Physical and digital design of the BlueBiobiomonitoring system prototype, to be used in

emergency medical responseGunnar Kramp

Aarhus School ofArchitecture

Margit Kristensen, and Jacob Fr0lund Pedersen,Department ofComputer Science, University ofAarhus, Denmark

BlueBio. BlueBio is intended to monitor injured persons,Abstract-This paper presents the physical and digital design especially in the prehospital work, in both major incidents

of a wireless biomonitoring system meant to be used especially in (MI) and everyday emergency response situations.the prehospital medical emergency response. The handling of The description of BlueBio will include current status ofmany patients with a minimum of ressources at major incidentsis an immense challenge for the emergency personnel at work on t

an accident site. New technology such as the BlueBio description of the prototype will be centered around thebiomonitoring system, can help emergency personnel monitor the physical and digital design.patients and support them in making priorities of treatment and The project, in which BlueBio is developed, is called "ITtransport of patients. However, if new technology is to be support in Major Incidents." It is a subproject of two largerintroduced in such a complex and stressed situation it must relate projects: PalCom [23] and Understandable IT [33]. The mainto the palpable aspects of pervasive computing. It must be able to proje[3a U standesit[ Themancomply with the scale of the situation and still be understandable. purpe ofthtee projects suto rinvestiat requir nIt must also be able to comply with change of location and users for an open software architecture supporting user's interactionand yet still be stable, and it must comply with the shifting with computers in a pervasive computing world. Thisrequirements from the users regarding automation and user possibility to interact with pervasive computer systems iscontrol. Our design framework, in relation to the BlueBio called palpability and is described in more details below.monitoring system, approaches these challenges from theperspective of familiarity, medical assessment and personID/registration of data. We present how this informs and hasimplications for both the physical and digital design of thecurrent prototype.

Index Terms- Biomedical Computing, Biomedical Monitoring,Design Methodology, Emergency services. .1a ty t u be

D. INTRODUCTION

Jc1HIS paper describes the digital and physical design of a1-wireless and mobile biomonitoring system, called

Manuscript received June 15, 2006. This work has been funded in part bythe European Union, project 002057 "PalCom" [23]; and by ISIS Katrinebjergproject "Understandable Computing." [33]. Paper title: Physical and Digital

Denmark. (phonel: +4589360000; fax:+484261;e-al:gapJacobr Kramun Peesewishwthetahes UieSityoof Aarchitcuse, 82000 Aarhus N, Tepoeti are u yueo h atcptr

Denmark. e-mail: folund@daii.au.dk).Design Method (PD) [13], [14]. In PD user's and researchers

1-4244-1 086-X/07/$25.OO ©2007 IEEE

obtain a common knowledge about the application area. They establish the cordoning off around the emergencyTogether they identify needs for new technologies, which are area and are responsible for providing and obtaining routesdesigned and built through iterative and experimental for transportation of equipment and people. They are alsoprocesses. The current design is the outcome of in depth responsible for the registration of all people involved -investigation of the user field, including comprehensive field both injured and non-injured. Moreover they have to takestudies and series of meetings and experimental workshops care of the public, the media and the relatives.with groups of end users participating. These activities have 3) The ambulance people are responsible for obtaining andprovided the data, leading to the prototype design described in maintaining enough available ambulances for thethis paper. The used method and primary results are described transportation of injured persons to hospital(s) and also forin more details in [18], [19], and will not be described in the transportation itself.further detail in this paper. On the basis of field studies, 4) The medical professionals are responsible for handling themeetings and workshops we have obtained detailed injured persons - the initial assessment, treatment andknowledge and understanding of the emergency response transfer to hospital. This is often carried out in closeefforts and structures as such. Furthermore, we have gained an cooperation with the ambulance staff.in depth understanding of how the medical emergency B Use oftechnologies in prehospital medical emergencyresponse work is carried out in Denmark today. In the* ~~respuonsefollowing section we give a brief overview of the emergency lresponse process and structure in Denmark, followed by a 1) Medical assessmentmore detailed description of the technologies at stake in the One of the man ssues n the medical emergency response

work concerns the medical assessment of victims at theDanish medical emergency response. incident site.A. Emergency response in Denmark Medical assessment is carried out by the medical staffThe response to an emergency is in almost all cases and/or the ambulance staff (normally physicians or

initiated immediately after receiving an alarm call (through paramedics) - who examine the patient.the public well known alarm system) (calling 112). When In incident situations the primary focus is on the individualactivated, the emergency response resources on stand-by are injured person's condition regarding: If the injured person'sallocated to the incident site and assisting sites (e.g. hospitals) (A)irway is free, if the person is (B)reathing and if there isthrough the use of the country and/or region specific code of blood (C)irculation. This is called the ABC treatment concept,practice. Each emergency situation is unique in the way that and all injured persons are observed and treated with thesethe specific incident situation is assessed regarding needs for three parameters as the starting point [10].resources, initially by the receiver of the call in the alarm At the hospitals or at the general practitioner's clinic,centre, and later on by the response manager(s) at site. physicians use many different kinds of investigative

