You are on page 1of 11

Automation in Construction 48 (2014) 5363

Contents lists available at ScienceDirect

Automation in Construction
journal homepage: www.elsevier.com/locate/autcon

A user-centered information and communication technology (ICT) tool


to improve safety inspections
Ken-Yu Lin a,, Meng-Han Tsai b, Umberto C. Gatti c, Jacob Je-Chian Lin d, Cheng-Hao Lee d, Shih-Chung Kang d
a

Department of Construction Management, College of Built Environments, University of Washington, 120 Architecture Hall, Box 351610, Seattle, WA 98195, United States
Center for Weather Climate and Disaster Research, National Taiwan University, Taipei, Taiwan
Department of Construction Management, University of Washington, Seattle, WA, United States
d
Department of Civil Engineering, National Taiwan University, Taipei, Taiwan
b
c

a r t i c l e

i n f o

Article history:
Received 14 August 2013
Received in revised form 16 August 2014
Accepted 27 August 2014
Available online 16 September 2014
Keywords:
Construction safety
Safety inspection
Site inspection
Field inspection
Safety audit
Field data collection
User-centered design
Information and communication technology
Safety technology
Research to practice

a b s t r a c t
Occupational safety is imperative in construction, and safety inspection is among the most common practices
that help enforce job safety on site. The safety inspection process, however, suffers from several drawbacks
that hinder the efciency, effectiveness, and analytical learning capacity of the process. Dedicated tools for
user-centered information and communications technology could signicantly reduce such drawbacks. This
paper discusses the use of an original two-step user-centered design approach to develop and evaluate an iPad
application that aims to address such drawbacks and improve the day-to-day practices and management of safety inspections. Evaluation results indicate the usefulness and practicality of the application and identify innovative uses not previously envisioned. Furthermore, the developed tool allows consistent data collection that can
eventually be used to aid the development of advanced safety and health data analysis techniques.
2014 Elsevier B.V. All rights reserved.

1. Introduction
Workforce safety is an important topic across global construction industries [1]. According to the United States (US) Bureau of Labor Statistics, more than 1000 work-related fatalities took place each year
between 1994 and 2011 in the US construction industry on average.
Furthermore, construction consistently ranks as one of the top three
most dangerous industries in the US, with the greatest total number of
work-related fatalities among all industries.
In the US, safety regulations, in particular the General Duty Clause
from the Occupational Safety and Health Administration (OSHA), require that employers provide their employees with safe and healthy
working environments free from recognized hazards. To meet this requirement, contractors typically adopt a mix of safety approaches,
such as regular safety meetings, substance abuse programs, task specic
safety training, and pre-project safety planning. Among these common
approaches, conducting regular and frequent construction site safety
Corresponding author. Tel.: +1 206 616 1915.
E-mail addresses: kenyulin@uw.edu (K.-Y. Lin), menghan@caece.net (M.-H. Tsai),
ucg@uw.edu (U.C. Gatti), jacob@caece.net (J. Je-Chian Lin), piggyhoward@caece.net
(C.-H. Lee), sckang@caece.net (S.-C. Kang).

http://dx.doi.org/10.1016/j.autcon.2014.08.012
0926-5805/ 2014 Elsevier B.V. All rights reserved.

inspections is particularly important [2,3]. Abudayyeh et al. [4] conducted a survey concluding that the injury and illness incidence rates of
companies that performed safety inspections were signicantly lower
than those that did not. By analyzing the total OSHA recordable injury
rate of 59 projects, the Construction Industry Institute [5] concluded
that the practice of checking safety inspection records on a regular
basis is generally associated with projects that have better safety performance. Kaskutas et al. [7] determined that safety inspections could
measure the risks of observed projects. Aksorn and Hadikusumo [6] suggested that safety inspections are very effective in preventing accidents.
Although safety inspections are a successful and widely used strategy for improving safety in construction, the inspection process lacks a
comprehensive and structured procedure and is accompanied by ineffectiveness and inefciency throughout. For instance, during a typical
safety inspection, a safety specialist looks for violations on site and
takes notes to record observed issues. However, inspection notes
taken by different safety specialists may vary greatly for the same type
of issues, making it difcult to have a systematic understanding of the
observed issues. Current practices do not take advantage of the time
and resources that safety professionals have already committed during
site inspections. Therefore, repetitive steps are taken to transform eld
notes into ofce les and then administrative reports. In addition,

54

K.-Y. Lin et al. / Automation in Construction 48 (2014) 5363

inspection results are rarely analyzed further to serve as performance


indicators for administrative and management use or to reveal the unsafe patterns on site.
These limitations may severely reduce safety specialists' effectiveness and efciency in collecting and compiling eld observations, and
as a result may hamper their ability to monitor site safety performance.
However, dedicated information and communication technology (ICT)
tools, such as portable tablets capable of retrieving applicable safety
procedures, rules and regulations, and software capable of automating
recurring activities (e.g., creation of violation statistics and reports),
could signicantly improve the day-to-day practices and management
of safety inspections.
Numerous ICT tools have been benecial for the construction industry. Goodrum and Haas said, Many industries have spent considerable
time and money studying how technology inuences productivity.
These studies have led to sizeable gains in productivity and prot margins [8]. Furthermore, the Construction Industry Institute [9] stated,
Advances in technology have many benets. Among the most often
cited are improved quality and productivity. ICT tools also provide benets to workforce safety and well-being. For instance, ICT innovations
allowing the industrialization and automation of work tasks were
considered to be one of the main factors preventing a signicant increase of injury rate in the US construction industry during the 1990s
[10]. The importance of ICT-enabled automation in improving safety
was also supported by Kim and Cho [11] and Cinkelj et al. [12]. Han
et al. [13] and Sulankivi et al. [14] implemented building information
modeling (BIM) and a 4D model for safety planning. Chi et al. [15],
Teizer [10], and Walia and Teizer [16] employed 3D imaging sensors
to reduce the occurrence of collisions within a construction site. Wu
et al. [17] used a radio-frequency identication (RFID) sensor network
to create an autonomous real-time tracking system of near-miss
accidents and Yang et al. [18] applied the same technology to identify
accident precursors.
However, many innovative technologies that have been proven to be
benecial are not commonly adopted by construction practitioners.
There are different reasons for this. First, it is well known that the construction industry is considered reluctant regarding the adoption and
implementation of innovations [19,20]. Koningsveld and van der
Molen commented that As we look at the pace of innovation in other
branches of industry, the building and construction industry should be
characterized as most conservative [21]. Furthermore, ICT tools have
traditionally been developed by adopting a technology-centered design
[22]. A technology-centered design occurs when researchers develop a
new technology or apply an existing technology to a different eld,
without considering users' needs and capabilities. Thus, a technologycentered design forces users to adapt to the new technology and eventually fosters the occurrence of errors. According to Rogers [23], a
technology-centered design also implies that innovative technologies
are developed without considering factors that can signicantly affect
innovation diffusion and acceptance. Such factors include relative advantage (the degree to which an innovation is perceived as being better than the idea it supersedes [23]), compatibility (the degree to
which an innovation is perceived as consistent with the existing values,
past experiences, and needs of potential adopters (p. 15)), complexity
(the degree to which an innovation is perceived as relatively difcult to
understand and use (p. 15)), trialability (the degree to which an innovation may be experimented with on a limited basis (p. 16)), and observability (the degree to which the results of an innovation are
visible to others (p. 16)). Several studies have demonstrated that
when an innovation has relatively higher advantage, compatibility, simplicity, trialability, and observability, the innovation also has a higher
chance to be extensively and rapidly accepted [24]. In fact, Mitropoulos
and Tatum [25] suggested that uncertainty in obtaining benets or competitive advantages from using innovative technologies is among the
barriers that impede the introduction and development of innovative
technologies in construction.

