International Construction Specialty Conference of the Canadian Society for Civil Engineering (ICSC) (5th : 2015)

A conceptual accident causation model based on the incident root causes Pereira, Estacio; Taghaddos, Hosein; Hermann, Rick; Han, SangUk; Abourizk, Simaan Jun 30, 2015

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
52660-Pereira_E_et_al_ICSC15_115_Accident_Root_Causes.pdf [ 389.99kB ]
52660-Pereira_E_et_al_ICSC15_115_Accident_Root_Causes_slides.pdf [ 1.14MB ]
Metadata
JSON: 52660-1.0076393.json
JSON-LD: 52660-1.0076393-ld.json
RDF/XML (Pretty): 52660-1.0076393-rdf.xml
RDF/JSON: 52660-1.0076393-rdf.json
Turtle: 52660-1.0076393-turtle.txt
N-Triples: 52660-1.0076393-rdf-ntriples.txt
Original Record: 52660-1.0076393-source.json
Full Text
52660-1.0076393-fulltext.txt
Citation
52660-1.0076393.ris

Full Text

5th International/11th Construction Specialty Conference 5e International/11e Conférence spécialisée sur la construction    Vancouver, British Columbia June 8 to June 10, 2015 / 8 juin au 10 juin 2015   A CONCEPTUAL ACCIDENT CAUSATION MODEL BASED ON THE INCIDENT ROOT CAUSES Estacio Pereira1, Hosein Taghaddos2, Rick Hermann2, SangUk Han1, Simaan Abourizk1, 3  1 University of Alberta, Canada 2 PCL Industrial Construction, Canada 3 abourizk@ualberta.ca Abstract: The measurement and control of incident root causes allows for proactive actions to mitigate risk in advance. In practice, however, it is difficult to identify and collect data that represent the root causes due to the complexity of incident occurrence processes. Despite previous studies on incident causation modelling, the identification of root causes in practice still relies on the investigator’s subjective opinion. This research presents a conceptual model that explains the causal relationships between the root causes and the site unsafe level, and eventually assesses incident investigation processes. A case study was conducted to evaluate the 13 root causes in a company’s investigation practice. The causal relationship between the root causes was observed based on the company safety database, interviews, and literature review. Then, the detailed model, which explains the incident occurrence process, was explored. Additionally, a hypothetical simulation model that allows for evaluation of the influence of each root cause on the safety level was built and tested to discuss the potential use of the conceptual model. Based on the company database, this paper also suggests and discusses the types of data to measure the root causes in practice. The model demonstrates that not only do safety personal and safety strategies affect the site unsafe level, but other factors also do, such as procurement, engineering, human resources, etc. As a result, the proposed model can be used to help identify the root cause in incident investigation practice and to develop strategies to improve safety performance. 1 INTRODUCTION Incidents in the construction industry can influence project cost, schedule and quality. According to the Association of Workers’ Compensation Boards of Canada (2012), the incident rate in the construction industry is 30% higher than in any other industry. Moreover, the fatality rate of the construction industry is approximately three times higher than the industry average. Incidents can affect the worker’s family, the community, and will also decrease the amount of worker resources available to the industry.  Incidents can generate accidents. According to Bird and Germain (1996), an accident is an event that results in unintended harm or damage, and when it is related to the worker, can result in injury. Any accident can be avoided; however, preventing accidents is difficult, mainly due to the difficulty of understanding accident causes, since several factors, such as worker and management commitment, schedule, and training, can affect it. 115-1 Construction companies usually perform an incident investigation to identify the root causes leading to an incident. Based on this investigation, the companies take actions (e.g. safety training, audits) that allow proactive management of safety performance by mitigating the risk in advance. Although several studies have developed accident causation models, the identification of the root causes in practice relies on investigator experience.  Besides identification of the root causes, the measure and control of the incident root causes can also contribute to improvement of the risk mitigation process. However, construction companies have difficultly identifying and collecting relevant data that represent the root causes due to the complexity of the incident occurrence process. Moreover, relevant data could be used to produce simulation models to better predict or estimate the site unsafe level. The difficulties in identifying, measuring and controlling incident root causes could be due to the difficulty of understanding the causal relationship between them. Nevertheless, the relationship between the root causes should be determined, since projects usually have a limited safety budget, and better results can be achieved if the company can identify the best safety strategy to allocate the resources available (Wirth and Sigurdsson, 2008). The objective of this research was to develop a conceptual accident causation model in order to explain the causal diagram between the root causes and the site unsafe level. 2 BACKGROUND Accident causation models aim to “understand the factors and processes involved in accidents in order to develop strategies for accident prevention” (Arboleda and Abraham, 2004; Mitropoulos et al., 2005). According to Hovden et al. (2010), the main reasons for discussing the accident causation models are to: (1) create a common understanding of the accident phenomena; (2) help structure and communicate risk problems; (3) guide investigation on data collection and accident analyses; and (4) analyze the relationship between the factors.  Researchers have developed methodologies to identify incident root causes. Wagenaar and Schrier (1997) developed the TRIPOD model. This model classifies the causes for an incident into 11 General Failures Groups (e.g. design and training). Abdelhamid & Everett (2000) developed the Accident Root Causes Tracing Model (ARCTM). This model uses a decision tree to identify the main root cause of an incident. Suraji et al. (2001) developed a model that classifies the factors that cause an incident into distal and proximal factors. Leveson (2004) developed the Systems-Theoretic Accident Model and Process (STAMP). In this model, the accident occurs when external disturbances, component failures or dysfunctional interactions are not adequately controlled. However, these models are only able to pinpoint the main factors that cause the incident, not support the dynamic relationship between them. As the previous models are not able to deal with the dynamic relationship between the factors, researchers have developed system dynamic models to understand how factors cause an incident. Cooke & Rohleder (2006) focus on how worker risky behavior and the learning process can cause an incident. Han et al. (2014) verified how the production pressure is related to incidents. Jiang et al. (2015) and Shin et al. (2014) developed models to understand the influence of the worker’s unsafe behavior on the incidents. It is possible to verify that these models are not able to deal with different root causes specified in practice by construction companies. Moreover, these models are generally conceptual and it is difficult to apply them to company safety routines. The models and techniques presented have difficulties measuring the root causes that influence incidents. In practice, the incident investigation is usually only able to classify the occurrence of a pre-established root cause as Yes/No. The incident investigations utilized by construction companies usually collect information to describe the incident, but do not collect data to measure the influence of each root cause on the incident. Therefore, the companies have difficultly finding preventive actions to avoid further incidents. 