direction the case will take. For example, in the real clinical situation Frank Hall would be subject to various health crises during this critical period of his illness. Namely, he would face the distinct possibility of a rebleed, delirium tremens, or a myocardial infarction. Scenes for these crises and their associated probabilities of occurrence are built into the Frank Hall case study. At a designated time during the second day following admission, the TIME system generates a random number and compares it to the probability values for each of these three crises. Should the random number fall within the probability range of one of the crises, that crisis will occur and the user is responsible for spontaneous intervention and management of the situation. This crisis intervention is accomplished through uncued verbal commands by the user to which the system responds immediately. As in the actual clinical situation, appropriate intervention and crisis management increase the probability of a successful outcome, while inadequate or inappropriate intervention decreases that probability. There is no guarantee of success or failure, only the opportunity to influence their probabilities.

Discharging the Patient

Ultimately, if Frank Hall survives his hospital stay, he must be discharged. When the group decides to discharge him, the instructor gives the command "discharge the patient." The system then asks for the final diagnoses and a discharge plan. This process involves the same user-system interaction as described previously for "admit the patient." Once these decisions have been provided, information as to what happened to the patient after discharge can be obtained by issuing the command "outcome." This command causes the system to compare a randomly generated number to three probability values and, thereby, determine the out- come for the patient. He may: 1) recover and become gainfully employed and productive; 2) return to the hospital as a "revolving door patient"; or 3) leave and be found dead sometime later. The outcome chosen by the system is the one seen by the class. The instructor may obtain, through vocal command, the probability values associated with the out- come decision made by the system. These values will support class discussion and intellectual dissection of the group’s performance concerning the management of the patient, the discharge plan, and the outcome.

Feedback

The interactive case study concludes following the out- come scene. At this point the instructor may request feedback from the system, which is then presented in the form of charts and graphs describing the group’s performance with the patient. The feedback parameters include:

  1. correctness of final diagnoses;
  2. correctness of discharge plan;
  3. percentage of critical information obtained during interaction with the case (according to expert committee), with breakdowns;
  4. danger/discomfort index for tests and procedures ordered, with breakdowns;
  5. costs incurred by the patient based on procedure ordered and length of hospital stay;
  6. appropriateness, efficiency and effectiveness of crisis management.

The Distinctive Characteristics Each of the following characteristics distinguish the interactive drama model as a unique concept in educational technology.

  1. Interactive drama is uncued in that there are no menu from the inner system to prompt the "external character" (user) in terms of action choices, dialog, o other interventions which are possible. The external character is, therefore, left to individual resources t< assume the role of problem solver in the scenario.
  2. The random access capability of the interactive dram program allows the user (external character) to intervene in the drama at will, while also permitting the system to respond with the appropriate information response, or scene. The external character, therefore must take the initiative in the scenario and determine appropriate pathways through the situation.
  3. The dialog between the external and internal characters is accomplished through voice recognition technology, which allows the external character to usi individual natural language to query and converse with characters in the scenario.
  4. Decisions concerning occurrences of various situations within the drama, and outcomes related to those and other situations, are determined by probability algorithm which is an integral part of the interactive drama model. At predetermined decision points during the course of the drama, the system generates a random number which is compared to the probability values associated with the possible occurrences and/or outcomes for that situation. These probability values are dynamically influenced by the decisions, actions and interventions of the external character(s) as the drama unfolds.

Summary

Interactive drama is a method for classroom education which is distinguished from the lecture, seminar, discussion groups, study halls, and the like. Interactive drama takes the form of a dynamic computer-driven, video-based scenario in which: 1) the events and outcomes are not predictable; 2) the primary problem solving character in the scenario is played by the students in the classroom; and 3) the interaction is uncued and accomplished through natural spoken language by the user. The process involves active participation by the student: in making decisions, with discussion whenever necessary to clarify and expound on issues which arise during the scenario. The interactive drama is interruptible at any time by the users. It will simply stop on command and wait until the user commands it to continue. This and other features o this method allow the teachable moment to be exploited by instructors and students.

--Journal of Computer-Based Instruction, Volume 13, Number 4, Autumn 1986 (p. 113-116)
Copyright © 1986 Journal of Computer-Based Instruction
With thanks to the Association for the Development of Computerized Instructional Systems

 

questions? info@idrama.com

If you have any questions or comments about our site, send an email to webmaster@idrama.com
Copyright © 1996-2008 Interactive Drama Inc. All rights reserved.
Interactive Drama Inc. 7900 Wisconsin Avenue, Suite 200, Bethesda, MD 20814
 

Join our mailing list to receive IDI announcements and special
discounts for Virtual Conversations® products.
Enter your email address below and click the "Opt-in" button:

We respect your privacy; IDI will never share or sell addresses.