Supporting Medical Ward Rounds Through
Mobile Task and Process Management
Rüdiger Pryss, Nicolas Mundbrod, David Langer, and Manfred Reichert
Institute of Databases and Information Systems, Ulm University, Germany
{ruediger.pryss, nicolas.mundbrod, david.langer, manfred.reichert}@uni-ulm.de
Abstract.
In a hospital, ward rounds are crucial for task coordination
and decision-making. In the course of knowledge-intensive patient treat-
ment processes, it should be possible to quickly dene tasks and to assign
them to clinicians in a exible manner. In current practice, however, task
management is not properly supported. During a ward round, emerging
tasks are jotted down using pen and paper and their processing is prone
to errors. In particular, sta members must manually keep track of the
status of their tasks. To relieve them from such a manual task man-
agement, we introduce the MEDo approach. It transforms the pen and
paper worksheet to a digital user interface on a mobile device. Thereby,
MEDo integrates process support, task management, and access to the
patient record. Interviews of medical sta members have revealed that
they crave for a mobile process and task support. This has been further
conrmed in a case study we conducted in four dierent wards. Finally,
in user experiments, we have demonstrated that MEDo puts task acqui-
sition on a level comparable to that of pen and paper. Overall, MEDo
enables users to create, monitor and share medical tasks based on a mo-
bile and user-friendly platform.
Key words: knowledge-intensive process, healthcare process, mobile task
management
1 Introduction
The increasing number of options to streamline production and administrative
services based on advanced information technologies have given rise to a struc-
tural shift of value-adding jobs. Standardized tasks are more and more supported
by IT or outsourced to cut operative costs [1,2]. Along this trend, the support of
knowledge-intensive processes has gained more attention as well [3]. Respective
processes are considered as the strong opposite to production processes. In par-
ticular, knowledge-intensive processes face unplanned and complex situations.
Further, they strongly depend on expertise and skills of the stakeholders in-
volved. Generally, knowledge-intensive processes can be characterized as follows
[4]:
Uncertainty
: Knowledge-intensive processes face complex and unplanned sit-
uations. An unmanageable set of inuencing factors, intertwined via dynamic
2 Pryss, Mundbrod, Langer, Reichert
correlations, makes it impossible for involved stakeholders to foresee the pro-
cess in all details before starting its execution.
Goal Orientation
: Struggling with uncertainty, a common goal is used as the
solid basis for all stakeholders in knowledge-intensive processes. Therefore,
all activities and resources are aligned to achieve the dened goal. To ease
the achievement of a goal, sub-goals may be dened.
Emergence
: A knowledge-intensive process usually comprises mutual phases
of planning and working. Having performed a number of activities, involved
stakeholders usually evaluate work results rst in order to properly focus on
the planning of activities to be performed in the following. Such an agile
planning and working let knowledge-intensive processes gradually emerge
towards the dened goal.
Growing Knowledge Base
: Participants of a knowledge-intensive processes
leverage their expertise and experiences to eectively achieve a common
goal. Hence, they expand their knowledge base during process execution.
Further, they create explicit knowledge through intermediate and nal work
results.
These characteristics expose core challenges with respect to the support of
knowledge-intensive processes. In particular, the latter require a high degree of
exibility and agility not provided by contemporary process-aware information
systems.
1.1 Problem Statement
In previous work, we analyzed various processes from the healthcare domain
[5]. We characterized them as
knowledge-intensive
as they are driven by many
factors and expose complex interdependencies, and recurring context switches.
Particularly challenging are the processes taking place in the context of clinical
ward rounds.
For ward rounds in a hospital, there is a high demand for improving the interac-
tion and communication among healthcare professionals [6]. Medical problems
of inpatients are becoming increasingly complex and patient data needs to be
managed at the bedside. Although information technology, which is tailored to
ward round demands, has reached a mature level, there is a lack of process-aware
task support during ward rounds [7]. However, this is not surprising consider-
ing the fact that ward rounds constitute knowledge-intensive processes. In turn,
this type of process is not properly supported by existing process management
technology as its steps and resources cannot be completely foreseen. Instead,
processes are driven based on the expertise and experiences at run time.
In order to understand how an adequate support for knowledge-intensive pro-
cesses could be smoothly integrated into medical practice, we attended a multi-
tude of ward rounds, interviewed medical sta, and analyzed existing solutions
in a case study. In turn, these activities revealed two major issues:
First
, the
current paper-based task worksheet needs to be transferred to a digital, mobile
Supporting Ward Rounds Through Mobile Task and Process Management 3
variant. Note that clinicians use worksheets as their personal "information sys-
tem" to organize daily tasks. In current practice, however, data gathered with
paper sheets is not consistently linked to the one managed by the hospital infor-
mation system in use. When adding a task to her sheet during a ward round, for
example, a physician makes notes like
"Mrs.Richards: X-ray request arranged,
monitor status and check images today"
. In the pen and paper version, basi-
cally, each task description solely consists of the patient's name and free text.
Second
, another demand posed by clinicians is to better align task management
with healthcare process support (e.g., to keep track of medical orders) as well as
integrated access to patient folders.
1.2 Contribution
This paper shows how such a process-aware, mobile task support has been re-
alized in our
MEDo
1
approach. The paper presents basic concepts as well as
the information architecture underlying MEDo. Furthermore, it describes the
methodology applied to design, develop and implement MEDo. We show in de-
tail how requirements were elicited and in how our participation in dierent
kinds of medical ward rounds contributed to this. Although MEDo focuses on
the process-aware, mobile support of clincial tasks, we believe that similar solu-
tions can be provided in the context of other knowledge-intensive processes as
well.
This paper provides a signicant extension of previous work we presented in
[8]. In particular, in [8], we have neither provided fundamental insights into
the method applied when developing MEDo nor the information architecture
underlying MEDo. In addition, this paper provides an in-depth description of
the phases of general ward rounds in hospitals in order to give detailed in-
sights in real-world scenarios for mobile task and process management as well
as existing challenges. While [8] only presents the ward rounds in internal and
orthopaedic medicine, this paper additionally gives insights into ward rounds in
trauma surgery and emergency. This includes a discussion of ward round char-
acteristics as well as the way we addressed them in MEDo. For example, we
provide an in-depth description of a ward round for which a mobile access to the
hospital information system is already in use. Overall, the analysis of the four
ward rounds has revealed that current approaches targeting at task and process
support often neglect the way how clinicians organize their daily work. In partic-
ular, established working procedures, even those which more or less bypass the
hospital information system, must be carefully analyzed and understood. We
believe that such detailed insights are helpful to better understand the needs
of knowledge-intensive processes in respect to mobile task support. Finally, we
provide more details on features implemented by MEDo and discuss feedback of
clinicians in this context.
The remainder of this paper is organized as follows: Section 2 summarizes the
1
MEDo stands for MedicalDo. A video demonstrating MEDo can be viewed at:
http://vimeo.com/channels/434999/54762040
4 Pryss, Mundbrod, Langer, Reichert
methodology we applied in the context of MEDo. In turn, Section 3 gives de-
tailed insights into medical ward rounds and describes the method we applied to
elicit fundamental requirements for a exible support of medical ward rounds.
In Section 4, we present these requirements in detail. Section 5 discusses the de-
sign, development and implementation of MEDo. In Section 6, we present results
from applying MEDo in clinical practice and Section 7 summarizes evaluation
results. Finally, Section 8 discusses related work and Section 9 concludes with a
summary and outlook.
2 Methodology
Regarding the design of our research on mobile task support in the context of
medical ward rounds, we consider two major aspects:
Case Study:
A detailed case study shall be conducted in order to properly un-
derstand how ward rounds are actually performed in clinical practice. In this
context, the knowledge-intensive character of ward rounds shall be demon-
strated and the understanding of knowledge-intensive processes in general
be enhanced.
Design Science:
Based on the lessons learned from the case study, require-
ments are derived to design, develop, implement and validate an approach
for mobile task support in the context of ward rounds. This approach shall
be also assessed regarding its benets in respect to the support of knowledge-
intensive processes in other domains.
Overall, we choose and apply the following methodology (cf. Fig. 1) ensuring
replicability and, hence, valid results.
Problem Areas Specific Problems &
Existing Solutions
Requirements Approach &
Styleguide
Prototype Suggestions
Mockups
User Feedback
Detailed Design
Define
Requirements
Elaborate
Approach
Observations
Design of
Process Models
Process Analysis
Literature
Research
Interviews
Framework
Preconditions
Programming
Application
Interviews
Usability Tests
Foundation Process Focus
on Ward Rounds
Requirements
Analysis
Detailed
Approach Implementation Evaluation Conclusion
Section 3 Section 4 Section 5 Section 6 Section 7 Section 9
Limitations
Fig. 1.
Methodology
Our methodology comprises seven phases (cf. Fig. 1). In this paper, we
present results of all phases (except the rst one) in detail. In an initial phase, we
analyzed existing work and interviewed clinicians from dierent medical disci-
plines (e.g., internal medicine and surgery). In particular, we wanted to elaborate
Supporting Ward Rounds Through Mobile Task and Process Management 5
major demands regarding mobile task support in the context of ward rounds. The
results we obtained have conrmed that such a mobile task and process support
is highly needed, but has not been properly addressed by existing approaches yet.
All other phases are presented in the following. Furthermore, Section 5 renes
our methodology to a more detailed level in order to indicate how the design and
development of MEDo was accomplished in close interaction with clinicians.
