Interactive Video-training for Medical Professionals

Marilisa Amoia

Department of Applied Linguistics, Saarland University, Saarbr

ucken, Germany

Keywords:

Information Technologies Supporting Learning, e-Learning Platforms, Interactive Language-training, ICall,

Lifelong Learning, Medical Applications.

Abstract:

This paper describes an innovative component of an e-learning platform targeted at second language training

of medical professionals. This component provides interactive video learning material in three European lan-

guages: English, German and Spanish. The implemented prototype integrates storytelling technology with

NLP and dynamic web engineering. Starting by the automatic annotation of relevant vocabulary and con-

cepts in the domain of healthcare, our e-learning component associates the videos with external open domain

resources, such as thesauri and specialized dictionaries thereby providing a comprehensive infrastructure sup-

plying medical professionals with a situated learning environment as well as with open domain look-up and

information retrieval facilities in three languages.

1 INTRODUCTION

Over the last few years, Europe has witnessed an in-

creased mobility of its citizens and professionals. Af-

ter the removal of the barriers for services such as

education and healthcare in particular, more people

move abroad to study, reside and work or for medical

treatment. In order to facilitate this mobility, the Eu-

ropean Union is promoting research for the develop-

ment of e-learning resources and platforms for second

language learning of all European languages.

A good command of the second language, in fact,

is particularly relevant in disciplines such as health-

care (cf. (Leroy et al., 2010)), where a missunder-

standing between patients and doctors may yield un-

safe services, e.g. wrong diagnoses. Recently, much

research has addressed the problem of providing high

quality multilingual services in the medical domain.

Most of this work addressed three main issues:

(i) Providing a uniﬁed database/ontology of health-

care terminology that can be instantiated in dif-

ferent languages. MeSH

, the Medical Subject

Headings maintained by the U.S. National Li-

brary of Medicine and its counterparts in other

languages (e.g. German and French) are promi-

nent examples of such a line of research (see (Nel-

son, 2009), (N

eol et al., 2007)).

(ii) Providing ad-hoc tools for the multilingual

http://www.nlm.nih.gov/mesh/

retrieval of scientiﬁc literature and medical

databases. Most of this work addresses the prob-

lem of creating centralized catalogues and seman-

tic indexing facilities that can facilitate the search

of multilingual scientiﬁc material for reference or

study (see (Kreuzthaler et al., 2011), (Grosjean

et al., 2012)).

(iii) Providing translation tools for healthcare profes-

sionals. MedSLT (Bouillon et al., 2008) is an ex-

ample of a multilingual spoken language transla-

tion system specialized on the medical domain. It

was developed with the aim of aiding the commu-

nication between patients and doctors in real situ-

ations. It supports speech recognition and transla-

tion facilities but its coverage is restricted to a few

domains (headache, chest and abdominal pain).

However, few e-learning resources exist which in-

tegrate all these research trends into a unique plat-

form. In this paper, we describe a component of an

e-learning platform speciﬁcally developed for teach-

ing second language to medical professionals provid-

ing a comprehensive resource that supports individual

learning and can be used as well as an assistive equip-

ment in the every-day patient-doctor communication.

The e-learning component presented here implements

interactive video training material.

A ﬁrst prototype has been realized based on sto-

rytelling technology. In the design of this component

we exploit multi-media and dynamic web technology.

In particular, we provide links between videos, pro-

122

Amoia M..

Interactive Video-training for Medical Professionals.

DOI: 10.5220/0004386701220127

In Proceedings of the 5th International Conference on Computer Supported Education (CSEDU-2013), pages 122-127

ISBN: 978-989-8565-53-2

 2013 SCITEPRESS (Science and Technology Publications, Lda.)

Figure 1: Interactive Video Training: the GUI.

duced with storytelling and text-to-speech technology

on the basis of real transcript of patient-doctors dia-

logues, with external resources allowing the user to

interact with the content of the video and to access

external knowledge.

This paper is organized as follows. In Section 2,

we present the teaching methodology behind the de-

sign of the interactive video training learning compo-

nent. Then we introduce the technical framework on

which the implementation of the e-learning compo-

nent is based. We describe the storytelling technol-

ogy used to realize the video material as well as the

system-user interaction model supported by the com-

ponent. Finally, we show how the videos are linked

with external knowledge resources. In Section 3 we

conclude with some ﬁnal remarks and pointers for fu-

ture work.

