A MOBILE COLLABORATION AND DECISION SUPPORT SYSTEM

FOR THE MEDICAL EMERGENCY DEPARTMENT

Dominique Brodbeck

, Markus Degen

and Maximilian Reiß

University of Applied Sciences Northwestern Switzerland, Olten, Switzerland

Technical University of Munich, Munich, Germany

Keywords:

Mobile technologies for healthcare applications, Medical informatics.

Abstract:

A hospital emergency department is a complex work environment, where the availability of the right informa-

tion at the right time is crucial for efﬁcient and safe operation. The current technology in use for communica-

tion and information management is mostly based on telephones and stationary personal computers. Modern

smartphones with their computational power, voice, image, and video capabilities have the potential to play

a signiﬁcant role in improving the ﬂow of information in the emergency department. We developed a system

that explicitly supports the work ﬂows of an emergency department. In addition to mobile access to patient

data and notiﬁcations about the availability of diagnostic ﬁndings, it provides the possibility to supply media

captured on-site to the patient record, and directly supports the consultation process.

1 INTRODUCTION

The treatment of emergency patients is an important

function of hospitals. The operation of an emergency

department is different from that of other departments

in several ways. Events are not foreseeable and ac-

tions are often time-critical. Furthermore, doctors

typically treat emergency patients in a parallel way.

After the anamnesis, diagnostic test (x-ray, lab test,

MRI, etc.) are administered. Patients need to wait un-

til the results are ready and then a doctor sees them

again to decide on treatments. Meanwhile, the doctor

is caring for other patients. During this process, the

doctors on duty in the emergency room need to con-

sult with experts from specialized disciplines (e.g. or-

thopedics). The experts are typically difﬁcult to reach,

not present on site or in a mobile situation. Another

aspect of hospitals is that doctors’ competence devel-

opment is not based on formal acts of education, but

rather embedded in daily practice and brought on by

cooperation and communication with colleagues and

experts. Because of these characteristics, the efﬁcient

operation of an emergency department is highly de-

pendent on the availability of the right information at

the right time, and the availability of multimedia com-

munication channels.

At the University Hospital Basel, we encountered

several issues where the ﬂow of information is not op-

timal. The problem with standard diagnostics is that

processes are still paper-based or server-centric. Doc-

tors don’t receive information about the results - or

even about the availability of the results - at the time

it is available at the place where they currently are, but

instead have to remember and go and look for it. The

problem with expert consultations is that experts are

often not available (especially at off-hours) or remote.

The consultations however are carried out by phone or

paper-based channels (fax) with no support for media

such as images or video. All these processes are not

integrated well, and the generated knowledge can not

be captured or managed to support doctors’ learning.

We presume that the situation is similar in other

hospitals as well. We therefore set out to conceptu-

alize, develop, and evaluate a distributed, mobile de-

cision support system for the emergency department.

The system has the following goals:

• Improve the ﬂow of information by pushing no-

tiﬁcations about new information to doctors and

making it available in mobile contexts

• Support the consultation process by providing

workﬂow-aware and multimedia enriched com-

munication

• Provide integrated access to the system on differ-

ent mobile and stationary platforms

In the following sections we describe related

work, the techniques we used to develop the concept,

and the system that we built based on this concept.

237

Brodbeck D., Degen M. and Reiss M..

A MOBILE COLLABORATION AND DECISION SUPPORT SYSTEM FOR THE MEDICAL EMERGENCY DEPARTMENT.

DOI: 10.5220/0003780102370242

In Proceedings of the International Conference on Health Informatics (HEALTHINF-2012), pages 237-242

ISBN: 978-989-8425-88-1

 2012 SCITEPRESS (Science and Technology Publications, Lda.)

2 BACKGROUND

The main technology for communication in a typical

hospital today is still the voice-based telephone.

With smartphones (e.g., iPhones) becoming more

and more widespread among medical doctors work-

ing in hospitals, they not only begin to replace the

”phone-only” devices for every-day work use, but

with the possibility to tailor them with specialized ap-

plications, they are now also becoming an information

access tool.

There are a number of medical applications avail-

able for the iPhone today. They range from electronic

versions of compendia (PSPsmartsoft, 2011), over

mobile viewers for DICOM images (Pixmeo SARL,

2011), to specialized software packages for emer-

gency doctors (Pepid, 2011).

Mor

an et al. showed (Mor

an et al., 2007) that a

typical hospital worker is using more time to man-

age information than on direct patient care. Also their

work showed, that there is a constant need for access

to relevant information in a heterogenous device en-

vironment.

