A novel system for the optimal selection of finishes for healthcare facilities

Full text

265
A NOVEL SYSTEM FOR THE OPTIMAL SELECTION OF
FINISHES FOR HEALTHCARE FACILITIES
M. Kishk, R. Laing, J. Scott and M. Edge
The Scott Sutherland School of Architecture and the Built Environment,
The Robert Gordon University, Aberdeen, AB10 7QB
E-mail: m.kishk@rgu.ac.uk
Abstract:
This paper reports the development of a novel decision support system for the
optimal selection of finishes for healthcare facilities. The system has been
developed by integrating a recently developed extended whole-life costing (WLC)
application with two generic databases using a user friendly interface. The first is
a resource database that houses data for several options for various finishes
suitable for hospital buildings. The second is a project database which
accommodates finishes data for a specific building. The developed system has
being tested and validated in three phases. First, the usability of the system’s
interface has been tested in a laboratory environment. A second phase has been to
demonstrate the system to a panel of experts to get feedback on further
refinements to the system. The third phase is to use a case study to demonstrate
the efficacy of the system.
Keywords:
Decision Support, Design Decisions, Finishes, Whole-Life Management.
1. Introduction
It has long been argued that professionals should adopt a whole-life attitude
regarding the design and management of buildings. The main reason has been the
dramatic shift in the balance between the initial capital cost and the running costs
of buildings towards a substantial increase in the running costs. Because most of
these running costs can be directly linked to the finishes, it is common sense to
adopt a whole-life approach in the selection of finishes. This fundamental desire
has not been realised for two main reasons. Firstly, the lack of appropriate,
relevant and reliable historical information and data needed for a whole-life
costing (WLC) exercise. Even if this data exist, the time needed for its collection
and analysis may leave inadequate time for the essential dialogue with the
decision-maker and the re-run of alternative options. According to Griffin (1994),
this is why a computerised model can be valuable. The second reason relates to
the way design decisions are made. The design or component selection decisions

266
can often be taken based on non-financial factors, e.g. strength of materials, fire-
protection, hygiene, health and environmental protection, safeguarding of use,
sound isolation, energy saving and thermal isolation, durability and utilisation
(Bogenstatter, 2000). Most of these factors, however, cannot be assessed in a strict
WLC framework because they are mostly intangible or . In many cases, these
intangibles are also in conflict with results of WLC (Wilkinson, 1996).
These arguments are especially true in the complex environment of healthcare
buildings. Hospital design is notably very complex, which makes the selection of
finishes within a hospital also complex. Currently there is an attitude of ‘one fits
all’ when it comes to the selection of finishes, it is a tried and tested approach
where the finish used in the past is the same as for the present. The desire to
reduce variation for economic reasons has to be balanced against a wide variety of
specialist uses and a large number of user groups with widely differing needs.
Reducing variation may help to control maintenance and cleaning as well as
capital costs. On the other hand the use of different spaces varies from office and
general use to very high wear circulation areas and indoor ‘streets’, to ward areas,
to highly specialised theatre areas. Within many of these spaces a range of issues
distinguishes healthcare environments from most other building types and needs
to be considered in the development of the proposed WLC tool. Perhaps the most
important of these issues relates to the control of infection. Hospital environments
in particular are subject to spillage of a range of potentially dangerous substances,
in areas of general use such as circulation areas, as well as in wards. Here the
choice of finishes is not only important in determining cleaning regimes, but may
for example incorporate resistance to the spread of infection through the use of
antimicrobial agents, fungicides etcetera, as additives to applied finishes (Gelder,
2003). Another issue is the impact and disruption of the application process, given
that many finishes are applied in areas in hospitals which continue in use during
the application period. As well as these important physical issues, research has
shown that design and specification, even down to the level of choice of finishes,
may have implications for psychological well-being and hence, again, for
recovery rates (e.g. Lawson and Wells-Thorpe, 2002; Lawson and Phiri, 2003;
Dalke, et al., 2004).
In a previous paper (Laing et al., 2006), major non-financial selection criteria for
hospital finishes have been identified through meetings with strategic team
members in the industry and through a literature search of, for example, health
building notes (HBNs). In another paper (Kishk et al. 2006a), an integrated
framework for the selection of hospital finishes has been outlined. The
development of a novel application to facilitate effective WLC based selection of
hospital finishes has been reported in a subsequent paper (Kishk et al., 2006b).
The logic of the application has been designed around two generic whole-life
costing (WLC) databases. The first is a resource database that houses data for