The larger an incident is (regarding physical spread and/or technologies to support the medical assessment, such as bloodnumber of casualties), the more resources (personnel and tests, scannings, x-rays and monitors, to support andequipment) are needed and the more complex and difficult it substantiate their examination and monitoring of the patient.becomes to organize the overall situation and with that also However, in the prehospital medical work it is difficult to usethe emergency response. Many different types of emergency such equipment, and today almost only one type of monitoringresponse professionals are involved (especially police, fire equipment is typically used - the Life Pack 121. Thisfighters, medical staff and ambulance staff), and follow the technology is developed primarily for diagnostics, treatmentbeforehand planned and known Incident Command System and surveillance of heart failure patients, but is also used for(ICS) [9]. In the ICS the roles of the different professionals monitoring people with other types of injuries - e.g. caused byand their mutual routes of communication are specified, a car accident. It is important to stress that it is very stable andtogether with directions for physical configuration of the trustworthy equipment, but use of it holds some centralrescue area. The action and communication happen within a problems: LifePack 12 consists of sensors to measure a fullcertain structured hierarchy, both within each profession and ECG signal (meaning that several electrodes have to be placedacross the different professions. Working and communicating on the body), oxygen saturation in the blood (meaning that awithin this hierarchy should have the effect that they all can sensor has to be placed on a finger, and a sensor for measuringrecognize what to do, who to refer to and collaborate with. blood pressure (meaning a cuff has to be places around theEach person can concentrate on exactly his/her task in close arm). These different sensors, placed on the injured person'scooperation with other professionals involved. Briefly body, have only wired connection to the monitor, containingexplained the overall division of work between professional the display. All the wires complicate the work, simply becauseskills is as follows: of their presence, but also the one to one connection between1) The firefighters are the primary rescuers - they are sensor and display means that the monitored data can only be

responsible for getting people out from the primary seen if you are just beside the "not mobile" display, which inemergency area and for securing the incident area.

2) The police are the overall responsible professional group. 1htp/wweeenmdiaehoo o/lfak2t1

practice means being beside the patient. This has one main everyday response work.advantage in the sense that while observing the monitored C Pervasive computing, palpability and medicaldata the injured person him/herself can also be directlyobserved, but the situation also entails constraints. Especially emerec rsovnsin an incident situation with many injured persons there is not Asnitnapeasbove an li [8 [1] us eou -teenough~~~~~~~~~~~~~~~~eqimn.osrelaltoewoaeijrd n information technology is limited at the incident site in thetherehareqtpeoenotosurveiallthbesie everr injured,pond medical emergency response work, carried out. However wethere are not people enough to sit beside every injured person

mov toad oit hr evsv optnto obev th moioe data. Threo, montoin move towards a society where pervasive computing

e p ise ye d foresupervisionrofothe mos sevren technologies are becoming integrated in our everyday life,equipment iS only used for supervision Of the most severetho xesv s fepcal tl mlecau*is through extensive use of especially still smallercasualties.

2) Person id and registration ofdata microprocessors, wireless technologies, increase of

Another important issue concerns identification of persons bandwidth and growing quantity of different mobile devicesand registration of different data related to a victim's injuries and displays. So, when considering and discussing futureor the accident as such. technologies to be used in medical emergency response the

Depending on the extent of the incident situation, different pervasive computing angle is important to implicate anddocuments come into play for registrating of relevant data. In investigate.any case the document contains - beside space for the relevant 1) Palpabilitydata - a unique id that should count as the temporary person id Palpable computing is a new perspective in pervasiveuntil the real person id is known. computing. The concept was introduced with the PalCom

The document, (called Incident Card) is intended to be project [23] and deals with the use aspects of pervasivefilled out continuously for each injured person, starting at the computing, in the way that palpability indicates that IT-

moment they are located and until they are handed over at the systems must be capable of being noticed and mentallyhospital. Thus the document has to be tied to the injured apprehended, that palpable systems support people in

person. But in practice the document is not used in MI understanding what is going on at the level they choose and

situations - there is simply not enough time to fill it out. that palpable systems support control and choice by people.When several persons are injured triage 2 is carried out More specific palpability of pervasive computing systems

Triage is a process where injured persons are sorted into can be characterized by the following concept pairs, where the

(four) groups, depending on their needs for or likely benefits left-side concepts characterize ambient computing and thefrom immediate medical treatment. Coloured cards are used to right side concepts characterize palpable computing [23].indicate to which triage category each injured persons belong:Red colour indicates need for immediate help; yellow colour Invisibility <-> Visibilityindicates that the person can wait a while; green colour scalabiit <-> Derstanindicates that the injuries are not critical and black colour Cntron <-> Denctindicates that the person is dead. As with the above mentioned Heterogeneity <-> Coherencedocument for registration of data, the triage card also should Change <-> stabilitybe tied to the injured person, but in practice this is also notcarried out as intended. The only paper document that

proessonasrall stive tofil ou inMI itutios i an The notion of palpability is in different aspects furtherprofessionals really strives to fill out in MI situations iS aneaoae n[] 4,[] 1] 2] 2]

Incident Log. The incident log contains information about elaborated in [1], [4], [5], [16], [24], [29].which patient is transported to which hospital. This means that 2)riTspo ior incident palabltyit is very limited information that is documented about each Regarding the work in the current project - "IT support ininjured person - an (interim) ID, two words that describe the Major Inids, -iwe foci rinjuries, how the person is therapeutically prioritized (the chal abilistedsabove:trae an whc hoptlh .rsei ruh o 1) Scalability <-> Understandability, where our focus is to

3e) T ailthsprinlhe ir . . . investigate how the same applications can be used (andLast - but not least - is the third issue concerning the useful) in both small and large scale situations, in a way

importance of the users being familiar with the technology thatith understanda of wha n thethey use. This is an important issue for technologies in all application is mainaied esituations, but the bigger an incident is the more stressful and application iSscaled up.chaotic the situation is. In MI situations there is simply no 2

time fo.tilswthtchoogeadosikl- s how stability of systems of technologies and connections... . . ' ~~~~~~~~betweensystems of technologies are obtained andand familiarity with technology can only unfold through usingyg

it in the everyday work. So, the technology that has to be used maniednlotoswthaigdgrefdymc.in M siuatonsshoud aso e ued anduseul inthe Not only regarding mobility and amount of people and

technologies, but also in situations which are stressful,unexpected, and where the settings are unknown.