2 INTERACTIVE VIDEO

TREANING

In this section, we present the framework behind the

interactive video training component implemented in

our e-learning platform. We describe how video train-

ing is implemented and which features support the

system-user interaction. Further, we give an overview

of the external resources which are linked to the video

material.

2.1 The Teaching Model

Situated learning (cf. (Lave and Wenger, 1990),

(Brown et al., 1989)) is acknowledged among the best

methodologies for enhancing effective second lan-

guage learning, particularly in the case of adult learn-

ers. A distinctive feature of situated learning is that

it presupposes that learning is linked to a real, every-

day situation in the target culture and that the learner

is immersed in a context where he/she can learn the

language in use in that particular situation.

Further, there are several studies that highlight the

role of multimedia input in enhancing learner perfor-

mance and engagement. (Amoia et al., 2011)), for

instance, show that children learn new vocabulary in

a second language faster, i.e. more effectively and

remember more words if they are exposed to both

audio and textual/visual input. Video training mate-

rial can engage learners more than simple text mate-

rial (Zhang et al., 2006) and it represents an efﬁcient

means to implement situated learning. It further al-

lows to fulﬁll several tasks required by the learning

component presented in this paper, and namely:

(i) to improve listening comprehension skills of

medics and healthcare givers,

(ii) to enhance cross-cultural understanding, an essen-

tial feature to assure satisfying doctor-patient in-

teraction and safe medical services.

As we focus on language learning in communica-

tion, our learning material mainly addresses spoken

interaction between patients and doctors. The e-

learning component provides interactive video mate-

rial in three European language, e.g. English, German

and Spanish. The target learners accounted for in our

framework include the following groups: (i) medical

InteractiveVideo-trainingforMedicalProfessionals

123

Figure 2: Automatic Links to External Resources.

professionals (doctors and nurses) working in a coun-

try other than their own, (ii) patients enjoying medi-

cal treatments in a country other than their own, (iii)

medical professionals working in their own country

but with patients speaking a foreign language.

The learning material aims at allowing self-study

but also at being used as an assistive resource for

reference during real patient-doctor interaction in the

clinic or in the daily care work. In order to satisfy the

needs of all three groups of learners in the learning

component, we distinguish between the clinical lan-

guage used by doctors to communicate with expert

colleagues and the more common, every-day medical

language used to communicate with the patients that

are generally laypersons.

Doctors should be able to express themselves in

the second language in such a way that the patient

can understand them. Thus in the interactive video

learning component, we have tried to ﬁnd a balance in

terminological complexity. The dialogues displayed

in the videos include a language complexity that cor-

responds to layperson, every day medical language

knowledge and specialized terminology is reduced to

a minimum. So for instance in the examples below,

we prefer the formulation in (1) to that in (2).

(1) Do you have any shortness of breath, or breath-

lessness?

(2) Do you have dyspnea?

Consequently the material can be used for learning by

both medical experts and laypersons. Of course, doc-

tors will use the term dyspnea in the report for the hos-

pital and talking with expert colleagues. This more

speciﬁc terminology can be accessed in our platform

through the knowledge resources linked to the videos.

2.2 Video Material and Dialogues

The video material included in our e-learning compo-

nent aims at recreating real situations in the healthcare

domain and typical patient-doctor interactions.

The videos are based on a corpus of texts includ-

ing typical patient/doctor dialogues gathered by ex-

pert physicians in Spain. The topics addressed in

these dialogues includes:

• putting the patient at ease,

• collecting patient demographics,

• taking a patient’s medical history,

• describing symptoms,

• explaining a treatment to the patient,

• describing the effect of a prescribed medication,

etc.

These dialogues have then been translated by expert

translators into English and German and reviewed by

doctors for checking property of language and natu-

ralness. The implementation of interactive video ma-

terial based on these dialogues includes two steps:

• creation of the video material,

• making the video interactive.

CSEDU2013-5thInternationalConferenceonComputerSupportedEducation

124

Figure 3: Thesaurus and Dictionary Look-up.

We start by describing the ﬁrst step. We use sto-

rytelling technology as provided by the Xtranormal

platform. This software is partially freely available.

This system supports text-to-speech facilities and pro-

vides animated video generation from written texts,

where animated characters act the patient-doctor in-

teraction. Xtranormal allows to set the value to differ-

ent features. The system

• supports 20 different languages including those

relevant for our e-learning component, i.e. En-

glish, German, Spanish,

• provides different male, female and children

voices,

• allows to chose between different scenarios (e.g.

ofﬁce, school, hospital) for the animated dialogue

and different background sound effects,

• allows to set the camera angle and provides zoom-

ing facilities,

• supports different animated characters (avatars)

that can walk, sit and even make up to 80 different

gestures.