Several publications researched the use of Per-

sonal Digital Assistants (PDAs) and smartphones in

hospitals (Gamble, 2010), (Fischer et al., 2003) and

identiﬁed a big potential for their use in clinical daily

work. The work of (Carroll et al., 2002) discusses

some important issues that can arise when using PDA

devices to access patient records and inputing data,

primarily due to the physical limitations of the devices

(e.g. small screen, input of long texts is tedious).

Banitsas et al. showed in (Banitsas et al., 2006)

and (Banitsas et al., 2004) the potential of the use

of mobile video conferencing devices in the process

of medical consultations. They also showed, that

the current available speed of wireless transmission

networks is sufﬁcient for most tele-consultation pro-

cesses.

Two major trends seem to emerge in the mobile

device market:

• With the smartphones adapting more and more of

the traditional PDA (Personal Digital Assistant)

functionality, they will replace the PDA-only de-

vices eventually.

• Because telecom providers offer products that al-

low the use of mobile phones on a campus set-

ting for unlimited cost-free internal calls, they will

take over the role of the traditional cordless phone

systems

The system described in (Hameed et al., 2008)

is somewhat similar to the one presented here, with

one major difference: Instead of focusing on plat-

form independence and using web technologies, our

approach is to make best use of the native user inter-

face elements available on the client target platforms.

This approach allows to create an optimal user expe-

rience on every platform.

(Choi et al., 2006) proposed a mobile information

system that is integrated in the clinical IT environ-

ment, and uses the HL7 message stream to generate

notiﬁcations (e.g., when new lab results for a given

patient are available).

3 DEVELOPING THE CONCEPT

In order to account for the transformative potential

of the system and the wide range of stakeholders in-

volved, we chose a decidedly user-centered approach

to developing the concept and building the system.

Contextual observations and inquiries were per-

formed by spending several days in the emergency

department, following doctors as they performed their

duties. This resulted in a good understanding of the

existing user tasks and workﬂows, and formed a ba-

sis for the development of scenarios. We developed

both text and video scenarios that show a series of en-

visioned use cases for the new system (Pimmer et al.,

2011).

The scenarios were presented to the doctors and

their feedback captured in semi-structured interviews.

Analysis of the interviews showed (Pimmer et al.,

2011) that features such as

• mobile access to patient data and ﬁndings,

• notiﬁcations about the availability of diagnostic

ﬁndings,

• possibility to supply media captured on-site to the

patient record,

• possibility to ”bookmark” cases with associated

data for later use in different contexts,

• and dedicated support of the consultation work-

ﬂow

were considered the most useful by the doctors. Other

features such as three-way video conferencing, white-

board functionalities, or video streaming were as-

signed less potential.

From these insights we were able to extract a list

of prioritized requirements. For the implementation

of these requirements we chose an agile development

process with four-week iteration cycles. The mobile

and desktop user interfaces were conceptualized and

veriﬁed using paper prototypes. The prototypes were

evaluated with doctors on duty in the emergency de-

partment using think-aloud protocols.

HEALTHINF 2012 - International Conference on Health Informatics

238

3.1 Consultation Process

Consultation is one of the key processes to be sup-

ported by the system. Emergency departments are of-

ten staffed with junior doctors. If they encounter a

situation outside of their medical expertise, they need

to consult with a senior doctor or a medical expert

from a specialized discipline. At the hospital that we

observed, this consultation process is currently per-

formed through phone calls and paper forms. Obser-

vations about injuries or symptoms need to be com-

municated with written or spoken words.

We analyzed the consultation process and identi-

ﬁed the four different states shown in Figure 1.

!"#$%&"'"(

)**"+&"(

,"($*-./(0*&0'/'"12"%&%

*03%2.&-&$03

4$3$%5"(

603%2.&-&$03

(0*27"3&-&$03

8$."(

,"($*-./*03%2.&-3&

'"-(%/12"%&$03/

,"($*-./*03%2.&-3&

-**"+&%/*03%2.&-&$03

9""3

,"($*-./*03%2.&-3&

&'-3%8"'%/*03%2.&-&$03/&0/

0&5"'/*03%2.&-3&

,"($*-./*03%2.&-3&

*.0%"%/*03%2.&-&$03

Figure 1: The consultation process modeled as a state-

machine.