To mitigate technology-centered design issues and limitations,


innovative ICT tools should be designed and developed through a
user-centered (also known as human-centered) approach [26]. When
a user-centered design is used, researchers develop a new technology,
or apply an existing technology to a different eld, by adapting it to
users' needs and capabilities and understanding how users interact
with it [27,28]. The employment of user-centered design principles
has been shown to be successful [29] and the importance of usercentered design is also emphasized by the fact that the International
Organization for Standardization [30] issued a specic standard for
user-centered design. The limitations of current safety inspection procedures can be addressed by dedicated ICT tools, but introducing innovative ICT tools in construction cannot be successful without a usercentered approach.
To this end, the objectives of this study are twofold. First, through a
user-centered approach, this study aims to verify the technological requirements that correspond to the safety inspection procedures and
their management implications. Second, based on the veried technological requirements, this study intends to develop an iPad (Apple) application for potential users to experience how the tool can effectively
support the day-to-day practices and management of safety inspections. Echoing Mitropoulos and Tatum's suggestion [25], adopting a
user-centered approach as the main research strategy is expected to reduce user uncertainty about the technology and illustrate the competitive advantages of using the technology.
2. Background: The safety inspection process
Based on eld observations, the safety inspection process can be divided into three phases. These are the project information collection
phase, the recording of observed violations phase, and the administration of inspection results phase. The conceptualized safety inspection
process is illustrated in Fig. 1.
After the details of a project are mostly settled and as the project
commences, safety specialists responsible for the project begin to conduct inspections at the project site. Generally, the site inspection
frequency depends on the scale and importance of the project. During
a site inspection, a safety specialist typically takes notes of any violations
and safety issues identied and communicates with workers on site to
express observed concerns. Upon returning to the ofce, the specialist
then recaps and compiles the inspection results into a report. Inspection
results are often further discussed during regular project management
meetings to prevent similar issues from recurring, to target specic
areas for training, and to raise safety awareness among all employees
[31]. Eventually, the inspection results can be used to identify strong indicators for safe (or unsafe) projects and improve site safety performance by identifying and understanding the trend of unsafe working
conditions/behaviors. The inspection results can also potentially be
used to establish relationships between project safety and other aspects,
such as schedule, productivity, and cost of the project. An integrated approach that examines the results of safety inspections and productivity
has been explored [32] and such efforts could inform organizations on
how project management factors inuence each other. In addition, integrating safety inspection results with other aspects of the project has
proved to be benecial in reducing accident rates and improving productivity [33].
The safety inspection process, however, is ineffective and inefcient
because of several drawbacks in current practices. The ve main drawbacks are described in the following paragraphs.
Drawback 1: Lack of Process Standardization
A safety inspection is expected to identify safety issues related to the
various trades, means, methods, and materials of construction, after
considering all applicable safety standards [34]. However, the volume of applicable safety standards is sizable and it is impossible
for safety specialists to verify whether all applicable standards are

K.-Y. Lin et al. / Automation in Construction 48 (2014) 5363

Phase 1: Collecng
Project Informaon

Phase 2: Recording
Observed Violaons

Phase 3: Administrang
Inspecon Results

Bid informaon

Site Inspecon

Violaons and issues are


compiled in a report and/
or tracked

Project and
project
parcipant
informaon

Idenfy and record


violaons and
issues
Discuss violaons
and issues with
workers

55

The collected data can be:


Analyzed to improve
safety performance
Integrated with other
project data (e.g.,
producvity)
Etc.

Fig. 1. The process of safety inspection.

satised. Alternatively, safety specialists use their experience to


check the overall site surroundings and identify the most alarming
issues on site. The inspection process lacks standardized procedures
and is prone to errors and bias. This could explain why the inspection results are not always recorded and why corrective actions are
not taken every time [31].
Drawback 2: Lack of Standardized Documentation
In discussing how to develop effective quality, environmental, and
safety management systems, Abudayyeh et al. [4] suggested that
preparing and implementing standard inspection checklists can be
extremely benecial in improving poor safety performance. Without
standardization of inspection documentation, ineffectiveness can
occur owing to arbitrary descriptions of observed safety issues
(e.g., face and eye protection violation versus missing welding
shield) or inconsistent references to the same project information
(e.g., Project at Pine Street versus Pine Tower). Consequently,
the quality of inspection outputs is reduced, making it difcult to
compile inspection records for analysis.
Drawback 3: Restricted Access to Information
Safety specialists have to walk around the construction site to check
for safety issues. However, research indicates that the inspection
process is not well-prepared [35]. On-site personnel and/or documentation often cannot provide timely information (e.g., the name
of a particular subcontractor or trade involved), thus requiring additional efforts to locate the required information and making the
recording process inefcient. Furthermore, it may be difcult for
safety specialists to provide specic regulations to workers when
discovering safety issues on site as applicable safety rules or training
materials might not be immediately accessible or retrievable.
Drawback 4: Repetitive Data Preparation
Poor accident recordkeeping and reporting systems are the major
problems that cause ineffectiveness for safety management [36].
During safety inspections, there are multiple recordings of the
same inspection results across several mediums (e.g., paper notebooks, inspection reports, and spreadsheets). This makes the process
more prone to documentation errors, such as those introduced by
multiple and manual entries of the inspection results. In addition,
pictures taken on site separately from the eld inspection notes
have to be integrated with the notes. Such integration generates an