115-2 3 METHODOLOGY A case study was conducted to evaluate the root cause in a company’s incident investigation practice. The incident root causes were identified. Although the root causes were established based on Bird and Germain (1996), there was no definition about how to classify each root cause during the incident investigation procedure. Therefore, the root causes were defined based on literature review and the company incident investigation. After identifying and describing the root cause used by the construction company, the causal diagrams were developed. These diagrams were built based on the company’s incident investigation, safety database, interviews, HSE Manual and further literature review. The last step was to define empirical equations and build a hypothetical simulation model to understand the model behavior and evaluate the influence of each root cause on the site unsafe level. Moreover, data types were suggested to measure each root cause based on the safety database and the incident investigation. 4 IDENTIFY AND DEFINE THE ROOT CAUSES According to the company safety policies, for every incident that occurs on the construction site, an incident investigation should be conducted. The company established 13 root causes of incidents, and the investigator should choose at least one cause based on his/her experience. A short description for each root cause is shown in Table 1. Besides the incident root causes, the incident investigation defined by the construction company also collects information about the date and time of the incident, weather and lighting conditions, worker information, worker schedule, injury details, activity type, tools and equipment utilized in the incident, substandard act, substandard conditions, witness statement, etc. 5 CONCEPTUAL MODEL The conceptual model established two main categories as the cause of the site unsafe level: worker behavior and site conditions. These categories were defined based on the incident investigation and literature review (Lingard and Rowlinson 2005). The site unsafe level can cause an incident. An incident, in this research, is every occurrence likely to lead to grave consequences. Accidents are every occurrence that decreases worker availability in the project. Therefore, incidents and accidents are positively correlated.  Three main loops were identified in the conceptual model. Loop R1 is related to the site condition. The company and some researchers (Mitropoulos et al. (2005) and Han et al. (2014)) stated that the accident affects the schedule pressure causing congestion, and increasing the site unsafe condition. Moreover, factors such as temperature, project type, activity type (Lee et al. 2012), and site layout (Anumba & Bishop, 1997), can also affect the site unsafe condition.  The other two loops (B1 and B2) are related to the worker behavior. The schedule pressure can affect the worker intention to work safe (Mitropoulos et al. 2005), and consequently, the worker safe behavior. Moreover, incident investigations can increase worker knowledge and also the perception of risk (Construction Industry Institute, 2002), improving the worker safe behavior (Han et al. 2014). Figure 1 shows a conceptual model of the influence of the worker safe behavior and site conditions on the site unsafe level.  115-3 Table 1: Incident root causes description N Root Cause Description 1 Hazard Identification and Control Worker characteristics influence the identification and control of hazards. 2 Human Resource / Professional Development (HR/PD) The hiring process was not able to verify the workers’ skills and knowledge. 3 Standard Operating Procedures Practices  The safety procedures to perform a task in a safe manner were not defined. 4 Leadership and Administration Attitudes from the management do not demonstrate commitment to safety. 5 Inspection and Audits The inspection and audits of equipment, processes, and workers were not defined/realized. In this research, the worker perspective of the inspection and audits will be considered. 6 Orientation and Training The orientation/training was not able to transfer knowledge to the worker.  7 Site Specific Safety Plan There is no recommendation about the safety procedures that should be followed in the construction site.  8 Communication Systems The communication system was not able to inform the worker about the risks on the site.  9 Security/Emergency Response There are no procedures to follow if an incident occurs.  10 Engineering Verify problems related with the project design. 11 Procurement Verify errors in the procurement process, such as lack of material specification and delay in delivery. 12 Sub / Trade - Contractor Management Verify problems related with the sub/trade training and commitment to safety. 13 Environment Verify the climate conditions that can influence an incident.   Figure 1: Basic conceptual model The root causes defined by the construction company were categorized between the worker safe behavior and the site conditions categories. Each loop is explained in further detail below. Site Condition (R1): Figure 3 shows the influence of the incident root causes on the site unsafe conditions. The site unsafe level increases the quantity of incidents and accidents. According to Han et al. (2014) and Mitropoulos et al. (2005), an accident can cause delays, increasing the schedule pressure. To compensate for the delay, the company can hire new workers. However, these workers increase the site congestion. The congestion increases the site unsafe condition because it increases workers’ exposure to struck-by or struck-against incidents (Fortunato et al. 2012). According to the company safety investigation, the site safety conditions can also be affected by the root causes Environment (e.g. IncidentSite UnsafeConditionsWorker PerceptionAccident+WorkerIntentionSiteConditionSite UnsafeLevel++Worker SafeBehavior+ +-WorkerKnowledgeWorkerIntention R1B1B2Schedule Pressure+ -Congestion+++115-4 temperature, lighting, and wind), Standard Operating Procedures, Site Specific Safety Plan and Security Emergence Response.  Figure 3: Influence of the root causes on the site unsafe conditions  The site conditions are also affected by the root causes Engineering and Procurement. Both of these root causes can also contribute to the schedule pressure. Procurement can lead to material delay and poor design can increase rework.  Hazard Identification (B1): Figure 4 shows the influence of worker knowledge on the site unsafe level. If the investigation is able to identify the root causes and the results are shared with the workers, they will increase their knowledge. Workers' previous experience can also affect worker knowledge. According to the company safety database, worker experience and incidents are negatively correlated. Therefore, duirng the hiring process, it is important to identify workers with more experience. Furthermore, according to the company safety database, the quantity of pre-task meetings is negatively correlated with the quantity of incidents because it increases worker hazard perception (Construction Industry Institute, 2002). In this model, the root cause Safety Comunication represents the pre-task meeting.   Figure 4: Worker knowledge influence on the site unsafe level The improvement of workers’ knowledge facilitates worker perception of hazards (Jiang et al., 2015). However, worker perception can be affected by the root cause Hazard Identification and Control. This Site UnsafeLevelSite UnsafeCondition -IncidentAccident++SchedulePressure+Crew Size +Congestion++ErrorRework+++-ProcurementEngineeringEnvironmentSecurity EmergenceResponseStandard OperatingProceduresSite SpecificSafety PlanR1SiteConditionDesign+-Delay indelivery-+Materials notattend the safetyspecifications+Design safetymeasures+-Activities safety risk-+Defective tools andequipments-+Inadequatewarning system--Exposureto risk+Site UnsafeLevelWorker SafeBehavior-Incident+SafetyInvestigationHazardIdentification andControl+SafetyCommunicationWorker Perception+WorkerExperienceHR/PD+Orientation andtrainingB1WorkerKnowledgeWorkerknowledge+++Worker phisicalcondition--115-5 root cause represents worker physical conditions such as work shift, worker’s age, health condition and other personal characteristics that can prevent the worker from recognizing a hazard. Worker Intention (B2): Figure 5 shows the influence of the worker intention on the site unsafe condition. Because of the particularity of the worker intention, it was divided in two sub-loops: Fatigue (B2.1) and Safety Climate (B2.2).   