Process-Oriented e-Learning System for Training Healthcare

Professionals on Big Data Usage

Vassiliki Karabetsou and Flora Malamateniou

Department of Digital Systems, University of Piraeus, 80 Karaoli & Dimitriou Str., 18534 Piraeus, Greece

Keywords: Big Data, Healthcare, Process-Oriented, Workflow, Learning Analytics, e-Learning.

Abstract: Big data technology promises to transform the way in which medical care is delivered and help the

healthcare industry to address problems related to variability in healthcare quality and escalating healthcare

costs. However, integrating Big Data use in healthcare professionals’ daily practice seems to be a

challenging task as they are accustomed to making treatment decisions independently, using their own

clinical judgement, rather than relying on protocols based on big data. Taking medical decisions based on

Big Data - combined with physicians’ valuable clinical knowledge and experience - can lead them to safer

and more accurate diagnosis and focused treatments. In order to support this transformation in medical

practice healthcare professionals (e.g. physicians, nurses, pharmacists) will need to be trained in the

collection, integration and analysis of large data sets. To this end, this paper presents a process-oriented e-

learning system which aims at making healthcare professionals understand how to use big data tools and

giving them the necessary skills to improve operations. The system uses workflow technology and Learning

Analytics which has been specifically planned for learners’ custom needs.

1 INTRODUCTION

Big Data in healthcare have given great propulsion

to research and to medical care as well. They have

contributed to predicting diseases by studying genes

while their use leads physicians to quickly targeted

and effective diagnosis and treatments (Groves et al.,

2013). They can stimulate innovation by identifying

new treatments and approaches for the provision of

medical care using data coming from clinical trials.

They also promise to cut down the cost of the health

system through the decrease of medical errors.

Correct diagnosis will reduce unnecessary exams

and limit wrong treatment.

However, healthcare organizations still have

difficulties in fully taking advantage of big data's

capabilities. This is because of the fact that they

don't know where to «start from». Actually,

everything begins from data. The use of big data in

healthcare requires a set of many heterogeneous

sums of data. The bigger the sum, the more the

chances of finding a correct answer on the questions

about medical practice. The use of the World Wide

Web has made possible the collection of clinical

data and the observation of epidemics on a global

scale. The ever-increasing number of data will

gradually lead to finding useful data, in better

decisions and more effective attempts (Murdoch and

Detsky, 2013).

Unfortunately, physicians don't have the totality

of data related to patient healthcare (e.g. clinical

data, electronic patient records’ data, sensor data,

emergency care data) which makes up Big Data

(Raghupathi and Raghupathi, 2014) concentrated

somewhere in order to access all of them. In

addition, physicians are accustomed to making

decisions based only on their judgment and not on

data. The result is that one of five medical decisions

is wrong or imperfect and the third reason for most

fatalities is medical errors that occur due to unlucky

medical decisions (Groves et al., 2013; Sun and

Reddy, 2013; IBM, 2015). However, taking the fact

that medical knowledge doubles every three years

into consideration, it will be impossible for

healthcare professionals to read all that information

to be informed (Kohn, 2012).

Some organizations and academic medical

centers have already started estimating capabilities

that big data can offer in clinical practice and

research and have adopted the use of big data

analytics systems. Unfortunately, professors in

medical schools have not been trained in order to

Karabetsou, V. and Malamateniou, F.

Process-Oriented e-Learning System for Training Healthcare Professionals on Big Data Usage.

DOI: 10.5220/0005828205230528

In Proceedings of the 9th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2016) - Volume 5: HEALTHINF, pages 523-528

ISBN: 978-989-758-170-0

523

manage and analyze data. Training in this sector is

considered to require acquiring technological skills

beyond a doctor's training capabilities. As a

consequence, physicians cannot make use of big

data analytics software and take advantage of their

use (Moskowitz et al., 2015; Waxer, 2014).

Process oriented e-learning systems have

recently received much attention, as they utilize

workflow technology, to support highly structured

teaching/learning processes. Furthermore, they allow

not only to design and evaluate the effectiveness of

processes, but also to easily redesign and improve

them. Especially, when training healthcare

professionals, redesigning of processes is required

because of the constantly increase of medical

knowledge (Lenz and Reichert, 2006).