2http:llwww arnswers corr/topic/tria e

3) Automation <-> User control, where our focus is on what theoretically elaborated, in e.g. [28].can happen and has to happen automatically in differentextents of dynamic, stressful, unexpected, unknown III. DESIGN OF BLUEBio - DESIGN FRAMEWORKsettings and what can be/has to be controlled by the users. Given the recognized issues concerning the medical

emergency response work, described above, and the researchII. RELATED WORK on palpability within the PalCom project [23], our research

In this section we present related work within the three group - in close cooperation with the users - decided to getareas: Medical Assessment, Person-id and registration of into development of the BlueBio biomonitoring system.data and the Familiarity Principle. The two main objectives are:

Within the first area - the use of monitors for medical * From a medical user perspective to get aassessment at site -The CodeBlue hardware and software biomonitoring system that is easy to handle and toplatform for medical sensor networks is of special interest, work with and can collect data that informsand is similar with our project, since this work is also related primarily about the most vital parameters (meaningwith emergency response and is intended to be used both at indicating an injured persons ABC status). Datasite and in related hospitals. Another similarity is that the should be available over distances and for allCodeBlue work focus on development of a wireless persons who might need them and when theycommunication infrastructure, but CodeBlue especially might need themaddresses the robustness and security challenges and * From a research perspective to develop arequirements that are of vital importance within emergency prototype, that enables us to carry out experimentsmedical settings [21], [22], while we focus on the palpability with the PalCom challengesaspects - or PalCom challenges - as described below in III. scalability/understandability, change/stability andMoreover we do both experiment with use of data from automation/user-control, as input to the opendifferent types of sensors, e.g. pulse oximetry and PalCom Architecture [23].electrocardiogram (ECG) sensors [12]. Also [30] investigate This has lead to the design framework, shown in Table Ion how to do continuously medical monitoring by use of below, where each cell is further decribed in the following.wireless microsensors, and here the key issue is the portabilityaspect, meaning wires and heavy batteries are not desired. [8] TABLE Ihas investigated the use of 3G networks for wireless DESIGN FRAMEWORKtransmission of different types of data (patient's video, Scalability / Change/ Automation/medical images and (ECG) signals) between an incident site understandability stability user control.and hospitals, to enhance the possibilities for trauma Medical A D Gspecialists at hospitals in supporting the prehospital trauma assesmentcare. Person ID/ B E H

Regarding the person-id and registration of data several Registrationstudies focus on these aspects: Today's problems with the use of dataof paper-based technologies are described via concrete Familiarity C F Iincident situations in [31], [32]. [15] Describes a speech inputand voice output prototype, to be used in the prehospital care, A) Medical assessment / Scalability - understandabilitywith a syntax that enable the speech recognition system is The daily use situation for the biomonitor is for monitoring alldescribed, to interpret patient conditions, treatments, triage patients at smaller accidents, where the biomonitor transmitsand patient transport requests. [7] gives an overview of information regarding ECG, Oxygen saturation and breathing.research on, among other factors, emerging information In these situations, detailed data is requested by the medical

technologies to support patient tracking, monitoring and staff both at site and at the receiving hospital. However atmedical care. Of smajor incidents where many patients are monitored, the level

systemedi e.elofedtomspecialdintererst,isa vitalhhdatmontor of information detail should be able to be downscaled, to thesystem,developed to monitor soldiers,butwhichalsocould benefit of monitoring more patients and thus supporting the

be used within the healthcare area, especially in disaster overview by for instance the medical coordinator on sitemedicine. This monitoring equipment, called the Personnel and/or the staff at the Acute Medical Coordination centre andStatus Monitor (PSM) is combined with a global positioning in the emergency department.system telling where the monitored person is [27]. B) Person ID and registration ofdata /

Regarding the familiarity principle several studies stress Scalability - understandabilitythat those who work within emergency response - and Person ID is a difficult task in emergency situations ifespecially in mass casualty or disaster situations - have to people are unconscious, confused or dead. Personal ID'sknow what to do and how to do it. Many studies take concrete found in wallets can be borrowed or even stolen, and witnessexperiences from mass casualty incidents as their starting descriptions can be biased by the turmoil of the situation.point, e.g. [11], [20], [25], [26], [3 1], [32] and are However it is necessary to have a unique ID system to support

continuity through the emergency response.