additional complicated and time-consuming task that may foster


mistakes and inaccuracies during data compilation.
Drawback 5: Limited Availability of Safety Specialists
The number of safety specialists that a contractor employs depends
on the company's size. For a small- to medium-sized company, its
sole safety specialist often has to oversee many projects at the
same time. It was found that there are not enough on-site safety
personnel to bear the workload [6] and safety specialists can only
conduct site inspections sporadically [34]. This may decrease the capacity of safety inspections when it comes to the identication of issues and violations. For example, the limited availability of a safety
inspector spread across multiple projects might lead to simplication of the inspection process or ignorance of some violations. To further challenge the limited availability of safety specialists, there is a
generation gap between novice and experienced safety specialists.
This may be due, for instance, to economic decline and the loss of
talented employees who turned to other industries for employment.
Experience is a key factor for recognizing violations [37] and therefore improving the availability of experienced safety specialists
through better work efciency and training is critical.
These ve main drawbacks negatively affect the inspection process
and its mission to identify and record eld issues and violations. The
lack of standardized documentation and the repetitive data preparation
particularly affects the management of inspection results by hindering,
for example, the further investigation of safety performance and the integration of safety data into other aspects of project performance. There
is an obvious need to streamline the safety inspection process and maximize the effectiveness and efciency of safety inspections.
2.1. Information and communication technology tools for safety inspections
Given the issues and limitations affecting the safety inspection process, contractors have turned to ICT for remedial solutions. Existing
ICT tools for safety inspections are mostly developed on mobile devices,
from PDAs and smart phones to tablets, for portability purposes and
eld applications. Existing ICT tools can be classied into three main categories: (1) safety auditing tools, such as iAuditor from Safety Culture;
(2) eld management tools, such as BIM 360 Field from Autodesk, Pervidi
from Techs4Biz, and Pocket Jobsite Inspector from PDAge; and (3) data
analysis tools, such as SafetyNet from Predictive Solutions. Although all

56

K.-Y. Lin et al. / Automation in Construction 48 (2014) 5363

these tools are intended to support safety specialists in recording observed issues and violations during site inspections and administrating
inspection results, they provide signicantly different functions.
Safety auditing tools mainly provide functions for recording the
inspection results. For instance, iAuditor presents templates of site inspection checklists for user selection and customization, and creates inspection reports that can be printed or saved in a digital format. A typical
inspection report contains an overall safety score to indicate the safety
performance of a project. However, iAuditor does not support integrated
communication with subcontractors, and cannot analyze the collected
data (e.g., to identify trends of unsafe working conditions/behaviors) or
merge them with other project performance measurements.
Field management tools provide comprehensive functions for quality
and safety management. For instance, BIM 360 Field consists of a series
of eld data management applications that allow construction project
activities to be recorded and stored in a centralized cloud database.
Since project participants can access this database, it can support integrated communication between project managers and subcontractors.
BIM 360 Field can also store the collected data and generate summary
reports to assess delays, rework, and punchlist items. However, BIM
360 Field does not have an integrated application to further analyze
the collected inspection records.
Data analysis tools provide powerful applications to perform comprehensive and statistical analysis. For instance, SafetyNet provides

Phase 1: Collecng
Project Informaon

2.2. Key functions of information and communication technology


By analyzing (1) the specic steps involved during safety inspections, (2) the ve drawbacks in current practices, and (3) the available
functions of existing ICT tools, the authors conclude that any ICT tool
must provide a series of key functions to effectively support the safety
inspection process. The authors used a bottom-up approach to link
major functions of existing ICT tools with the three inspection phases.
Information from tool vendors' websites and available trial versions
provided insights into the existing ICT tools. The identied key functions
are described below and divided according to the three main safety inspection process phases, with Fig. 2 further illustrating how these key
functions correspond to the recognized drawbacks:
Phase 1. Collecting project information:
Pre-congured lists of project and project participant information. Information about a project (e.g., project name and location) and its
participants (e.g., subcontractor names and performed tasks) can
be incorrectly reported in the documents generated during safety

Phase 2: Recording
Observed Violaons

Project and
project parcipant
informaon precongured lists
(For drawbacks 1, 3 & 5)

data learning algorithms to analyze the collected data in order to identify and examine safety-related issues. This tool specically aims to predict workplace injuries by determining leading indicators and can be
used to record issues and violations identied during site inspections.

Hardware
portability

Phase 3: Administrang
Inspecon Results

Safety score

(For drawbacks 3, 4 & 5)


Integraon with
other project data

Photo capturing

(For drawbacks 4 & 5)

Integrated
communicaon
with project
parcipants

Pre-congured
safety checklists
(For drawbacks 1, 2 & 5)
Access to
applicable safety
standards

(For drawbacks 4 & 5)

(For drawbacks 1 & 5)


Analysis of
violaons trends/
paerns

(For drawbacks 3 & 5)


Reminder for
related violaons

(For drawbacks 1 & 5)


Legend
BIM 360 Field

iAuditor

SafetyNet

Project
administraon
communicaon
(For drawback 5)
of the tools
provide the funcon
None

Fig. 2. Comparison of current ICT tool functions and key functions.

