A Soft Approach to Survey Design
L. Weston, R. Whiddett and B. Jackson
Department of Information Systems
Massey University
Palmerston North
New Zealand
Tel: 06-3569099
LizW @rds.co.nz,
r.j.whiddett @ massey.ac.nz,
b.x.jackson @massey.ac.nz
Abstract
The healthcare sector is a large, complex and information rich environment. For
many years Information System professionals have developed systems which attempt
to meet the information needs of the stakeholders. A key item, which has recently
gained prominence, has been the need for a comprehensive set of I.S. standards. In
particular, considerable activity by organisations such as C.E.N., 1.8.0. and the HL7
Foundation has been directed towards the development of standards for the storage
and exchange of clinical information. This paper discusses a novel approach that
was used to gain a better understanding of factors which influence the adoption of LS.
standards within healthcare. The approach utilised aspects of Soft Systems
Methodology and conventional survey methodology to gain a better understanding of
the processes involved in developing and adopting standards. This understanding
was then used to guide the development of a questionnaire which was then used to
elicit further information from the stakeholders.
Introduction
This paper discusses the way in which some of the techniques of Soft Systems
Methodology (SSM) was used in research which aimed at gaining a better
understanding of the issues involved in the development and introduction of standards
for Electronic Medical Records (EMRs) to record the treatment of patients. The paper
begins with an overview of SSM and how it was used in this project. The paper than
discusses some of the background issues relating to the domain of EMRs and LS.
standards and shows how the use of SSM helped the researchers to gain a better
understanding of the issues.
Soft Systems Methodology
Soft systems methodology was developed with the specific intention of addressing
complex issues associated with human activity systems. It has been applied with success
in a number of areas: public utilities, health services, industry and education (Watson
and Smith, 1988: Checkland and Scholes, 1990). A key feature of SSM is that it enables
the analyst to embark on a process of learning about the domain under scrutiny and seek
ways of improving the problem situation.
The methodology is described schematically in Figure 1 as a three phase process
(Identification of issues and tasks, Modelling possible ways of improving the problem
situation, Taking action that is culturally desirable and feasible). However, it is important
to stress that each problem domain is unique and the three phase cycle may not be
appropriate.
1. In the first phase, the would-be improvers enter the problem domain and attempt to
understand and describe the problem situation. It is expected that the need for
intervention has been flagged by some prior experience or analysis (History). In this
Finding Out phase the key players, and their roles, are identified together with a list
of issues thought to impact on the situation. The results of the Finding Out phase are
frequently represented in a Rich Picture. Such a picture is likely to bring together the
structural and dynamic features perceived relevant by the players within the domain.
2. In the second phase, issues thought relevant to the problem domain are then subject
to two streams of enquiry. A logical stream of enquiry is carried out which attempts
to seek ways of addressing the issues by developing systems, definitions and models
(hypothetical) that might be applied to the problem domain, and a cultural stream of
enquiry which modifies, reflectively and dynamically, the logical stream by
bringingto bear factors associated with the social and political characteristics of the
domain. The cultural stream brings into play the roles, norms and values of the
stakeholders and the other players associated with the situation.
3. In the third and final phase, SSM concludes with a set of proposals for feasible and
culturally appropriate action for change with a view to improve the problem
situation. Of course, there is an implication that the process is cyclic and as such is a
neverending process of learning and improvement.
A would-be improvers of the problem situation
History
—
real-world
problem situation
Nae and tasks
situation as a culture
Vv Relevant models situation
2 2 F systems
intervention analysis eae
social system analysis ] 9g QO
> iy) 4
political system analysis G@ QO
J] eS
— Differences between model and the real world
Change: Systematically desirable and culturally
feasible rf
Action to improve the situation
Figure 1: Soft Systems Methodology (Checkland & Scholes, 1990)
Application of SSM to the problem domain.
In seeking to develop Electronic Medical Record standards the researchers felt that there
were four key areas of activity where the application of soft systems methodology could
be useful:
Gathering of pertinent historical data relating to EMRs
Identification of key issues concerned with the introduction of IS standards.
Identification of key players in the problem domain, their roles and values
Development and format of questionnaires that encourage the recipients to respond
in a meaningful and responsible manner.