Using the Biomonitor for ID is an obvious possibility, important knowledge for the ambulance people and/or thehowever it requires adequate precautions to secure that a staff at hospital.specific persons ID is tied to a specific biomonitor. This can Storing of information and conveying of the information ofbe done by the medical professionals specifically pairing ID availability should be accomplished by the systemand Biomonitor by a specific interaction. automatically. Nevertheless, in case of breakdowns, user

Furthermore, there is the question of what kind of ID is control is essential for re-establishing connections and forpreferable. On site the medical personnel use descriptions like assessment of the extent of the breakdown level or for"the one in the red shirt", "person lying next to the blue car" checking out if signal disruptions stem from the technology oror "guy with the chest injure and the broken leg". However the patient.this information does not stand for larger scaling up, and Breakdowns should not be handled by the medical staff at theimplicates other solutions to be found such as uniquely incident site, but by persons remotely, e.g. at the hospital.generated numbers or precise positioning with the ability to These can survey the biomonitors in use and will be able tokeep track of the patient. recognize and handle breakdowns (contingencies).

Attaching other types of data like speech annotations, H) Person ID and registration of data /Automation - userpictures and/or video streams could also contribute both to the controlmedical assessment and to the scalability/understandability The ID should be transmitted continuously together with thedimensions. monitored data or other registered data. However, if doubtC) Familiarity principle /Scalability - understandability occurs regarding whether the patient has the correct ID or

In order to be operational at major incidents, regarding whether the monitor is transmitting from the right patient, itworking structures and processes, equipment, tools and should be possible for staff and personnel in the vicinity of theinstruments must be familiar to the medics through daily use. patient to inspect the ID and biomonitor to ensure the correctThis has profound implications for interfaces and interaction correlation between ID, patient and biomonitor.design which must be able to scale and be understandable and I) Familiarity principle /Automation - user controluseful for both smaller accidents and major incidents. When the biomonitor is mounted, the transmission of data andD) Medical assessment! Change - stability ID should begin instantly. Reading of ID and data can be done

The biomonitor is intended for use not only on site, but for on various devices ranging from cell phones, pda's, laptops,monitoring the patient during the whole period from the televisions etc., containing our open architecture. Themonitoring starts at the accident site, during transportation to biomonitor software should be able to detect what kind ofthe hospital and at the initial interventions at the hospital. This display is requesting to be used, while on the other hand, therequires that even though the patient is moved and the user should be able to choose the format and level ofwireless connections change during transportation, the information she/he wishes displayed. This kind of interactioninformation flow should be stable, and in the eventuality of must be familiar to the users in relation to scale and mobility.periods with no connections, the history of data should still beavailable when connection is established again. IV. BLUEBIO - A USE-SCENARIOE) Person ID and registration ofdata / Change - stability The BlueBio biomonitor is intended for daily use by the

The applied person ID must keep stable throughout the medic(s) on the accident site, for monitoring the condition ofwhole treatment period ranging from the accident site until . . . 'initial treatment at the hospital - and preferably during the pient regadn airwaybreathing a c ion The

whole stay. ~~~~~~~~~~biomonitor iS wireless and approximately has a size of 150 xThole IDay. must remain thesame,eventhoughthe 100 x 18 mm. The underside of the biomonitor consists of an

surouDingschne.ain hen paxi, patentsocan the selfadhesive strip which when moved, reveals 3 electrodesapplied several ID numbers during the different inteentntions and activates the BlueBio monitor.The selfadhesive strip offrom the accident site until treatment is finished. the packaging has an REID tag with an ID number whichF) Familiarity principle / Change - stability couples the strip to the biomonitor, so a medic with a scanner

The interaction with the biomonitor must also comply to the can read the RFID strip on the patient, and receive thefact that different professionals ranging form the medics on corresponding data from the BlueBiomonitor. (The RFID stripsite to for instance a nurse at the hospital, need to be familiar is used for placing on the patient in the same way as luggagewith the biomonitor. strips are placed on suitcases at the check-in desk at airports).G) Medical assessment /Automation - user control When the medic arrives by a patient on the accident siteWhen monitoring patients on site, the medical staff will regardless of whether it is a small accident or a major incident,

want to read information from specific patients. This calls for she takes out the biomonitor from her medic bag. She unsealsexplicit user interaction, where the medical staff can choose the packaging surrounding the biomonitor, takes of thesource and level of information from the biomonitor. selfadhesive strip andplaces it on the chest ofthe patient. TheHowever, when monitoring of a specific patient is not ID strip with the REID tag is placed on the hand wrist of therequired, it is still desirable that the streamed information from patient if possiblethe patient is stored and thus accessible at a later period of Tepteti o oioe.I h ei at htime. Automation of this process is required, as well as an seii edn fteptetsdt,temdcue nRIexplicitly indication that this information iS available iS.. scanner connected to her/his display to read the ID tag on the

patient and the data appears on her screen. The detail level of As the patient is carried into the ambulance, the ambulanceinformation is high in the default setting, but can be reduced manager on the accident site annotates the ID and to whichby the user. If it's a major incident the patients will be hospital the patient is transported. In the ambulance, thegathered at the triage and treatment area and placed according ambulance personnel can monitor the biosensor signals onto the triage principle, earlier described. In this situation the their monitor when having read the RFID-ID tag. Thebiomonitor supports the overview for the people watching biomonitor information is simultaneously monitored at theover the triage area, by alarming the personnel if major AMC, who are also notified when the ambulance people readchanges in the monitored data take place. the RFID-ID tag, thus indicating that the specific

A seffadhesvive stripwith patient/biomonitor is in the specific ambulance.a writen ID and containing an At arrival at the emergency ward, the RFID-ID tag is readRFD tag is fastented to the back by the staff at the ward, indicating for the AMC the positionof the blomonitor

of the patient and making it possible for the staff at the wardto monitor the biosensor information.