3 Medical Ward Rounds
3.1 General Procedure of a Ward Round
Regardless of the medical discipline, the basic procedure of ward rounds in (Ger-
man) hospitals is more or less the same. We sketch it and illustrate its major
characteristics.
In general, a ward round comprises four phases (cf. Fig. 2).
Prepare
Ward Round
Prepare
Patient Visit
Conduct
Patient Visit
Perform
Post-Processing
use memorized
information from
this phase
use memorized
information from
this phase
Fig. 2.
Characteristic phases of a ward round
P repare W ard Round.
In the preparation phase, patient data relevant for
the ward round is gathered. This is either accomplished by nurses or assis-
tant physicians. In turn, the collected data is then made available to medical
sta depending on the IT infrastructure of the hospital. Usually, three op-
tions are common in practice: First, paper-based records are provided on a
ward round trolley that is carried from room to room. Second, paper-based
records are carried by the ward round team. Third, the ward round trol-
ley is equipped with a computer enabling access to the hospital information
system. In current practice, the second option is the predominant one.
P repare P atient V isit (Ante P ortas).
During this phase, the ward round
team discusses the patient's health status outside the patient's room. If two
or more patients are accommodated in the same room, their status is conse-
quently discussed outside. This communication is either managed by a senior
or an assistant physician responsible for the patient. Two kinds of medical
orders may be made at this stage: First, orders may be assigned to nurs-
ing sta with the purpose to support the patient visit. Second, the physician
may make notes about consultations required with other medical disciplines.
Conduct P atient V isit.
For each patient, the physician in charge decides
whether or not the patient must be physically examined. Following this,
medical ndings are directly discussed with the patient. In particular, the
6 Pryss, Mundbrod, Langer, Reichert
patient is asked about current complaints. Finally, changes concerning the
treatment of the patient are discussed and determined.
P erform P ost−P rocessing.
In the last phase, the physician in charge re-
ects on the steps of the two preceding phases. Following this, she may
schedule additional treatment procedures for the patients visited, i.e., she
may sign medical orders (e.g., laboratory examinations) and external consul-
tations, prescribe changes in medication, or make appointments with other
experts (e.g., physical therapists). Note that this phase might be postponed
to a later stage when exceptions occur or emergency situations interrupt the
physician. As memorizing all information collected during the second and
third phase is hardly possible, physicians rely on the (paper-based) docu-
mentation they made during the ward round.
3.2 Observations
We observed that frequently occurring exceptions, ad-hoc coordination eorts,
and time pressure constitute major challenges for healthcare professionals in the
context of their daily ward routines. Considering the high complexity and un-
certainty of patient treatment processes (cf. Section 2), healthcare professionals
rely on ward rounds as a reliable source regarding the current status of their in-
patients [7]. Hence, a ward round is the common environment in which decisions
about patient treatment can be synchronized and communicated, or changes
in patient treatment (e.g., medications) be made. Considering this complexity
and the tasks emerging in the context of ward rounds, clinicians have developed
their own way for coping with uncertainty as well as the drawbacks of contem-
porary hospital information systems. In particular, the use of a paper-based task
worksheet is an established practice for eectively organizing daily work. While
paper-based worksheets show many advantages, a number of drawbacks can be
observed as well:
Using a task worksheet comprises two phases. In the rst one, task acquisi-
tion is accomplished in order to prepare the second phase, during which the
tasks are actually performed. In the rst phase, problems might occur if the
physician dening the task is not the same as the one making corresponding
notes on a worksheet. In turn, the second phase reveals problems if task
performance is done later, i.e., with a signicant delay in relation to task
acquisition. The most critical issue, we observed in this context, is loss of
information.
Before a shift change takes place, the physician in charge usually makes
a neat copy for her colleague responsible for the next shift. On one hand,
this way of handing over patient information shows several problems. For
example, a physician may use her own "codes" for notes, which is not useful
for the next physician in charge. Therefore, she changes her own codes in
order to make them available for her colleague as well. On the other, this
procedure is error-prone as well as time-consuming.
Supporting Ward Rounds Through Mobile Task and Process Management 7
Handwritten notes might be unreadable or ambiguous, causing additional
errors or at least time-consuming checks.
The task worksheet is usually stored in the ward oce to enable access to
all ward members. As a result, physicians must visit the oce in order to
access the task worksheet. In turn, this is time-consuming and inecient.
Tasks and their status on paper-based worksheets are not traceable and the
performed work can not be properly archived.
These obvious drawbacks have motivated our research. Section 3.3 gives de-
tailed insights into four ward rounds we analyzed in dierent medical depart-
ments. Particularly, we discuss their characteristics and particularities in respect
to the dierent medical disciplines they refer to.
3.3 Ward Round Analysis
We started our clinical investigations with an initial survey. In particular, we
evaluated how physicians perceive the current management of ward rounds when
using pen and paper. First, we studied how they perceive task acquisition. Sec-
ond, we asked them how they currently manage ward rounds in general, i.e.,
how they handle communication with other healthcare professionals as well as
the access to patient data provided by hospital information systems. Fig. 3 de-
picts the results of this survey. Later on, we will compare the use of pen and
paper during ward rounds with the one of MEDo.
Question MeanValue StandardDeviation
TaskDefinition 2.22 1.30
WardRoundManagement 2.50 1.12
|Scalerangehas6valueswithintheinterval:verygood–inadequate|
Fig. 3.
Task denition and overall ward round management without IT support
To elicit fundamental requirements and to better understand how these are
currently addressed, we analyzed characteristic ward rounds. More precisely, we
participated in four wards rounds at dierent clinical departments. The basic
facts related to these ward rounds are summarized in Fig. 4. Interestingly, only
one of the four clinical departments already provides an IT support for accessing
patient data during ward rounds; i.e., imaging data and laboratory results may
be accessed during ward rounds using a tablet PC. As can be seen from Fig. 4,
the ward rounds we analyzed vary in respect to their basic characteristics.
Based on these insights, we extracted procedures performed or triggered in
the context of a ward round. In particular, we were interested in how they can
be smoothly integrated with mobile task support. Additionally, we identied pa-
tient data, physicians want to access in the context of their task lists.
In order to identify required procedures and needed patient data, we partici-
pated in the dierent ward rounds several times and then transferred the gained
8 Pryss, Mundbrod, Langer, Reichert
InternalMedicineWardOrthopaedicWard
(paraplegicpatients)
TraumaSurgeryWard
(withheadphysician)
EmergencyWard
DiscussedinSection3.3.1Section3.3.2Section3.3.3Section3.3.4
HospitalUniversityHospitalUlmRehabilitationHospitalUlm UniversityHospitalUlmUniversityHospitalUlm
NumberofBeds2535>10012
AveragePeriodof
Hospitalization
days,weeksweeks,monthsacoupleofdaysupto
onemonth
hours
FrequencyofWardRounds
twiceadayonceadayonceaweekthreetimesaday
PurposeofWardRound
dailyoverviewdailyoverviewprovidinganoverview
fortheheadphysicianon
allwards
sharinginformation
withtheteam
responsibleforthenext
shift
InvolvedPartiestwowardphysicians,
twonurses
threewardphysicians,
uptothreenurses,
severaltherapists
headphysician,
substituteofhead
physician,
wardphysicians
uptoeightward
physicians,
oneseniorphysician,
additionallyneeded
experts(optional)
VisitingTimeperPatient
(measuredduringinvestigation)
7+3.5minutes
(preparation+treatment)
4minutes3minutes4minutes
HospitalInformationSystemERPSoftware(SAP)ERPSoftware(Meierhofer)ERPSoftware(SAP)
ERPSoftware(SAP)
MobileDataAccessto
VitalData(nursingdocumentation) (nursingdocumentation)(nursingdocumentation) (nursingdocumentation)
Medication(nursingdocumentation) (nursingdocumentation)(nursingdocumentation) (nursingdocumentation)
ImagingDiagnostics(tablet) (printout)
LaboratoryFindings(tablet) (printout)
|():available|():notavailable|
Fig. 4.
Basic facts characterizing four dierent ward rounds
insights to a more formal IT representation. In the latter context, explicitly den-
ing the identied procedures in terms of BPMN process models has turned out
to be useful. Thereby, a core set of BPMN elements has turned out to be su-
cient for the discussions with clinicians. As a result of these discussions, relevant
patient data physicians want to access is depicted in Fig. 5a. Additionally, Fig.
5b shows the data privileges required by physicians and nurses in this context.
Note that physicians emphasized the need to share task sheets with colleagues
and nurses in order to improve overall communication.
(a) Patient data
DataPhysician
rw
NursingStaff
rw
VitalData
Medication
MedicalReports
Appointments
()
Diagnostics
|:needed|():sometimesneeded|r:readaccess|w:writeaccess|
(b) Access privileges
Fig. 5.
Patient data and access privileges
Supporting Ward Rounds Through Mobile Task and Process Management 9
In the following sections, we give insights into the ward rounds we investi-
gated.
3.3.1 Ward Rounds in Internal Medicine
Fig. 6 shows the process co-
ordinating the steps of a ward round in a clinical department from internal
medicine. This type of ward round is common to many hospitals encompassing
two stages. Each of them comprises a number of tasks. In the rst stage, the
physician analyzes the cases of all inpatients without facing them immediately.