Each of the above areas require an appreciation of not only the key issues but how they
reside within the wider culturally rich external environment. In essence each is a separate
problem and could be supported by the separate application of SSM, however, only the
top-level application of SSM is described here. The steps taken relate to the first two
phases of SSM, the outcome for the third phase is yet to be developed.
Historical data
Most human activity problems have associated with them a history of success and failure
and the development of Electronic Medical Records is no exception. As shown in
Figure 1, SSM begins by drawing together the various, and often wildly/widely
conflicting, historical accounts relating to the problem domain. In this investigation the
principal source of ‘history’ was the literature review and the knowledge of experts in
the field. This history is discussed below in the “Context” section below.
Key players
These perspectives were gathered from a range of sources primarily through the use of a
literature review. Care was taken during this phase to ensure that perspective from a
wide range of actors were obtained. Entering the problem domain with the intention to
improve the situation are the analysts who bring their own perspectives and opinions.
The researchers identified, through detailed examination of the literature reviews and
knowledge of the healthcare community, those actors within the problem domain whose
contribution would be meaningful and valuable. Such participants are not passive but
have opinions, perspective and motives that differ from one another and can often
dominate the debate. The participants operate within their own social, political and
cultural environment and may have quite distinct roles, values and norms to each other.
Issues
A rich picture of the problem domain was drawn. Such a picture not only captures the
basic formal structural relationships between the various actors but also the social and
political undercurrents. The rich picture also defines the boundary of the problem
domain excluding those organisations and people who are outside the domain but who
behaviour impact upon the problem but includes those groups whose behaviour can be
influenced and modified through discussion and debate. The rich picture stimulated
debate and discussion which led ultimately to formulation of a list of issues and primary
tasks that need to be resolved. (Note: In SSM an issue is concerned with the ability of the
system to develop a ‘solution’ in an effective and efficient manner. A primary task
consists of those activities that must be completed for the system to achieve its goal.)
The key issue for this project was to identify what the key players thought were
characteristics of a good standard. However, since there were a lot of key players, it was
decided that the most appropriate approach would be to conduct a survey, which is a
novel approach for SSM.
Questionnaire
The successful distribution and analysis of a questionnaire was deemed essential by the
researchers. It became apparent that there was no ‘typical’ respondent and that care
needed to be taken to design questionnaires for specific target groups. For each target
group a questionnaire was constructed that respected the respondents’ level of
understanding of the problem domain and their roles and values within the wider health
sector environment.
The following sections of this paper will illustrate the process by following the
evolution of the project .
Context of Healthcare and the Pressures for Electronic Medical Record
Standards
The delivery of healthcare services is an inherently complex task. It involves the
interaction of many of the natural systems of the patient with the scientific tools of
diagnosis and therapy of the physicians. These interactions often involve multiple
actors and are extremely information intensive and often time-critical. For many years
now the healthcare industry has been struggling to use IT to manage the vast amount
of data that constitutes the medical records of patients. The difficulties in developing
systems arise from the fact that the information is poorly structured, it involves many
different media (test results, X-rays etc) and it gets distributed around the country as
patients move, change doctors or are referred to different hospitals. Issues of privacy,
confidentiality and the legal status of the record further complicate the problem.
The delivery of care takes place within the complex political and economic
environment of healthcare delivery. Throughout the course of a clinical episode the
patient may be under the care of many clinicians working within primary, secondary
and long-term care institutions. They may use facilities of other parties to provide
laboratory and radiological investigations and be treated with drugs from pharmacies.
The payment for these services may come from a number of sources, in New Zealand
these might be private health insurance or throught the government bodies such as
Health Benefits LTD (HBL) or the Accident Compensation Corporation (ACC).
In recent years there has been a considerable degree of change introduced into the
New Zealand Health Sector (NIPB, 1992; NZDoH, 1990; NZMoH, 1991, 1992). The
main emphasis of these policy changes is resource management, which has separated
the funding of health services from their provision; introduced contestability into the
process of allocating resources; increased the private supply in the market and
increased the interface between former state health service providers and new private
sector providers.
These desired changes are similar to the ones proposed in the United Kingdom
(UKDoH, 1989) and in Australia (NHSU, 1993). This period of change has had a
major impact on the development of health information systems (HIS), since there is
now increased pressure to obtain of more timely and accurately management
information.