Figure 5: Worker intention influence on the site unsafe level Fatigue (B2.1): The schedule performance can make the company increase the workers’ shifts. According to Alvanchi et al. (2012), prolonged working hours can produce fatigue due to decrease in the muscular strength and mental stress. Fatigue can make the worker take shortcuts, not follow the safety recommendations, and consequently, decrease the worker’s intention to work safely (Jiang et al., 2015). Moreover, mental stress can cause distraction and decrease the worker’s capacity for hazard recognition (Hinze, 1997). Safety Climate (B2.2): In this sub-loop, accidents increase the safety pressure and consequently increase management’s commitment to safety. However, Mitropoulos et al. (2005) stated that the schedule pressure may prevent management from providing and maintaining required safety measures, decreasing efforts to control the worker behavior. Moreover, management commitment is affected by the Leadership and Administration. According to the company HSE manual, the Leadership and Administration considers factors such as lack of discpline, lack of enforcement, lack of safety recources and lack of safety planning. The Management Commitment consequently affects the safety climate (Chinda & Mohamed, 2008). Although not specified as a root cause, safety climate is also affected by the Foreman Behavior (Choudhry and Fang, 2008). The root cause Sub-Contractor Management also affects Safety Climate. The worker perception of safety (Han et al. 2014) is influenced by the safety climate and inspection and audits. One example of inspection is the Behavior-Based Observation (BBO) Card. The BBO improves worker safe behavior because the worker feels that he/she is being watched by the safety personnel (Vaughen et al., 2010).  Figure 6 shows the complete conceptual model. Site UnsafeLevel WorkerBehavior-IncidentAccident++Safety Pressure+SchedulePressure+Work Overload+Fatigue+Worker Intention-+Safety overschedule--+ ManagementCommitment+Leadership andAdministration+Safety Climate++Inspection andAuditsB2.1FatigueB2.2Safety ClimateWorkerobservation++Sub-ContractorManagement3rd part commitmentwith safety++Foreman BehaviorForemanCommitment++115-6  Figure 6: Site unsafe level conceptual model  6 MODEL EXPERIMENTS AND DISCUSSION A hypothetical simulation model was built and four scenarios were tested to evaluate the influence of each root cause on the site unsafe level. In the first three scenarios, three different root causes were tested individually: 1) Environment, 2) Orientation and Training, and 3) Inspection and Audits. To better understand the influence of the root cause in each scenario, its value was set to 0 (worst condition), 0.5 and 1 (best condition). The other root causes had their values set at 0.5. The last graph compares the site unsafe level when all root causes are equal to 0.1 and 1. The time defined to visualize the root causes’ influence on the site unsafe level is 90 days. Figure 7 shows the site unsafe level obtained in each scenario.   Figure 7: Effect of different root causes on the site unsafe level Site Unsafe LevelSite UnsafeConditionWorker Safebehavior-+IncidentAccident++SafetyInvestigation+SafetyCommunicationWorkerPerception+WorkerExperienceHR/PD+Safety PressureSchedule PressureCrew Size+Congestion++Work Overload+Fatigue+WorkerIntention-ErrorRework++ Safety overSchedule++ManagementCommitment+Leadership andAdministration+Safety Climate+ForemanBehaviorInspection andAuditsProcurementEngineeringEnvironmentSub-ContractorManagementStandard OperatingProceduresSite SpecificSafety PlanSecurity EmergenceResponseOrientation andtraining+-B2.1FatigueB2.2SafetyClimateB1WorkerKnowledgeR1Site ConditionWorkerKnowledgeWorker phisicalconditionHazard Identificationand ControlForemannCommitment3rd par commitmentwith safetyDesignDelay indeliverMaterials notattend the safetyspecificationsDesign safetymeasuresExposureto riskInadequatewarning systemDefective Tolls andEquipment Activities safetyrisk0510152025300 20 40 60 80 100Site Unsafe LevelDaysEnvironment effects on Site Unsafe Level0 0.5 10510152025300 20 40 60 80 100Site Unsafe LevelDaysTraining effect on Site Unsafe Level1 0.5 00510152025300 20 40 60 80 100Site Unsafe LevelDaysInspection and Audits effect on Site Unsafe Level  1 0.5 0-30-20-10010203040500 20 40 60 80 100Site Unsafe LevelDaysInfluence of Root Causes on Site UnSafe LevelAll Root causes = 1 All root causes = 0.1115-7 It is possible to verify that the root causes defined by the construction company can affect the site unsafe level. The root causes defined in the model are inversely proportional with the site unsafe level. Moreover, it is possible to verify that after day 40, the site unsafe level is almost constant. This behavior is due to the schedule pressure, since the work hour overload and the crew size can compensate for the delay caused by incidents and rework. The similarity between the results of the three first graphs demonstrates that different root causes should be improved concurrently to decrease the site unsafe level (graph 4).  To improve the root causes, it is necessary to measure them. Furthermore, Table 2 suggests types of data to measure each root cause. Table 2: Incident root causes description N Root cause Suggested types of data to measure the root causes 1 Hazard Identification and Control Work shift; worker’s experience on the project; worker’s age 2 Human Resource / Professional Development (HR/PD) Average of workers’ experience on project; worker’s previous ability 3 Standard Operating Procedures Practices  Activity risk level 4 Leadership and Administration Management site inspection; participation in safety meetings  5 Inspection and Audits Quantity of BBO filled per month; quantity of workers per foreman 6 Orientation and Training Worker training hours; evaluate of workers’ learning of the course content 7 Site Specific Safety Plan Equipment and tool maintenance per month; safety program level of maturity 8 Communication Systems Quantity of pre-job inspections completed per month  9 Security/Emergency Response Escape route facilities (clear, indicated and shorter path)  10 Engineering Engineering quality by discipline 11 Procurement Procurement quality by discipline  12 Sub / Trade - Contractor Management Evaluate observation of safety practices in the project 13 Environment Temperature; wind speed; noise  - Foreman Foreman skill level; foreman age; safety supervisor experience Besides the data types presented in Table 2, the company can also collect information about other factors used in the model, such as congestion (worker ramp up and ramp down), schedule pressure (delays), rework (project quality) and safety pressure (total recordable incident rate – TRIR). The incident investigation can be improved to collect the data type suggested. Moreover, as some of the company’s incident investigations were not fully completed, the model could reinforce the importance of collecting all data requested by the investigation. In this case, the investigation will not be utilized just to describe an incident, but also to measure and control the incident root causes. The definition of the root cause can also help to better identify the incident causes, especially for those investigators who have to conduct the investigation. The conceptual model was developed to identify the relationships between the root causes, but it is not recommended to be used to predict the site safety level. For this purpose, other simulation techniques, such as hybrid models combining discrete event simulation with system dynamics, or agent-based models, can achieve better results. According to Sawhney et al. (2003), an agent-based model “can be used to mimic the construction environment in which the worker [is] performing [his/her] work, along with 115-8 heterogeneous set of agents representing these workers to study various aspects of safety.” Furthermore, root causes such as Environment, Procurement and Engineering can change values during the simulation and improvements are necessary to better predict the site unsafe level. However, the relationship between the root causes identified in this research can be used on other simulation models to improve the results. Based on the model and the data type suggested to measure the incident root causes, construction companies can adopt strategies to improve the site safety level, such as improve the selection of engineering, suppliers, and sub-contractors in aspects related to safety; implement inspection procedures such as the BBO card and measure the supervisor’s commitment to safety. 