In this paper a process-oriented e-learning

system for training physicians on big data usage is

presented. In particular, the prototype system uses

workflow technology and aims at training physicians

on how to search, gather, visualize and store medical

data from their daily practice. Furthermore, Learning

Analytics are incorporated into the system in order,

not only for physicians to monitor their progress, but

for the whole learning process to be evaluated.

2 BACKGROUND

The need to support teaching/learning processes,

rather than simple tasks, in recent times has resulted

in a new type of process-oriented, educational

technology. By using workflow-based e-learning

systems, it is possible to introduce flexible start and

finish times for each task (unit of content) based on

user’s needs and progress. Technically, this is made

possible by the coordination mechanism used by

workflow technology. Furthermore, it is possible to

increase flexibility of the curriculum. This means the

introduction of flexible learning pathways so

students can progress through the content in a

variety of ways based on their needs and preferences

(Marjanovic, 2007).

The use of digital technology in learning

processes though, leads to the collection of more and

more data. Data create awe with their breadth and

heterogeneity but they can contribute to education

development with the help of Learning Analytics.

Learning Analytics can collect and process data with

volume, variety and velocity, they can process “Big

Data”. The question is which educational data

should be gathered and analyzed? Who should

choose them? Which are the suitable criteria for this

purpose? Ellaway et al., (2014) believe that

professional medical education should use analytics’

techniques which are proper for the temperaments

and special needs of the health sector. Another

question that someone should take in consideration

is if learners must be aware of the whole procedure

of gathering and analyzing educational data and if

this awareness can affect the training outcome.

Finally, it is important to examine how analysis’

results should be used in order to identify best

practices and to provide educational process success.

As far as physicians’ special characteristics are

concerned, someone could end up with the following

conclusions:

 Despite technology’s extensive growth and the

fact that almost everyone owns modern devices,

physicians are not quite familiar with all these.

As a result, they still prefer studying printed

documents (medical articles and books) to get

informed. They also prefer storing data on paper

(Raghupathi and Raghupathi, 2014).

 Their demanding schedules don’t allow them to

get informed and to digitize the printed data they

own. Moreover, the lack of spare time makes it

impossible for them to attend educational

programmes that take place in traditional classes.

 During their daily clinical practice a lot of

questions come up. These questions are about

medicine of course, not only older but also new

medical knowledge, and sometimes they are

relative to other specialties. In all of these cases,

physicians turn to their colleagues to ask for a

piece of advice. What is more, there are asked to

offer their services to e-patients, who have

already searched for information online and have

a lot of questions to ask.

 They use the internet to a limited extent, because

they don’t feel quite safe when they use it and

they don’t actually trust it. For this reason,

although they use to sign in social networks, they

don’t use to search for medical information.

 A lot of physicians would choose distance

learning to get educated, but they can’t always

find the proper e-learning program made for this

purpose (Cortelyou-Ward et al., 2013; Ellaway et

al., 2014).

As regards to the way of their training, they need to:

 Have support, guidance, confirmation and

feedback during the whole process

 Have flexibility of space and time

 Attend a program that allows them to follow

their own learning pace

 Make their own selections in order to draw their

own learning path

HEALTHINF 2016 - 9th International Conference on Health Informatics

524

 Get educated quickly, easily and economically.

As regards to the subject of their training, they need

to:

 Get more familiar with new technology

 Use internet safely

 Have easy and quick access to the data they need

in order to make decisions regarding medical

diagnosis and treatment

 Get informed of Big Data usage benefits

 Use Big Data in their everyday practice, in order

to improve medical care provided to patients

 Meet healthcare system’s needs for reducing

medical errors, unnecessary tests and cost

 Get informed about changes and improvements

in medical field

 Share best practises with the medical

community.

3 RELATED WORK

A lot of e-learning systems for training healthcare

professional have been implemented. Some of them

are based on Workflow technology. For example,

Puustjärvi and Puustjärvi’s (2010) proposal for

physicians distance learning. They suggest a solution

that provides physicians’ daily duties coordination

and the necessary learning material as well. Another

example is the complete framework that Chodos et

al., (2009) suggested for healthcare professionals

training by video and virtual world simulations.