The functionality to support the consultation pro-

cess in our system was modeled after this concept,

and is fully integrated with the rest of the function-

ality. Requests for consultation can be initiated and

received directly on the smartphone. The inquiring

doctor is able to see the current state of a consultation

at all times. The medical consultant is reminded by

the system if an action is outstanding. Consultations

can be enriched by media ﬁles to facilitate the voice

communication.

4 HIGH-LEVEL ARCHITECTURE

OF THE SYSTEM

Figure 2 shows the high level architecture of the sys-

tem. The communication between the backend sys-

tem and the end-user clients was realized using a cus-

tom message protocol, based on an efﬁcient wire pro-

tocol (See Section 4.5).

One of the non-functional key requirements was,

that the system should be accessible by different

clients. We therefore deﬁned a message protocol

and implemented several different clients for the sys-

tem. In order to increase acceptance and usability,

!"#$%&

'(!&%)

*%+,$!+

'(!&%)

-&./)$0

'(!&%)

123/&%+

45.)&6

784

-94

"1:

7$.0!)2(

8%;,2.),<3)<,&

Figure 2: High-level architecture of the system, and inte-

gration into the hospital IT infrastructure. The backend sys-

tem connects to the Hospital Information System (HIS), the

Document Management System (DMS) and the Telephone

System (PBX, planned)

the client applications were developed to have the

look&feel of their respective platforms, rather than a

”one for all GUI”.

4.1 Backend

The backend provides several functions:

• Aggregation of the data of several hospital IT sys-

tems

• Bi-directional data transformation for the hospital

IT systems’ data from and to the mobile devices

• Authentication for the mobile devices via the hos-

pital IT authentication system

• Support for the discussion conference feature as

well as the enhanced consultation feature

• Sending of notiﬁcations to the clients, e.g., if new

data is available or if some action is overdue

The reason for the data aggregation of the different

IT systems of the hospital is that the mobile devices

should present one aggregated access to the patient,

in contrast to the legacy desktop approach in which

every data supplier has its own interaction user inter-

face. On a mobile device where one app is a men-

tal equivalent to a tool for a speciﬁc task, different

apps to access the patients’ data would break with the

users’ expectation. Given the two choices to aggre-

gate the data either on the mobile device or in the

backend, the system was designed to aggregate in the

backend. The reason for this decision was the heavy

load (CPU, memory and storage) that this operation

causes.

The second backend part, the bi-directional data

transformation, is needed for several reasons. There

are several data formats that the mobile device cannot

display directly, and even if it can display the format,

it might have issues with the large size of the data

A MOBILE COLLABORATION AND DECISION SUPPORT SYSTEM FOR THE MEDICAL EMERGENCY

DEPARTMENT

239

(e.g., there is a limitation on the mobile device for

how large an image can be). Another reason for the

data adaptation is, that depending on the network con-

nection (and other factors) the data might not be avail-

able fast enough on the mobile device, even though a

reduced version of this data (that can be transferred

faster) would still be sufﬁcient for the current use case

of the doctor. In the other direction, the mobile de-

vices might produce data in a format that the IT sys-

tems of the hospital do not understand. In this case

the backend would transcode the data into a format

that the hospital systems can understand.

Since the backend is the one place to monitor all

events and notiﬁcations in the system, it observes

these and can issue special notiﬁcations for a mo-

bile device of a speciﬁc doctor, e.g., ”New MRI im-

ages are available for patient X”. In case that a mo-

bile device client is not reachable, most notiﬁcations

are queued until the mobile device becomes available

again. If the notiﬁcation passes an ”importance” ﬁl-

ter, a short version of the notiﬁcation is additionally

sent to the mobile device as push notiﬁcation, via the

platform’s push notiﬁcation service.

4.2 Mobile Client on iOS (iPhone)

The primary mobile client was developed for the iOS

platform, mainly focused on the iPhone. With the

iPhone client it is possible to participate in the three

main activities that our system supports:

• It is possible to view patient record data aggre-

gated from several backend systems (document

management system, RIS, etc.),

• it is possible to actively supply diagnostic ﬁndings

(images, videos, audio recordings, etc.),

• and it is possible to participate in the medical de-

cision making process by helping with the con-

sultation workﬂow, and by providing audio con-

ferences with the possibility to share and discuss

patient data in this conference.

Figure 3(a) shows the typical listing of a doctor’s

current patients. In this view and the views reachable

from this list, the doctor can access the patient record

data, the patient master data, the previous cases with

their diagnostics and documents, and the lab record

and current consultations, as shown in Figure 3(b).

The third main activity is the discussion support.