267
several options for various finishes suitable for various spaces of hospital
building. The second database accommodates project specific data required for the
analysis including economic, geographical, space and other data. Detailed
explanation of the structure and content of both databases will be reported in two
future papers because of space limitations.
The objective of the research work reported in this paper is to develop a decision
support system by integrating the extended WLC application and the two WLC
databases through an interactive interface. In the following section, the integration
logic is introduced. Next, the system interface is reported in detail. Then, the
usability and efficacy of the system is demonstrated through an example
application. Finally, directions for further research are introduced.
2. The Integration Logic
2.1 The Space Concept
A hospital building can be conveniently defined as a collection of spaces. Within
a healthcare environment, the use of different spaces varies from office and
general use to very high wear circulation areas and indoor ‘streets’, to ward areas,
to highly specialized theatre areas. Within many of these spaces a range of issues
distinguishes healthcare environments from most other building types and needs
to be considered in the development of the proposed system. Each space is
defined by up to three boundaries: its floor, its walls and its ceiling. In this way,
three types of finishes are identified for each space: an internal wall finish, a floor
finish, and a ceiling finish. Within a typical project, spaces should be clearly
defined and coded to facilitate the design and management of the building. The
use of the ‘space concept’ will facilitate varying the selection criteria and their
relative weights of importance for various spaces as discussed above. This view
has been supported by a recent survey undertaken by the research team to get the
view of hospital professionals in various NHS trusts throughout the UK.
2.2 The Integration Framework
Figure (1) shows the integration logic of the system. As shown, the interface is
designed around the space concept. For each space, the selection criteria for each
of the space finishes and their relative weights of importance are defined. Based
on these criteria, a number of competing alternatives are generated and passed to
the extended WLC application. The optimum alternative is then identified based
on a value for money analysis taking into account both financial and non-financial
criteria. Data of the optimum set of the current space is inserted into the project
database. Various WLC profiles can then be generated by the profiler.

268
Alternatives
WLC Application
Optimum set
Project
Database
Criteria
Profiler
Interface Analysis
Space
Resource
Database Alternatives
WLC Application
Optimum set
Project
Database
Criteria
Profiler
Interface Analysis
Space
Resource
Database
Figure 1. Simplified diagram of the integration logic
3. System Design and Testing
3.1 The Resource Database Interface
In MS Access, two ways of data entry are possible. In the first method, the
targeted table is opened in datasheet view and new data can be entered directly.
However, this method is not convenient especially when dealing with large
amounts of data and/or with data related through two or more tables. This method
has been adopted in adding data for simple auxiliary tables. In the second method,
data can be entered through a customized form that includes a selection of fields.
This method has three advantages. First, it is user-friendlier. Secondly, additional
information on the required data can be displayed. Thirdly, and more importantly,
data for fields in two or more tables can be simultaneously entered and
automatically included in these tables.
A single user-friendly form has been designed within MS Access to facilitate data
editing of the resource database. As shown in figure 2, data can be easily edit on
four levels of detail: (1) the space options level; (2) the option activities level and
(3) the activities cost items level; and (4) the cost items detailed components level.
The form is well organized. In addition, a concise yet useful help text is
automatically shown to assist the user during data entry. Besides, combo box
controls are used, whenever possible, to further facilitate the input process.
Moreover, event procedures written in Visual Basic® have been written for all
controls to make the form more intelligent. These features will be highlighted in a
separate paper.

269
Figure 2. The resource database interface
3.2 The System Interface
3.3 The Main System Interface
Figure 3 shows the main system interface. As shown, the system functionality is
achieved through various menus accessed through the menu bar. In addition, the
current project summary is displayed. The space and criteria deal with defining,
saving, and open saved spaces and their criteria, respectively. The view menu
allows the visualization of various spaces’ WLC profiles. The Database menu
allows configuring and editing the current resource and project databases as
usable data sources. The system is designed around Open Database Connectivity
(ODBC) to access data from both the resource and project databases. So, to use
the system the user must add an MS Access driver to the operating system and
define both databases as data sources. The Help menu provides information about

270
the current version of the system and it is intended to provide detailed help index
on the system in future versions of the system.
Figure 3. Main system interface
3.4 The Space Dialog Box
Figure 4 shows the space dialog box that can be accessed from the space menu.
As shown, it includes several controls required to defining a space and selecting
its finishes. Data required includes identity, type and physical data. Identity space
data includes a unique number and/or name and is required to uniquely define a
space within a given hospital building.

