VALUE DRIVEN KPI DESIGN FOR HEALTH CARE
Paul Johannesson and Erik Perjons
Department of Computer and Systems Sciences, Stockholm University/Royal Institute of Technology
Forum 100, SE-164 40, Kista, Sweden
Keywords: Performance management, KPI, Enterprise modelling, Service quality, SERVQUAL.
Abstract: A core component in any welfare society is an equal and efficient health care system. Large resources are
spent on health care, but a number of problems still remain, such as deficiencies in service quality and long
patient queues. A main reason behind these problems is the complexity of health care organisations which
make them difficult to govern. One possible approach for improving the governance and management of
health care organisations is the use of performance management, which is a management approach aiming
at optimising the performance of an organisation. The main instrument in performance management is the
Key Performance Indicator (KPI). However, designing effective KPIs is a complex and time consuming
task that requires substantial efforts. Therefore, there is a need for methods and tools that assist
organisations in designing KPIs. This paper proposes a value driven method for identifying patient oriented
KPIs. The method is illustrated using a case study in eye health care.
1 INTRODUCTION
Large resources are spent on health care in the
European welfare society, but a number of problems
still remain, including unequal access to health care,
large variations in outcomes of treatments,
deficiencies in service quality, and inefficient
resource use. A main reason behind these problems
is the complexity of the health care sector which
makes it problematic to govern. For example, health
care needs to fulfil several and often contradictory
goals, such as equal, high quality and efficient health
care. At the same time, a large number of
stakeholders need to interact with each other in order
to ensure the delivery of high-quality health care.
Furthermore, European health care faces a period of
potentially profound changes in social attitudes,
demographic structure, economic conditions and
medical technologies.
In order to better govern the health care sector, there
is an increasing interest in performance
management, which is a management approach
aiming at optimising the performance of an
organisation, i.e. optimising its business processes
and outcomes (Walburg, 2006). The main activities
in performance management are identifying business
goals, developing key performance indicators
(KPIs), monitoring and measuring the performance
using the KPIs, analysing the results of the
measurements, and acting in order to better fulfil the
goals of the organisation (Eckerson, 2006). Thus,
performance management can be an approach for
managing efficiency, cost, and quality in any
business, including health care.
The main instrument in performance management is
the KPI, which we here define as a property of a
phenomenon that can be used to measure the
performance of an organisation. Examples of KPIs
are “sales revenue”, “patient satisfaction”, “patient
throughput”.
KPIs operationalise business goals, i.e. vague and
high-level goals can be expressed in terms of KPIs,
thereby making them concrete and easy to
understand (Eckerson, 2006). For example, a goal
like “Patients should get fast treatment for serious
diseases” can be operationalised as “Waiting time
for hip replacement surgery should be less than one
month during 2010”, which contain a KPI (“Waiting
time for hip replacement surgery”) and a target
(“less than one month during 2010). Such a
statement can easily be monitored and measured.
KPIs are powerful instruments for governing an
organisation, as they are easily understandable,
actionable, and can be effectively monitored.
However, this also means that the use of
121
Johannesson P. and Perjons E. (2010).
VALUE DRIVEN KPI DESIGN FOR HEALTH CARE.
In Proceedings of the Third International Conference on Health Informatics, pages 121-131
DOI: 10.5220/0002749801210131
Copyright
c
 SciTePress

inappropriate KPIs can have significant negative
effects on the performance of an organisation
(Eckerson, 2006). For example, using KPIs focusing
only on quantity instead of quality can divert
employees from paying attention to quality in
products and services. Thus, it is essential to design
and implement effective KPIs that actually improve
the performance of an organisation in different
aspects, such as efficiency, cost and quality.
Designing effective KPIs is a complex and time
consuming task, which requires substantial effort
from performance management experts as well as
domain experts, such as physicians and nurses (Jun
et al, 2006). In particular, it is difficult to ensure
completeness in KPI design, i.e. when a set of
potential KPIs have been identified, there is still a
risk that relevant and effective KPIs have been
omitted. Another issue in KPI design for healthcare
is the fact that KPIs aimed at measuring service
quality in many cases cannot be directly understood
by patients because of the complex relationship
between diagnostic and therapeutic services and
their results (Jun et al, 2006). This is different from
most other industries, where service quality almost
always can be directly perceived by service
customers.