For this purpose, he accesses the hospital information system in order to retrieve
the information required. Furthermore, he is assisted by a nurse who provides
him with relevant patient information as well. Finally, relevant issues and al-
ternatives regarding patient treatment are discussed. In the second stage, the
physician visits the inpatients, makes notes using pen and paper, and veries
her decisions. If new tasks (e.g., medical orders) become necessary, she makes a
note on her worksheet and adopts this change in the hospital information system
afterwards. Examples of tasks and medical orders, respectively, emerging in the
context of the second stage include requests for X-ray examinations, laboratory
tests, consultancies, and changes in patient medication. Making appointments
with remote hospital departments constitutes another kind of task that might
emerge during a ward round.
Compared to the other ward round scenarios, the internal medicine department
must cope with larger numbers of orders per patient in the context of a ward
round. Reason for this phenomenon is that inpatients are often multimorbid (i.e.,
patients suering from multiple deceases). Hence, task worksheets based on pen
and paper are crucial for eectively coordinating treatment changes during ward
rounds. As a result, in internal medicine, physicians crave for a quick access to
monitoring data during ward rounds (i.e., vital signs like pulse or oxygen level)
using smart mobile devices. Note that the amount of data recorded per patient
and day might be high. Anyway, it is indispensable for making decision during
ward rounds. Another requirement raised by the physicians relates to the ability
to access patient monitoring data in real time. Finally, in internal medicine only
one physician usually performs the ward round. As a consequence, she must ef-
fectively hand over patient data to the next shift or to colleagues. Accordingly,
any mobile assistance must consider this requirement.
3.3.2 Ward Rounds in Orthopaedic Medicine
Fig. 7 shows the process
coordinating the steps of a ward round in a clinical department from orthopaedic
medicine. This ward round encompasses the stages described in the following.
Before the ward round starts, in the rst stage, one or more assistant physi-
cians prepare relevant patient documents on a ward round trolley (e.g., X-rays
or paper-based reports). In general, the ward round team consists of assistant
physicians, a senior physician, and a physical therapist. In certain cases, the
head physician participates as well. In the second stage, the following procedure
is accomplished for each patient during the ward round. First, the patient sit-
uation is discussed outside his room based on the patient documents from the
10 Pryss, Mundbrod, Langer, Reichert
Fig. 6.
Ward round in internal medicine
trolley. Second, the ward round team enters the room and a nurse determines
the current vital signs of the patient. Meanwhile, the senior or head physician
talks to the patient and decides about additional treatment.
Compared to the other scenarios, ward round management diers in one aspect:
inpatients usually have a longer length of stay. Accordingly, the need to request
additional examinations is high.
To minimize eorts for follow-up tasks, in the analyzed scenario a mobile access
to the hospital information system has been already established. Interestingly,
this mobile system does enable an access to the complete electronic patient
record. Instead, the preparation of specic patient-related tasks can be accom-
plished using a tablet PC. In addition, the latter ensures mobile access to specic
contents of the hospital information system.
Although the functions provided by the hospital information system have not
been designed with the use of tablet PCs in mind and are hence not adequately
mapped to the latter, physicians may order certain examinations through a mo-
bile access to the hospital information system. Examples include request for
X-rays, and consultancies as well as changes in patient medication. Making ap-
pointments with physical therapists constitutes another important task to be
accomplished by physicians. During a ward round, the senior or head physi-
cian makes notes about upcoming tasks, which are then processed by assistant
physicians after the ward round. Recall that several documents related to these
tasks may have already been prepared during the ward round. Therefore, mobile
access to the hospital information system is leveraged enabling the physicians
Supporting Ward Rounds Through Mobile Task and Process Management 11
to quickly work on their tasks afterwards. Still, most orders are processed in a
paper-based fashion.
Fig. 7.
Ward round in orthopaedic medicine
As a result, physicians considered the mobile acquisition of tasks and med-
ical orders as being useful. Particularly, they emphasized that it contributed
to reduce error rates (e.g., omissions). Again, the paper-based task worksheet
constituted the most prevalent instrument for memorizing and communicating
upcoming tasks. Furthermore, being already experienced with a mobile access
to specic contents of the hospital information system, the physicians revealed a
number of aspects they were still missing. First, the user interface of the mobile
device was not properly mapped to the specic situation during a ward round.
For example, context switches were not properly considered and task acquisition
took too much time. Second, not all tasks could be captured using the mobile
system, which resulted in a mixture of electronic and paper-based task notes.
Third, no vital patient signs (e.g., heart rate, temperature, etc.) were displayed.
Fourth, no status information about tasks concerning medication changes was
provided by the mobile access to the hospital information system.
In the orthopaedic ward, we further observed a specic situation that we con-
sidered when designing MEDo [9]. In practice, a mobile access to the hospital
information system with a tablet PCs is only possible if the respective hospital
area provides a reliable wireless network. Actually, this has not been the case
12 Pryss, Mundbrod, Langer, Reichert
for one of the orthopaedic wards. Hence, the physicians of this ward could not
use their tablet PC, but had to work with pen and paper worksheets instead.
Accordingly, the setting of the orthopaedic ward allowed us to directly compare
wards with and without mobile access to digital patient data. Fig. 8 presents
obtained results [9].
WardAWardB
Numberofbeds6Numberofbeds29
Typeofrooms1x2‐bedroom
1x4‐bedroom
Typeofrooms1x1‐bedroom
6x2‐bedroom
4x4‐bedroom
W‐LANaccessnotavailableW‐LANaccessavailable
Visitedpatients5Visitedpatients33
Figuresforwardroundsrelatedto
wardAandB
Duration:2hoursand21minutes
PrintednumberofX‐rays:70
Numberofinvolvednurses12
Numberofinvolvedphysicians23
Numberofinvolvedphysicaltherapists33
Overalldurationofnursingtasks20minutes10minutes
Overalldurationofphysiciantasks26minutesand25seconds88minutes
Averagetimeneededperpatient5minutesand17seconds2minutesand40seconds
Numberofquestionsfrompatientstophysician 414
Numberofquestionsfromphysiciansto
patients
46169
Numberofquestionsfromphysicianstonurses319
Numberofquestionsfromphysiciansto
physicaltherapist
68
Amountoftimewhichisdirectlydedicatedto
thepatientduringhisvisitingtime
92%72%
Fig. 8.
Time measurements at two orthopaedic wards
Regarding Fig. 8, two issues are of particular interest in the context of our
work. First, the average time needed per patient is decreasing for both physi-
cians and nurses when using mobile devices. Second, the time directly dedicated
to a patient (i.e., the physician is not distracted and talks with the patient) is
decreasing when using a mobile device. While the rst aspect shows that mobile
assistance could relieve the medical ward round team in respect to task man-
agement, the second one indicates that such assistance might have drawbacks
as well. Several patients pointed out that, even without the active usage of mo-
bile devices during ward rounds, the physicians are somewhat distracted. Using
mobile assistance bears the risk to further emphasize this drawback.
3.3.3 Ward Rounds in Trauma Surgery
Fig. 9 shows the basic procedure
of a ward round in trauma surgery. As a peculiarity, the head physician is present
during the whole procedure. Usually, her attendance requires to visit all wards
during a limited period of time due to a tight schedule. Interestingly, if the
head physician is not present, only a subset of all wards is visited. In general,
coordination eorts will increase if the head physician participates in the ward
round. The reasons for this are as follows:
More patient data must be prepared in advance. In turn, missing patient
data would impair the ward round signicantly.
Supporting Ward Rounds Through Mobile Task and Process Management 13
The time that may be spend for a patient per average signicantly decreases
when visiting all wards during a limited time period.
Changes related to patient treatment frequently become necessary. Taking
the tight schedule into account, it is a challenging task to perform all medical
orders eectively and accurately.
During the ward round, a senior physician having specic knowledge about
the respective patient case, might have to be consulted to discuss further
treatment of the patient. If no senior physician is available, the respective
post-procedure will be both time-consuming and error-prone.
Fig. 9.
Ward round in trauma surgery
The procedure of a surgery ward round is organized along the following
stages: Before starting the ward round, in the rst stage, assistant physicians
prepare all relevant patient data. In particular, X-rays are needed for each pa-
tient. Since no mobile access to electronic patient data stored in the hospital
information system is provided, X-rays are carried in paper-based form. For ex-
ample, if four wards with 10 patients per ward must be visited, about 40 X-rays
must be carried during the ward round.
After the preparation phase, in the second stage, the ward round starts. For
each patient the procedure is the same. First, X-rays are discussed. Then, the
head physician evaluates the patient`s health status and decides about further
treatment activities. Meanwhile, an assistant physician takes notes about orders
and tasks using a pen and paper task worksheet. Tasks emerging in this context
may refer to X-ray examinations, appointments with physical therapists, or lab
tests. Particular attention will be required if a patient will undergo a surgery
soon. For this case, it must be ensured that all tasks required for preparing the
surgery are completed in time. In turn, all treatment changes, medical orders,
and other tasks will be processed afterwards by one of the assistant physicians.
14 Pryss, Mundbrod, Langer, Reichert
Note that the assistant physician making the notes and the one performing the
changes might be dierent persons. Worst case, the latter did not attend the
ward round at all.