K.-Y. Lin et al. / Automation in Construction 48 (2014) 5363

Safety score. ICT tools should provide a site safety performance


comprehensive measure, such as a safety score, based on the inspection results. In fact, such a measure could provide the project
management team with an easy-to-relate indication of the site
safety level and eventually suggest directions for improvement.
Integration with other project data. The collected safety data can be
used to analyze the relationship between safety and other project
aspects, such as schedule, productivity, and cost. Therefore, ICT
tools should allow the data to be exported to other tools capable
of integrating and synthesizing the different data.
Integrated communication with project participants. Integrating inspection results into the communication protocols among project
participants (e.g., general contractor and subcontractors) can increase participant awareness of the safety issues and ensure that
no issues are overlooked. Therefore, ICT tools should support
such integration.
Analysis of violations trends/patterns. The analysis of historical data
collected across several sites may identify trends/patterns between
specic causal factors and the occurrence of violations, accidents,
and injuries. Understanding these casual factors helps to minimize
or prevent the occurrence of related violations, accidents, and injuries. Although this function does not specically address any of the
ve drawbacks, it is essential to the establishment of a proactive
safety and health culture and should be included in the intended
ICT tools.
Project administration communication. Providing corporate leadership with a high-level overview of the inspection results is imperative to the success of overall business management, but it is often
deferred until monthly or quarterly meetings with the leadership.
Therefore, ICT tools should allow safety specialists to communicate
the results obtained to the project management in a timely
manner.
As shown in Fig. 2, not all the key functions are fully supported by
the current ICT tools even if they are integrated into one solution and
there does not seem to be reported efforts on the unfullled functions.
Some tools are also too costly or bulky for small- to medium-sized contractors. Furthermore, by considering the general reluctance of construction practitioners in adopting innovations and the lack of ICT
tools providing a user-oriented experience, it is not surprising to see
that when the authors surveyed 20 medium-sized general contractors

in the Seattle area in late 2012, 85% of them still resorted to the penand-paper approach for conducting safety inspections, even though integrated project or quality management solutions generally have a
placeholder for safety records. As such, the safety inspection process
continues to suffer from the drawbacks outlined above.
3. Methodology
In this section, the authors explain the user-centered approach they
adopted in the development of an ICT tool, which incorporates the
technological requirements from Section 2 so that potential users can
experience how the tool can support the day-to-day practices and management of safety inspections. Although user-centered design has been
extensively discussed in several publications, no preferred method has
been advocated. For instance, the International Organization for Standardization [30] determines the overarching process and phases
(Fig. 3) but does not detail the exact methods.
By applying user-centered design principles and the ow chart in
Fig. 3, two main development phases were performed (Fig. 4). First,
the authors designed a paper prototype based on actual safety inspection procedures and administered two evaluations (i.e., interim and expert evaluations) using mock-up violation scenarios and an evaluation
questionnaire. Second, the authors developed an application prototype
based on the rened paper prototype and consulted safety specialists
to evaluate the application prototype through eld tests. The development and evaluation stages were heavily informed by user inputs and
are described in detail in the following sections.
3.1. Informed design of the template
The rst step of the development and evaluation of the ICT safety
inspection tool was to initiate a paper prototype that can help communicate and reect the process of safety inspection. The draft design of the paper prototype was informed by the experience of one
of the co-author faculty internship with a general contractor from
the greater Seattle area [38]. A single internship cannot serve as the
only point of reference but is a good starting point with the exibility

START
Idenfy need for usercentered design

User-centered design phases

inspections, or require additional efforts for verication. Therefore,


ICT tools should provide pre-congured lists to contain reusable
project and participant information.
Phase 2. Recording observed violations:
Hardware portability. To allow safety specialists to use an ICT tool
during site inspections and instantly record the observed violation,
ICT tools should run on a portable device.
Photo capturing. Pictures of identied violations can be a powerful
instrument in recording and describing the violations. Therefore,
ICT tools should integrate a photo capturing function.
Pre-congured safety checklists. Having pre-congured safety
checklists stored in the ICT tool would ease site inspection procedures and improve process and documentation standardization.
Access to applicable safety standards. It is extremely unlikely that a
safety specialist will carry a paper copy of all the applicable safety
standards and regulations during inspections. Therefore, ICT tools
should store a digital copy of such safety standards and regulations,
and allow safety specialists to quickly locate applicable rules for
discussing with workers on how a violation or issue should be
addressed.
Reminder for related violations. Some violations are generally concurrent with other violations. Therefore, ICT tools should automatically remind safety specialists to investigate certain violations
when related violations are identied.
Phase 3. Administrating inspection results:

57

Understand and specify


context of use
Specify user and
organizaonal requirements
Produce design soluons
Evaluate design against
requirements
Does the design sasfy
user and organizaonal
requirements?

YES
END
Fig. 3. General user-centered design ow chart.

NO

58

K.-Y. Lin et al. / Automation in Construction 48 (2014) 5363

Fig. 4. Development and evaluation process.

to allow subsequent changes. Specically, formats, content, and


recording logic of desirable data for the preservation of site inspection results were identied.
Safety issues of different types, as indicated by the host company's internal site inspection records, were clustered to reveal major types
(e.g., Personal Protective Equipment (PPE) related violations) and subtypes of safety issues under each theme (e.g., missing safety glasses).
Safety inspection resources from different general contractors were

also reviewed to reveal additional common precautions (e.g., policy)


for inclusion in the paper prototype. Finally, the topological structure of
the US OSHA's construction safety standards was applied to adjust and
complete the base paper prototype design. Fig. 5 illustrates the violation
types (e.g., Scaffold and Fall Hazards) and subtypes (e.g., Guardrail)
dened in the base design. At this point, the paper prototype was essentially a static inspection form for organizing and recording all possible
safety issues on site.

Tool Safety
Powder Actuated Tool

Scaffold
Fall Hazards
Guardrail
Fall Protection
Other, please specify:

Power Tool
Machine Guard
Saw Guard
Other, please specify:

Trenching
Policy
Smoking

Electrical
Cord
Other, please specify:

Headphones
Other, please specify:

Ladder Safety
PPE (Personal Protective Equipment)
Face Shield
Safety Glasses
Hard Hat
Gloves
Other, please specify:

House Keeping
Equipment Operation
Crane
Scissor Lift
Forklifts
Vehicles
Other, please specify:

Hazardous Chemicals
Fuel Gas

Work and Public

Other, please specify:

Other, please specify:

Fig. 5. Initial draft of the inspection form.