The goal of the paper is to propose a method for
constructing a complete set of patient oriented KPIs.
(KPIs measuring internal efficiency are not
addressed.) Section 2 gives an overview of related
work, in particular performance management and
service quality analysis. Section 3 presents the
method that addresses both technical and functional
service quality. Section 4 provides an illustration of
the method based on a case study from the eye care
health sector in Stockholm. Finally, Section 5
summarises the results of the paper and gives
suggestions for future work.
2 RELATED WORK
Performance management, i.e. measuring
performance of an organization, has been an
important management instrument in business for
the past 20 years (Adair et al, 2003). In recent years
the use of performance management has accelerated,
as evidenced by the proliferation of performance
management methods, key performance indicators
(KPI), and IT support such as business intelligence
and performance dashboard systems. Health care is
no exception to this trend; performance management
has attracted substantial attention among
governmental and private funders of health care,
health care providers, patient organizations as well
as researchers in health care management (Adair et
al, 2003). This interest has emerged in parallel with
the interest for other, and closely related, health care
improvement approaches, such as accreditation,
service evaluation, quality improvement, external
auditing, outcome research and evidence-based
medicine (Adair et al, 2003). However, there is no
clear demarcation line between these approaches and
performance management, as different practitioners
and researchers use different definitions and
concepts.
Measuring performance is about measuring the
outcomes and quality of business processes. In
health care, there are two main categories of quality:
technical service quality, i.e. clinical results of
health services, and functional service quality, i.e.
patient assessments of the quality of care (Jun et al,
2006), see also (Gronroos, 1984). In general,
technical service quality is difficult to understand
and measure for both health care providers, health
care funders, and patients (Jun et al, 2006).
Therefore, many performance management
initiatives in healthcare have focused on functional
service quality.
A main research theme within functional service
quality has been to find the right balance of service
quality dimensions. As a starting point, a well-
known service quality framework, SERVQUAL
(Parasuranam et al, 1988), has frequently been used.
SERVQUAL measures the gap between perceived
and expected service quality, using five quality
dimensions: reliability, assurance, tangibles,
empathy and responsiveness. Often used is also a
predecessor to SERVQUAL (Parasuranam et al,
1985), which includes ten quality dimensions. In
SERVQUAL these ten dimensions were condensed
to five, each with a set of sub dimensions, called
items. SERVQUAL was meant to be a service
quality framework independent of business area.
However, when used in health care, further
dimensions are often added, commonly created
using focus groups with physicians, nurses, health
care managers and patient representatives, see for
example (Gupta, 2008), (Kilbourne et al, 2004), and
(Karassavidou et al, 2009). The relationship between
service quality in health care and patient satisfaction
is also a complex issue, discussed in several research
papers, see, for example (Gill and White, 2009).
For identifying KPIs, the most common tool is the
use of goals and objectives (Eckerson, 2006).
Usually the goals and objectives of an organization
HEALTHINF 2010 - International Conference on Health Informatics
122

are structured graphically in goal models, visualizing
an hierarchy of strategic, tactical and operational
goals.
Goal models are examples of enterprise models,
while other examples are value models, process
models, conceptual and information models. In
general, enterprise models offer graphical
representations of the structure, goals, processes,
information, resources, people, and constraints of an
organisation. They provide compact and graphical
descriptions of an organization and its environment,
which makes them ideal for supporting
communication between different stakeholders.
They can work as a base for creating a common
understanding of the organization, for change
management, as well as IT system design. Process
and conceptual models have been used extensively
both in health management research and health care
standards and specifications, such as SAMBA,
HISA, HL7, CONTSYS. Recently, value models
have gained increased attention (Kartseva et al,
2007), (Johannesson et al, 2009). In this paper, value
models and conceptual models are used as the basis
for identifying KPIs. To our knowledge, this has not
been done in previous work.
3 METHOD FOR VALUE BASED
KPI IDENTIFICATION
The purpose of the method proposed is to assist a
KPI designer in creating a set of KPIs that can be
used to assess the quality of health care services
from a patient point of view. An overview of the
method is depicted in Figure 1.