271
Figure 4. The Space definition dialog box
Type data includes its department, e.g. outpatient or surgery, and its category, e.g.
administration, patient room. Type data is used to facilitate the criteria definition
process. Physical data includes the effective space horizontal area, its wall area,
number of internal and external windows and number of door sets. Physical data
are needed to define the space and are later used by the extended application to
calculate various costs for competing alternatives using the rates stored in the
resource database. Besides, there are up to 6 controls to select various finishes of
the space. Obviously physical data define the space. Minimum physical data
includes the main and the wall areas. The controls for selecting internal, external
and door sets are dynamic and appear only if required, i.e. if the user enters a
positive number for any of them.
3.5 The Criteria Dialog Box
Figure 5 shows the criteria dialog box that can be accessed from the criteria menu.
As shown, it includes several controls for four panels of criteria and a group of
reference controls. Criteria panel 1 includes up to 3 generic quality criteria.
Criteria panel 2 includes up to 5 more specific technical criteria. For floor
finishes, these include slip resistance, anti static, permeability, flexibility and
heavy traffic suitability. Criteria panel 3 includes up to 6 construction/operating
criteria. For floor finishes, these include ease of maintenance, ease of installation,
practicality of repair, sound isolation, and suitability for frequent detergent or hard
cleaning. Criteria panel 4 is a flexible user defined group of up to 4 criteria and
can only be used to include additional trade-off criteria. Once a criterion is
defined, its importance level can be set using a 1 to 5 importance scale.

272
Within the first 3 panels, any criterion can be chosen to be a screening criterion, a
trade-off criterion, or both. Screening criteria are those attributes that cannot be
compromised because of limited resources, legal or minimum performance
requirements. They are mainly quantitative criteria that might include financial,
health and safety, statutory and technical criteria. In other words, they are yes/no
criteria used to exclude options early in the selection process. For example, a soft
floor finish is excluded if the hard criterion is checked as a screening criterion for
a patient room.
On the other hand, trade-off criteria are used in the weighted evaluation analysis
where each alternative is scored against these criteria and the scores are added to
calculate a total benefit score. Then, a benefit to cost ratio (BTC) is calculated for
the alternative by dividing the total score by its WLC measure, usually the net
present value (NPV) according to the methodology developed by Kishk (2002). A
screening criterion can still be used in the trade-off analysis after the screening
step to favor alternatives with high performance regarding screening attributes.

273
Figure 5. Criteria definition menu
Default values for criteria inclusion are automatically set for various types of
spaces. The user can, however, change these values if they are not legally or
technically binding. To help the user understand and change these values, it is
important for the user to have access to the right information. The tool therefore
has automated links, within a references section, to relevant information within

274
available technical and/or legislative literature, e.g. health building notes or
Building Standards. These links are automatically updated according to the
selected space. In this way, the user can open relevant literature to verify, inform
or simply to clarify certain finish related issues.
For example, HBN 57 has specific information on flooring, but additionally has
further information on the needs of finishes for disabled, elderly and children plus
more information on finishes in specific locations in the relevant spaces covered
in this document. The tool can be used to quickly link to relevant passages of this
document if further information is required by the user to make a more informed
decision of the selection of finishes.
3.6 The Cost Profiler Tool
Once the ideal finish has been selected, its activity and cost data is automatically
inserted within the space records in the project database. Figure (6) shows the tool
interactive interface where the user can select to view various WLC profiles of a
specific space. The interface includes 2 combo box controls and 5 push buttons.
The first combo box control sets the visualization level. This can be on the space,
the element (or finish), the activity or the cost item levels. The other combo box
control sets the visualization detail. The five push button controls are used to
display the discounted cash flow diagram (CFD), discounted cash flow diagram
(DCFD), cumulative whole life costs (CWLC), remaining whole life costs
(RWLC) and significance factors for the selected level and detail.
Bar charts are used to visualize cash flow diagrams. For cash flows of a single
generic cost, element, activity or cost item, a novel bar chart is employed whereby
low, best and high estimates are shown simultaneously (see figure 9 later). On the
other hand, stacked bar charts are used when the cash flows of multiple generic
costs, elements, activities or cost items are being compared as shown in figure 6.
Line graphs are employed to depict cumulative and remaining whole-life costs
whereby the past and future significance of various costs can be identified. Line
graphs are also used to identify the significance of the building space under
consideration on the generic cost, element, activity, and cost item levels according
to theory presented in Kishk et al. (2003).
3.7 System Testing
The developed system has being tested and validated in three phases. First, the
usability of the system’s interface has been tested in a laboratory environment. A
second phase has been to demonstrate the system to a panel of experts to get
feedback on further refinements to the system. The third phase is being carried out