University of California, Davis (UC Davis, 2015)

promises that applying clinical Analytics can

“improve health care, manage risks and improve

patient outcomes”. UC Davis Extension’s online

Healthcare Analytics Certificate Program designed

for working clinicians and IT professionals aims to

help them acquire comprehensive understanding of

the use and implementation of healthcare analytics.

A lot of Universities (University of Central

Florida, Rio Salado Community College, Northern

Arizona University, Purdue University, Ball State

University, University of Michigan, University of

Maryland, Graduate School of Medicine, University

of Wollongong) use Learning Analytics in order to

monitor and engage their students, provide them

support when they need it and enhance the learning

experience. In healthcare education though “learning

analytics is in its infancy” (Dietz-Uhler and Hurn,

2013).

Elsevier, a world-leading provider of information

solutions, which provides web-based courses for

nursing and health professions students, will use

Knewton’s infrastructure “to power personalized

digital solutions”. Elsevier offers the educational

content when Knewton provides analytics and

technology (Knewton, 2014).

Recognizing the significance of gaming and

learning analytics, the Stanford School of Medicine

developed two educational games, Septris and

SICKO, for medical students. The games were very

well received and this proved that “the application

of gaming and the collection of learning analytics

data offer many potential opportunities in

education”(Tsui et al., 2014).

4 THE LEARNING PROCESS

MODEL

A high-level model of the learning process

considered, which is based on constructivism

learning theory, is shown in Figure 1. The learning

process has been specifically designed for healthcare

professionals to enable them realize the benefits

from using healthcare analytics at their workplace.

The model consists of the following activities:

 Introduction. Learners watch a video

presentation with ultimate goal to help them

digitize printed data, collect medical data online,

organize, store and visualize them and to

evaluate data usage in decision-making process

according to diagnosis and treatment.

 Assignment Subject Selection. Learners are

asked to choose among ten different subjects.

They can also suggest their own subject. The free

selection is necessary, because it will give them

the possibility to collect data that they really

need and can use in their medical practice. The

fact that they find the subject interesting, will

motivate them and get them engaged. Teacher is

informed about their choice in order to be able to

support them when needed.

 File Creation. Learners create the files in which

they will store their data. They will store all the

data they will collect about their subject in these

files. The main file will be named after the

subject and a number of sub-files will be also

created. In the sub-files information about the

subject’s definition, symptoms, aetiology,

diagnosis, prevention, treatment, medications

and personal notes will be stored. This is a very

simple way for anyone to organize data that

doesn’t demand any programming skills.

Process-Oriented e-Learning System for Training Healthcare Professionals on Big Data Usage

525

Figure 1. High-level BPM Studio model of the learning process.

 1st Assignment Assessment. Learners are

informed about their performance. If they get

grade under 50%, they will watch a simulation

video in order to correct their assignment.

Otherwise, they are addressed to the next

activity.

 Data Collection. Learners gather all the data

they could find and then store them in the files

writing down useful metadata as the link, the

date they found them and a short description of

the content. Finding, collecting, organizing data

and creating metadata about them, is a simple

stage of analysis that can make data ready to be

used and easily updated. Data collection

constitutes a process activity consisting of three

sub-activities:

o Printed Material Management and

Digitation. Learners check first if the

material (e.g. articles, images, diagrams,

video, medical books) already exists. If it

doesn’t exist, they will digitize their printed

material. In any case, they should store data.

o Guided Research. Learners collect data

using suggested links of medical blogs and

portals.

o Free Research Online. Learners search for

data using suggested combinations of words

and directions to use internet in a safe way.

Learners are informed about the way their

assignments will be evaluated. The

assessment depends on the volume of the data

that they collect, the relevance of the data to

the subject and the way that these data are

organized.

 Assignments’ Assessment. Teacher assesses

each assignment separately. The average grade is

automatically calculated and presented to

him/her. If the average grade is less than 50%,

teacher makes suggestions in order to help

learners to complete their collection and submit

it again. Otherwise, they are driven to the next

activity.

 Data Collection’s Completion and Evaluation.