The first task to be carried out is to create a value
model scoping the health care scenario under
consideration, thereby determining the most
important health care services as well as other
services in the scenario. The second task aims at
generating KPIs for technical service quality that
concern the results of health services and the
consequences of health issues. For this task,
we introduce a supporting instrument for
representing relevant health care services and health
issues, in the form of a so called standard service
and issue model (see Section 3.2) that can be
modified and extended to fit the current health care
scenario. The third task aims at generating KPIs for
functional service quality that concern the manner in
which services are delivered to patients. For this
Task1: Develop
value model
Task 3: Generate
functional KPIs
Task 4: Select KPIs
Step1: Identify
complementary
services
Step3: Apply the
service dimension
model
functional KPIs
technical KPIs
complementary
services
services
health care
services
Step2: Adapt
standard service
dimension model
adapted SDM
Plug-in: Stan dard
service dimen-
sional model
Step 1: Adapt
standard service
and issue model
Task 2: Generate
technical KPIs
Step2: Apply
generic KPI
templates
Step3: Apply
specific KPI
templates
adapted SIM
specific KPI
templates
Plug-in: Standard
service and issue
model
technical
KPIs
Figure 1: An graphical overview of the method for value
based KPI identification.
purpose, we introduce a supporting instrument in the
form of a so called standard service dimension
model (see Section 3.3), based on the predecessor to
SERVQUAL (Parasuranam et al, 1985), which can
be modified and extended to fit the current health
care scenario. Finally, there is a reviewing and
filtering task where the KPIs suggested in the
previous tasks are evaluated and those deemed most
important are selected as candidates for
implementation.
3.1 Task 1: Develop Value Model
A first task is to delimit the domain for which the
KPIs are to be designed. For this purpose, a value
model is to be created. The value model shall
describe the actors of the value network under
consideration, such as health care providers and
VALUE DRIVEN KPI DESIGN FOR HEALTH CARE
123

patient. The value model shall also show the
resource exchanges between these actors, including
exchanges of services such as examinations and
treatments. The value model will be the basis for
determining health care services and other patient
related services for which KPIs are to be developed.
A simple example of a value model is shown in
Figure 2, which shows two actors, patient and
primary health care provider, and their resource
exchange.
Primary
health care
provider
Patient
Payment
Eye examination
Figure 2: A simple value model.
3.2 Task 2: Generate Technical KPIs
Technical service quality focuses on technical
accuracy and procedures, in particular the quality
and effectiveness of the diagnostic and therapeutic
intervention processes. In other words, technical
service quality is about the results of health services
and the consequences of health issues. A number of
top level goals for technical service quality are:
1. Health care services should have positive
effects on health issues
2. Secondary diseases should be avoided
3. Health issues should be treated by evidence
based health care services
4. Health care services should not give rise to
unwanted side effects
In order to measure the fulfillment of these goals, we
introduce four generic KPI templates, each
addressing one of the goals. When instantiated, these
templates will result in a set of KPIs.
1. Percentage of health care service x
1
that is
followed by health care service x
2
2. Percentage of health issue y
1
that is
followed by health issue y
2
3. Percentage of health issue y
1
that is
followed by health care service x
1
4. Percentage of health care service x
1
that is
followed by health issue y
1
In order to instantiate these templates, relevant
health care services and health issues need to be
chosen. As this choice depends on the health care
scenario under consideration, KPI designers need an
instrument for documenting and representing the
health care services and health issues they decide to
include. When they have done so, they can go on to
instantiating the KPI templates to arrive at KPIs. We
suggest that relevant health care services and health
issues be represented by a simple conceptual model,
called a service and issue model. Such a model
consists of two classes, Health Care Service and
Health Issue as well as a number of subclasses of
these. The subclasses included in a service and issue
model will depend on the health care scenario being
addressed and will, therefore, vary from scenario to
scenario.
Health Care Service
Health Issue
Symptom Diagnosis
Examination Treatment
Problem
Figure 3: The standard service and issue model.
In Figure 3, we suggest a standard service and issue
model, based on notions from CONTSYS,
(CONTSYS, 2007), that should be applicable in
many scenarios with no or minor modifications.
The method for designing KPIs for technical service
quality can now be formulated in three steps.
Step 1: Adapt Standard Service and Issue Model
Construct a service and issue model (SIM). This can
be done by modifying the standard service and issue
model of Figure 3. Figure 4 show an example of
such a SIM, adapted from the standard service and
issue model.