7 CONCLUSION The accident conceptual model developed in this research was able to demonstrate the relationship between the incident root causes defined by the construction company and the site unsafe level. It is also possible to conclude that not only are safety procedures, safety personnel, and field workers responsible to improve safety performance, but other company departments are as well. In this way, it was possible to conclude that, root causes such as project design, procurement, and HR/PD can affect the site unsafe level. Moreover, all departments that can cause project delays or influence the quality can influence the site unsafe level.  The company can improve the incident investigation procedures based on the conceptual model. Factors to measure each root cause were suggested and it is recommended to collect them during the incident investigation. Although this conceptual model cannot be used to predict the safety level, it is believed that the relationship defined in this research can be used in the future to develop a simulation model to measure and forecast how different safety strategies can affect the incident root cause and consequently the site unsafe level.  Ongoing research on the company safety database has been developed to validate the relationship between the root causes. Further steps are to define to which degree the different root causes affect the site unsafe level and validate the data type to measure the root causes with the safety managers. Acknowledgements This research was made possible in part by Coordination for the Improvement of Higher Education Personnel (CAPES) - Ministry of Education of Brazil (Proc. no 0393-12-6). It was also supported by the NSERC Industrial Research Chair in Construction Engineering and Management, IRCPJ 195558-10. The authors also want to acknowledge PCL Industrial Construction for their assistance. References Abdelhamid, T. S. and Everett, J. 2000. Identifying Root Causes of Construction Accidents. Journal of Construction Engineering and Management, 126(1): 52–60. Alvanchi, A., Lee, S. and Abourizk, S. 2012. Dynamics of Working Hours in Construction. Journal of Construction Engineering and Management, 138(1): 66–77.  Anumba, C. and Bishop, G. 1997. Safety-integrated Site Layout and Organization. Annual Conference of Canadian Society for Civil Engineers, CSCE, Montreal, QC, Canada, 147–156. Arboleda, C. A. and Abraham, D. M. 2004. Fatalities in Trenching Operations — Analysis using Models of Accident Causation. Journal of Construction Engineering and Management, 130(4): 273–280. Association of Workers’ Compensation Boards of Canada (AWCBC). 2012. Summary Tables of Accepted Time-loss Injuries/Diseases and Fatalities by Jurisdiction. Table 282-0008: Labour force survey estimates. Bird, F. E. and Germain, G. L. 1996. Practical Loss Control Leadership. International Loss Control Leadership. 115-9 Chinda, T. and Mohamed, S. 2008. Structural Equation Model of Construction Safety Culture. Engineering, Construction and Architectural Management, 15(2): 114–131. Choudhry, R. M. and Fang, D. 2008. Why Operatives Engage in Unsafe Work Behavior: Investigating Factors on Construction Sites. Safety Science, 46(4): 566–584.  Construction Industry Institute. 2002. Safety Plus: Making Zero Accidents Reality. Austin TX.: CII Project Rep. 160. Cooke, D. L. and Rohleder, T. R. 2006. Learning from Incidents: From Normal Accidents to High Reliability. System Dynamics Review, 22(3): 213–239.  Fortunato, B. R., Hallowell, M. R., Behm, M. and Dewlaney, K. 2012. Identification of Safety Risks for High-performance Sustainable Construction Projects. Journal of Construction Engineering and Management, 138: 499–508.  Han, S., Saba, F., Lee, S., Mohamed, Y. and Pena-Mora, F. 2014. Toward an Understanding of the Impact of Production Pressure on Safety Performance in Construction Operations. Accident Analysis and Prevention, 68: 106–116. Hinze, J. W. 1997. Construction Safety. Prentice-hall. Hovden, J., Albrechtsen, E. and Herrera, I. A. 2010. Is there a Need for New Theories, Models and Approaches to Occupational Accident Prevention? Safety Science, 48(8): 950–956.  Jiang, Z., Fang, D. and Zhang, M. 2015. Understanding the Causation of Construction Workers ’ Unsafe Behaviors Based on System Dynamics Modeling. Journal of Construction Engineering and Management, in press.  Lee, H., Kim, H., Park, M., Teo, E. A. L. and Lee, K. 2012. Construction Risk Assessment using Site Influence Factors. Journal of Computing in Civil Engineering, 26: 319–330. Leveson, N. 2004. A New Accident Model for Engineering Safer Systems. Safety Science, 42(4): 237–270.  Lingard, H. and Rowlinson, S. 2005. Occupational Health and Safety in Construction Project Management. Taylor & Francis. Mitropoulos, P., Abdelhamid, T. S. and Howell, G. A. 2005. Systems Model of Construction Accident Causation. Journal of Construction Engineering and Management, 131(7): 816–825.  Sawhney, A., Bashford, H., Walsh, K. and Mulky, A. R. 2003. Agent-based Modeling and Simulation in Construction. I 2003 Winter Simulation Conference. IEEE, Hoboken, NJ, USA. 1541–1547. Shin, M., Lee, H., Park, M., Moon, M. and Han, S. 2014. A System Dynamics Approach for Modeling Construction Workers ’ Safety Attitudes and Behaviors. Accident Analysis and Prevention, 68: 95–105. Suraji, A., Duff, A. R. and Peckitt, S. J. 2001. Development of Causal Model of Construction Accident Causation. Journal of Construction Engineering and Management, 127: 337–344. Vaughen, B. K., Lock, K. J. and Floyd, T. 2010. Improving Operating Discipline through the Successful Implementation of a Mandated Behavior-based Safety Program. Process Safety Progress, 29(3): 192–200. Wagenaar, W. A. and Schrier, J. Van Der. 1997. Accident Analyses - The Goal, and How to Get There. Safety Science, 26(l): 25–33. Wirth, O. and Sigurdsson, S. O. 2008. When Workplace Safety Depends on Behavior Change: Topics for Behavioral Safety Research. Journal of Safety Research, 39(6): 589–98.   