At this point, it might be relevant for the hospital personnelBefb

Ip ngthe biomon itor

to exchange the BlueBio biomonitor with more specializedfAf 86~~~~~Bfoe pla<cing t;he biomonitoron the patietthe strip i s equipment; however it is possible to keep monitoring theremoved (Which activates the patient with the BlueBio biomonitor, if wanted.Bluetooth transmission ftom theBueB4 monitOor.)

V. THE DIGITAL DESIGN

The BlueBio system is designed for use in Major Incidents;this is reflected in the digital design by allowing as much

and he strip is placed <flexibility as possible and at the same time providingthe patient.. robustness when used. The biomonitor collects biomedical

signals from the victim's body when placed on the victim. Atthe moment the biomonitor used in the prototype measuresECG.

..whereafter the bomoinoris The biomonitor is comprised of the following mainf ~~~~~~~placedon the patient. components:

* Bluetooth module [3]* ATMEGA8L [2] microcontroller including a A/D

Al/ > converter* Analogue amplifier and filer used to normalize the

Corresponding Strip containing ECGID No.on strip RFID tag. The biomonitor is constructed to use as little power asand B'Iomonitor

possible. This power restriction has influenced the choice ofFigure 2: Placement of the BlueBio monitor components. The microcontroller is only 8MHz and thus do

not have the power to handle the infrastructure, further moreThe personnel is notified with both a visual indication from the Bluetooth devices available when we began the project

an alarming device on the patient and a notification on the were not able to handle scatternets and we did not want tomedical professional's personal device. The alarming device, limit our prototype to 7 devices, as is the limit in a piconet.a device with a significant blink, is placed on the patient on The biomonitor communicates using Bluetooth as it provide aarrival at the triage and treatment area and is paired to robust transport medium in the 2.4 GHz band and thebiomonitor by using the RFID tag on the patient. throughput of Bluetooth meets the requirements to transmit an

In most incidents the majority of patients are transported by ECG signal. The current biomonitor used in the prototypeambulance to the hospital within the first hours. However, it is provides less good support for scaling up the number ofcrucial to organize and structure the distribution of patients biomonitors used, due to the Bluetooth hardware used.between the hospitals to prevent overburdening the emergency Bluetooth as a communication was chosen because it was thedepartments. This is done from the AMC (Acute Medic only descent standardized low power RF technology availableCoordination). at the time when the prototype were constructed.

The assessment regarding which patients are to be moved In a Major Incident the normal communicationfirst is done by the Medical Coordinator at the accident site infrastructure might not be available. For this reason, thetogether with the AMC. At the AMC they can monitor the BlueBio prototype system is built in such a way that thepatient/biomonitor on large displays, and together with the professionals involved in the rescue effort automatically bringMedical Coordinator the AMC can coordinate which hospital the infrastructure to the incident site. The infrastructure for theis relevant and has capacity for the specific patient. BlueBio system consists of base stations. These base stations

communicate internally using WIFI to form an ad hoc information read by the RFID-reader is send to the distributor,network, and are equipped with Bluetooth radios. The base as the RFID tag contains the address of the biomonitor it camestations are used by the biomonitor as a contact point making with, the system can identify the patient and send theit possible for a much further distribution of biomedical physiological signals to the paramedic.signals. In this way the base stations have two responsibilitiesat the incident site. They create an ad hoc infrastructure that VI. THE PHYSICAL DESIGNhandles transmission of biomedical signals and they act as The aim of the project it not to end up with a physicalcontact point for the biomomitor. An example can be seen in product. However, for two reasons we are focusing on thefigure 3. physical design:

1) Health care personnel are very "hands on" in theirapproach to technology. Consequently in order totest out use situations, we have prioritized to

RFID Fk-adOr develop very high end prototypes.2) The developed software prototype is intended to

. X be used in very dynamic settings. The users are

7r / / busy using their hands while doing their medical.ueLth job in a very stressed situation, and "normal"

/// / / / ~~~~~~~~~~~handhelddevices might not be the best solutions.So, focus also has to be on the physical aspect,"simply" to get new ideas for future physicalffi ffi 1 > ^ , v ~~~~~~~~~design.

BASESTaT/N

VW/We describe prototypes in terms of [17] horizontal and

vertical prototypes, in relation to the graph shown in fig. 3.l 1g{4SEON ~~~~~luelt 0MN

} -4-~ "' "=X" J High FidelityFUNCTIONAL PRODUCT

I g - W ~~~~~~~~~~BKONITORII<~~~~~~

Fig. 3: Deployment figure. |X

The base stations have two purposes at the incident site. ZFirst they act as contact points for the biomonitor, and since U'

the biomonior is using Bluetooth, the range is limited to Horizontal prototypesaround lOOm. To avoid using scarce power resources, such as HHigh Fidelitybattery, it is desirable not to let the biomonitor handle the Low Fidelity FEATURESinfrastructure even though it might be possible. Fig. 4: Vertical and horizontal prototypes

Second, the base stations at the incident site create an adhoc network, through which they are able to distribute the A vertical prototype is a prototype with functional qualities,signals received from the different biomonitors and thus make while a fully horizontal prototype has features such as surfacethem available on each of the base stations. The availability of and material qualities. In the project we have worked with athe data on each base station, allows the professionals at the parallel process between fully functional (vertical) prototypesincident site to view data from all biomonitors at the incident and high end horizontal prototypes. As the project has movedsite. This can either be done using Bluetooth or WIFI. on, more and more functional properties have been added toTo identify patient RFID is used, as previously mentioned. the horizontal prototypes and more and more material

The RFID hardware consists of the following parts: qualities have been added to the vertical prototypes.* F2MO3AC23 In the following, the basic issues regarding the physical* KS-RDH-M14 design such as material qualities, surface qualities, interface

The RFID scanner at the moment connects to the base and interaction are related to medical assessment, person IDstation using Bluetooth. Though the base station the and registration of data and the familiarity principle.