Step 2: Apply Generic KPI Templates
For each generic KPI template T and every subclass
X of Health Care Service in SIM and every subclass
Y of Health Issue in SIM, introduce a new specific
KPI template by replacing every x
i
with X and every
HEALTHINF 2010 - International Conference on Health Informatics
124

Health Care Service
Health Issue
Symptom Diagnosis
Examination Treatment
Problem
Treatment
ToCu re
Medicine
Treatment
Emergency
Treatment
Primary
Prevention
Secondary
Prevention
Emergency
Investigation
Figure 4: A service and issue model, adapted from the
standard service and issue model from Figure 3.
y
i
with Y. Each such template will have the form
“Percentage of X that is followed by Y”, where X
and Y are classes in SIM.
From KPI template 1 and the SIM in Figure 4:
• Percentage of primary prevention
that is
followed by secondary prevention
From KPI template 3 and the standard SIM in Figure
4:
• Percentage of diagnosis that is followed by
a medicine treatment
These specific KPI templates can also by themselves
be used as KPIs.
Step 3: Apply Specific KPI Templates
For every specific KPI template from Step 2,
“Percentage of X that is followed by Y”, replace X
and Y with instances of these classes. This replacing
can be carried out in two ways:
a) A domain expert examines every specific KPI
template and replaces X and Y based on her
knowledge.
b) A domain expert populates the SIM. A list of all
possible KPIs is then generated automatically.
Finally the domain expert selects relevant KPIs from
the list.
From the specific KPI template “Percentage of
diagnosis that is followed by medicine treatment”
• Percentage of flu diagnosis that is followed
by Tamiflu treatment
• Percentage of stroke diagnosis that is
followed by warfarin treatment
After having applied these steps, the KPI designer
will have arrived at a number of KPIs that focus on
the results and consequences of health care services
and health care issues. These KPIs are objective in
the sense that they do not directly depend on the
assessments of patients.
3.3 Task 3: Generate Functional KPIs
While technical service quality focuses on the
effects of health care services, functional service
quality refers to the manner in which services are
delivered to the patient. Functional service quality
includes aspects like facilities, hospital food,
employee attitudes, responsiveness, and cleanliness.
While technical service quality is typically difficult
to judge for the individual patient, functional service
quality is usually directly visible to the patient.
Therefore, functional service quality often has more
impact on a patient’s service quality perception than
technical service quality.
In contrast to technical quality, functional quality
does not depend only on health care services
themselves but also on complementary services, like
ordering, information and complaint services. We
have identified the following complementary service
types, based on an adaptation of open-EDI
(UN/CEFACT Modeling Methodology, 2008).
Identification. An identification service offered by a
health care provider is an information service that
provides information about the provider's services
and helps a patient in identifying and selecting
among health care providers and their services.
Negotiation. A negotiation service offered by a
health care provider is an interactive service where
the patient and the health care provider negotiate in
order to arrive at an agreement on a future health
care service including its cost, scheduling, location
and personnel.
Pre-actualisation. A pre-actualisation service
offered by a health care provider is an information
service where the health care provider informs the
patient on adequate preparations for a health care
service.
Post-actualisation. A post-actualisation service
offered by a health care provider is either
VALUE DRIVEN KPI DESIGN FOR HEALTH CARE
125

a) an information service where the health care
provider informs the patient on the result of a health
care service
b) an information service where the health care
provider informs the patient on adequate behaviour
to be observed after the performance of a health care
service
c) a service where the health care provider accepts
and addresses complaints on a health care service
Both health care services and their complementary
services should be assessed along a number of
service dimensions. Thus, a KPI designer is to select
a set of service dimensions and for each service
dimension, one or several service items. Which
service dimensions and items to investigate depend
on the health care scenario under consideration. We
envisage that in most cases, the service dimensions
will be variations or extensions of those in
SERVQUAL or its predecessor. Furthermore, for
each service item and each type of service (including
complementary service types), the KPI designer
shall specify how significant the service item is for
that type of service. A service item is significant for
a service type if it is important for a patient’s
perception of services of that type and if it is
difficult for the health care provider to score well on
the item. Only if both these conditions are satisfied,
it becomes interesting to measure the item through
KPIs.