Furthermore, there was no mobile access to patient data during a surgery ward
round. Since X-rays are usually captured in a digital format, accessing them
by using mobile devices would ease the management of ward rounds. In addi-
tion, requesting X-rays using a mobile device was considered as fundamental
requirement by the physicians we interviewed. Regarding task management, we
observed that quick and easy task acquisition is crucial. We further observed,
that in absence of the head physician, certain issues became easier to handle.
Nevertheless, their handling based on pen and paper remains costly and error-
prone.
As a result, the most important requirements with respect to mobile task sup-
port are as follows: physicians asked for a digital and mobile access to X-rays and
wanted to be able to schedule X-ray examinations immediately at the patient's
bedside. Finally, physicians wanted X-rays to be organized in folders style, which
shall be easily browsable.
3.3.4 Ward Rounds in Emergency Medicine
Fig. 10 shows the process
coordinating the steps of a ward round in a clinical department from emergency
department. This type of ward round shows specic characteristics that distin-
guishes it from the other three ward rounds we analyzed. Usually, patients stay
in the emergency department only for a short period of time; afterwards they are
either discharged or transferred to another ward. Although patients only have a
short stay, they are visited at least twice during ward rounds. Considering the
emergency situation of the patients, ward rounds take place at the beginning of
a shift. We observed three specic aspects compared to the other ward rounds
described.
The situation of all patients is discussed when exchanging treatment details
in the context of a shift change. Corresponding treatment changes are applied
afterwards by the physicians who have just started their shift.
For exchanging patient data among physicians, a whiteboard is used. This
whiteboard allows exchanging patient information between dierent shifts.
Further, it is used for making changes in patient treatment. Thereby, basic
updates of patient data are made after a ward round including the option to
erase data. Partial updates are made between two ward rounds. During this
period, information may be added or pending tasks be marked as nished.
No data may be erased from the whiteboard during this period.
In many cases, patients staying at the emergency ward are sedated or nar-
cotized, and hence are unable to communicate with physicians and nurses.
The emergency ward round comprises the following stages: In the rst stage,
an assistant physician collects data of all inpatients. Thereby, he may access
three data sources. First, he may consult the whiteboard, which is updated after
each ward round. Second, he uses his task worksheet. Third, he either may use
Supporting Ward Rounds Through Mobile Task and Process Management 15
Fig. 10.
Ward round in emergency medicine
patient data already transferred to the hospital information system or provided
by the emergency medical services that transported the patient to the hospital.
After the preparation phase, in the second stage, the ward round is performed
by the physician who prepared it, a senior physician in charge, and all physicians
taking over the shift. In this context, the senior physician is also responsible for
treatment changes. Furthermore, he may order X-ray examinations, lab tests,
and changes of the patient's medication. During this phase, the physicians having
just started their shift are only observers. However, they will perform all acquired
tasks afterwards and update the whiteboard accordingly.
As a result, physicians emphasized three crucial aspects to be considered when
replacing their pen and paper task worksheet by a digital and mobile one. Most of
them are related to the high coordination eorts and the rather critical situation
of inpatients.
Task acquisition must be easy and quick. In addition, it must be possible to
prioritize tasks according to their importance.
Gathering vital signs based on sensors (see [10] for a respective approach)
and storing recorded data in an aggregated way on the task worksheet would
be highly welcome by physicians. In addition, certain kinds of tasks should be
automatically recommended to physicians based on the vital signs collected.
The physicians crave for an easy and quick way to exchange patient data
with colleagues from other shifts.
3.3.5 Discussion
According to the ward round phases presented in Section
2, the results of our ward round analyses can be summarized as depicted in Figs.
11 and 12. We contrast the phases of the four ward rounds presented in this
section. In turn, the obtained results provide the basis for designing MEDo and
hence for replacing the paper-based task worksheet.
16 Pryss, Mundbrod, Langer, Reichert
InternalMedicineWard OrthopaedicWard
Preparewardround
Onenurse
Onephysician
Accesshospitalinformation
systemwithadesktop
computertoorganizetasksin
advance
Usewardroundtrolleywith
paper‐baseddocuments
Onenurse
Onephysician
Preparewardround
trolleywithdesktop
computer
Preparemobilehospital
informationsystem
Preparepatientvisit
(anteportas)
Notneeded,onlyfew
cliniciansparticipateand
everythingispreparedin
advance
Seniororheadphysician
usescomputeronthe
trolleyandadditionally
gathersinformationfrom
wardteam
Visitpatient
Physicianonlymakesnotes
tothoseparticulartasks,
usinghispaper‐basedtask
worksheet,whichmustbe
changedcomparedtothe
preparationdecision.He
mustmemorizewhathehas
doneduringpreparation
Seniororheadphysician
coordinatesthepatient
phaseandusesa
combinationofpaper‐
basedworksheetand
mobilehospital
informationsystemto
managethepatient
treatment
Performpost‐processing
Taskchangeswillbe
performedinthisphase
basedontheworksheet
Task,whichcannotbe
accomplishedbymobile
hospitalinformation
systemmustbe
performed
Probabilityofdelayed
performingofthepost‐
processingphase
highlow
Probabilitythatthe
physicianwhocreatesand
performstasksdiffers
highlow
Observedmedicalorders
{labtests,consultations,
medicationchanges}
{labtests,medicationchanges,
X‐rayexaminations,
appointmentswithphysical
therapists}
Generalissues
Memorizetaskchanges
Nodigitalsupportduring
patientphase
Onlypartialaccesstohospital
informationsystemwithmobile
device
Mainelectronicworksheet
demands
(inorderofimportance)
Providestatusofmedicalorders
Providehandovertonextshift
Providereal‐timevitalsigns
Enablequicktaskacquisition
Providestatusofmedicalorders
Providereal‐timevitalsigns
Fig. 11.
Analyses results for internal and orthopaedic wards
In particular, our case study revealed that quick task acquisition is demanded
by all clinicians. Furthermore, except for the trauma surgery ward, all clinicians
asked for a quick access to the status of their medical orders. Interestingly, the
Supporting Ward Rounds Through Mobile Task and Process Management 17
handling of orders only slightly diers between the dierent wards. Altogether,
based on these insights and the experiences we made in the context of our case
study, we are able to dene fundamental requirements to be met by any mobile
task support during ward rounds.
TraumaSurgeryWard
EmergencyWard
Preparewardround
Onenurse
Onephysician
CollectallneededX‐rays;
eithertheyareprintedor
thephysicianmemorizes
information.
Collectlaboratoryresults
Onenurse
Onephysician
Memorizeorcopy
whiteboardinformation
Preparewardroundtrolley
withdesktopcomputer
Preparepatientvisit
(anteportas)
Assistantphysiciansreport
toseniororheadphysician
X‐raysarenotviewedin
thisphase.
Tighttimescheduleandbig
wardroundteam(new
shiftisparticipating)makes
thisphaseimpossibleto
perform
Visitpatient
Seniororassistant
physiciancoordinatesthis
phase.Alltreatment
changesaredocumented
usingthepaper‐basedtask
worksheet
Seniorphysician
coordinatesthisphase.An
assistantphysicianmakes
notesonpaper‐based
worksheet
Performpost‐processing
Alltasksrelatedtopatient
treatmentwillbe
performedinthisphase
Alltasksrelatedtopatient
treatmentwillbe
performedinthisphase
Probabilityofdelayed
performingofthe
post‐processingphase
averagehigh
Probabilitythatthe
physicianwhocreatesand
performstasksdiffers
averagehigh
Observedmedicalorders
{labtests,medicationchanges,X‐
rayexaminations,consultations,
appointmentswithphysical
therapists}
{labtests,medicationchanges,X‐
rayexaminations,consultations}
GeneralissuesNodigitalaccesstoX‐rays
Nodigitalaccesstolaboratory
Exchangeinformationbetween
threedifferentmedia
Mainelectronicworksheet
demands
(inorderofimportance)
Enablequicktaskacquisition
ProvidedigitalaccesstoX‐rays
Providedigitalaccessto
laboratory
Enablequicktaskacquisition
Providereal‐timevitalsigns
Providestatusofmedicalorders
Fig. 12.
Analyses results for trauma surgery and emergency wards
18 Pryss, Mundbrod, Langer, Reichert
4 Requirements for Mobile Task and Process Assistance
This section summarizes the major requirements that can be derived from our
case study, i.e., from the analysis of the four ward rounds. First, we discuss
the requirements related to digital task creation and management. Second, we
discuss requirements of the processes supporting mobile task management.
4.1 Task Creation and Management
We rst summarize requirements raised by physicians with respect to digital
worksheets enabling mobile task management. Usually, physicians and nurses
organize their daily tasks based on paper worksheets. Thereby, task denition
is accomplished the same way. For example, physicians make handwritten notes
on their worksheet. In turn, each of these notes consists of a patient name and
descriptive text. Creating and managing tasks this way means working quickly
and being exible. Thus, tasks may be acquired in dierent context, while task
acquisition may be interrupted at any time. Exactly these two aspects have been
mentioned by physicians as major reasons for still using pen and paper. Hence,
when transferring task management to a mobile IT application, we must ensure
that its use is time-ecient, intuitive, and oers the same exibility as pen and
paper. Note that no mobile task support will be accepted by medical sta if
its use is more cumbersome and time-consuming than just using paper-based
worksheets. Fig. 13 summarizes major requirements in respect to mobile task
creation and management.