275
where a case study is used to demonstrate the efficacy of the system. This process
is illustrated in figure 7.
Figure 6. Profiles’ visualization menu

276
`
Testing of prototype Refinement following expert panel assessment
System usability
System effectiveness
Application in the design stage
Application in other stages
Final
System
Final workshops
Testing of prototype Refinement following expert panel assessment
System usability
System effectiveness
Application in the design stage
Application in other stages
Final
System
Final workshops
Figure 7. Process for system testing
The first stage of the testing process required that various constituent databases
operated efficiently, and in a manner allowing for data transfer between parts of
the system. The resource database was rigorously tested to ensure that all data
pertaining to individual materials could be stored by the system, and that the
criteria and options available to users complied with current guidelines. Similarly,
dummy data was inserted in the project database (for each individual building to
be included), to ensure that the system raised the correct ‘prompts’ for users, and
to ensure that the resulting data analysis was mathematically correct.
The second stage of refinement, which involved presentation to a panel of experts
acting as a project steering group, resulted in two key alterations to the system.
The first of these involved contents of the resource database, which it was
determined should be populated with generic material ‘types’ at this stage, which
could potentially be overwritten or superseded by actual product data once the
system came into use. Secondly, the decision was also taken to establish two case
studies, each of which would contain resource and project data pertaining to two
hospital spaces, these being a patient room (a ward space) and a bathroom
(constructed using prefabricated technologies). The rationale for this was to
provide a basis upon which participants in the third stage workshops can assess
the application of the system in simulated design and maintenance situations.
Stage three of the testing involved running a series of interrelated workshops,
each focused on a different aspect of the system. The first considered usability,
and concentrated on the user interface, transparency of the system, the ability to
generate useful information prompts (including HTMs) and the effect of user
choice on outputs. The second considered the potential effectiveness of the
system, which also required a discussion of the form of outputs, but which also

277
considered the manner in which the system could best remain current when in use,
bearing in mind that available data will change over time, and that processes and
guidelines will likewise be subject to alteration and refinement. The final two
workshops used the case studies to simulate the design and maintenance stages for
hospital spaces, and allowed users to better understand how the system could be
of use to them in practice as an effective design and management tool.
4. An Example Application
To illustrate more aspects of the developed system, example outputs for the
selection of a floor finish for a patient space using illustrative dummy data are
shown in figures 8 and 9. To demonstrate the efficacy of the system, all data used
were represented either by probability distribution functions or fuzzy numbers.
Figure 8 shows the final ranking of the four competing alternatives taking into
account non-financial criteria using benefits to cost ratios (BTCs) according to the
theory represented in Kishk (2002). As shown, the ideal alternative is the finish
with the maximum BTC.
00.2 0.4 0.6 0.8 11.2 1.4
x 10
-6
0
0.2
0.4
0.6
0.8
1
↓
FL1 Linoleum sheet
Ranked first
BTC
µ
ALT E R NAT I VE BTC
ORD
FL1 Linoleum sheet 5.200519e-007
Anti-slip Vinyl sheet 4.328888e-007
Anti-static Terrazo 3.637694e-007
FL2 Vinyl rolls 3.033722e-007
BTC
ORD
RAN KING
Figure 8. Final ranking based on BTC ratios
Figure 9 shows the remaining whole-life costs whereby the past and future
significance of various costs can be identified. Besides, the prevailing recurring
type of contributing costs can be easily spotted. For example, the smooth RWLC