Summarising, the KPI designer shall define a service
dimension model, consisting of a set of service
dimensions; for each service dimension, a number of
service items; and for each combination of service
item and service type, the significance of the service
item for the service type. A service dimension model
can be represented as a matrix, see Figure 5.
We do not believe that there exists a canonical
service dimension model that is optimal for each
health care scenario. However, we envisage that
most service dimension models will share large
parts. Therefore, it is worthwhile to introduce a
standard service dimension model that can serve as a
starting point when a KPI designer develops her own
model for a particular scenario. For this purpose, we
suggest the service dimension model in Figure 5,
based on the predecessor to SERVQUAL, where we
have indicated significance on a three level scale:
zero, one or two + marks. The significance marks of
the model are justified as follows:
Reliability is less important for identification and
negotiation services partially because no promises
have been made before these services, thereby
making the question of faithful service execution
irrelevant. Consistency of performance and prompt
attention to defect service are particularly important
for actualisation, as these directly influence the
health state of patients.
Responsiveness is important in all phases, as
customers value prompt services and willingness
during their entire episodes of care.
Competence is particularly important for
actualisation, as it directly influences the health state
of patients. It is also important for pre and post
actualisation phases as it also here has a strong
impact on health state. Competence is less important
for the identification and negotiation phases, as the
competence needed for these activities is relatively
easy to achieve.
Access is particularly important for identification
and negotiation services, as the use of these services
potentially may be stressful and take a long time.
Good access may reduce stressfulness as well as
time consumption for patients.
Courtesy is important in all phases, as customers
value a positive attitude and privacy during their
entire episodes of care.
Communication is especially important in pre- and
post-actualisation phases, as these involve much
patient interaction and may have substantial impact
on the health state as well as the feeling of safety of
the patient.
Understanding the Customer is important in all
phases, as health issues are complex and dependent
on the individual patient. Therefore, customised
services are often needed in health care.
Tangibles are of interest when services are given at
healthcare facilities by health care personnel, i.e. in
the actualisation, as well as in the pre- and post-
actualisation phases. Visual appeal of physical
facilities and cleanliness of employees will indicate
that the healthcare provider offers high quality
services. However, a well-structured web page is
also of importance in the identification and
negotiation phases.
The method for designing KPIs for functional
service quality can now be formulated in three steps.
Step 1: Identify Complementary Services
For each service from the value model, introduce
complementary services according to the
complementary service types above.
HEALTHINF 2010 - International Conference on Health Informatics
126

Actualisation Identification Negotiation Preactualisation Postactualisa -tionresult Postactualisa- tionbehaviour Postactualisa- tioncomplains
RELIABILITY
consistency of performance ++ + + +
correct keeping of records + + + +
correct billing N/A N/A N/A N/A N/A N/A +
on-time (promised) delivery + N/A N/A + + + +
prompt attention to defect services ++ + + + +
RESPONSIVENESS
prompt service delivery + + + + + + +
exact information on service delivery time + N/A N/A + + + +
employee willingness to help + + + + + + +
employee willingness to help when busy + + + + + + +
calling the customer back quickly + + + + + + +
COMPETENCE
knowledge and skill of the personnel ++ ++ ++ ++ ++
education level of the personnel ++ + + +
experiences of the personnel ++ + + +
research capability of the organisation + N/A N/A N/A
ACCESS
easy access to services by phone, web
and mail
++ ++
short waiting time to receive service ++ ++ ++ ++ ++ ++ ++
convenient hours of operation + ++ + + + + +
convenient location of service facility +
COURTESY
attitude of employee in performing the
service
+ ++ ++ + + + ++
high level of privacy ++ ++ ++ N/A ++ N/A
COMMUNICATION ++
understandable explanation of the service + ++ ++ ++ ++
understandable explanation of cost and
trade-offs between service and costs
N/A N/A ++ N/A N/A N/A ++
patient interaction + + + ++ + +
continuous state updates + N/A N/A + + + +
UNDERSTANDING/
KNOWING THE CUSTOMER
learning the customer’s specific
requirements
++ + ++ ++ ++ ++ +
providing individualised attention ++ + ++ ++ ++ ++ +
TANGIBLES
up-to-date equipment + N/A N/A N/A N/A N/A N/A
visual appeal of physical facilities + +
well dressed and neat appearance of
employees
+ +
consistency between the appearance of
physical appearance and type of services
provided
+ +
cleanliness of employees ++ +
cleanliness of physical facilities ++ +
web page ++ +
Figure 5: The standard service dimension model.