The introduction of Independent Practitioner Associations (IPAs) into the New
Zealand primary healthcare arena has increased the awareness and need for general
practice to be computerised. This has been a ‘push’ towards increasing the electronic
transfer of data between the GPs, the IPAs and the Health Funding Authorities and
other government organisations. There is also considerable political pressure to
improve the efficiency and effectiveness of the health service by increasing the
integration of the primary and the secondary care sectors. Improved communication
and information flow has a major part to play in this integration. Electronic transfer
systems, which involve computer linkages between providers, are being considered as
a way of implementing this integration.
The basis for these transfer systems is often seen to be and integrated patient-oriented
record system (or electronic medical record - EMR) which integrates all of a patient’s
medical, administrative and financially related information. This approach was
recommended by the Institute of Medicine of the National Academy of Sciences
(Dick and Steen, 1991) which identified EMRs as priority the improvement,
efficiency and cost-effectiveness of the health services. They argue that EMRs should
also provide better clinical decision support, aid in the monitoring of the quality of
care and of services and provide more complete information for research.
Despite the long-acknowledged need for EMRs, unfortunately, there are a number of
barriers to their increased adoption, both in primary care and secondary care. Some of
these barriers include the lack of national and regional direction, co-ordination,
planning and advice and the lack of the provision of the required infrastructure
including the policies, standards, networks and communications.
In resonse to the need for standards there has been considerable activity by
organisations such as the European Committee for Standardisation (C.E.N.), the
International Standards Organisation (I.S.O.) and the Health Level 7 Foundation
(HL7) directed towards the development of standards for the storage and exchange of
clinical information. Within New Zealand, activities in this area are undertaken by a
number of working groups which are coordinated under the Health Information
Standards Committee of Standards New Zealand.
The members of these working groups find themselves at the center of nexus with
competing demands. They need a better understanding of their clients needs in order
to prioritise their efforts and maximise benefits they can deliver with the limited
resources that are available for our local standards development initiatives. What the
standards developers really needed to know was:
What are important characteristics of standard that impact on its chances of
adoption? i.e. what constitutes a good standard?
What Constitutes A Good Standard?
Time seems to be one of the major faults of standards; they either take too long to
create (Aiken and Cavallini 1994, Gritzalis 1997 and Scott-Hill 1996) or they don’t
remain useful and applicable within the desired arena for a long enough time frame.
This means, that people are spending considerable time and cost creating standards,
and they are in turn, being rendered useless. Cargill (1998) states that the participants
in standardisation activities must realise that they are not there to protect the
standardisation process; they are there to get standards out. These standards must be
deployable by, and useful to, businesses that are producing products. “Perfect
standards two years late are worthless; tremendously imperfect standards are also
worthless.”
Morrell and Stewart (1996) take a somewhat different stance and present a list of
problems that can occur when utilising or up-taking standards. Some of these include:
¢ The standards-making world is subject to strong forces over which it has little or
no control, including de facto standards, new technologies, national positions on
trade policy and the market positions of existing vendors.
e¢ Most company representatives to standards committees approach their task from a
technical rather than business perspective.
¢ Representatives to standard groups are often ‘volunteers’ rather than in dedicated
paid positions which militates against the rapid development of standards.
¢ The work of different standards groups is often related, but those relationships are
not always recognised. To complicate matters, the fusion of various technologies
may generate connections between previously unrelated standards.
Oksala et al (1996) raise an issue, which few other articles have. They note that
standards are developed within a cultural milieu. Standards are developed by a
community that has a set of values and a particular perspective on information
technology. The people, developing and using the standard have a long involvement
in their particular field and the paradigm of the field may work as a kind of selection
paradigm.
This is very important as it is acknowledging that although standards are for use
(ideally) everywhere, standards may often be created, with a specific environment
(indeed, their own environment) in mind. It may, in some situations, be far too hard
to try and understand what another company, competitor, or even country will be
using, or doing with the same tools. Indeed this is one of the problems of standards
development.
Hovenga, Kidd and Cesnik (1996) briefly discusses this point by noting that the
adoption of standards may be mandatory or voluntary, and various types of standards
exist. The type of standard is determined by who has developed or adopted the
standards or by the purpose for which the standards was developed.