TaskRequirementsDescription
R1Managethetaskdigitallywithmobileassistance Providepatient’snameandfreetexttophysiciansinorderto
managetheirdigitaltaskentries(todo‐items)comparableto
thatofpenandpaper.
R2Accesspatientinformationproperly Providequickaccesstopatientdata,e.g.,vitaldata, medical
reports,externalappointments,medications,and
diagnostics.
R3Ensureahighinputspeedfortaskacquisition
R31Providetexttemplatesfortaskcreation
R32Enablevoicerecordingfortaskcreation
R33Enablecreationofnewtexttemplates
Enableahighinputspeedfortaskdefinition.
Providepre‐definedtexttemplatestophysiciansforcreating
tasksquickly.
Provideavoicerecordingfeaturetophysiciansfordefining
taskscomfortably.
Enablephysicianstocreateneworpersonallytailoredtext
templates.
R4Enable barcodescanningformedications,plasters,
bandages,andsoforth
Provideabarcodefeaturetoscanmedications,plasters,
bandages,ortodressmaterialusedforpatients.
R5Providefilterfunctionsfortasks Provideafilterfunctiontosavetimewhilemanagingtask
entries.
R6Organizetasksaccordingtotheirimportance Providephysicianswithafeaturetoorganizetasksaccording
totheirimportance.
Fig. 13.
Requirements for mobile task management
When meeting the requirements summarized in Fig. 13, tasks can be created
quickly. Furthermore, task management and patient data can be properly inte-
grated. However, when building our rst MEDo prototype, it turned out that
Supporting Ward Rounds Through Mobile Task and Process Management 19
task acquisition had been perceived as not being fast enough. To reach a level
of usability comparable to that of pen and paper, therefore, additional improve-
ments became necessary (cf. Requirements R
31
-R
33
in Fig. 13). Among others,
we realized a collection of text templates that may be used for creating tasks
digitally. Our tests with clinicians have shown that these text templates have in-
creased task acquisition speed signicantly. Furthermore, tasks may be acquired
and created using a voice recording feature, which we integrated with MEDo as
well.
4.2 Process Management
Flexible process support (see [11] for a survey of existing concepts and tech-
niques) is another key requirement raised by physicians. By integrating mobile
task assistance with process support, it shall become possible to keep track of
the status of their tasks (e.g., medical orders), to get aggregated overview lists,
or to receive immediate notications if a problem occurs. For example, the sta-
tus of an X-ray order or blood examination, requested during a ward round, can
be easily monitored based on corresponding processes. Figs. 14 and 15 depict
examples of characteristic processes we identied in this context.
Fig. 14.
Process for handling X-ray examinations
Regarding process-based task support, we gathered three major requirements
from physicians. First, they require a quick overview of their tasks and related
processes (cf. R
9
). Second, these processes shall be automatically triggered and
be properly coordinated by a process-aware hospital information system after
creating respective tasks (cf. R
7
). Third, the current state of a task and its
underlying process, respectively, shall be accessible and changeable based on
dierent user screens (cf. Requirements R
8
and R
9
). For example, if the results
of a lab test order arrive, the physician shall get immediate access to them.
Afterwards, he may want to directly change the state of this task and its related
20 Pryss, Mundbrod, Langer, Reichert
Fig. 15.
Process for handling blood examinations
process to "seen". In particular, such an interaction must be possible via dierent
user screens. For example, a physician may want to change the status of her task
by either using the task overview or the laboratory screens. Fig. 16 summarizes
the requirements related to such a process-aware task support.
Process Requirements Description
R7Useprocessestocoordinatetheprocessingofatask
andtokeeptrackoftheirstatus
Supportprocesses forhandlingandprocessingtasks.
Integratetheseprocessessmoothlywiththehospital
informationsystem.
R8Allowinterruptingtasksandcontinuing themlater Theprocesses aswellastheuserinterfacemustproperly
supporthealthcareprofessionalsininterruptingtheircurrent
workatanypointintimeandcontinuingitlateron.
R9Provideproperviewsonprocesses totheuser Processesshould beintuitivelypresentedtohealthcare
professionals.
Fig. 16.
Process requirements
5 Designing, Developing and Implementing MEDo
Drawing upon the presented requirements, we sketch the procedure applied for
designing, developing and implementing MEDo. In particular, we detail the sec-
ond part of the methodology presented in Section 2. Fig. 17 shows this procedure
and its seven steps.
In the rst step, we elaborated task management scenarios together with
physicians and nurses. Thereby, the results of the ward round investigation phase
were incorporated, i.e., the created process models constituted an important ba-
sis for discussing relevant issues with physicians and nurses. Based on the insights
we gained from these discussions, the decision for a high-delity prototype repre-
senting the rst MEDo application was made in a second step. Such prototypes
are functionally similar to a nal application and already implemented on the
target platform. Regarding the latter, we evaluated existing mobile tablet PC
frameworks to select one, meeting the requirements of clinicians best. In this
Supporting Ward Rounds Through Mobile Task and Process Management 21
Problem Areas Specific Problems &
Existing Solutions
Requirements Approach &
Styleguide
Prototype Suggestions
Mockups
User Feedback
Detailed Design
Define
Requirements
Elaborate
Approach
Observations
Design of
Process Models
Process Analysis
Literature
Research
Interviews
Framework
Preconditions
Programming
Application
Interviews
Usability Tests
Foundation Process Focus
on Ward Rounds
Requirements
Analysis
Detailed
Approach Implementation Evaluation Conclusion
Section 3 Section 4 Section 5 Section 6 Section 7 Section 9
1Section 5.1
Task Management Scenarios
2Section 5.2
Prototyping Method
3Section 5.3
Information Architecture
Feedback
4Section 5.4
Navigation Concept
Feedback
5Section 5.5
Application Architecture
6Section 6
End User Perspective on MEDo
7Section 7
Evaluation
Limitations
Fig. 17.
Designing, developing and implementing MEDo
context, aspects like size, weight, display resolution, viewing angle dependency,
usability and input sensibility were considered. Taking these properties into ac-
count, we decided to develop MEDo for the iPad. In a third step, the overall
information architecture for MEDo was designed, determining which activities
are provided by the application (i.e., control ow structure). To validate this
architecture, we provided it to the medical sta and asked for feedback. After-
wards, we designed mockups for MEDo's navigation concept and screen design
based on the developed information architecture. Following this, we discussed the
mockups with physicians and nurses in order to obtain feedback from them. In
this phase, several cycles have become necessary in order to meet the clinicians'
requirements. In a fth step, we designed the nal architecture and functionality
of MEDo. Sixth, we implemented the MEDo prototype for the iPad. Finally, we
evaluated the resulting prototype with physicians and nurses.
Due to lack of space, we only present selected aspects of the explained proce-
dure in detail. We focus on aspects important for understanding MEDo. Finally,
we emphasize particular challenges that emerged when implementing MEDo.
5.1 Task Management Scenarios
In order to identify a minimal set of needed screens, we identied and analyzed
frequent task management scenarios covering task acquisition, types of tasks,
task execution, and situations demanding context switches. Fig. 18 exemplarily
depicts ve of these scenarios in which tasks are primarily created and inspected.
Altogether, we identied 20 dierent scenarios in this context.
22 Pryss, Mundbrod, Langer, Reichert
Scenario Description
ScenarioIBeforeenteringtheroomofapatientduringawardround,aphysicianwantstogetaquickoverview
ofallexaminationresultsthathaverecentlyarrivedaswellastheoverallhealthconditionofthe
patient.Therefore,sheusestheMEDoscreenshowninFig.26.
ScenarioII
Afterawardround,thephysicianwantstoseeallupcomingtasksofthedayataglance.Hethenuses
thisoverviewforplanninghisworkday(cf.Fig.27a).
ScenarioIII
Duringawardround,thephysicianwantstoknowwhetheranX‐rayexaminationhas alreadybeen
requestedandwhatstatustheexaminationcurrentlyhas.Fig.27bindicatesthattheX‐rayhasarrived
andthephysiciansmayfinishtheprocessbysettingthestatetoMarkasseen.
ScenarioIV Duringawardround,asetoftasksshallbecreated.Often, aparticularphysicianmakesorders,while
anotheroneiscollectingthem.Fig.28ashowsthecreationofataskusingtexttemplates,whereas
Fig.28bshowsthecreationofalaboratoryrequestusingaspecializedlaboratorycreationscreen.
ScenarioVDuringawardround, aphysicianwantstousethevoicerecordingfeatureofhertabletPCtoquickly
createatask(cf.Fig.28a).
Fig. 18.
Examples of task management scenarios
5.2 Prototyping Method
In usability engineering, two dierent ways of prototyping can be dierentiated
in principle: low- and high-delity prototyping [12]. While the rst one is often
dened by pen and paper or coarse-grained digital representations, the latter
targets more on detailslike a nal application. A lesson we had learned in
previous projects is that experiences with smart devices like tablet computers
can hardly be imagined by clinicians. Since we had not found any comparable
studies about the usage of contemporary smart devices in clinical environments,
we preferred designing a high-delity prototype for obtaining more profound
feedback on our approach and to provide a working experience similar to the
nal application. While working on task management scenarios with medical
sta, we had learned that two particular perspectives need to be implemented:
a patient-data perspective and a combined task process perspective. As a result,
in addition to the basic screen and navigation concept, we implemented a basic
patient data management and a lightweight process engine with MEDo. Finally,
usability can be improved when making use of results from cognitive science and
applying design techniques from usability engineering, like choosing the right
colours or realizing a comprehensible and useful segmentation of the application
screen. Both interviews and usability tests with medical sta helped us to gure
out which user interface elements are intuitive for them.