VALUE DRIVEN KPI DESIGN FOR HEALTH CARE
127

Step 2: Adapt the Standard Service Dimension
Model
Construct a service dimension model. This can be
done by modifying the standard service dimension
model of Fig. 5.
Step 3: Apply the Service Dimension Model
For each health care service and complementary
service from Step 1, apply the service dimension
model from Step 2, i.e. identify for each service item
zero or more KPIs. The significance of a service
item for a service type will assist in determining
whether to introduce a KPI or not.
4 METHOD APPLICATION
In this section, the proposed method is applied for
illustration purposes. The method is applied on the
results from a research project (REMS) in the eye
health care domain (Henkel et al, 2007). The main
aim of the project was to develop and evaluate a set
of e-services that could be used to create, manage
and transfer health care referrals between primary
health care and eye specialist providers. In the
project, a set of enterprise models were constructed
to support the design and evaluation of the e-
services. Some of these constructed models are used
in this section.
Task 1: Develop Value Model
The first task in the proposed method is to delimit
the domain for which the KPIs are to be designed,
i.e. specify which services need to be measured
using KPIs. For that purpose, a value model is
constructed. In this case an excerpt from a
constructed value model in the REMS project is
used, see Figure 6.
The value model in Figure 6 shows three actors -
patient, primary health care provider and eye care
specialist provider - and the transfers of resources
between them. The background for the value model
is the following: When a patient experiences an eye
health problem, she/he will visit the primary health
care provider. The basic resource this provider offers
is an eye examination service. In order to receive
that service, the patient needs to pay a certain fee for
the service. The fee transfer is carried out via a
payment service. If the patient needs further
treatment, either the primary care provider will carry
out the treatment (which is not shown in Figure 6) or
the provider refers the patient to an eye care
Primary
health care
provider
Patient
Eye care
specialist
provider
Eye treatment
Payment
Eye examination
Payment
Referral service
Referral service
Referral answer
service
Figure 6: A value model from the REMS case.
specialist provider who is able to provide advanced
treatment services. To do this, the primary health
care provider transfers a referral via a referral
service to the patient, which allows her to be treated
by an eye care specialist provider. The primary care
provider will also transfer referral information via a
referral information service to the eye specialist
provider, which will be used for scheduling a
treatment service for the patient, as well as being the
base for resource allocation. The eye care specialist
provider offers a treatment service to the patient and
the patient needs to transfer a fee via a payment
service. After the treatment service is carried out, the
eye care specialist provider will transfer a referral
answer service to the primary health care provider,
informing it about the result of the treatment.
In the method proposed in this paper, only KPIs
from the perspective of the patient are to be
generated. Therefore, only resourced transferred to
and from the patient are included. This will delimit
the value model further, see figure 7.
Primary
health care
provider
Patient
Eye care
specialist
provider
Eye treatment
Payment
Eye examination
Payment
Referral service
Figure 7: The value model from the perspective of the
patient.
HEALTHINF 2010 - International Conference on Health Informatics
128

Task 2: Generate Technical KPIs
In order to generate the KPI for measuring technical
service quality, three steps are to be carried out.
Step 1: Adapt Standard Service and Issue Model
In this step, the standard service and issue model,
see Figure 3, is to be adapted to the value model
constructed in task 1. In other words, the standard
service and issue model has to include all the health
care services from the value model, but not other
types of services, such as administrative and
resource allocation services. For example, the value
model in Figure 7 includes two health care services,
eye examination and eye treatment, which are to be
included in the service and issue model, see Figure
8.
Step 2: Apply Generic KPI Templates
In this step, the four generic KPI templates presented
in section 3.2 are to be instantiated by the concepts
(subclasses) of the service and the issue model from
step 1, in order to create a set of specific KPI
templates. There are seven concepts in the adapted
service and issue model in figure 8, but the more
generalized concepts examination and treatment will
not be used in this step, only their specialized
concepts, eye examination and eye treatment. For
example, the generic KPI template 4: “Percentage of
health care service x
1
that is followed by health issue
y
1
” can be instantiated by the health care service eye
treatment and the health care issue symptom from
the adapted service and issue model. This will result
in a specific KPI template: “Percentage of health
care service eye treatment that is followed by health
issue symptom”.