Cargill (1998) depicts standardisation as running on urban myths; elaborating on the
idea that there is no coherent, widely held and widely accessible body of literature on
the nature, rationale, or practice of standards. There are no standards for
standardisation. Cargill (ibid.) continues that standards are the fundamental agents of
change, and yet the knowledge of how they work, inside the discipline and the
market, is only vaguely understood. Until this vague understanding is made into a
coherent knowledge base, standards will continue to exist in the twilight realm.
Aiken and Cavallini (1994) continue this point and were one of the few authors
willing to admit, that they themselves are sure of the need for standards, yet find that
they have no agreement about a number of basic issues, like the following:
¢ How to identify which standards exist and which need to be developed and
enforced?
e How should standards be chosen? To what extent should the choice be influenced
by industry, or the purchasing power of the federal government, or the actions of
formal standards bodies and consortia?
¢ Who are the people actually developing and mandating the standards? Do they
have real-life operational experience in the area they so greatly influence?
e¢ What are the professional and ethical responsibilities of those persons who set
standards? Are short-term cost benefits and conformity more important than
diversity and competition?
¢ Should multiple standards be allowed to coexist? For example, at the network
layer, are IP and OSI allowed to coexist?
¢ What is the real practical life cycle of a technology and/or standard and how is it
phased out or replaced when appropriate?
It is indeed this list of questions that rest on the mind of many people when using
standards. Part of the goal of this research project is to answer pieces of the first two
questions, with regard to New Zealand healthcare.
The Standards Adoption Framework (SAF)
To ascertain which standards are selected over and above another standard, it is
essential to set up a form of criteria. The literature presented many different sets of
criteria, with many having common elements. The following section, lists the
different criteria found, and presents the criteria that shall be used within the research.
Six relevant articles were identified with regard to evaluation criteria for standards.
One of these articles, National Institute of Standards and Technology (NIST, 1998)
presented a criterion that was used to evaluate a range of existing standards. The five
other articles were giving a hypothetical list of questions that should be answered
when selecting standards, these were: Batik (1989), Morrell and Stewart (1996)
Oksala et al (1996), Baldo et al (1997), Aden and Harris (1993). It is the combining
of these different forms of criteria that are used to formulate the underlying model of
this research. The Standards’ Adoption Framework aims to specify what individual
elements are essential in the decision making process when selecting a specific
standard.
It was decided that the combining of these would create the most useful model to
work within. It is important to note that National Institute of Standards and
Technology (1998) presents the most thorough criteria as this model has already been
successfully used and accepted as a standard criterion.
Firstly, it was decided that it would be appropriate to pool the different questions
(delete any identical questions), then group them into similar terms, and discover how
they then relate to the NIST model.
This model is useful because it shows the individual questions in relation to the
already tested and acknowledged NIST categories. It is important to note that once
the groupings were made it became obvious that another category had to be appended
to the NIST model. ‘Interoperability’ was important to the different literature pieces,
but was not accounted for with the NIST model, consequently it was added as a
specific category.
Another change was also created with the NIST model. In the original model
‘Maturity’ and ‘Stability’ were two separate categories, it was decided that it would
be appropriate to combine the categories into one, as they were both concerned with
the issue of time.
The final model, shown in Figure 2, called the “Standards Adoption Framework”
(SAF) provides a general theoretical model which identifies and integrates the factors
which are required for an IS standard to be successful, and therefore adopted.
In summary, the review of the literature related to standards revealed two main issues.
e Firstly, there are many criteria regarded as being important for the adoption of
standards, which could be integrated into the SAF.
¢ Secondly, as Oksala et al (1996) noted, the processes of development and
adoption of standards is located within a cultural milieu and the processes may
therefore be influenced by a variety of stakeholders who may have differing
priorities and perspectives on information technology.
Therefore, while the SAF provided a checklist of issues that needed to be addressed in
standards development, it did not provide any guidance as to the relative importance
and priority of these issues. Furthermore, it would seem likely that these issues are not
generalisable and the priorities are likely to be different for the different stakeholders,
and possibly the priorities may change for different application areas. Further
empirical investigation was therefore required before the model would be able to
assist the standards developers.
Refining our Understanding
As discussed in the previous sections, both the areas of healthcare and of standards
development are very complex areas with many interested stakeholders who have
different backgrounds and perspectives. It was felt that as far as the adoption of IS
standards was concerned, one of the most significant stakeholders was the IS Manager
since they would have considerable influence over the systems that are actually
selected for deployment.