The videos included in the prototype of the e-learning

component contain short dialogues (about 5 min) il-

lustrating a typical situation in the medical practice

in three European languages. In the next section, we

describe the second step, i.e. how interactivity is man-

aged in our component.

http://www.xtranormal.com/

2.3 System-user Interaction

In order to create interactive videos, we exploit the

dynamic web technologies integrated into Mozil-

las HTML5 media framework and Javascript. This

framework allows us to manage video, audio and

text material dynamically and to link our compo-

nent to open data resources and web services such as

Wikipedia, Google Map, etc.

In order to create an immersive and interactive

video experience, we proceed in the following way.

Starting from the transcript of a real patient-doctor

dialogue provided by medical professionals, we pro-

duce the video, and then we automatically generate

links to external resources. The written dialogues are

parsed and aligned with their translations, and key-

words relevant to the medical domain are extracted.

Thus, those keywords are highlighted in the textual

output of the component and automatically linked to

external resources such as Wikipedia and MedLine in

all three languages. The timeline managing the in-

teraction with the video is synchronized with the di-

alogue content. This allows us to extend the video

content with hypermedia experiences and links to ex-

ternal resources.

Thus the system supports the following sys-

tem/user interactions. The user can:

• stop/start videos on demand,

• ask for annotation/transcript of the dialogue being

played in the given video,

InteractiveVideo-trainingforMedicalProfessionals

125

• ask for translation of the dialogue being played in

the video,

• click on highlighted words in the text transcript

of the dialogue and thus access external vocabu-

lary information, translation or linguistic knowl-

edge relevant for these words,

• ask for further information relevant to the text

topic, such as information from a thesaurus or an

external dictionary,

In this way, the interactive video training component

of our e-learning system provides a situated learn-

ing experience and further allows the learner to freely

navigate through the linked open domain resources

providing external, probably more specialized con-

tent.

2.4 External Resources

The external resources linked to the video in the pro-

totype platform are the following:

Wiktionary. At the moment the dictionary lookup is

implemented as a Wiktionary query. This choice

is motivated by the fact that this resource is freely

available and provides dictionaries in the three

languages required by our learning component.

Wiktionary also supports common medical vocab-

ulary that is relevant to a standard patient-doctor

interaction. In order to support physicians at writ-

ing specialized reports, we plan to include look-

up facilities in specialized terminology resources

in our component, such as MeSH.

MedLine, Wikipedia. At the moment, the dia-

logues are linked with the MedLine

database that

includes a medical domain thesaurus targeting pa-

tients or laypersons, and additionally links to rel-

evant articles, books in both English and Spanish.

For German or in the case the keywords are not

found in MedLine, our component uses Wikipedia

lookup.

Google Translate. The system further provides

translation facilities. At the moment, this service

is implemented as a pop-up facility that opens a

link to the Google Translate

service and can be

used by the user for translating new words into

his/her native language or in any language sup-

ported by the Google service.

MedLine is a service of the U.S. National Li-

brary of Medicine National Institutes of Health. http://

www.nlm.nih.gov/medlineplus/

http://translate.google.com/

2.5 The GUI

The user can interact with the video training compo-

nent through the system GUI shown in Figure 1. In

a typical learning session while watching the video,

the learner can interact with the system by asking for

a translation in one of the supported languages (En-

glish, Spanish or German). Further, the user is aided

in learning new vocabulary. In Figure 1 for instance,

the word womb is highlighted, as it is recognized by

the system as a relevant keyword in the medical do-

main. The system provides links to (Wiktionary) dic-

tionary entries (in the bottom left slot) and thesaurus

deﬁnitions of relevant concepts (e.g. Voltaren, right

canvas on the top). The user can decide to access

this information by clicking on the highlighted word.

These links are also available if the video is stopped.

By clicking on the Translate button, a popup win-

dow appears providing a link to the Google Translate

utilities that can be used to translate a term or a phrase.

The user can additionally decide to look up a new con-

cept (e.g. sistema linf

atico) in a medical thesaurus by

using the thesaurus lookup slot that is provided by the

system in all three languages of the project (see Fig-

ure 3).