Step 3: Apply Specific KPI Templates
In this step, the resulting specific KPI templates
from step 2 need to be instantiated by a set of
instances from the service and issue model. This
requires that the service and issue model in Figure 8
be instantiated. Examples of instances of the service
and issue model can be found in Figure 8. These 9
instances are applied on the specific KPI templates,
resulting in a number of KPIs. For example, the
template constructed in the previous step,
“Percentage of health care service treatment that is
followed by health issue symptom”, can be
instantiated by the treatment eye pressure test and
the symptom veiling glare. This will result in the
KPI “Percentage of health care service treatment eye
pressure test that is followed by health issue veiling
glare”.
Health Care Service
Health Issue
Symptom Diagnosis
Examination Treatment
 Cronic
glaucoma
Loss of visual
acuity
Loss of
contrast
sensitivity
Veiling glare
Eye drops
Laser surgery
Canaloplasty
Vision screening
(eyesight test)
Eye pressure
test
Problem
Eye
Examination
Eye
Treatment
Figure 8: The service and issue model adapted based on
the information from the value model.
Task 3: Generate Functional KPIs
In order to generate KPIs for measuring functional
service quality, three steps are to be carried out.
Step 1: Identify Complementary Services
In this step, the complementary services are applied
on each service in the value model, both the health
care services and the other services. For example,
the service eye examination will have four
complementary services: identifying an eye
examination service, negotiating an eye examination
service, pre-actualising as well as post-actualising an
eye examination service.
Step 2: Adapt the Standard Service Dimension
Model
In this step, the standard service dimension model,
see Figure 5, is adapted. This is done by adding,
deleting or changing the dimensions and items of the
standard service dimension model. In this example
VALUE DRIVEN KPI DESIGN FOR HEALTH CARE
129

case, we do not make any changes to the standard
service dimension model but keep it as it is.
Step 3: Apply the Adapted Service Dimension Model
For each service identified in Step 1, we consider all
the service dimensions and service items in the
service dimension model from Step 2. For each item,
we determine whether to include some KPI or not.
For example, for the (complementary) service “pre-
actualising an eye examination” and the service item
“understandable explanation of the service”, we
introduce a KPI based on the following question to a
patient “Did you get an understandable explanation
of how to prepare for your eye examination?”
5 CONCLUDING REMARKS
In this paper, we have proposed a method for
designing KPIs for a health care organisation with a
focus on patient oriented KPIs. The starting point of
the method is a value model that delimits the domain
under consideration by identifying health care actors
and resource exchange, including services. Based on
the services identified, the method helps to generate
KPIs for both technical and functional service
quality. Technical quality KPIs measure the results
of health services and the consequences of health
issues and are generated by considering a small
number of top level goals for technical quality.
Functional quality KPIs refer to the manner in which
health care services are delivered to patients and are
generated by identifying complementary services
and applying SERVQUAL dimensions. This way of
generating KPIs can also be seen as goal driven, as
the SERVQUAL dimensions and items express
subgoals of customer satisfaction.
The proposed method provides four main
advantages:
Ease of Design. Designing a set of KPIs for a health
care scenario is a complex and time consuming task.
The proposed method alleviates this task by offering
a designer a large number of potential KPIs from
which she can choose the most relevant ones.
Completeness. A problem in KPI design is to ensure
that all relevant KPIs have been identified, i.e. to
ensure the completeness of the KPIs. The proposed
method addresses this problem by systematically
generating relevant KPIs for technical as well as
functional quality.
Traceability. KPIs should be possible to justify by
relating them to the goals whose fulfilment they
measure, i.e. KPIs should be traceable to goals. The
proposed method provides traceability by directly
generating potential KPIs from goals, which in the
case of functional quality are expressed through
SERVQUAL dimensions and items.
Flexibility. The health care scenario under
consideration will influence the choice of KPIs,
meaning that a KPI designer needs to be able to
describe the specifics of a certain scenario in a
convenient way. The proposed methods supports this
description through the use of value models, health
and issue models and service dimension models,
where the KPI designer can describe the most
important health care services and health issues as
well as relevant service items and their
prioritization.