K.-Y. Lin et al. / Automation in Construction 48 (2014) 5363

3.2. Mock-up violation scenarios and evaluation questionnaire


By utilizing pictures of common site violations, 20 mock-up violation scenarios were compiled and equally divided into two sets
(i.e., basic and advanced). Violations in the basic set were more
straightforward and easier to spot whereas violations in the advanced set were generally more ambiguous or involved more than
one major hazard.
In addition to the violation scenarios, a questionnaire for evaluating
the paper prototype was drafted. In particular, the questionnaire
assessed four areas that the paper prototype could support in the site inspection process, including the form's effectiveness (i.e., does the form
help to describe the given violation scenarios?), coverage (i.e., does
the form cover most of the on-site violations?), ease of use (i.e., is the
form easy to understand and work with?), and other potential use
cases of the form.
Once the violation scenarios were compiled and the evaluation
questionnaire was drafted, they were pre-tested by two University
of Washington construction management graduate students. The
students had more than 3 years of experience in the construction industry and were not involved in drafting the violation scenarios or
the evaluation questionnaire. As a result of this process, unforeseen
evaluation problems (e.g., violation scenario pictures being unclear)
and confusing questions in the questionnaire were identied and
addressed.
3.3. Interim evaluation
In order to obtain a consensus on how various safety issues should
be grouped in the paper prototype to make sense in the eyes of potential
users, an interim evaluation of the general organization of prototype
content was conducted with 65 University of Washington undergraduate students taking a class on construction safety. The students were
tasked with performing mock site inspections for the basic set of violation scenarios using the paper prototype. This helped verify if the grouping structure of the safety issues in the prototype was logical. If most
students in the class could assign the same violation to a particular category, then the category was considered logical. Since not all students
had construction experience prior to the test, those who recorded a
given violation scenario under an obviously inapplicable type ve or
more times were considered outliers and their responses were not considered. The outlier determination threshold came from the most common number of incorrect violation recordings conducted by students
who possessed no prior construction experience. Out of the 65 students
who participated and the 650 violations recorded, only 52 violations
(about 10%) were recorded under inappropriate categories, with scaffold, housekeeping, and tool safety being the top three problematic
categories for inexperienced students. The interim evaluation results
implied that the structure of the violation categories was logical for
the most part.
After the evaluation, students further completed the evaluation
questionnaire and their responses indicated that the form was effective,
comprehensive, easy to deploy, and additionally served as a checklist.
After the interim evaluation, three main changes were made to rectify
or clarify the terminologies used. Specically, Face Shield was replaced by Respirator under PPE, Fencing and Other were incorporated into the subtype Work and Public, and the subtype Policy was
renamed to Company Policy. The entire interim evaluation lasted
about 30 min.
3.4. Expert evaluation
The purpose of the expert evaluation was to conduct a more indepth examination on the paper prototype in order to further rene
and conrm how safety issues should be organized in a standard
recordkeeping form that makes sense to industry practitioners. For

59

this purpose, both the basic and advance sets of violation scenarios
were used. Six eld practitioners from ve companies, including one
safety consulting rm and four general contractors from the greater Seattle area, participated in the expert evaluation. Among the six participating practitioners, one was a superintendent and ve were safety
directors. Together, these practitioners had over 60 years of experience
in construction site safety, with one practitioner having between 5 and
10 years of experience and the others each having more than 10 years of
experience. Each expert recorded the 20 given violation scenarios using
the paper prototype, responded to the evaluation questionnaire, and answered additional open-ended questions. The additional questions
were intended to collect concerns about (1) the current safety inspection practices, the forms used for such practices, and existing analytical
applications of the inspection data records; (2) the applicability of the
classication and terminologies used in the form; and (3) prototype
revision and improvement required. It took each practitioner about
70 min to complete the process.
Although the number of participating practitioners was limited
owing to the nature of the interview approach, the researchers were
able to obtain consistent feedback from the six practitioners. The most
signicant input from these experts was about the coverage of safety
issues. This contradicted the student input, which could be explained
by the safety knowledge these experts had accumulated as their frame
of reference when they evaluated the comprehensiveness of the paper
prototype. As a result, additional violation types and subtypes were
identied. Two new types, Fire Protection and Required Documents,
and their related subtypes were added to the paper prototype. New subtypes under existing violation types including Scaffold, House Keeping, Trenching/Excavation, and Ladder Safety were also added.
Finally, the names of some violation types and subtypes were also revised (e.g., Work and Public to Worker and Public Safety). For the
process-related recommendations, expert feedback further indicated
the hidden relationships between some violation types (e.g., if the
violation is related to Electrical, then it has a high chance of also
being a Tool Safety violation). These relationships, if built into the prototype, could help potential users perform site inspections more rigorously. Some also suggested that the violation types and subtypes be
numbered for easy reference. These recommendations were more related to the inspection process and were considered during the tool prototype development and evaluation. Compared to the initial paper
prototype, the total number of violation types increased from 13 to 15
and the total number of subtypes more than doubled, increasing from
25 to 51. The rst 14 types categorize the most common violations,
whereas type Other (15) allows the recording of rare issues. As a result of this process, the framework of how safety issues should be organized and described in order to practically support the recordkeeping
requirement during site inspections was shaped. Two requirements on
what an ideal inspection tool should do to streamline the inspection
process were also identied.

3.5. Application prototype development


The purpose of the application prototype is to embed critical
functions in the safety inspection process in order to carry out useroriented testing and solicit feedback in the anticipated environment of
the application. The application prototype was developed as an iPad application since, according to Johnson [39], an iPad is still by far the most
prevalent mobile technology for eld applications in the US construction industry. In particular, the application prototype was developed
using FileMaker Pro from FileMaker.
The authors developed the application prototype based on the nalized paper prototype and it included all the ICT tool key functions identied and listed in Fig. 2, except for access to applicable safety
standards. The missing function is planned to be added in a future version of the application as the function itself calls for a substantially