A. Medical assessment3 http://www.datasoft.se/class2modulewithantennaf2mO3ac2.htm4http://www.meshedsystems.com/3_Produkte/prod modul_ml .htm For medical assessment the biomonitor must be attached to

the patient. In order to achieve the ECG signal, there must bea close connection between the electrodes and the body, whichis typically established by a gel. The biomonitor has threeelectrodes placed with a mutual interspaced distance of approx10 cm. on line or as a triangle. It is important that the materialis flexible enough to follow the organic curves of the bodyranging from small children to grownups. Also, there aresignificant problems with placing the electrodes due to body bFhair, moisture, wounds etc.

Fig. 7: An example of the biomonitor prototype on a patient.

To obtain the most optimal ECG signal the biomonitors~~~~~shoul be placed on the chest region, but because thebiomonitor then is hidden beneath clothes as well as blanketsit may not be visible for others. This implicates the need foranother visible device placed on or beside the patient, in orderto show that this specific patient has a biomonitor and also

Fig. 5: The biomonitor with high end surface and material qualities. maybe indicate visually changes in the patient's conditionand/or give visual alarms. One of the PalCom challengesconcerns automation/user control. When mounting the

_ biomonitor on the body of an injured person the user - by thisaction turns the biomonitor on - an explicit action, controlledby the user, prompts an automatically action.One of the PalCom challenges is that palpable devices

should comply with scalability. The same device used forsmaller incidents such as for example traffic injuries, shouldalso be used at major incidents with maybe fifty or morepatients. This has consequences regarding transportationcapacity if a medical professional has to bring 20 or morebiomonitors wit her/im. Ths a biomonitor shoul be aslight as possible, take up as little space as possible and be easytransportable.

Another issue is if the biomonitor should be disposable orreused. If the biomonitor is to be reused, the surface must becleanable and the electronics should then be well protected.

Fig. 6: The biomonitor with sensors, batteries, Bluetooth and processor. Preferably, a composite device would make it possible toreuse the core electronics while the rest would bedecomposable. The external visible device which notnecessarily gets in direct contact with wounds should have asurface suitable for cleaning and reuse implicating a smoothsurface without nooks and crannies.

B. FamiliarityThe familiarity principle was mentioned previously. One of

the strongest features regarding the existing LifePack 12 isthat the professionals are familiar with this instrument. Theyknow what to expect, how to react when it malfunctions andthey are familiar with the technology - and not only inconnection with use at accidents, but also at the hospitalwards. The healthcare personnel mention this as a mainfeature, regarding use of equipment at major incidents. In an

ultimate stress situation they must be perfectly accustomed to would be natural. However, the patient might move the headthe devices and instruments which they are using. As either in response to pain or in order to keep an eye on what'smentioned earlier, there is not time for learning use of new going on, thus risking covering the line of sight, or provokingequipment, or to tamper with even simple things as for the device to fall of by fast shaking movements. In respect toexample new packaging systems for sterilized instruments. this, a placement on the body and most likely on the surface of

The notion of familiarity relates to three issues regarding a blanket seems be the natural choice.the biomonitor. First and ideally, the biomonitor shouldbecome a device which is routinely placed on the patient from VII. CONCLUSIONS AND FUTURE WORKthe initial assessment and treatment and kept on the patient In this paper we have shown how we through amaybe even through the whole treatment period, including participatory design process in a multidisciplinary researchstay at the hospital ward. Thus, the biomonitor must be very team, have identified some significant challenges regardingeasy to handle and to place on the body (in order to become the physical and digital design of the BlueBio biomonitor.routine). Packaging of the monitoring device should be easy By taking point of departure in three main issues in theto do and a successful placement of the biomonitor should be medical emergency response (medical assessment, ID andexposed through visual feedback directly, before the registration of data and familiarity) and focussing on three ofbiomonitor is covered with clothing and blankets again, the PalCom challenges (Scalability-understandability;Successful placement is in these terms related to receiving a Change-stability; Automation-user control), we havereadable ECG signal and that the biomonitors Bluetooth introduced a biomonitor with promising features for use indevice is transmitting the signal properly. both small and large scale accidents. The research has also

Secondly, the biomonitors signal device (placed externally provided important input for the PalCom open architectureon clothes or blankets) should have the same qualities software necessary for making palpable (healthcare) systems.regarding handling, placement and material qualities as the The biomonitor is still work in progress. The currentcore biomonitor device. BlueBio can read and transmit a 3 point ECG and connect to

Third the visual feedback from the signaling device should WIFI, GPRS, EDGE and other Bluetooth devices through ahave a character recognizable by the personnel, meaning that locally established infrastructure of basestations. Throughthey should be familiar with the feedback from similar experiments with our vertical and horizontal prototypesequipment. together with end users, such as emergency medical