Learners are informed not only about the way

they could use their data collection, but also

about the way world medical community uses

Big Data. They are informed about the medical

databases that are created and furthermore about

the Big Data analytics systems that are adopted

by physicians, their advantages and their

conditions, way and scope of use. They could

collect and use Big Data alone or they could

adopt a system that provides them all of the

useful data they need in order to answer

complicated questions.

 Learning Process’ Evaluation. Learners are

asked to evaluate the learning process by

completing an evaluation rubric. The data that

will come up from this procedure will enable the

quality control of the process and will contribute

to it’s improvement through redesign.

5 IMPLEMENTATION ISSUES

The proposed system was implemented with

Business Process Management Oracle BPM Studio

10g Release 3 (10.3.0). BPM Studio was used both

for designing and implementing the learning process

model and for implementing learning analytics.

As shown in Figure 2, learners are able to choose

the subject of their assignment. Ten different

subjects are proposed to them. They can choose one

of them or they can submit their own subject. In

addition, it offers educational designers the ability to

review the workflow, when needed. For example, if

HEALTHINF 2016 - 9th International Conference on Health Informatics

526

they notice that learners don’t submit an assignment

on time, they suspect that this assignment is too

difficult and they can modify it properly.

Figure 2: Assignment subject selection.

Learning Analytics is used to keep track of the

learners’ progress and intervene in cases when it is

needed. As shown in Table 1, Learning Analytics

are being used in two different ways, “Micro

interventions” during the whole learning process and

educational data gathering for future use. As “Micro

interventions” are concerned, Chris Brooks’ opinion

was adopted which suggests giving support to

learners at the right moment (Diaz and Brown,

2012). The proposed model includes a dashboard

which presents the elapsed time between activities

and another one where trainer can watch the average

engagement time for participants to each activity.

There are also two charts for students where they

monitor the learning process progress. At the same

time, learners’ privacy is being protected, because

information is displayed anonymously. Teacher’s

intervention is immediate in order for learners not to

abandon the process. Also, frequent assessment is

applied for the teacher to will be able to detect

problems and intervene immediately.

During the learning process lots of educational

data are collected that are used to extract

valuable

conclusions. These data include learners’

assignments, dashboards references, assignments

comments and grades, learning process evaluation.

The material that is gathered consists of texts,

images, graphs, questionnaires and e-mail. Data

volume and velocity depend on learners’ number.

6 CONCLUDING REMARKS

While the use of big data shows exciting promise for

improving health outcomes and controlling costs,

healthcare professionals should increase their

reliance on big data technology for triage, diagnosis

and decision-making. To this end, training

Table 1: Learning Analytics Implementation.

Goal Actions

Prediction of

learners’

progress

o 1st Assignment

o Learning pace

Early detection

of learners’

difficulties

o Date of assignments’ delivery

o Elapsed time between

assignments (Dashboard)

o Automated reminder of

deadlines and detection of

problems

Intervention

when needed

o Support learners providing

extra material individually

o Evaluate assignments using

comments to guide and support

o Support and encourage learners

through e-mails

Motivating

learners

o Graphs presenting active work

items in one activity and in the

whole process

o Inform learners about their

assignments’ evaluation criteria

Helping and

improving

learning

process

o Dashboard that visualize the

average time that learners spend

in each activity

o Teacher’s inform every time

that a learner uploads an

assignment with e-mail

o Provided support evaluation

o Completion of an assessment

rubric

healthcare professionals on big data usage

constitutes a rather challenging task. In this paper a

process-oriented e-learning system was implemented

with the objective to train physicians on big data

usage. The system uses Learning analytics for

monitoring learners’ progress and for, improving the

learning process.

With today’s huge patient loads, treating patients

sooner saves both lives and healthcare costs. To this

end, the proposed system aims at helping healthcare

professionals to deliver much more precise and

personalized care by quickly making informed

medical decisions.

System evaluation is a task to be undertaken in

the near future aiming at determining the system

usability. Thus, its potential weaknesses may be

revealed suggesting alterations in the system design.