115-10  5th International/11th Construction Specialty Conference 5e International/11e Conférence spécialisée sur la construction    Vancouver, British Columbia June 8 to June 10, 2015 / 8 juin au 10 juin 2015   A CONCEPTUAL ACCIDENT CAUSATION MODEL BASED ON THE INCIDENT ROOT CAUSES Estacio Pereira1, Hosein Taghaddos2, Rick Hermann2, SangUk Han1, Simaan Abourizk1, 3  1 University of Alberta, Canada 2 PCL Industrial Construction, Canada 3 abourizk@ualberta.ca Abstract: The measurement and control of incident root causes allows for proactive actions to mitigate risk in advance. In practice, however, it is difficult to identify and collect data that represent the root causes due to the complexity of incident occurrence processes. Despite previous studies on incident causation modelling, the identification of root causes in practice still relies on the investigator’s subjective opinion. This research presents a conceptual model that explains the causal relationships between the root causes and the site unsafe level, and eventually assesses incident investigation processes. A case study was conducted to evaluate the 13 root causes in a company’s investigation practice. The causal relationship between the root causes was observed based on the company safety database, interviews, and literature review. Then, the detailed model, which explains the incident occurrence process, was explored. Additionally, a hypothetical simulation model that allows for evaluation of the influence of each root cause on the safety level was built and tested to discuss the potential use of the conceptual model. Based on the company database, this paper also suggests and discusses the types of data to measure the root causes in practice. The model demonstrates that not only do safety personal and safety strategies affect the site unsafe level, but other factors also do, such as procurement, engineering, human resources, etc. As a result, the proposed model can be used to help identify the root cause in incident investigation practice and to develop strategies to improve safety performance. 1 INTRODUCTION Incidents in the construction industry can influence project cost, schedule and quality. According to the Association of Workers’ Compensation Boards of Canada (2012), the incident rate in the construction industry is 30% higher than in any other industry. Moreover, the fatality rate of the construction industry is approximately three times higher than the industry average. Incidents can affect the worker’s family, the community, and will also decrease the amount of worker resources available to the industry.  Incidents can generate accidents. According to Bird and Germain (1996), an accident is an event that results in unintended harm or damage, and when it is related to the worker, can result in injury. Any accident can be avoided; however, preventing accidents is difficult, mainly due to the difficulty of understanding accident causes, since several factors, such as worker and management commitment, schedule, and training, can affect it. 115-1 Construction companies usually perform an incident investigation to identify the root causes leading to an incident. Based on this investigation, the companies take actions (e.g. safety training, audits) that allow proactive management of safety performance by mitigating the risk in advance. Although several studies have developed accident causation models, the identification of the root causes in practice relies on investigator experience.  Besides identification of the root causes, the measure and control of the incident root causes can also contribute to improvement of the risk mitigation process. However, construction companies have difficultly identifying and collecting relevant data that represent the root causes due to the complexity of the incident occurrence process. Moreover, relevant data could be used to produce simulation models to better predict or estimate the site unsafe level. The difficulties in identifying, measuring and controlling incident root causes could be due to the difficulty of understanding the causal relationship between them. Nevertheless, the relationship between the root causes should be determined, since projects usually have a limited safety budget, and better results can be achieved if the company can identify the best safety strategy to allocate the resources available (Wirth and Sigurdsson, 2008). The objective of this research was to develop a conceptual accident causation model in order to explain the causal diagram between the root causes and the site unsafe level. 2 BACKGROUND Accident causation models aim to “understand the factors and processes involved in accidents in order to develop strategies for accident prevention” (Arboleda and Abraham, 2004; Mitropoulos et al., 2005). According to Hovden et al. (2010), the main reasons for discussing the accident causation models are to: (1) create a common understanding of the accident phenomena; (2) help structure and communicate risk problems; (3) guide investigation on data collection and accident analyses; and (4) analyze the relationship between the factors.  Researchers have developed methodologies to identify incident root causes. Wagenaar and Schrier (1997) developed the TRIPOD model. This model classifies the causes for an incident into 11 General Failures Groups (e.g. design and training). Abdelhamid & Everett (2000) developed the Accident Root Causes Tracing Model (ARCTM). This model uses a decision tree to identify the main root cause of an incident. Suraji et al. (2001) developed a model that classifies the factors that cause an incident into distal and proximal factors. Leveson (2004) developed the Systems-Theoretic Accident Model and Process (STAMP). In this model, the accident occurs when external disturbances, component failures or dysfunctional interactions are not adequately controlled. However, these models are only able to pinpoint the main factors that cause the incident, not support the dynamic relationship between them. As the previous models are not able to deal with the dynamic relationship between the factors, researchers have developed system dynamic models to understand how factors cause an incident. Cooke & Rohleder (2006) focus on how worker risky behavior and the learning process can cause an incident. Han et al. (2014) verified how the production pressure is related to incidents. Jiang et al. (2015) and Shin et al. (2014) developed models to understand the influence of the worker’s unsafe behavior on the incidents. It is possible to verify that these models are not able to deal with different root causes specified in practice by construction companies. Moreover, these models are generally conceptual and it is difficult to apply them to company safety routines. The models and techniques presented have difficulties measuring the root causes that influence incidents. In practice, the incident investigation is usually only able to classify the occurrence of a pre-established root cause as Yes/No. The incident investigations utilized by construction companies usually collect information to describe the incident, but do not collect data to measure the influence of each root cause on the incident. Therefore, the companies have difficultly finding preventive actions to avoid further incidents. 115-2 3 METHODOLOGY A case study was conducted to evaluate the root cause in a company’s incident investigation practice. The incident root causes were identified. Although the root causes were established based on Bird and Germain (1996), there was no definition about how to classify each root cause during the incident investigation procedure. Therefore, the root causes were defined based on literature review and the company incident investigation. After identifying and describing the root cause used by the construction company, the causal diagrams were developed. These diagrams were built based on the company’s incident investigation, safety database, interviews, HSE Manual and further literature review. The last step was to define empirical equations and build a hypothetical simulation model to understand the model behavior and evaluate the influence of each root cause on the site unsafe level. Moreover, data types were suggested to measure each root cause based on the safety database and the incident investigation. 4 IDENTIFY AND DEFINE THE ROOT CAUSES According to the company safety policies, for every incident that occurs on the construction site, an incident investigation should be conducted. The company established 13 root causes of incidents, and the investigator should choose at least one cause based on his/her experience. A short description for each root cause is shown in Table 1. Besides the incident root causes, the incident investigation defined by the construction company also collects information about the date and time of the incident, weather and lighting conditions, worker information, worker schedule, injury details, activity type, tools and equipment utilized in the incident, substandard act, substandard conditions, witness statement, etc. 5 CONCEPTUAL MODEL The conceptual model established two main categories as the cause of the site unsafe level: worker behavior and site conditions. These categories were defined based on the incident investigation and literature review (Lingard and Rowlinson 2005). The site unsafe level can cause an incident. An incident, in this research, is every occurrence likely to lead to grave consequences. Accidents are every occurrence that decreases worker availability in the project. Therefore, incidents and accidents are positively correlated.  