In this paper, we have only addressed technical and
functional service quality. However, a third aspect of
service quality is the effect a health service can have
on the quality of life of a patient, including the
physical, psychological and social functioning of
the patient. A direction of future work is to design
guidelines for identifying KPIs for this service
aspect. Another direction of future work is to
provide further support in formulating KPIs for
functional quality. The method presented only
suggests service items for different services that
should be measured by KPIs but does not suggest
formulations of these KPIs. We believe that it is
possible to find such formulations for a large
proportion of the service items. In this paper, only
KPIs from the patient’s perspective are generated.
Future work will also include generation of KPI
from the health care providers’ perspective, such as
revenue and cost efficiency.
REFERENCES
Adair, C.E, Simpson, L., Birdsell, J.M., Omelchuck, K.,
Casebeer, A.L., Gardiner H.P., Newman, S., Beckie,
A., Clelland S., Hayden K.A., Beausejour, P. (2003)
Performance Measurement Systems in Health and
Mental Health Servicws: Models, Practices and
Effectiveness. A State of the Science Review, The
Alberta Heritage Founddation for Medical Research
CONTSYS (2007), EN-13940-1:2007, Health Informatics
– Systems of concepts to support continuity of care.
Part 1: Basic Concepts
Eckerson W. (2006) Performance Dashboards. Measuring,
Monitoring and Managing Your Business, John Wiley
Gill L., White L. (2009) A critical review of patinet
satisfaction, In: Leadership in Health Services, vol. 22,
no 1, pp 8-19, Emerald Group Publishing
HEALTHINF 2010 - International Conference on Health Informatics
130

Gronroos, C. (1984) "A Service Quality Model and Its
Marketing Implication", European 18, no 4, pp 36-44
Gupta, H. D. (2008) Identifying Health Care Quality
Constituents: Service Providers' Perspective, Journal
of Management Research, vol 8, no 1
Henkel, M., Johannesson, P., Perjons, E., Zdravkovic, J.,
"Value and Goal Driven Design of E-Services", The
IEEE International Conference on e-Business
Engineering (ICEBE'07), Hong Kong, China, October
24-26.
Johannesson, P. and Perjons, E. (2009) "Managing
Complexity through Value Modeling", In: Healthcare
IT Management, vol. 4, issue 2
Jun, M., Peterson R.T., and Zsidisin G.A. (1998) The
identification and Measurment of Quality Dimension
in Health Care: Focus Group Interview Results,
Health Care Managemnt Review, 23(4): 81-96
Karassavidou E., Glaveli, N., Papadoupoulos, C.T. (2009)
Quality in NHS hospitals: no one knows better than
patients, In: Measuring Business Excellence, vol. 12,
no. 1, pp. 34-46
Kartseva V., Hulstijn J., Gordijn J. and Tan Y-H (2007).
Control Patterns in a Healthcare Network. In Alan
Brnong and Marijn Janssen editors, Proceedings of the
8th Annual International Conference on Digital
Government Research (DG0'2007), Pages 66-75,
Digital Government Research Center,
Kilbourne, W.E, Duffy, J.A., Duffy, M., and Giarchi, G.
(2004) SERVQUAL in cross-national measurements of
health care quality, In: Journal of Service Markeing,
Volume 18, No 7 pp.524-533
Parasuraman A., Zeithaml, V.A., and Berry, L.L. (1985)
"A conceptual model of service quality and its
implication for future research", Journal of Marketing,
vol. 49, no. 4, pp 41-50
Parasuraman A., Zeithaml, V.A., and Berry, L.L. (1988)
"SERVQUAL: a multiple-item scale for measuring
consumer perception of service quality", Journal of
Retailing, vol. 64, no. 1, pp 12-40
UN/CEFACT Modeling Methodology (2008) User Guide.
Accessed Nov, 2008, http://www.unece.org/cefact/
Walburg, J. (2006) "The outcome quadrant". In: Walburg
J., Bevan H., Wilderspin J. and Lemmens K. (eds)
Performance Management in Health Care. Improving
patient outcomes: an integrated approach, Routledge,
UK.
VALUE DRIVEN KPI DESIGN FOR HEALTH CARE
131