Whereas the main focus point is indeed the IS manager, it has been acknowledged
that those around the manager also hold relevance and influence, and accordingly see
standards in a different way. Soft Systems Methodology (SSM) allows for the concept
that people have different views of the same situations because people see events
occurring in genuinely different ways (Open University, Block IV). This can be
represented graphically in a tool of the rich picture utilised by SSM. The rich picture
for the IS Manager is shown in Figure 3 below.
Although the focus of the picture is the IS Manager, there still remain a number of
people who can be considered as being relevant. This group of people include the
equipment suppliers, the end users, the service providers as well as the information
consumers (Ministry of Health, Health Funding Authority) to name but a few.
Are vendors building products?
How large is the installed base?
Does the standard provide common
ground for acceptance?
Is the standard consensus based?
How much choice do you have when
adopting standards?
Does the standard enjoy broad
support and demonstrated
effectiveness?
|
NIST CATEGORY
Level of
Consensus
Is_ the standard available for
implementation from a cross section
of vendors?
Has the standard had an impact on
the viability of the vendors industry?
NIST CATEGORY
Product
Availability
Is the standard based on a well
understood technology?
v
Does the standard provide
compatability with other standards?
Can the standard be tested to prove
compliance?
Does the standard meet the required
need and functionality?
Is there freedom from legal issues?
Is the document clear, accurate and
otherwise useable
NIST CATEGORY
Completeness
Were standards conformant products
on the market before the standard was
finalised?
Has it matured enough to ensure that
no major changes will occur
immediately?
NIST CATEGORY
Maturity/
Stability
How significant are the problems and
limitations?
NIST CATEGORY
Problems/
Limitations
Standards’
Adoption
Framework
)
)
?
Does the standard allow applications
that are portable to different
platforms?
Scalable in size?
Interoperable with other applications
Can the standard easily add new
technologies?
|
NEW CATEGORY
Interoperability
Figure 2 -Standards' Adoption Framework
Ss. 8
Equipment cusica
Ministry of Health
geo IS MANAGER
RS / Vo &
$$$$.
End Users
Organisation? Health Funding
Size Money Authority
Staff Need
Stage of
Service Providers oe
Figure 3 —- IS Manager Rich Picture
Other issues which influence the IS manager include, the size of the organisation, the
number of staff, the difficulty of the standard to implement, the level of computer
usage, and the installed computers. This latter point is also extremely important in the
context of healthcare in New Zealand, since there are three different types of
organisations that are relevant here. Each of which will have their own characteristics
in terms of equipment and expertise.
1. Public hospitals, which are generally large organisations with a large amount of
patient throughput and a large amount of data.
2. IPAs which are much smaller than public hospitals, and a lot more physically
disperse, and finally,
3. Private hospitals which are generally small and local healthcare agencies.
It should also be noted that although each of these organisations may have a role of an
IS Manager, there will be considerable differences in the nature of the task and in the
skills and expertise required of the person fulfilling that role.
In this way some of the tools of SSM helped us to gain an insight into context of the
problem situation and to identify some of the major stakeholders and the roles that
they played.
Surveying the Stakeholders
It was now appropriate to undertake some empirical investigation to solicit the views
and opinions of the stakeholders regarding the priorities for the work on standards
development. The traditional approach would be to develop and administer a
standardised questionnaire based on the SAF. However, this approach would be
unlikely to be satisfactory because of the difficulty of designing a generalised
questionnaire that would be appropriate for a sample population with such a diverse
range of perspectives and levels of expertise.
These problems can be overcome by developing the questionnaire as a series of
variants based around the central theme and addressing the same basic issues, i.e.
those of the SAF. Each variant of the questionnaire can then be targeted to the
context of a particular class of stakeholder and address the specific context in which
they work. The questions and wording would therefore differ depending on whether
the intended recipient worked for a Healthcare organisation of for an IT service
provider.
Outcome and Results
This approach was used to develop a range of questionnaires that were administered
to a range of stakeholders, a more detailed description of the process and outcomes
can be found elsewhere (Weston, 1999, Weston and Whiddett 1999).
The findings confirmed the complexity of the subject area. Whilst all the factors
identified in the SAF were seen to be important to all of the stakeholders, groups of
stakeholders accorded significantly different priorities to different aspects of the SAF.
For example, healthcare providers rated ‘level of consensus’ very highly but IT
service providers ranked it as the least important aspect.