REFERENCES

Chodos D., Stroulia E. and Naeimi P., 2009. An integrated

framework for simulation-based training on video and

in a virtual world. [Online] Pedagogy, Education and

Process-Oriented e-Learning System for Training Healthcare Professionals on Big Data Usage

527

Innovation in 3-D Virtual Worlds: Journal of Virtual

Worlds Research, Vol.2. No.1. Available:

https://journals.tdl.org/ [4 Jan 2015]

Cortelyou-Ward K., Noblin A. and Kahlon S., 2013.

Competencies for Global Health Informatics

Education: Leveraging the US Experience. Ahima,

online research Journal. Available:

http://eduperspectives.ahima.org/ [2 Oct 2014]

Diaz V. and Brown M., 2012. Learning Analytics.

[Online] A Report on the ELI Focus Session.

Available: https://net.educause.edu/ [9 Dec 2014]

Dietz-Uhler B. and Hurn J., 2013. Using Learning

Analytics to Predict (and Improve) Student Success: A

Faculty Perspective. Journal of Interactive Online

Learning, Vol. 12, N. 1. Spring 2013.

Ellaway R., Pusic M., Galbraith R. and Cameron T., 2014.

Developing the role of big data and analytics in health

professional education. Pubmed, 36(3):216-22. doi:

10.3109/0142159X.2014.874553.

Groves P., Kayyali B., Knott D. and Van Kuiken S., 2013.

The ‘big data’ revolution in healthcare. [Online]

Available: http://www.studymode.com/ [10 Aug 2014]

IBM, 2015. The computing system that won “Jeopardy!”

is helping doctors fight cancer. [Online]. Available:

http://www.businessinsider.com/ [28 Feb 2015]

Knewton, 2014. Elsevier To Collaborate with Knewton to

Build Adaptive Learning Products for Nursing, Health

Professions, and Medical Students Globally. [Online].

Available: https://www.knewton.com/ [9 Sep 2015]

Kohn M, 2012. Putting IBM Watson to work in

healthcare. [Online] Available: http://

www.tripleringtech.com/MedTechPresentations/Kohn

_TRT.pdf [6 Jan 2015]

Lenz R. and Reichert M., 2006. IT support for healthcare

processes – premises, challenges, perspectives. Data

& Knowledge Engineering 61 (2007) 39–58.

doi:10.1016/j.datak.2006.04.007.

Marjanovic, O., 2007. Using process-oriented, sequencing

educational technologies: Some important pedagogical

issues. Computers in Human Behavior (Impact

Factor: 2.69). 23(6):2742-2759. doi: 10.1016/

j.chb.2006.08.005.

Moskowitz A., McSparron J., Stone D. and Celi L., 2015.

Preparing a New Generation of Clinicians for the Era

of Big Data. [Online] Available:

http://www.hmsreview.org/ [28 Jan 2015]

Murdoch T. and Detsky A., 2013 The inevitable

application of big data to health care. [Online] JAMA.

Vol 309, No. 13. Available: http://

jama.jamanetwork.com/article.aspx?articleid=1674245

[20 Aug 2014]

Puustjärvi J. and Puustjärvi L., 2010. Integrating

Medicinal Learning Objects with Daily Duties. In: El

Sabah Mohammed, ed. 2010. Ubiquitous Health and

Medical Informatics. IGI – Global. pp.367-372.

Raghupathi W. and Raghupathi V., 2014. Big data

analytics in healthcare: promise and potential. Health

Information Science and Systems. 2:3.

Sun J. and Reddy C., 2013. Big data analytics for

healthcare. [Online] Available: https://www.siam.org/

meetings/sdm13/sun.pdf [5 Nov 2015]

Tsui J., Low J. and Shieh L., 2014. Septris and SICKO:

Implementing and Using Learning Analytics and

Gamification in Medical Education. [Online].

Available: https://net.educause.edu/ir/library/pdf/

ELIB1401.pdf [5 Nov 2015]

University of California, Davis, 2015. Improving Health

Care Through Analytics. [Online] Available:

https://extension.ucdavis.edu/certificate-program/

healthcare-analytics-certificate-program. [9 Sep 2015]

Waxer, C., 2014. Health Industry Turns to Big Data to

Improve Patient Care, Cut Costs. [Online] Available:

http://data-informed.com/ [28 Jan 2015]

HEALTHINF 2016 - 9th International Conference on Health Informatics

528