5.3 Information Architecture
A fundamental aspect of any mobile smart device application is its information
architecture, i.e., the fundamental control ow structure of its user interface.
For example, in our context, it is crucial to allow medical sta members to
interrupt the processing of a task at any point in time and to continue it later.
Therefore, users should be enabled to switch between dierent screens of the
mobile application. Next, we utilized the collected task management scenarios
(cf. Section 5.1) in order to dene the basic functions to be covered by MEDo
and its information architecture (cf. Fig. 19). In this context, functions IAF
1
-
IAF
3
refer to the collaboration among medical sta members in the context of
Supporting Ward Rounds Through Mobile Task and Process Management 23
medical tasks. Obviously, collaboration is crucial during ward rounds and hence
should be smoothly integrated with any task management component.
InformationArchitecture Functions Description
IAF1Sharethetasklistwithotherphysiciansandnurses Sharetaskswithotherphysiciansandnurses.
IAF2Provideamulti‐usermodefortasks Providedifferentscreens forphysiciansandnurses.
IAF3Considertasklistswhenswitchingbetweentwo
shifts
Electronicworksheetshavetobeeasilyhandedover
betweenthestaffofashift.
IAF4Providetheuserwithspecialscreens showing
upcomingtasks
Thehandling ofupcomingtasksshallbeeased.Accessto
themmustbequickandeasy.
IAF5Providetheuserwithscreens summarizing
examinationresults
Afastandadequateaccesstoexaminationresultsis
required.
Fig. 19.
Information architecture functions
Based on these insights, we came to the conclusion that tasks shall constitute
the predominant paradigm for user interaction in MEDo's information architec-
ture (cf. Fig. 20). Therefore, a physician does not need to switch screens if she
wants to create a new task, but she can do this at any point during interaction.
For instance, a request for an X-ray examination may be entered using the task
screen or alternatively through every other screen (e.g., the screen showing a
patient's lab results).
5.4 Navigation Concept
An information architecture denes the user screens to be designed and imple-
mented as well as the control ow between them. In order to illustrate how we
designed this information architecture, we present selected aspects of the patient
screen. The rst issue to be addressed concerns the visual design and layouting
of the screen. Since dierent types of patient information shall be displayed, a
proper design is crucial for later user acceptance. Together with physicians and
nurses, we elaborated three dierent options how this patient screen may look
(cf. Fig. 21). After discussing the options with the clinicians, it turned out that
the third one ts best to their needs (see the arrow marked with 3 in Fig. 21). In
particular, they perceived this option as the most intuitive one since it is similar
to a paper-based patient record. In turn, in order to properly realize this design,
we decided to use
tabs
for accessing the dierent kinds of patient information
(e.g., laboratory, vital signs, and medical reports).
Our rst mockup related to tabs is depicted in Fig. 22. Note that we use one
mockup to illustrate tabs in general and one to show how tabs can be realized
on an iPad. By default, tab controls are placed on the bottom of an iPad screen.
Clinicians perceived this tab positioning as not being very useful. Taking their
feedback into account, we designed tab controls in a dierent way according to
the mockup shown in Fig. 23. As a result, we had to implement new tab control
elements for MEDo.
24 Pryss, Mundbrod, Langer, Reichert
Fig. 20.
MEDo information architecture
Fig. 21.
Patient screen options
Supporting Ward Rounds Through Mobile Task and Process Management 25
Fig. 22.
First mockup of a patient screen with tabs
Fig. 23.
Final mockup of a patient screen with tabs
In general, every MEDo screen was designed and implemented stepwise through
feedback provided by ten clinicians (3 male and 3 female physicians; 4 nurses).
Altogether, it took us two months to accomplish this design of the MEDo ap-
proach. In future work, we will also measure how long physicians and nurses
spend on working with the dierent screens in order to enhance overall usability.
5.5 Application Architecture
As described in Section 5.2, MEDo contains a basic patient data management
as well as a lightweight process engine. For this purpose, a data model provides
a solid foundation. Fig. 24 exposes this data model featuring dierent semantic
areas: data related to task and process management, data related to patients,
and data related to ward and user management.
Regarding the implemented process engine, we only present selected aspects
due to lack of space. Fig. 25 depicts the interface of the engine as well as an
XML snippet showing a process denition in MEDo.
26 Pryss, Mundbrod, Langer, Reichert
Fig. 24.
MEDo data model
A MEDo process comprises three important components: phrases, tasks, and
actions. Phrases identify a particular process. Hence, if a physician uses one of
these phrases, the respective process will be started. Then, the tasks will be
performed in sequence as dened (i.e., from the top to the bottom). Finally,
the
action
command allows interrupting users to display a pop-up screen. In
response to that, a user must actively conrm the action. Note that this feature
has been rated as quite important by clinicians.
6 End User Perspective on MEDo
To give an idea how MEDo works in practice, we present selected features.
For this purpose, we introduce user interface screens related to the scenarios
we described in Fig. 18 and discuss how they have been realized in MEDo.
Regarding
Scenario I
(cf. Fig. 18), Fig. 26 shows a screen displaying all relevant
information about a particular patient at a glance. This screen was used by
physicians most often when using MEDo. Physicians pointed out that getting
such an overview at a glance is very crucial for them. Further, they conrmed
that the option to quickly switch to a particular patient screen is very useful for
their daily routines. Fig. 27a shows the MEDo entry screen and
Marking 1
refers
to the list of all patients. In MEDo, patients may either be listed alphabetically
or according to the rooms they are assigned to.
Supporting Ward Rounds Through Mobile Task and Process Management 27
<Processes>
<Processid=”BloodAnalysis”title=”BloodAnalysis”>
<Phrases>
<Phrase>LaboratoryBloodTest</Phrase>
<Phrase>TakeBloodSample</Phrase>
<Phrase>BloodAnalysis</Phrase>
<Phrase>BloodExamination</Phrase>
<Phrase>BloodTest</Phrase>
</Phrases>
<Tasks>
<Taskid=”1”role=”nurse”>
Preparepatientforbloodexamination,
preparetesttubes
</Task>
<Taskid=”2”role=”physician”>
Takebloodsample
</Task>
<Taskid=”3”role=”laboratory”>
Performbloodanalysis
</Task>
<Taskid=”4”role=”physician”action=”labResult”>
Evaluatebloodtestresults,
confirmanalysis
</Task>
</Tasks>
</Process>
<Processid=”XrayExamination”title=”XrayExamination”>
…
</Process>
<Processid=”Council”title=”Council”>
…
</Process>
</Processes>
@interface MDProcessManager :NSObject
+ (Task*)advanceProcessInstance:(ProcessInstance*)processInstance
advanceAutomatically:(BOOL)automatically
withDelay:(BOOL)withDelay;
+ (Task*)advanceProcessInstance:(ProcessInstance*)processInstance;
+ (id)processInstanceWithTask:(Task*)task
andPatient:(Patient*)patient;
@end
Action
Fig. 25.
MEDo XML process denition
In Fig. 27a,
Marking 2
refers to both upcoming and recently completed tasks.
This screen is used by the physicians in the context of
Scenario II
(cf. Fig. 18).
In turn,
Marking 3
shows the processing state of a selected task (X-ray request).
For example, when an X-ray image arrives, a respective symbol indicates the
status change to the physician. In order to study the results obtained, in turn,
the physician may switch to the imaging screen (cf. Fig. 27b).
Marking 4
refers
to an element allowing the physician to change the state of the X-ray request
to "nished" (cf.
Scenario III
in Fig. 18). Accordingly, the corresponding pro-
cess completes. In Fig. 28a,
Marking 5
shows the pre-dened text templates,
the physician may use when dening a task (cf.
Scenario IV
in Fig. 18). Note
that this feature has turned out to increase task acquisition speed signicantly.
In addition, related processes can be automatically derived from task denitions
based on the text templates provided. For instance, if a physician uses text tem-
plate
Request Council
for dening a task, a corresponding process will be started
28 Pryss, Mundbrod, Langer, Reichert
Fig. 26.
Patient detail screen
in the background. In addition to text templates, a voice feature has been inte-
grated (cf. Marking 6), i.e., the physician may use her voice for recording tasks
(cf.
Scenario V
in Fig. 18). However, it is not yet possible to derive respective
processes directly from these voice recordings. Fig. 28b depicts the form of a lab
test order (cf.
Scenario IV
in Fig. 18).
Marking 7
exemplarily shows dierent ways of creating tasks in the context of
the laboratory screen. When interviewing medical sta, we analyzed how many
task creation buttons are useful for a particular MEDo screen and how they
shall be positioned on it. Moreover, the user interface was optimized in several
cycles based on the feedback provided by medical sta. Finally, Fig. 29 shows
how to use the
action
command of the process engine (cf. Fig. 29a) and how to
smoothly switch between dierent patients (cf. Fig. 29b). In detail, in Fig. 29a, a
physician must conrm the results of a blood examination, while in Fig. 29b, he
may access the list in the patient screen to see to which patients he may switch.
Supporting Ward Rounds Through Mobile Task and Process Management 29
(a) Patient overview list (b) Evaluation of X-ray image
Fig. 27.