60

K.-Y. Lin et al. / Automation in Construction 48 (2014) 5363

different scope of investigation and expertise in text classication. Key


functions included in the application prototype are as follows:
Pre-congured lists of project and project participant information. After a
project commences, trades involved on site, names of subcontractors
employed, project title and prole number all become available information. Therefore, the application prototype allows the input of such
information in pre-congured lists (Fig. 6) so that users do not have to
memorize or key in the information when they are conducting inspections.
Hardware portability. Since the most important function is portability
during site inspections, the authors developed the application prototype in a way that it can be hosted on iPad tablets.
Photo capturing. The photo capturing function utilizes an iPad's camera, allowing users to take and save violation pictures and include
them as a part of the inspection records. In particular, a user can
take up to four photos of each observed violation (Fig. 7).
Pre-congured safety checklists. The authors incorporated the paper
prototype's pre-congured safety checklists into the application
prototype (Fig. 8). The checklist interface design required thoughtful
consideration on how to best organize the checklist information
while minimizing the number of clicks it takes to work through the
checklists.
Reminder of related violations. Different violations may be strongly related to each other. Therefore, the application prototype shows potentially related violations if certain violations are entered. For example,
when an electrical violation is chosen, the bottom of the interface
will show Please check Tool (09) for related violation in red to
remind users that there could be a potentially related hazard. This
function differentiates the application prototype from existing software and could eventually be congured to benet and learn from
collected inspection records over time.
Safety score. This function allows users to enter an overall score for the
conducted inspection.
Integration with other project data. To allow integration with other data
sources, the application prototype can export the data into a spreadsheet le, which is compatible with many common software applications (e.g., Microsoft Excel). Although a shared schema for integrating

Fig. 7. Photo capturing interface.

inspection records with other project data was not explored, as this
topic was outside the user-centered focus of this study, the existing
data-exporting capability engages end-users to brainstorm potentially
benecial data fusion scenarios.
Integrated communication with project participants. The application
prototype provides cloud capabilities (Fig. 9) by hosting an online database operable through FileMaker's services. Furthermore, the system allows safety specialists to email inspection reports to related
project participants (e.g., subcontractors).
Analysis of violation trends/patterns. To identify possible violation
trends, the application prototype is capable of presenting different
charts about the violations. Users can choose to view violation distributions by projects or by subcontractors and understand the violations over different aspects.
Project administration communication. This function supports users in
keeping track of the violation statistics and provides the related violation pictures in the cloud. The application prototype shows the overall
violation distribution of all the projects, and the site safety statistics
displayed in the charts would change immediately when a new violation is recorded. This gives administrators or the leadership immediate feedback.

3.6. Field evaluation

Fig. 6. Project information interface.

After the application prototype was completed, it was evaluated to


see if it satised the practical needs of site inspection and at the same
time exhibited the potential to support future data analysis. Three safety
experts from different construction companies participated in the eld
evaluation. Two of the three experts were the heads of their respective
safety departments while the third expert was a senior safety ofcer.
Two safety experts used the application prototype for 2 weeks whereas
one used it for over 6 months. Each safety expert was instructed to, at a
minimum, use the application prototype to record violations in the eld
(Fig. 10) and administer inspection results (e.g., generate violation reports for meetings and investigations). Members of the research team
participated in several site inspections with the safety experts to

K.-Y. Lin et al. / Automation in Construction 48 (2014) 5363

61

Fig. 10. A safety expert using the prototype tool for site inspection.

Fig. 8. Scaffold safety checklist interface.

introduce them to the application prototype, assess their user experience, and collect their feedback through interviews.
As a user provides feedback, the application prototype improved the
safety inspection process by minimizing repetitive data preparation
tasks and increasing work efciency, thus addressing issues caused by
the limited availability of safety specialists. For instance, many of the application prototype functions helped to reduce the overall amount of
paperwork required (e.g., integrated communication with project participants) and the efforts that safety specialists take to collect project information (e.g., pre-congured lists of project and project participant
information), record observed violations (e.g., hardware portability,
photo capturing), and administer inspection results (e.g., project
administration communication). Repetitive data preparation-related issues were reduced by the functions related to communication, because
once inspection records are generated, they can be reused in different
communication formats, such as emails.

Fig. 9. Data management interface.

The eld evaluation was also extremely benecial because it enabled


the authors to envision new uses of the application prototype that had
not previously been considered. First, the prototype application allowed
safety experts to provide immediate feedback to the superintendents.
For instance, safety experts can show violation pictures to site superintendents before leaving the inspected sites and discuss with them ways
to prevent similar violations in the future. Second, a construction
company decided to use the data collected through the application prototype to rate subcontractors' safety performance as a basis for future
project prequalication. The company also used the data to assess superintendents' safety performance in order to determine their yearly
bonus.
Furthermore, through eld evaluations, the authors identied three
problems in the application prototype. First, the iPad 2 was used as a
platform and the iPad camera was not good enough for taking violation
pictures in low-light situations. As pictures serve as good evidence for
recording violations, the issue with the tablet camera can only be improved when its manufacturer upgrades the specications. Second,
even though the iPad is a portable device, use of the tablet on site is
still far from ideal. Users had to be mindful with the device to avoid
rain, scratches, and falls. Providing a full protection case for the device
could minimize this problem, but also makes it difcult for users to
type or check the violation box on the device. Third, the existing process
of site inspection is mostly paper-based and different companies have
different paper templates for their safety inspections. Therefore,
connecting the application prototype with the existing processes required additional research (i.e., a study of data schema) to integrate
safety and health data into the larger business management practices.
Additional evaluation was performed in the form of an expert workshop in the summer of 2013, with 10 participating subject experts
whose main job tasks are related to construction safety and health management in the industry. In the workshop, experts learned about the
application prototype through hands-on interactions and then lled
out a two-page survey to indicate their level of agreement with given
evaluation criteria. A 5-point Likert scale measured the level of agreement with 1 being Disagree and 5 being Agree. Results from
the survey are summarized in Table 1. Since the workshop, 10 private
entitiesfour specialty contractors, four general contractors, one
consulting rm, and the Associated General Contractors of America
(Western Washington District)have signed the evaluation license
agreement with the University of Washington and obtained free access
to the tool.
Although a quantitative eld evaluation was not conducted during
the reported research stage, the authors considered the collected qualitative feedback to help support the intended research objectives and
guide the research into its next phase. In the future, the research team

62

K.-Y. Lin et al. / Automation in Construction 48 (2014) 5363

Table 1
Survey results for the application prototype evaluation.
Criteria

Avg. agreement
level

1. The prototype can reduce the challenges you face when


performing safety inspections.
2. The prototype can improve your current approach to safety
inspection.
3. The prototype is compatible with your company's business
value and needs.
4. The design of the prototype aligns with your inspection
practices.
5. The prototype is easy to understand and use.
6. The benet of using the prototype can be easily
communicated to your peers.
7. The benet of using the prototype can be easily
communicated to your boss.