Three main loops were identified in the conceptual model. Loop R1 is related to the site condition. The company and some researchers (Mitropoulos et al. (2005) and Han et al. (2014)) stated that the accident affects the schedule pressure causing congestion, and increasing the site unsafe condition. Moreover, factors such as temperature, project type, activity type (Lee et al. 2012), and site layout (Anumba & Bishop, 1997), can also affect the site unsafe condition.  The other two loops (B1 and B2) are related to the worker behavior. The schedule pressure can affect the worker intention to work safe (Mitropoulos et al. 2005), and consequently, the worker safe behavior. Moreover, incident investigations can increase worker knowledge and also the perception of risk (Construction Industry Institute, 2002), improving the worker safe behavior (Han et al. 2014). Figure 1 shows a conceptual model of the influence of the worker safe behavior and site conditions on the site unsafe level.  115-3 Table 1: Incident root causes description N Root Cause Description 1 Hazard Identification and Control Worker characteristics influence the identification and control of hazards. 2 Human Resource / Professional Development (HR/PD) The hiring process was not able to verify the workers’ skills and knowledge. 3 Standard Operating Procedures Practices  The safety procedures to perform a task in a safe manner were not defined. 4 Leadership and Administration Attitudes from the management do not demonstrate commitment to safety. 5 Inspection and Audits The inspection and audits of equipment, processes, and workers were not defined/realized. In this research, the worker perspective of the inspection and audits will be considered. 6 Orientation and Training The orientation/training was not able to transfer knowledge to the worker.  7 Site Specific Safety Plan There is no recommendation about the safety procedures that should be followed in the construction site.  8 Communication Systems The communication system was not able to inform the worker about the risks on the site.  9 Security/Emergency Response There are no procedures to follow if an incident occurs.  10 Engineering Verify problems related with the project design. 11 Procurement Verify errors in the procurement process, such as lack of material specification and delay in delivery. 12 Sub / Trade - Contractor Management Verify problems related with the sub/trade training and commitment to safety. 13 Environment Verify the climate conditions that can influence an incident.   Figure 1: Basic conceptual model The root causes defined by the construction company were categorized between the worker safe behavior and the site conditions categories. Each loop is explained in further detail below. Site Condition (R1): Figure 3 shows the influence of the incident root causes on the site unsafe conditions. The site unsafe level increases the quantity of incidents and accidents. According to Han et al. (2014) and Mitropoulos et al. (2005), an accident can cause delays, increasing the schedule pressure. To compensate for the delay, the company can hire new workers. However, these workers increase the site congestion. The congestion increases the site unsafe condition because it increases workers’ exposure to struck-by or struck-against incidents (Fortunato et al. 2012). According to the company safety investigation, the site safety conditions can also be affected by the root causes Environment (e.g. IncidentSite UnsafeConditionsWorker PerceptionAccident+WorkerIntentionSiteConditionSite UnsafeLevel++Worker SafeBehavior+ +-WorkerKnowledgeWorkerIntention R1B1B2Schedule Pressure+ -Congestion+++115-4 temperature, lighting, and wind), Standard Operating Procedures, Site Specific Safety Plan and Security Emergence Response.  Figure 3: Influence of the root causes on the site unsafe conditions  The site conditions are also affected by the root causes Engineering and Procurement. Both of these root causes can also contribute to the schedule pressure. Procurement can lead to material delay and poor design can increase rework.  Hazard Identification (B1): Figure 4 shows the influence of worker knowledge on the site unsafe level. If the investigation is able to identify the root causes and the results are shared with the workers, they will increase their knowledge. Workers' previous experience can also affect worker knowledge. According to the company safety database, worker experience and incidents are negatively correlated. Therefore, duirng the hiring process, it is important to identify workers with more experience. Furthermore, according to the company safety database, the quantity of pre-task meetings is negatively correlated with the quantity of incidents because it increases worker hazard perception (Construction Industry Institute, 2002). In this model, the root cause Safety Comunication represents the pre-task meeting.   Figure 4: Worker knowledge influence on the site unsafe level The improvement of workers’ knowledge facilitates worker perception of hazards (Jiang et al., 2015). However, worker perception can be affected by the root cause Hazard Identification and Control. This Site UnsafeLevelSite UnsafeCondition -IncidentAccident++SchedulePressure+Crew Size +Congestion++ErrorRework+++-ProcurementEngineeringEnvironmentSecurity EmergenceResponseStandard OperatingProceduresSite SpecificSafety PlanR1SiteConditionDesign+-Delay indelivery-+Materials notattend the safetyspecifications+Design safetymeasures+-Activities safety risk-+Defective tools andequipments-+Inadequatewarning system--Exposureto risk+Site UnsafeLevelWorker SafeBehavior-Incident+SafetyInvestigationHazardIdentification andControl+SafetyCommunicationWorker Perception+WorkerExperienceHR/PD+Orientation andtrainingB1WorkerKnowledgeWorkerknowledge+++Worker phisicalcondition--115-5 root cause represents worker physical conditions such as work shift, worker’s age, health condition and other personal characteristics that can prevent the worker from recognizing a hazard. Worker Intention (B2): Figure 5 shows the influence of the worker intention on the site unsafe condition. Because of the particularity of the worker intention, it was divided in two sub-loops: Fatigue (B2.1) and Safety Climate (B2.2).   Figure 5: Worker intention influence on the site unsafe level Fatigue (B2.1): The schedule performance can make the company increase the workers’ shifts. According to Alvanchi et al. (2012), prolonged working hours can produce fatigue due to decrease in the muscular strength and mental stress. Fatigue can make the worker take shortcuts, not follow the safety recommendations, and consequently, decrease the worker’s intention to work safely (Jiang et al., 2015). Moreover, mental stress can cause distraction and decrease the worker’s capacity for hazard recognition (Hinze, 1997). Safety Climate (B2.2): In this sub-loop, accidents increase the safety pressure and consequently increase management’s commitment to safety. However, Mitropoulos et al. (2005) stated that the schedule pressure may prevent management from providing and maintaining required safety measures, decreasing efforts to control the worker behavior. Moreover, management commitment is affected by the Leadership and Administration. According to the company HSE manual, the Leadership and Administration considers factors such as lack of discpline, lack of enforcement, lack of safety recources and lack of safety planning. The Management Commitment consequently affects the safety climate (Chinda & Mohamed, 2008). Although not specified as a root cause, safety climate is also affected by the Foreman Behavior (Choudhry and Fang, 2008). The root cause Sub-Contractor Management also affects Safety Climate. The worker perception of safety (Han et al. 2014) is influenced by the safety climate and inspection and audits. One example of inspection is the Behavior-Based Observation (BBO) Card. The BBO improves worker safe behavior because the worker feels that he/she is being watched by the safety personnel (Vaughen et al., 2010).  