Patient overview list and X-ray diagnostic screen
(a) Create task view (b) Request for a laboratory
Fig. 28.
Task creation and laboratory diagnostic screen
30 Pryss, Mundbrod, Langer, Reichert
(a) Conrm blood examination re-
sults (b) Smooth switch to another pa-
tient
Fig. 29.
Conrm blood examination results and smooth switch to another patient
7 MEDo Evaluation
In this section, the results of using MEDo are compared with the ones from
Section 3, in which we evaluated task management performed with pen and
paper. To evaluate the usage of MEDo, the procedure shown in Fig. 30 was
applied. Overall, 16 physicians (8 female, 8 male) and 9 nurses participated in
the evaluation. Thereby, half of the participants had already worked with MEDo,
while the other half did not. Based on the experiences reported in comparable
work [13] as well as the knowledge-intensive character of ward rounds, we decided
to conduct the evaluation procedure in a qualitative manner. Consequently, the
test, which was not an experiment but rather a simulation, was performed in
three phases: rst interview phase, ensuing testing phase, and a second interview
phase. The rst interview round was based on general questions, while the last
one comprised specic questions regarding the use of MEDo. The testing phase
was accomplished in the following way: First, we explained four rules to the
participants.
1. Participants may nish or interrupt the test whenever they want.
2. Participants shall think aloud.
3. If a participant is not able to nish a task, she may skip it.
4. The observer does not assist or help the participants.
Second, we presented a ward round scenario to the participants. Then, a
participant was asked to accomplish nine tasks regarding the described scenario.
Six tasks are related to the situation during a ward round, while three refer to
the situation after a ward round.
Supporting Ward Rounds Through Mobile Task and Process Management 31
Fig. 30.
Evaluation procedure
Due to limited space, we only present selected results of the evaluation (cf.
Figs. 31-33). Fig. 31 shows how the participants perceived the overall information
architecture as well as screen navigation provided by MEDo (on a scale with 6
values within the interval from 1 (very good) to 6 (inadequate)).
Question MeanValue StandardDeviation
ScreenNavigation 2.00 0.47
OverallImpression 1.90 0.50
|Scalerangehas6valueswithintheinterval:verygood–inadequate|
Fig. 31.
Information architecture evaluation
In Section 3 (cf. Fig. 3), we presented results to questions how clinicians
perceive overall task and ward round management without using any digital
support. In turn, Fig. 32, shows results when using MEDo for task and ward
round management.
Question MeanValue StandardDeviation
TaskDefinition 1.90 0.50
WardRoundManagement 1.80 0.53
|Scalerangehas6valueswithintheinterval:verygood–inadequate|
Fig. 32.
Task denition and overall ward round management without MEDo support
Question MeanValue StandardDeviation
PatientCommunication 3.67 1.66
|Scalerangehas6valueswithintheinterval:verygood–inadequate|
Fig. 33.
Patient communication evaluation when using MEDo
Finally, Fig. 33 shows how clinicians perceived patient communication using
MEDo. In summary, physicians classied the task creation with MEDo on the
same production level as the one based on pen and paper. Furthermore, they
32 Pryss, Mundbrod, Langer, Reichert
appreciated the use of MEDo as process-aware assistance during ward rounds.
Finally, they considered the close integration of (patient) data and process as
one of the major benets of using MEDo.
However, when interviewing with physicians, we also learned that patient com-
munication was aected when using MEDo; i.e., intensively focusing on the
mobile application might distract users from inpatients (cf. Fig. 33). Further
research is needed to deal with this important issue. We further learned that
the status of a process should be presented more intuitively to clinicians. Many
sta members proposed a status notication similar to trac lights, i.e., using
three dierent colours for representing status information. Furthermore, physi-
cians missed subcategories regarding the pre-specied text templates. Another
demand was to integrate additional patient information in MEDo, e.g, reports
from practitioners or surgery appointments. Finally, MEDo is supposed to be
smoothly integrated into existing hospital information systems. Overall, physi-
cians were satised with using MEDo (see and compare results from Figs. 3 and
32). After completing this test period, physicians pointed three additional issues
not been considered so far.
The buttons for creating tasks shall be individually placeable on MEDo
screens. For creating a task, each physician has specic preferences depending
on her context.
Tasks should be better integrated with vital data signs and medication
schemas. For example, if the medication of a patient expires, a reminder
task shall be automatically added to the physician's worksheet in order to
indicate that another medication might have to be created.
Documentation becomes more and more complex in hospitals. Thereby, many
activities are necessary which are not related to patient treatment. For ex-
ample, legal requirements require to maintain statistical records. Physicians
therefore asked for features enabling them to nish such activities.
Regarding the rst issue, the described information architecture must be ex-
tended accordingly. The second issue can be solved through the integration of
advanced process support features. Finally, the third issue raised needs to be
investigated in detail in future research.
Limitations.
Our evaluation on MEDo has the following limitations: First,
the corresponding questionnaire was designed without the help of experts on
survey design and therefore further evaluation is required to conrm the sat-
isfactory results. In particular, the likert-scale was not symmetric and further
eorts are necessary for the next evaluation cycle. Second, the number of subjects
(physicians and nurses) in our exploratory study is relatively low (25 subjects),
hampering result generalization. Nevertheless, it is noteworthy that the sample
size is not unusual for this kind of empirical investigation due to the substantial
eort to be spent per subject [14]. In particular, while evaluating MEDo with
professionals, it frequently happened that they were interrupted or distracted
by unforeseen situations around their patients, which made our investigation
Supporting Ward Rounds Through Mobile Task and Process Management 33
time-consuming, complex and challenging. Third, all participating subjects have
been either ward physicians or nurses. In addition, physicians not directly par-
ticipating in the ward rounds (e.g., physicians from the X-Ray unit), physicial
therapists, and other professionals (e.g., from the laboratory) have not been
involved so far, limiting result generalization. However, subjects indicated pro-
found background knowledge regarding the issues of other professionals. Hence,
we argue that they can be interpreted as proxies.
8 Related Work
In [1517], approaches adopting mobile information technology to ward rounds
are described. However, none of them focuses on mobile task assistance com-
bined with a process support component, as done in the context of the MEDo
approach.
The general importance of task support, in particular for scenarios characterized
by concurrent tasks, is emphasized in [18]. Moreover, [18] considers scenarios for
task support as well. Compared to MEDo, however, it does not consider process
support as enabler for sophisticated task assistance. Besides, the information
architecture used in [18] is not specically tailored towards medical demands.
Other related work deals with the support of clinicians at the point of care [19,
20]. However, an integrated view on patient and task status information, as pro-
vided by MEDo, is not available.
Another research eld related to MEDo concerns the replacement of the paper-
based patient record by an electronic surrogate [21,22]. Again, existing work
does not focus on how the physicians organize their daily work. Similar concerns
hold for approaches supporting ward rounds based on process management tech-
nology [5,2327] or transferring information technology to ward rounds [28,13].
In addition, industrial solutions need to be considered, like the one running in
the orthopaedic ward round (cf. Fig. 4) [29]. Either they focus on task man-
agement or mobile access to the hospital information system, but none of them
integrates task management into daily work like MEDo does.
MEDo can be related to research targeting at the support of knowledge-intensive
processes. Existing approaches like Declare [30], [31], ADEPT2, ProMInanD [32]
and AristaFlow, for example, allow for an ad hoc composition of a process. Typ-
ically, these approaches rely on proper specications of tasks and their man-
agement as well as their interdependencies. In contrast, MEDo is a lightweight
collaborative task management tool, which allows physicians to focus on what
tasks are supposed to be done instead of how. Hence, complex interdependencies
between tasks are excluded.
In the context of knowledge-intensive processes, emerging approaches belong
to the initiative of
Adaptive Case Management
and hence allow for a case and
data-centric support, respectively, of knowledge workers [2,33]. These approaches
dier from those dealing with traditional predened business processes. Hence,
MEDo can be regarded as contribution to this initiative as well.
Much research eort has been spent on examining how communication can be
34 Pryss, Mundbrod, Langer, Reichert
improved among (healthcare) professionals during ward rounds [34,35]. Finally,
many studies can be found which underpin the goals of MEDo [36,23,37], but
none of the approaches provided focuses on proper mobile task assistance in
combination with smart process support.
9 Summary and Outlook
Existing information technology does not consider requirements of hospital ward
rounds properly. With MEDo, we suggest mobile task management as a rst
step towards exible user support during a ward round in particular and im-
proved IT support of knowledge-intensive processes in general. Combining task
management with process support has proven to be benecial for healthcare
professionals in this context. However, to use processes eectively, they must be
well elaborated. We presented an in-depth analyses of four wards rounds and
discussed them regarding their suitability for process and mobile task support.
The insights we gained in this context have been used to tailor MEDo to the
practical demands of clinicians. In particular, we illustrated aspects we consider
as crucial regarding the replacement of paper-based worksheets. Finally, we have
shown that MEDo eases overall ward round management.
In future research, we will extend MEDo by adding features like the ones dis-
cussed in Section 8. Furthermore, we plan to integrate two additional features.
First, we want to integrate temporal aspects with process support of MEDo. To
dene, for example, when a process related to the planning of a surgery must
be nished, is crucial in the context of ward rounds. Second, MEDo does not
provide a feature to properly show all running processes at a glance. We have
learned that users would consider this feature to be quite useful in certain cases.