Differences in views were also found between the IS Managers of different types of
healthcare organisations. These differing views can be interpreted in the light of the
structure and maturity of the different types of organisations.
Furthermore, it was found that classes of stakeholders gave different priorities to
different aspects of a standard depending on the application area (i.e. clinical coding,
information exchange, technical infrastructure or IS management). For example, the
IPA managers ranked interoperability as very important for standards relating to
technology infrastructure but as relatively unimportant for standards for information
management.
Summary and Conclusions
Healthcare and standards development are both complex areas to understand and to
work in. Both have numerous stakeholders with contrasting perspectives. SSM
techniques can be useful in giving a clearer understanding of the dynamics such
situations, in particular it helped to gain an understanding of the development of
standards for IT in healthcare.
The use of SSM techniques as a precursor to the development of a questionnaire tool
helped to target the questions to the different perspectives and contexts of the
stakeholders. In this way the two approaches can be seen as complementing each
other in the attempt to resolve a problem situation that spans a large number of
stakeholders.
References
Aiken, R. J., & Cavallini, J. S. (1994). Standards: When Is It Too Much of a Good
Thing? StandardView, 2(2), 110-119.
Aden, M., & Harris, M. (1993). A Practitioner's Guide to Standards and the
Government. StandardView, 1(2), 25 — 34.
Baldo, J., Moore, J., & Rine, D. (1997). Software Reuse Standards. StandardView,
5(2)
Batik, A. L. (1989). A Guide to Standards. Philadelphia: Parker Colorado
Benjamin, B. (Ed.). (1977). Medical Records. London.
Cargill, C. F. (1998). Standardization: Art or Discipline. IEEE Micro, May/June 1998,
18 — 24.
Checkland, P. & Scholes, J. (1991). Soft Systems Methodology in Action. Wiley,
Chichester.
Dick RS & Steen EB (1991) The Computer-Based Patient Record. IOM (Institute of
Medicine) Report: National Academy Press, Washington, D.C.
Gritzalis, D. (1997). A baseline security policy for distributed healthcare information
systems. Computers & Security, 16(8), 707 — 719.
Hovenga, E., Kidd, M., & Cesnik, B. (1996). Standards in health informatics, Health
Informatics: An Overview (pp. 41 - 45). Melbourne: Churchill Livingston.
Morrell, J., & Stewart, S. (1996). Standards Development for Information
Technology: Best Practice for the United States. StandardView, 4(1), 42 - 51.
NIST, National Institute of Standards and Technology. (~1998, 5 Jul 1995).
Application Portability Profile. http:/Avww.nist. gov.
National Health Strategy Unit (1993) Health that Works - Workplace Reform and
Best Practice in the Australian Health Industry. In Thelander N (1994)
Transforming Health Care. Proceedings of the Second National Health
Informatics Conference, Gold Coast, Australia.
National Interim Provider Board (1992) Providing Better Health Care for New
Zelanders. National Interim Provider Board, Wellington, New Zealand.
New Zealand Department of Health (1990) Health in the 1990’s. Department of
Health, Wellington, New Zealand.
New Zealand Minister of Health (1992) Directions for Health and Independence.
Ministry of Health, Wellington, New Zealand.
New Zealand Minister of Health (1991) Your Health and the Public Health. Ministry
of Health, Wellington, New Zealand.
Oksala, S., Rutkowski, A., Spring, M., & O'Donnell, J. (1996). The Structure of IT
Standardization. StandardView, 4(1), 9 - 22.
Open_University. (Block IV). The Systems Movement, Management and Change.
Scott-Hill, B. (1996). Living standards - rethinking the standards development
process. Standards, 42, 6.
UK Department of Health (1989) Working for Patients. HMSO, London.
Watson, R. & Smith, R. (1988). Application of the Lancaster Soft Systems Methodology
in Australia. Journal of Applied Systems Analysis, 15, 3-26.
Weston, L. (1999) The Use of Standards for Information Systems within New Zealand
Healthcare M.Sc. Thesis, Department of Information Systems, Massey
University
Weston, L. & Whiddett, R. (1999) The Use of Standards for Information Systems
within Healthcare in New Zealand: Summary of Findings Technical Report
#2/99, Department of Information Systems, Massey University
CC BY-NC-SA 4.0