278
curve of the cleaning activity in figure 9 clearly shows that almost all costs of this
element are due to annual continuous activities; while the stepped curve of the
miscellaneous repair activity indicates it is a non-annual recurring activity.
5. Summary and Conclusions
A user-friendly system has been developed to facilitate the selection of hospital
finishes in order to meet health, safety, statutory and environmental needs over its
life cycle in a cost effective way. The system has been developed by integrating a
recently developed extended whole-life costing (WLC) application with two
generic databases using a user-friendly interface. The beneficiaries of the
developed system include NHS Trusts and other healthcare providers, patients,
academia and the construction industry. The NHS will benefit from testing and
implementing the system and the development of an ‘intelligent’ database that
would enable running costs to be monitored, measured and controlled. This can
improve the productivity and efficiency of material selection and maintenance
decisions. Furthermore, it will help the NHS to influence the uncertain
relationships between present and future costs and values of various finishes and
increase the transparency of the decision making process. All of these issues have
a direct bearing the economics of running healthcare buildings.

279
Figure 9. Remaining WLC contributions of the selected floor finish activities
Patients will benefit from reduced risk of falls and reduced severity of injuries
should falls occur, as well as improved recovery rates. The proper choice of
finishes and their cleaning regimes will help to increase the resistance to the
spread of infection through the use of antimicrobial agents, fungicides etcetera.
Furthermore, research has shown that good design, including well chosen finishes,
can have implications for psychological well-being and hence, again, recovery
rates. Academics will become aware of recent technological and process advances
of WLC applications. The industry will also benefit by sharing the information
available on the resource database. There is also a recognition that the system will
lead to a cultural changes in the approach to design decision making in the
industry.
It is hoped that the tool can act as a persuasive instrument to encourage good
design and thoughtful finishes specification. The wider market for the system will
be tested in a launch event aimed at all major UK healthcare organizations, which
will be followed up by training and feedback events by agreement with the
Department of Health.

280
Acknowledgement:
The research work that underpins this paper has been mainly funded by the UK
Department of Health Estates research and development fund.
References:
Bogenstatter, U. (2000) Prediction and optimization of life-cycle costs in early
design. Building Research & Information, 28(5/6), 376-386.
Dalke, Hilary, et al. (2004) Lighting and colour for hospital design: a report on
an NHS Estates funded project. London : Stationery Office.
Gelder, J. (2003) Specification Checklist: Hospital Finishes. RIBA Publications.
Griffin, J. J. (1993) Life cycle cost analysis: a decision aid. In Bull, J. W. (ed.),
Life cycle costing for construction. Blackie Academic & Professional,
Glasgow, UK.
Kishk, M. and Al-Hajj (2000) A fuzzy model and algorithm to handle subjectivity
in life cycle costing based decision-making. Journal of Financial Management
of Property and Construction, 5(1-2), 93-104.
Kishk, M. (2002) Towards effective consideration of non-financial factors in the
design and management of construction assets. Journal of Financial
Management of Property and Construction, 7(3), 163-174.
Kishk M., Al-Hajj A., Pollock R., and Aouad, G. (2003) Effective feedback of
whole-life data to the design process. Journal of Financial Management of
Property and Construction, 8(2), 89-98.
Kishk, M., Laing, R., Edge, M. & Scott, J. (2006a) An integrated framework for
the optimal selection of hospital finishes - system algorithm and methodology.
The Joint International Conference on Decision Making in Civil and Building
Engineering, Montreal, June 14-16, pp. 840-847.
Kishk, M., Laing, R., & Edge, M. (2006b) An extended whole-life application for
the selection of hospital finishes. The 22nd Annual Conference of the
Association of Researchers in Construction Management, The University of
Central England, Birmingham, 4-6 September, Vol. 2, pp.719-728.
Laing, R., Kishk, M., Scott, J., & Edge, M. (2006). On the selection criteria of
hospital finishes. The Joint International Conference on Decision Making in
Civil and Building Engineering, Montreal, June 14-16, pp. 822-829.
Lawson, B. and J. Wells-Thorpe (2002). The Effect of the Hospital Environment
on the Patient Experience and Health Outcomes. The Journal of Healthcare
Design and Development, March, 27-32
Lawson, B. R. and M. Phiri (2003) The Architectural Healthcare Environment
and its Effects on Patient Health Outcomes. London, The Stationery Office.
Wilkinson, S. (1996) Barriers to LCC Use in the New Zealand Construction
Industry. The 7th International Symposium on Economic Management of
Innovation, Productivity and Quality in Construction, Zagreb, 447-456.