Figure 6 shows the complete conceptual model. Site UnsafeLevel WorkerBehavior-IncidentAccident++Safety Pressure+SchedulePressure+Work Overload+Fatigue+Worker Intention-+Safety overschedule--+ ManagementCommitment+Leadership andAdministration+Safety Climate++Inspection andAuditsB2.1FatigueB2.2Safety ClimateWorkerobservation++Sub-ContractorManagement3rd part commitmentwith safety++Foreman BehaviorForemanCommitment++115-6  Figure 6: Site unsafe level conceptual model  6 MODEL EXPERIMENTS AND DISCUSSION A hypothetical simulation model was built and four scenarios were tested to evaluate the influence of each root cause on the site unsafe level. In the first three scenarios, three different root causes were tested individually: 1) Environment, 2) Orientation and Training, and 3) Inspection and Audits. To better understand the influence of the root cause in each scenario, its value was set to 0 (worst condition), 0.5 and 1 (best condition). The other root causes had their values set at 0.5. The last graph compares the site unsafe level when all root causes are equal to 0.1 and 1. The time defined to visualize the root causes’ influence on the site unsafe level is 90 days. Figure 7 shows the site unsafe level obtained in each scenario.   Figure 7: Effect of different root causes on the site unsafe level Site Unsafe LevelSite UnsafeConditionWorker Safebehavior-+IncidentAccident++SafetyInvestigation+SafetyCommunicationWorkerPerception+WorkerExperienceHR/PD+Safety PressureSchedule PressureCrew Size+Congestion++Work Overload+Fatigue+WorkerIntention-ErrorRework++ Safety overSchedule++ManagementCommitment+Leadership andAdministration+Safety Climate+ForemanBehaviorInspection andAuditsProcurementEngineeringEnvironmentSub-ContractorManagementStandard OperatingProceduresSite SpecificSafety PlanSecurity EmergenceResponseOrientation andtraining+-B2.1FatigueB2.2SafetyClimateB1WorkerKnowledgeR1Site ConditionWorkerKnowledgeWorker phisicalconditionHazard Identificationand ControlForemannCommitment3rd par commitmentwith safetyDesignDelay indeliverMaterials notattend the safetyspecificationsDesign safetymeasuresExposureto riskInadequatewarning systemDefective Tolls andEquipment Activities safetyrisk0510152025300 20 40 60 80 100Site Unsafe LevelDaysEnvironment effects on Site Unsafe Level0 0.5 10510152025300 20 40 60 80 100Site Unsafe LevelDaysTraining effect on Site Unsafe Level1 0.5 00510152025300 20 40 60 80 100Site Unsafe LevelDaysInspection and Audits effect on Site Unsafe Level  1 0.5 0-30-20-10010203040500 20 40 60 80 100Site Unsafe LevelDaysInfluence of Root Causes on Site UnSafe LevelAll Root causes = 1 All root causes = 0.1115-7 It is possible to verify that the root causes defined by the construction company can affect the site unsafe level. The root causes defined in the model are inversely proportional with the site unsafe level. Moreover, it is possible to verify that after day 40, the site unsafe level is almost constant. This behavior is due to the schedule pressure, since the work hour overload and the crew size can compensate for the delay caused by incidents and rework. The similarity between the results of the three first graphs demonstrates that different root causes should be improved concurrently to decrease the site unsafe level (graph 4).  To improve the root causes, it is necessary to measure them. Furthermore, Table 2 suggests types of data to measure each root cause. Table 2: Incident root causes description N Root cause Suggested types of data to measure the root causes 1 Hazard Identification and Control Work shift; worker’s experience on the project; worker’s age 2 Human Resource / Professional Development (HR/PD) Average of workers’ experience on project; worker’s previous ability 3 Standard Operating Procedures Practices  Activity risk level 4 Leadership and Administration Management site inspection; participation in safety meetings  5 Inspection and Audits Quantity of BBO filled per month; quantity of workers per foreman 6 Orientation and Training Worker training hours; evaluate of workers’ learning of the course content 7 Site Specific Safety Plan Equipment and tool maintenance per month; safety program level of maturity 8 Communication Systems Quantity of pre-job inspections completed per month  9 Security/Emergency Response Escape route facilities (clear, indicated and shorter path)  10 Engineering Engineering quality by discipline 11 Procurement Procurement quality by discipline  12 Sub / Trade - Contractor Management Evaluate observation of safety practices in the project 13 Environment Temperature; wind speed; noise  - Foreman Foreman skill level; foreman age; safety supervisor experience Besides the data types presented in Table 2, the company can also collect information about other factors used in the model, such as congestion (worker ramp up and ramp down), schedule pressure (delays), rework (project quality) and safety pressure (total recordable incident rate – TRIR). The incident investigation can be improved to collect the data type suggested. Moreover, as some of the company’s incident investigations were not fully completed, the model could reinforce the importance of collecting all data requested by the investigation. In this case, the investigation will not be utilized just to describe an incident, but also to measure and control the incident root causes. The definition of the root cause can also help to better identify the incident causes, especially for those investigators who have to conduct the investigation. The conceptual model was developed to identify the relationships between the root causes, but it is not recommended to be used to predict the site safety level. For this purpose, other simulation techniques, such as hybrid models combining discrete event simulation with system dynamics, or agent-based models, can achieve better results. According to Sawhney et al. (2003), an agent-based model “can be used to mimic the construction environment in which the worker [is] performing [his/her] work, along with 115-8 heterogeneous set of agents representing these workers to study various aspects of safety.” Furthermore, root causes such as Environment, Procurement and Engineering can change values during the simulation and improvements are necessary to better predict the site unsafe level. However, the relationship between the root causes identified in this research can be used on other simulation models to improve the results. Based on the model and the data type suggested to measure the incident root causes, construction companies can adopt strategies to improve the site safety level, such as improve the selection of engineering, suppliers, and sub-contractors in aspects related to safety; implement inspection procedures such as the BBO card and measure the supervisor’s commitment to safety. 7 CONCLUSION The accident conceptual model developed in this research was able to demonstrate the relationship between the incident root causes defined by the construction company and the site unsafe level. It is also possible to conclude that not only are safety procedures, safety personnel, and field workers responsible to improve safety performance, but other company departments are as well. In this way, it was possible to conclude that, root causes such as project design, procurement, and HR/PD can affect the site unsafe level. Moreover, all departments that can cause project delays or influence the quality can influence the site unsafe level.  