On a shared document, the doctor can mark regions,

as shown in Figure 3(c). These markers are visible to

the other doctors participating in the discussion con-

ference. In this way, our system uses a distributed

discussion model - each doctor can supply her own

data, and if she wants to show the data to others, then

she can draw their attention to the data supplied by

her. This can be done via the audio connection.

4.3 Mobile Client on Android

Since not only iOS-based smartphones but also

Android-based ones are present in the clinical land-

scape, it was a requirement to also implement a mo-

bile client for Android systems, at least as a ”proof of

concept”.

We opted to not simply port the user interface

of the iPhone to the Android platform, but reimple-

mented the interface using the look and feel of An-

droid applications. Because the implementation lan-

guage for the desktop client was also Java, much of

the code talking to the backend system could simply

be reused in the Android client.

4.4 Desktop Client

While doctors are frequently on the move, there are

still situations where they sit near a desktop computer,

or where larger input and output hardware is more

convenient for a task (e.g. text entry), or where pa-

tient cases are discussed in front of larger audiences

(e.g. morning status report). We therefore decided to

develop a desktop client for stationary use that takes

advantage of the increased computational resources

compared to the mobile clients.

In particular, there are less screen space limita-

tions for the desktop client, and the most important

information can always be shown on the main screen,

providing an ”at a glance” overview while at the same

time allowing quick and easy access to details.

Figure 4 shows the user interface of the desktop

client. On the left side, a list with all current patients

in the emergency department is shown. More infor-

mation can be obtained by simply hovering over a pa-

tient entry. In the lower left, a spatial overview of all

beds in the emergency department is shown, together

with the current state (red: occupied, green: free).

The right side shows a patient’s details, including the

master data, data from the current as well as from pre-

vious cases, and all the information about pending and

ﬁnished consultations. All the views are tightly linked

through interactions to provide the big picture.

4.5 Communication Protocol

The communication between the clients and the back-

end is twofold:

• A set of control messages were deﬁned to syn-

chronize the application data between the clients

HEALTHINF 2012 - International Conference on Health Informatics

240

(a) Patient Overview (b) Patient Case Data (c) Conference discussion

Figure 3: (a) Screenshot of the iPhone UI showing the patient overview (b) Screenshot of the iPhone UI showing the patient’s

data sections. The actual document group names are pulled from the hospital systems. (c) Screenshot of the iPhone UI

showing a media discussion in a conference.

Figure 4: User interface of the desktop client.

and the backend. These messages were transmit-

ted via socket connections, and Google Protocol

Buffers (Google Protocol Buffers Team, 2011)

were used for efﬁcient bit stufﬁng.

• Since there would be little gain to transfer media

data (images, audio, video etc.) via the bit stufﬁng

channel, it is retrieved via an HTTPS connection

by the clients. Sending JPEGs etc. encoded with

Protocol Buffers would only marginally improve

the message size with some conversion overhead.

4.6 Connection to the Hospital IT

Since our system relies on live data, it has to be con-

nected to the hospital’s IT systems. The connections

are made via ”adapters” in the backend that are in-

terchangeable, should the system be used in another

hospital with a different IT infrastructure.

4.6.1 Hospital Information System (HIS)

The ”information backbone” of a typical hospital con-

sists of an HL7 message stream connecting all subsys-

tems, and parsing this stream is the best information

source, if live data of a patient is needed. We wanted

to know the current location (i.e., bed) of every pa-

tient, in order to provide a graphical layout of the

load of the emergency department. In our case how-

ever, this information is not communicated via HL7

but handled within the HIS directly. We therefore had

to connect to the HIS instead of listening to the HL7

stream, using a database connection with read-only

access, and polling it regularly by the backend to de-

tect changes in patient location.

4.6.2 Document Management System (DMS)

A major part of the clients handles the visualization of

documents stored in the Document Management Sys-

tem (DMS). The DMS provides SOAP Web services

that can be used to upload documents and to search

and retrieve documents stored in the DMS. The doc-

uments retrieved from the DMS are stored in a smart

cache in the backend for some time. Also, lower size

versions of the images for the mobile clients are gen-

erated in the backend.

5 EVALUATION

The system will be evaluated in a trial in early 2012

in the emergency department of the University Hos-

pital Basel. It is planned to involve several medi-

A MOBILE COLLABORATION AND DECISION SUPPORT SYSTEM FOR THE MEDICAL EMERGENCY

DEPARTMENT

241

cal doctors that will be provided with smartphones

running the application. The participants will follow

their daily routine, treating patients in the emergency

department, and use our system to record and store

media ﬁles (photos, videos, audio ﬁles) and issue re-

quests for consultations via the functions of our sys-

tem.