To summarize, the user can stop the video at any

time and interact with the text transcript to gather in-

formation on highlighted keywords, for instance their

translation, or he/she can extract information on arbi-

trary medical terms from Wikipedia or other special-

ized databases. The innovative aspect of the compo-

nent proposed in this paper is that it provides a com-

prehensive platform for learning and for accessing

open domain resources that are directly linked with

the learning component, thereby allowing learners to

freely manage the language acquisition process.

3 CONCLUSIONS

In this paper, we described an interactive video train-

ing framework for teaching a second language to

medical professionals and healthcare givers. The

component presented here integrates self-learning fa-

cilities with external open domain resources, e.g.

thesauri and specialized dictionaries, and provides a

comprehensive infrastructure supplying medical pro-

fessionals with both a situated learning environment

and open domain look-up and information retrieval

facilities in three languages. We performed a prelim-

inary evaluation of the system mainly based on qual-

itative criteria and asked users to judge our system

in terms of easiness of GUI navigation, usefulness of

linked knowledge and enhancement of active learning

CSEDU2013-5thInternationalConferenceonComputerSupportedEducation

126

and engagement. We received encouraging feedback

and plan a broader evaluation of our system in hos-

pitals and universities. In future work, we want to

extend the range of linked resources to include more

specialized terminologies, e.g. MeSH. Further, we

plan to develop a self-assessment component and to

associate automatically generated training activities

to the existent videos, such as ﬁll-in-the-blank exer-

cises based on the same dialogue used to generate the

videos.

ACKNOWLEDGEMENTS

The research reported in this paper has been carried

out in the framework of the TELL-ME project (Ref.

517937-LLP- 2011-UK-LEONARDO-LMP), which

has been funded with support of the Lifelong Learn-

ing Program of the European Commission. This pub-

lication reﬂects the views only of the author, and

the Commission cannot be held responsible for any

use which may be made of the information contained

therein.

Many thanks to Igor A.L. da Silva for his support

in the realization of the video material and to Paul

Schmidt for useful comments.

REFERENCES

Amoia, M., Gardent, C., and Perez-Beltrachini, L. (2011).

Learning a Second Language with a Videogame. In

Proceedings of the International Conference on In-

formation and Communication Technologies for Lan-

guage Learning, Florence, Italy.

Bouillon, P., Flores, G., Georgescul, M., Halimi, S.,

Hockey, B. A., Isahara, H., Kanzaki, K., Nakao, Y.,

Rayner, M., Santaholma, M., Starlander, M., and

Tsourakis, N. (2008). Many-to-many multilingual

medical speech translation on a pda. In Proceedings of

The Eighth Conference of the Association for Machine

Translation in the Americas, Waikiki, Hawaii.

Brown, J., Collins, A., and Duguid, S. (1989). Situated

cognition and the culture of learning. Educational Re-

searcher, 18(1):32–42.

Grosjean, J., Merabti, T., Griffon, N., Dahamna, B., and

Darmoni, S. (2012). Teaching medicine with a termi-

nology/ontology portal. Stud Health Technol Inform.,

180:949–53.

Kreuzthaler, M., Bloice, M., Simonic, K.-M., and

Holzinger, A. (2011). Navigating through very large

sets of medical records: An information retrieval

evaluation architecture for non-standardized text. In

USAB’11, pages 455–470.

Lave, J. and Wenger, E. (1990). Situated Learning: Legit-

imate Periperal Participation. Cambridge University

Press, Cambridge, UK.

Leroy, G., Helmreich, S., and Cowie, J. R. (2010). The ef-

fects of linguistic features and evaluation perspective

on perceived difﬁculty of medical text. In 43rd Hawaii

International Conference on System Sciences, Hawaii.

Nelson, S. J. (2009). Making terminologies that work: The

example of mesh. In Proceedings of the 10th Interna-

tional Symposium on Pervasive Systems, Algorithms,

and Networks (ISPAN 2009), pages 380–384, Kaohsi-

ung, Taiwan.

eol, A., Shooshan, S., Mork, J., and Aronson, A.

(2007). Fine-Grained Indexing of the Biomedical Lit-

erature: MeSH Subheading Attachment for a MED-

LINE Indexing Tool. In Proceedings of the AMIA

Symposium. American Medical Informatics Associa-

tion.

Zhang, D., Zhou, L., Briggs, R., and Nunamaker, J. (2006).

Instructional video in e-learning: Assessing the im-

pact of interactive video on learning effectiveness. Inf.

Manage, 43(1):15–27.

InteractiveVideo-trainingforMedicalProfessionals

127