5.00
5.00
4.71
4.50
5.00
4.87
4.87

plans on recruiting a larger group of eld users and taking quantitative


measurements to document the user experience.

4. Discussions
The developed ICT tool has proved to be benecial in improving the
efciency and effectiveness of the safety inspection process for the
subject experts studied. In fact, the most noticeable benets identied
in the eld evaluation were the reduction of paperwork and faster communication between project participants. In addition to addressing the
limited availability of safety specialists and repetitive data preparation-related issues, implementing the ICT tool minimizes other drawbacks in the safety inspection process. For instance, the use of precongured lists of project and project participant information allows
safety specialists to store and quickly retrieve project data and, therefore, improve their access to project information. Then, functions such
as pre-congured safety checklists, reminders of related violations,
and integrated communication with project participants can improve
process standardization by making these steps an automatic part of
the process. Finally, the lack of standardized documentation can be
reduced by the use of pre-congured safety checklists.
Although the ICT tool functions were developed to support safety
specialists, some functions have both positive and negative impacts
and, as a result, not all the possible features were integrated in the application prototype. For example, the ability to edit the pre-congured
safety checklists is desirable for contractors who want to design safety
checklists specic to each project and/or special clients. However, a consistent schema for labeling violation data will be hard to create, making
it difcult to analyze violation data for trend and pattern recognition.
Thus, as an informed design decision, the application prototype does
not allow users to customize the pre-congured safety checklists. To
provide users with a certain level of customization without affecting
the quality and consistency of the collected data, the application prototype has a place in the pre-congured safety checklists where users can
add options that are not already in the list.
The collection of consistent and high-quality eld safety inspection
data can have a tremendous inuence on the development of advanced
data analysis techniques. For instance, a statistical report of the results
can reveal the most frequent violations in the eld and direct further
research to understand the relationships between weather conditions
(e.g., temperature) and violation features (e.g., trade type involved).
Eventually, it will be possible to develop ICT tools capable of autonomously identifying the priority inspection items and presenting the potential workplace safety issues based on the site weather, worker, and/
or trade information. Therefore, although additional uses of the collected data beyond recordkeeping applications still need to be explored,
this research established a structured safety inspection process and

included related functions in the preliminary process to facilitate advanced data analysis. This enhances our systematic understanding of
the inspection process and provides researchers as well as practitioners
the opportunity to explore and experience potential benets, echoing
Mitropoulos and Tatum's suggestions [25] on minimizing uncertainty
when introducing and developing innovative technologies for construction applications.
Regardless of all the possible benets, the integration of an innovative ICT tool with existing procedures and business practices remains
a critical challenge. In fact, one safety specialist who participated in
the eld evaluation decided to stop using the application prototype
tool because the integration was an additional workload for him. The
general resistance to change and technology in the construction industry remains an additional barrier. Although a user-centered design approach was used in developing the application prototype, another
safety specialist involved in the eld evaluation demonstrated a clear
reluctance to shift their inspection process from paper to ICT. Therefore,
best practices in the use of technological innovations for eld safety inspections should be established in order to showcase success stories and
provide incentives for adopting innovations within the industry. Such
best practices need to highlight how technological innovations benet
not only day-to-day practices and safety inspection processes but also
overall safety management.
Finally, it is important to point out that the implementation of a
user-centered design approach in the reported study was extremely
benecial. By working with industry practitioners closely, the authors
captured and understood the end-users' business needs, preferences,
practices, and objectives. Similarly, by participating in the research
study, the industry partners better understood the technological capabilities and were motivated to envision new ICT tool uses not previously
considered.
5. Conclusion
Safety inspection is imperative for reinforcing and promoting job
site safety. However, it is often undermined by the inefciency and
ineffectiveness of the process. The authors concluded that the safety inspection process is hampered by several drawbacks, including lack of
standardized processes and documentation, restricted access to information, repetitive data preparation, and limited availability of safety
specialists. Among other negative impacts, these drawbacks prevent
current practices from generating consistent and high-quality inspection records and therefore limit the potential of advanced safety and
health data analysis.
By analyzing the features of the safety inspection process, its drawbacks, and available ICT tools, the authors determined a series of key
ICT functions that can improve the safety inspection process. A twostep user-centered design approach was implemented to investigate
the requirements of these functions with the goal of embedding them
in an ICT tool for evaluation in the intended area of application. In the
rst step, the authors developed and evaluated a paper prototype. In
the second step, the research team developed and evaluated an ICT
application prototype running on an iPad and containing most of the
key functions. In particular, this research carried out user-oriented
testing and feedback solicitation in the anticipated environment of application. The use of such a user-centered design approach positively inuenced the research activities and enabled the authors to identify
safety specialists' needs and practices. The user-centered approach
also enabled the participating safety specialists to understand the ICT
tool's capabilities and to envision innovative ICT-based procedures for
the eld.
iSafe, the iPad application developed, can now be accessed for
free under the evaluation licensing agreement at the University of
Washington's Safety and Health Advancement through Research and
Education (SHARE) in Construction Management Lab's website at
http://cm.be.washington.edu/Research/SHARE.