Also several senior doctors and medical experts

will be equipped with the desktop version of the client

on their workstation. This ensures that they can be

reached by requests for consultation.

All involved medical doctors will be asked to par-

ticipate in a structured interview after the trial. The

data from the interviews will allow us to evaluate the

usability of the system as a whole, and of the speciﬁc

client user interfaces in particular.

6 CONCLUSIONS AND FUTURE

WORK

We have built a mobile collaboration and decision

support system that can enhance and speed up the pro-

cesses in an emergency department of a hospital.

The features of the system were deﬁned and devel-

oped in close collaboration with the end users (med-

ical doctors) and the ﬁrst feedbacks were encourag-

ing. The system will be evaluated in a trial in early

2012 in the University Hospital Basel. The data gath-

ered from log ﬁles during the trial and the ﬁndings of

the interviews conducted after the trial with involved

medical doctors, will allow us to examine and judge

the usefulness of the system.

ACKNOWLEDGEMENTS

This work was supported by funding from the

Swiss Innovation Promotion Agency CTI. The au-

thors would like to thank the University Hospital in

Basel, the Academy for medical training and simula-

tion AMTS and Agfa Healthcare for their support.

REFERENCES

Banitsas, K., Georgiadis, P., Tachakra, S., and Cavouras, D.

(2004). Using handheld devices for real-time wireless

teleconsultation. Conf Proc IEEE Eng Med Biol Soc,

Banitsas, K., Georgiadis, P., Tachakra, S., and Cavouras, D.

(2006). Mobile consultant: combining total mobility

with constant access. In Engineering in Medicine and

Biology Society, 2006. EMBS ’06. 28th Annual Inter-

national Conference of the IEEE, pages 5248 –5251.

Carroll, A. E., Saluja, S., and Tarczy-Hornoch, P. (2002).

The implementation of a Personal Digital Assistant

(PDA) based patient record and charting system:

lessons learned. Proceedings / AMIA ... Annual Sym-

posium. AMIA Symposium, pages 111–5.

Choi, J., Yoo, S., Park, H., and Chun, J. (2006). Mo-

bileMed: a PDA-based mobile clinical information

system. IEEE transactions on information technology

in biomedicine : a publication of the IEEE Engineer-

ing in Medicine and Biology Society, 10(3):627–35.

Fischer, S., Stewart, T. E., Mehta, S., Wax, R., and Lapin-

sky, S. E. (2003). Handheld computing in medicine.

Journal of the American Medical Informatics Associ-

ation, 10(2):139–149.

Gamble, K. H. (2010). Wireless Tech Trends 2010. Trend:

smartphones. Healthcare informatics the business

magazine for information and communication sys-

tems, 27(2):24, 26–27.

Google Protocol Buffers Team (2011). Google

Protocol Buffers Project Page. Avail-

able online at http://code.google.com/intl/de-

DE/apis/protocolbuffers/ Visited on September,

6th, 2011.

Hameed, S. A., Hassa, A., Shabna, S., Miho, V., and Khal-

ifa, O. (2008). An efﬁcient emergency, healthcare, and

medical information system. International Journal of

Biometrics and Bioinformatics (IJBB), pages 1–9.

Mor

an, E. B., Tentori, M., Gonz

alez, V. M., Favela, J., and

Mart

ınez-Garcia, A. I. (2007). Mobility in hospital

work: towards a pervasive computing hospital envi-

ronment. International Journal of Electronic Health-

care, 3(1):72–89.

Pepid (2011). iPhone Emergency Physicians SW. Available

online at http://www.pepid.com/products/iPhone-

Emergency-Physicians-Software.asp Visited on

September, 6th, 2011.

Pimmer, C. P., Pachler, N., and Genewein, U. (2011).

The potential of smartphones to mediate intra-hospital

communication and learning practices of doctors. pre-

liminary results from a scenario-based study. Pro-

ceeding oft he Mobile learning conference: Cross-

ing boundaries in convergent environments, Bremen,

pages 151–4.

Pixmeo SARL (2011). Osirix HD Product Page. Available

online at http://itunes.apple.com/app/id419227089

Visited on September, 7th, 2011.

PSPsmartsoft (2011). iKomp Product Page. Available

online at http://www.appannie.com/ikomp/ Visited on

September, 7th, 2011.

HEALTHINF 2012 - International Conference on Health Informatics

242