C. Feedback professionals, ambulance personnel, hospital staff and fireThe medical professionals want three types of information fighters, we have been able to identify important issues for our

from the signaling device: 1) The biomonitor is working current and future work. We are currently experimenting withproperly, 2) the state of the patient (stable - worsening - sensors providing information of blood oxygen level andgetting better), and 3) and they want the device to provide sound and frequency of respiration, necessary to carry out a

feedback when reading the display in order to make sure that full ABC medical assessment. However, we have alreadythey are receivin informationfromthecorrectpe. identified a range of problems connected to non visible

t e ip.. wireless connections, and are currently working with crucialFeedback in general at major incidents can be provided issues such as how users identify of connections betweenacoustically, tactile and/or visually. However, acousticfeedback iS at risk of not being heardl in the general turmoil Of deie,we' h oncin r ieesanedbaccid a at risk of *aravtnhen stressfu Furthermore we are working with issues regarding ID, forsituationo tent sandisalsoatheiskmdc stgafTactilhe sytremss example how to improve the identification of victims usinglituacsom of the samientscracdther ics,as justa.mntionedsuhems digital media and how this ID can be used effectively from the

lack some of th am haateitisaus einitial stage of the accident until treatment is provided at theas risk of not being noticed, or having to have a character sointrusive that the tactile response is more of a nuisance than hsieta. Early epermnts in the projectshowe tahelp. Visual feedback seems to be the most promising, but line

numbers. Numbers - or other types of unique ID's are usedof sight is a fundamental requirement and can only be used at nubs. berse- or o hery eseofuniqueIs re ueshort distances. (~~~~~and should be used), but, when several people work inLghot diodances.inaignalngdeice,ustbvisil collaborative, dynamic settings with injured persons, they doLight diodes in a signaling device, must be visible in strong

no reonz or spa abu nue esn yueodaylight indicating for example flash diodes. However, at the notbrecognieo se out inurD personsmbyusesame time it should be taken into consideration how to reduce ners th use other types of ID mechanisms, lkethe exposure of the light to the patient. The diodes need to be decripis of age, haircol damageand damage

.. . . ~~~~~mechanism and/or where the person was situated when found.protcte frmuintnde jots nd troes nd he oveing These observations - made during experiments - have lead toglas/acrylic must be resistant to dirt and scratches. Theultimate placement of a signaling device remains to be tried fou onueo. ifrnttpso Ds

* * * , * , , ~~~~~~~~Positioning iS also a core issue we have to deal with in ourout~~~~~~~~~ ~inoupatcptr deig prcssbtweaewokn future work - it does not make sense to be able to monitorwith two candidate placements. Firstly, the patients head Willlin~~~~..most cae be viil iniatn that aplaementher people over large distances if not being able to recognize

where they are. [16] M. Ingstrup, and K. M. Hansen, "Palpable Assemblies: Dynamic ServiceWe are currently working on the third iteration of the Composition for Ubiquitous Computing," in Proc. Sevententh

International Conf. Software Engineering andKnowledge Engineering.physical design which will result in 10 BlueBio monitors to be Taipei, Taiwan, Republic of China. July 2005.used in future laboratories [6], which will provide further [17] G. Kramp, "Bridging communication gaps with High Fidelityinput for the PalCom software open architecture [23]. In this prototypes". D2B conference, Shanghai 2006.process we will also get into connections and dependencies [18] M. Kristensen, M. Kyng, L. Palen, "Participatory design in emergencyacross the different palpable challenges pairs; e.g. how medical service: designing for future practice," in Proc.SIGCHI conf.Scalability-understandability and Automation-user control are Human Factors in computing systems, Montreal, 2006, pp. 161-170.mutual related. [19] M. Kyng, E. T. Nielsen, M. Kristensen, "Challenges in Designing

Interactive Systems for Emergency Response," to be published in Proc.DIS conf. Designing Interactive Systems, Pennsylvania, 2006.

AcKNOWLEDGMENT [20] D. J. Lockey, R. MacKenzie, J. Redhead, D. Wise, T. Harris, A. Weaver,We thank our colleagues in the PalCom project and the pre- K. Hines, G. E. Davies, "Major Incident report - London bombings July

hospital effort in Aarhus. We also thank the engineers S. 2005: The immediate pre-hospital medical response, " Resuscitation vol.hospital~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~66, pp. 9-12, 2005.Kucharski, L. Kubala and M. Byczuk, who developed the firstversion of the biomonitor system. [21] K. Lorincz, et al, "Sensor networks for emergency response: Challenges

and opportunities," in IEEE Pervasive Computing, oct.-dec. 2004, pp.16-23.

REFERENCES [22] D. Malan, T. Fulford-Jones, M. Welsh, S. Moulton, "CodeBlue: An ad[1] P. Andersen, J. E. Bardram, H. B. Christensen, A. V. Corry, D. hoc sensor network infrastructure for emergency medical care,"

Greenwood, K. M. Hansen, R. Schmid, "An Open Architecture for presented on Intl. Workshop on Wearable and Implantable Body SensorPalpable Computing - Some Thoughts on Object Technology, Palpable Networks, Apr. 2004.Computing, and Architectures for Ambient Computing," in Proc. [23] PalCom project web-page. Available: wwwist-palcom.oECOOP 2005 Object Technology.forAmbient Intelligence, Glasgow, [24] G. Rimassa, D. Greenwood, M. Calisti, "Palpable Computing and theU.K., 2005.