K.-Y. Lin et al. / Automation in Construction 48 (2014) 5363

Acknowledgments
The authors would like to acknowledge the nancial support for this
research received from the University of Washington's Royal Research
Fund. The authors would also like to recognize their construction industry partners from the greater Seattle area.
References
[1] W. Zhou, J. Whyte, R. Sacks, Construction safety and digital design: a review, Autom.
Constr. 22 (2012) 102111.
[2] F.Y.Y. Ling, M. Liu, Y.C. Woo, Construction fatalities in Singapore, Int. J. Proj. Manag.
27 (2009) 717726.
[3] E.A.L. Teo, F.Y.Y. Ling, A.F.W. Chong, Framework for project managers to manage
construction safety, Int. J. Proj. Manag. 23 (2005) 329341.
[4] O. Abudayyeh, T.K. Fredericks, S.E. Butt, A. Shaar, An investigation of management's
commitment to construction safety, Int. J. Proj. Manag. 24 (2006) 167174.
[5] Construction Industry Institute, The Owners' Role in Construction SafetyResearch
Summary 1901, Austin, TX, 2003.
[6] T. Aksorn, B.H.W. Hadikusumo, Critical success factors inuencing safety program
performance in Thai construction projects, Saf. Sci. 46 (2008) 709727.
[7] V.K. Kaskutas, A.M. Dale, H.J. Lipscomb, B.A. Evanoff, Development of the St. Louis
Audit of Fall Risks at Residential Construction Sites, Int. J. Occup. Environ. Health
14 (2008) 243249.
[8] P.M. Goodrum, C.T. Haas, Partial factor productivity and equipment technology
change at activity level in U.S. construction industry, J. Constr. Eng. Manag. 128
(2002) 463472.
[9] Construction Industry Institute, Leveraging Technology to Improve Construction
Productivity, Construction Industry Institute, Austin, TX, 2008.
[10] J. Teizer, 3D range imaging camera sensing for active safety in construction, J. Inf.
Technol. Constr. 13 (2008) 103117.
[11] D.I. Kim, H. Cho, Safety and productivity analysis on alternative steel column-girder
joint for automated construction, Proceedings of International Conference on Control, Automation and Systems (ICCAS 2010), Seoul, Korea, 2010, pp. 23382341.
[12] J. Cinkelj, R. Kamnik, P. Cepon, M. Mihelj, M. Munih, Closed-loop control of hydraulic
telescopic handler, Autom. Constr. 19 (2010) 954963.
[13] S. Han, F. Pena-Mora, M. Golparvar-Fard, S. Roh, Application of a visualization technique for safety management, 2009 ASCE International Workshop on Computing in
Civil Engineering, ASCE, Austin, TX, 2009, pp. 543551.
[14] K. Sulankivi, T. Makela, M. Kiviniemi, BIM-based site layout and safety planning, First
International Conference on CIB, Technical Research Center of Finland, Espoo,
Finland, 2009, pp. 125140.
[15] S. Chi, C.H. Caldas, J. Gong, A crash avoidance framework for heavy equipment
control systems using 3D imaging sensors, Special Issue Sensors in Construction
and Infrastructure Management, 13, 2008, pp. 118133.
[16] A. Walia, J. Teizer, Analysis of spatial data structures for proximity detection,
Tsinghua Sci. Technol. 13 (2008) 102107.
[17] W. Wu, H. Yang, D.A.S. Chew, S. Yang, A.G.F. Gibb, Q. Li, Towards an autonomous
real-time tracking system of near-miss accidents on construction sites, Autom.
Constr. 19 (2010) 134141.

63

[18] H. Yang, D.A.S. Chew, W. Wu, Z. Zhou, Q. Li, Design and implementation of an identication system in construction site safety for proactive accident prevention, Accid.
Anal. Prev. 48 (2012) 193203.
[19] J. Andresen, A. Baldwin, M. Betts, C. Carter, A. Hamilton, E. Stokes, et al., A framework
for measuring IT innovation benets, J. Inf. Technol. Constr. 5 (2000) 5772.
[20] P. Gonzalez de Santos, J. Estremera, E. Garcia, M. Armada, Power assist devices for
installing plaster panels in construction, Autom. Constr. 17 (2008) 459466.
[21] E.A. Koningsveld, H.F. van der Molen, History and future of ergonomics in building
and construction, Ergonomics 40 (1997) 10251034.
[22] M.R. Endsley, D.G. Jones, Designing for situation awareness: an approach to humancentered design, 2nd ed. CRC Press, Boca Raton, FL, 2012.
[23] E.M. Rogers, Diffusion of innovations, 5th ed. Free Press, New York, NY, 2003.
[24] A. Dillon, M. Morris, User acceptance of new information technology - theories and
models, in: M. Williams (Ed.), Annual Review of Information Science and Technology, vol. 31, Information Today, Medford, NJ, 1996, pp. 332.
[25] P. Mitropoulos, C. Tatum, Forces driving adoption of new information technologies,
J. Constr. Eng. Manag. 126 (2000) 340348.
[26] C.-H. Chen, K. Sato, K.-P. Lee, Editorial: human-centered product design and development, Adv. Eng. Inform. 23 (2009) 140-141.
[27] C.-C. Lu, S.-C. Kang, S.-H. Hsieh, R.-S. Shiu, Improvement of a computer-based
surveyor-training tool using a user-centered approach, Adv. Eng. Inform. 23
(2009) 8192.
[28] N. Dawood, D. Scott, E. Sriprasert, Z. Mallasi, The virtual construction site (VIRCON)
tools: an industrial evaluation, ITcon 10 (2005) 4354.
[29] H.N. Ahuja, S.P. Dozzi, S.M. AbouRizk, Project Management: Techniques in Planning
and Controlling Construction Projects, Wiley, Hoboken, NJ, 1994.
[30] International Organization for Standardization, Ergonomics of human-system
interactionPart 210: human-centered design for interactive systems, 2010.
[31] R. Choudhry, D. Fang, S. Ahmed, Safety management in construction: best practices
in Hong Kong, J. Prof. Issues Eng. Educ. Pract. 134 (2008) 2032.
[32] N. Hamidi, M. Omidvari, M. Meftahi, The effect of integrated management system on
safety and productivity indices: case study; Iranian cement industries, Saf. Sci. 50
(2012) 11801189.
[33] B. Fernndez-Muiz, J.M. Montes-Pen, C.J. Vzquez-Ords, Safety management
system: development and validation of a multidimensional scale, J. Loss Prev. Process Ind. 20 (2007) 5268.
[34] T. Toole, Construction site safety roles, J. Constr. Eng. Manag. 128 (2002) 203210.
[35] E.E. Koehn, N.K. Datta, Quality, environmental, and health and safety management
systems for construction engineering, J. Constr. Eng. Manag. 129 (2003) 562569.
[36] N.A. Kartam, I. Flood, P. Koushki, Construction safety in Kuwait: issues, procedures,
problems, and recommendations, Saf. Sci. 36 (2000) 163184.
[37] B. Hadikusumo, S. Rowlinson, Capturing safety knowledge using design-for-safetyprocess tool, J. Constr. Eng. Manag. 130 (2004) 281289.
[38] K.Y. Lin, Y.W. Kim, C. Dossick, Repositioning the faculty internship experience in
construction management, Proceedings of the Construction Research Congress
2010, Banff, Alberta, Canada, 2010.
[39] G. Johnson, An Evaluation of the Application for Tablet Computing in Construction,
University of Washington, Seattle (WA), 2012.