Furthermore, we plan evaluating MEDo under rigorous utilization. Thereby, we
focus on unfavorable aspects (e.g., limited battery) imposed by mobile technolo-
gies in general and the used iPad for MEDo in particular. These insights shall
be utilized to further improve MEDo.
Finally, the usage of best practice checklists represents an emerging and impor-
tant aspect of everyday life in hospitals [38,39]. Following this, we plan to inte-
grate corresponding results of dealing with paper-based checklists into MEDo.
However, as discussed, using mobile technology during ward rounds might dis-
tract physicians and nurses (see [40] for similar concerns), an issue to be inves-
tigated in future research.
References
1. Davenport, T.H.: Thinking for a living: how to get better performance and results
from knowledge workers. Harvard Business School Press (2005)
2. Swenson, K.D., ed.: Mastering the Unpredictable. Meghan-Kier-Press (2010)
3. Davenport, T.H.: Working knowledge: How organizations manage what they know
(excerpt) (23.03.2011)
Supporting Ward Rounds Through Mobile Task and Process Management 35
4. Mundbrod, N., Kolb, J., Reichert, M.: Towards a system support of collaborative
knowledge work. In: 1st Int'l Workshop on Adaptive Case Management (ACM'12),
Proc. BPM'12 Workshops. Number 132 in LNBIP, Springer (2012) 3142
5. Dadam, P., Reichert, M., Kuhn, K.: Clinical workows - the killer application for
process-oriented information systems? In: Proc. 4th Int'l Conference on Business
Information Systems (BIS'00), Springer (2000) 3659
6. Lenz, R., Reichert, M.: IT support for healthcare processes - premises, challenges,
perspectives. Data and Knowledge Engineering
61
(2007) 3958
7. Kennedy, M.M., Baker, D.J., Gurses, A.P., Pronovost, P.J.: Creating the ideal
ward round. Organisation and management of intensive care. Medizinisch Wis-
senschaftliche Verlagsgesellschaft, Berlin (2010) 259267
8. Pryss, R., Langer, D., Reichert, M., Hallerbach, A.: Mobile task management for
medical ward rounds - the MEDo approach. In: 1st Int'l Workshop on Adaptive
Case Management (ACM'12), Proc. BPM'12 Workshops. Number 132 in LNBIP,
Springer (2012) 4354
9. Eichwald, A.: Einsatz von prozess-management-technologie in der klinischen visite
mithilfe der aristaow bpm suite, Bachelor thesis, University of Ulm (2012)
10. Schobel, J., Schickler, M., Pryss, R., Nienhaus, H., Reichert, M.: Using vital sensors
in mobile healthcare business applications: challenges, examples, lessons learned,
9th Int'l Conference on Web Information Systems and Technologies WEBIST
2013 (2013) 509518
11. Reichert, M., Weber, B.: Enabling Flexibility in Process-Aware Information Sys-
tems: Challenges, Methods, Technologies. Springer, Berlin-Heidelberg (2012)
12. Nielsen, J.: Usability engineering. Access Online via Elsevier (1994)
13. Ammenwerth, E., Buchauer, A., Bludau, B., Haux, R.: Mobile information and
communication tools in the hospital. Medical Informatics
57
(2000) 2140
14. Costain, G.: Cognitive Support during Object-Oriented Software Development:
The Case of UML Diagrams. PhD thesis, ResearchSpace, Auckland (2008)
15. Carlsson, V., Klug, T., Ziegert, T., Zinnen, A.: Wearable Computers in Clinical
Ward Rounds, Berlin/Oenbach, Germany. (2006)
16. Lamber, P., Ludwig, B., Ricci, F., Zini, F., Mitterer, M.: Message-based patient
guidance in day-hospital. In: Proc. IEEE 12th Int'l Conference on. MDM '11, IEEE
Computer Society (2011) 162167
17. Kofod-Petersen, A.: Challenges in case-based reasoning for context awareness in
ambient intelligent systems. In: 8th European Conference on CaseBased Reasoning,
Ölüdeniz. (2006)
18. Camacho, J., Favela, J., Gonzalez, V.: Supporting the management of multiple
activities in mobile collaborative working environments. In: Groupware: Design,
Implementation, and Use. Volume 4154 of Lecture Notes in Computer Science.
Springer Berlin Heidelberg (2006)
19. Dawson, L., Fisher, J., Heslop, L.: Investigating the potential improvement of
patient anagement systems in hospital ward settings using mobile, wireless tech-
nologies. Mobile Government Consortium International LLC (2005)
20. Sawa, T., Okahara, M., Santo, M., Schmidt, U., Nakata, Y., Morita, S., Ohno-
Machado, L.: Preoperative information management system using wireless PDAs.
Proc. AMIA Annual Symposium Proc. (2003)
21. Fitzpatrick, G.: Understanding the paper health record in practice: Implications
for EHRs (2000)
22. Knaur, P., Pilz, J., Thalheimerc, M.: Can temporary paper-based patient records
sensibly complete an electronic patient record? In: Ubiquity: Technologies for
36 Pryss, Mundbrod, Langer, Reichert
Better Health in Aging Societies, Proceedings of Mie2006. Volume 124., IOS Press
(2006) 207
23. Heslop, L., Raikundalia, G.K.: Enhancing clinical nurse workow through redesign
of networked wireless laptop computers. Electronic Health Informatics
6
(2011)
e25
24. Fitterer, R., de Witte, B.: Enabling pervasive healthcare by means of event-driven
service-oriented architectures. the case of bed management in mid-sized to large-
sized hospitals. In: Pervasive Computing Technologies for Healthcare, 2009. Per-
vasiveHealth 2009. 3rd Int'l Conference on. (2009) 14
25. Ali Fareedi, A., Tarasov, V.: Modelling of the ward round process in a healthcare
unit. Volume 92 of Lecture Notes in Business Information Processing. Springer
Berlin Heidelberg (2011) 223237
26. Reichert, M.: What BPM technology can do for healthcare process support. In:
Proc. AIME'11. Number 6747 in LNAI, Springer (2011) 213
27. Wang, L.: An agent-based simulation for workow in emergency department. In:
Proc. 2009 IEEE Systems and Information, IEEE (2009) 1923
28. Bardram, J.E., Bossen, C.: A web of coordinative artifacts: collaborative work at
a hospital ward. In: Proc. GROUP '05, ACM (2005) 168176
29. SAP: Give caregivers secure mobile access to their patient data - with our
mobile patient record apps.
http://www54.sap.com/pc/tech/mobile/software/
industry-apps/patient-record-app/index.html
(2013) [Online; accessed 06-
Mai-2013].
30. Schonenberg, H., Weber, B., Dongen, B., Aalst, W.: Supporting exible processes
through recommendations based on history. In Dumas, M., Reichert, M., Shan,
M.C., eds.: Business Process Management. Volume 5240 of Lecture Notes in Com-
puter Science. Springer Berlin Heidelberg (2008) 5166
31. McCarthy, D.R., Sarin, S.K.: Workow and transactions in InConcert. IEEE Data
Eng. Bull.
16
(1993) 5356
32. Karbe, B., Ramsperger, N., Weiss, P.: Support of cooperative work by electronic
circulation folders. SIGOIS Bull.
11
(1990) 109117
33. Chiao, C.M., Künzle, V., Reichert, M.: Object-aware process support in healthcare
information systems: Requirements, conceptual framework and examples. Interna-
tional Journal on Advances in Life Sciences
5
(2013)
34. Alvarez, G., Coiera, E.W.: Interruptive communication patterns in the intensive
care unit ward round. I. J. Medical Informatics
74
(2005) 791796
35. Weber, H., Stöckli, M., Nübling, M., Langewitz, W.A.: Communication during
ward rounds in internal medicine. an analysis of patient-nurse-physician interac-
tions using RIAS. Journal Patient Education and Counseling
67
(2007) 343348
36. Moran, E.B., Tentori, M., Gonzalez, V.M., Favela, J.: Mobility in hospital work:
towards a pervasive computing hospital environment. Int'l Journal of Electronic
Healthcare
3
(2007) 7289
37. Ho, D., Xiao, Y., Vaidya, V., Hu, P.: Communication and sense-making in intensive
care: an observation study of multi-disciplinary rounds to design computerized
supporting tools. In: Proc. AMIA Annual Symposium. Volume 2007., American
Medical Informatics Association (2007)
38. Haynes, A.B., Weiser, T.G., Berry, W.R., Lipsitz, S.R., Breizat, A.H.S., Dellinger,
E.P., Herbosa, T., Joseph, S., Kibatala, P.L., Lapitan, M.C.M., Merry, A.F., Moor-
thy, K., Reznick, R.K., Taylor, B., Gawande, A.A.: A surgical safety checklist to
reduce morbidity and mortality in a global population. New England Journal of
Medicine
360
(2009) 491499
Supporting Ward Rounds Through Mobile Task and Process Management 37
39. Weiser, T.G., Haynes, A.B., Dziekan, G., Berry, W.R., Lipsitz, S.R., Gawande,
A.A.: Eect of a 19-item surgical safety checklist during urgent operations in a
global patient population. Journal Annals of Surgery
251
(2010) 976980
40. Alsos, O.A., Dabelow, B., Faxvaag, A.: Doctors' concerns of PDAs in the ward
round situation. lessons from a formative simulation study. Methods of Information
in Medicine
50
(2011) 190200