Tole of Agent Technology," in Proc. Multi-Agent Systems and[2] Atmel microcontroller description, available at: Applications IV, 4th International Central andEastern European

www.atmel.com/dyn/resources/proddocuments/2486S.pdf Conference on Multi-Agent Systems, CEEMAS 2005, Budapest,[3] Bluetooth description, available at: Hungary, September 2005

h /w gg o f af=2&th1&D=l300&subp'2OO [25] J. D. Roccaforte, "The World Trade Center Attack, Observations from[4] M. Buischer, G. Jensen, "Sound sight: Seeing with ultrasound," to be New York's Bellevue Hospital", in Critical Care, vol. 5, no. 6, Dec.

published in Health Informatics Journal, 2006 2001

[5] M. Biischer, M. Christensen, K. M. Hansen, P. Mogensen, D. Shapiro, [26] L. Romundstad, K. 0. Sundnes, J. Pillgram-Larsen, G. K. Roste, M."Bottom-up, top-down? Connecting software architecture design with Gilbert, "Challenges of Major Incident Management When Excessuse," in: Hartswood et al. Configuring user-designer relations: Resources are Allocated: Experiences from a Mass Casualty IncidentInterdisciplinary perspectives. Springer Verlag. (forthcoming - Copy after Roof Collapse of a Military Command Center, Prehospital andavailable from Preben Mogensen preben Ldaimi.audk). Disaster Medicine, vol. 19, no. 2, pp. 179-184, April-June 2004.

[6] Buischer M, Eriksen MA, Kristensen JF, Mogensen PH. Ways of grounding [27] R. M. Satava, "Virtual reality and telepresence for military medicine," inimagination. Proc. eighth conf participatory design: Artful integration: Annals, Academy ofMedicine, Singapore, vol. 26, 1997, pp. 118-120.Interweaving media, materials andpractices - volume lACM Press; 2004. [28] M. J. Schull, L. E. Ferris, J. V. Tu, J. E. Hux, D. A. Redelmeier,p 193-203. "Problems for clinical judgement: 3. Thinking clearly in an emergency,"

[7] T. C. Chan, J. Killeen, W. Griswold, L. Lenert, "Information technology in Canadian Medical Associan Journal, vol 164, no 8, 2001, pp. 1170-and emergency medical care during disasters," in Academic Emergency 1175.Medicine, vol. 11, no. 11, nov. 2004, pp. 1229-1236. [29] U. P. Schultz, E. Corry, K. V. Lund, "Virtual Machines for Ambient

[8] Y. Chu, A. Ganz, "A mobile teletrauma system using 3G networks," in Computing: A Palpable Computing Perspective," in Proc. ECOOP 2005IEEE Transactions on Information Technology in Biomedicine, vol. 8, Object Technologyfor Ambient Intelligence, Glasgow, U.K., 2005.no. 4, Dec. 2004, pp. 456-462. [30] E. Shih, V. Bychkovsky, D. Curtis, J. Guttag, "Demo abstract:

[9] The Danish National Level Incident Command System. Description Continuous medical monitoring using wireless microsensors," in Proc.available at: http:!/www.brs.dk/uk/danish preparedness acthtm 2ndInt. Conf. Embedded Network Sensor Systems, Baltimore, Maryland,

USA, Nov. 2004, p. 310.[10] P. Driscoll, D. Skinner, R. Earlam (eds), ABC ofMajor Trauma. BMJ

Books, 2.ed, 1996. [31] R. B. Vukmir, P. M. Paris, "The Three Rivers Regatta Accident: AnEMS Perspective, American Journal ofEmergency Medicine, vol. 9, no.

[11] E. R. Frykberg, "Commentary: Terrorist Bombings in Madrid," in 1, pp. 64-71. January 1991.Critical Care, vol. 9, no. 1, Febr. 2005, pp. 20-22

[32] L. Welling, R. S. G. M. Perez, S. M. van Harten, P. Patha, D. P. Mackie,[12] T. R. F. Fulford-Jones, G. Wei, M. Welsh, "A portable, low-power, R. W. Kreis, J. J. L. M. Bierens, "Analysis of the pre-incident education

wireless two-lead EKG system," in Proc. 26th Ann. Intl. Conf. IEEE and subsequent performance of emergency medical responders to theEMBS, San Fransisco, CA, USA, Sept. 2004, pp. 2141-2144. Volendam cafe fire, European Journal ofEmergency Medicine, vol. 12.

[13] J. Greenbaum, M. Kyng (eds), Design at work: Cooperative Design of no. 6, pp. 265-269, 2005.Computer Systems. Hillsdale, New Jersey, USA, Lawrence Erlbaum [33] Understandable IT web-page. Available (in Danish):Associates Inc, 1991. ht:/w~ssaeadakhatcr/rjke~t

[14] K. Gronbaek, M. Kyng, P. Mogensen, Toward a cooperativeexperimental system development approach. In: M. Kyng, L. Mathiassen(eds), Computers and design in context. MIT Press; 1997. p 201-38.

[15] T. G. Holzman, "Computer-human interface solutions for emergencymedical care, " in Interactions, vol. 6, no. 3, 1999, pp. 13-24.