The company can improve the incident investigation procedures based on the conceptual model. Factors to measure each root cause were suggested and it is recommended to collect them during the incident investigation. Although this conceptual model cannot be used to predict the safety level, it is believed that the relationship defined in this research can be used in the future to develop a simulation model to measure and forecast how different safety strategies can affect the incident root cause and consequently the site unsafe level.  Ongoing research on the company safety database has been developed to validate the relationship between the root causes. Further steps are to define to which degree the different root causes affect the site unsafe level and validate the data type to measure the root causes with the safety managers. Acknowledgements This research was made possible in part by Coordination for the Improvement of Higher Education Personnel (CAPES) - Ministry of Education of Brazil (Proc. no 0393-12-6). It was also supported by the NSERC Industrial Research Chair in Construction Engineering and Management, IRCPJ 195558-10. The authors also want to acknowledge PCL Industrial Construction for their assistance. References Abdelhamid, T. S. and Everett, J. 2000. Identifying Root Causes of Construction Accidents. Journal of Construction Engineering and Management, 126(1): 52–60. Alvanchi, A., Lee, S. and Abourizk, S. 2012. Dynamics of Working Hours in Construction. Journal of Construction Engineering and Management, 138(1): 66–77.  Anumba, C. and Bishop, G. 1997. Safety-integrated Site Layout and Organization. Annual Conference of Canadian Society for Civil Engineers, CSCE, Montreal, QC, Canada, 147–156. Arboleda, C. A. and Abraham, D. M. 2004. Fatalities in Trenching Operations — Analysis using Models of Accident Causation. Journal of Construction Engineering and Management, 130(4): 273–280. Association of Workers’ Compensation Boards of Canada (AWCBC). 2012. Summary Tables of Accepted Time-loss Injuries/Diseases and Fatalities by Jurisdiction. Table 282-0008: Labour force survey estimates. Bird, F. E. and Germain, G. L. 1996. Practical Loss Control Leadership. International Loss Control Leadership. 115-9 Chinda, T. and Mohamed, S. 2008. Structural Equation Model of Construction Safety Culture. Engineering, Construction and Architectural Management, 15(2): 114–131. Choudhry, R. M. and Fang, D. 2008. Why Operatives Engage in Unsafe Work Behavior: Investigating Factors on Construction Sites. Safety Science, 46(4): 566–584.  Construction Industry Institute. 2002. Safety Plus: Making Zero Accidents Reality. Austin TX.: CII Project Rep. 160. Cooke, D. L. and Rohleder, T. R. 2006. Learning from Incidents: From Normal Accidents to High Reliability. System Dynamics Review, 22(3): 213–239.  Fortunato, B. R., Hallowell, M. R., Behm, M. and Dewlaney, K. 2012. Identification of Safety Risks for High-performance Sustainable Construction Projects. Journal of Construction Engineering and Management, 138: 499–508.  Han, S., Saba, F., Lee, S., Mohamed, Y. and Pena-Mora, F. 2014. Toward an Understanding of the Impact of Production Pressure on Safety Performance in Construction Operations. Accident Analysis and Prevention, 68: 106–116. Hinze, J. W. 1997. Construction Safety. Prentice-hall. Hovden, J., Albrechtsen, E. and Herrera, I. A. 2010. Is there a Need for New Theories, Models and Approaches to Occupational Accident Prevention? Safety Science, 48(8): 950–956.  Jiang, Z., Fang, D. and Zhang, M. 2015. Understanding the Causation of Construction Workers ’ Unsafe Behaviors Based on System Dynamics Modeling. Journal of Construction Engineering and Management, in press.  Lee, H., Kim, H., Park, M., Teo, E. A. L. and Lee, K. 2012. Construction Risk Assessment using Site Influence Factors. Journal of Computing in Civil Engineering, 26: 319–330. Leveson, N. 2004. A New Accident Model for Engineering Safer Systems. Safety Science, 42(4): 237–270.  Lingard, H. and Rowlinson, S. 2005. Occupational Health and Safety in Construction Project Management. Taylor & Francis. Mitropoulos, P., Abdelhamid, T. S. and Howell, G. A. 2005. Systems Model of Construction Accident Causation. Journal of Construction Engineering and Management, 131(7): 816–825.  Sawhney, A., Bashford, H., Walsh, K. and Mulky, A. R. 2003. Agent-based Modeling and Simulation in Construction. I 2003 Winter Simulation Conference. IEEE, Hoboken, NJ, USA. 1541–1547. Shin, M., Lee, H., Park, M., Moon, M. and Han, S. 2014. A System Dynamics Approach for Modeling Construction Workers ’ Safety Attitudes and Behaviors. Accident Analysis and Prevention, 68: 95–105. Suraji, A., Duff, A. R. and Peckitt, S. J. 2001. Development of Causal Model of Construction Accident Causation. Journal of Construction Engineering and Management, 127: 337–344. Vaughen, B. K., Lock, K. J. and Floyd, T. 2010. Improving Operating Discipline through the Successful Implementation of a Mandated Behavior-based Safety Program. Process Safety Progress, 29(3): 192–200. Wagenaar, W. A. and Schrier, J. Van Der. 1997. Accident Analyses - The Goal, and How to Get There. Safety Science, 26(l): 25–33. Wirth, O. and Sigurdsson, S. O. 2008. When Workplace Safety Depends on Behavior Change: Topics for Behavioral Safety Research. Journal of Safety Research, 39(6): 589–98.   115-10  INDUSTRIAL RESEARCH CHAIR IN CONSTRUCTION ENGINEERING AND MANAGEMENT A CONCEPTUAL CAUSATION MODEL BASED ON THE INCIDENT ROOT CAUSESEstacio Pereira, PhD CandidateHosein Taghaddos, PhD, PEngRick Hermann, PEngSangUk Han, PhDSimaan AbouRizk, PhD, PEngICSC15 – International Construction Specialty ConferenceVancouver – June, 2015OUTLINE1. Introduction2. Problem description3. Objective4. Methods5. Results6. Conclusion7. Ongoing research2015-11-10 21. INTRODUCTION 2015-11-10 3Introduction Problem Description Objective Methods Results Conclusion Ongoing researchhttp://depletedcranium.com/were-steel-workers-really-this-reckless/http://racanellinc.com/about-us/safety/http://business.financialpost.com/news/fourth-oil-sands-worker-dies-at-suncor-site-this-year-in-deadliest-stretch-since-20091. INTRODUCTION 2015-11-10 4(Wanberg et al., 2013; Hallowell, 2011; Mitropoulos et al., 2005)Introduction Problem Description Objective Methods Results ConclusionOngoing researchAccidents are an issue in the construction industry in different countries such as USA, UK, China, Brazil (ILO, 2014; Bureau of Labour Statistics, 2014; Zou et al., 2007)The incident rate for construction industry is 30% higher than for other industries (AWCBC, 2012) Understanding incident causes is difficult due to project complexity1. INTRODUCTION2015-11-10 51919 1927 1935 1943 1951 1959 1967 1975 1983 1991 1999 2007TodaySystem modelof construction accident causationMitropoulos et al. 2005StampLeveson 2004ARCTMAbdelhamid, Everett 2000Leamon's human machine learningLeamon 1980Domino theoryAdjustment-stress theoryKerr 1957Accident proneness theoryVernon 1918 cited by Hinze 19971st Group2nd Group3rd GroupHeinrich 1959Accident Causation Models/Theories - Understand the factors and processes to develop strategies to avoid accidents(Khanzode et al., 2012;  McCabe et al., 2005; Brown et al., 2000)Person causeManagement cause ComplexsystemAct of GodIntroduction Problem Description Objective Methods Results ConclusionOngoing research2. PROBLEM DESCRIPTION2015-11-10 4Complex systemMeasure and control incident root causesincident root cause identification is subjectiveIdentify indicators to measure the root causes Introduction Problem Description Objective Methods Results ConclusionOngoing research3. OBJECTIVEDevelop a conceptual incident causation model to explain the causal diagram between the root causes and the site risk level.2015-11-10 4Introduction Problem Description Objective Methods Results Conclusion Ongoing research4. METHODS2015-11-10 8Hypothetical Simulation Model – System dynamicsEmpirical equations Model behaviorBuild conceptual modelsCompany documentation Literature reviewCase StudyIdentification of incident root causes Description Define indicatorsIntroduction Problem Description Objective Methods Results ConclusionOngoing research5. RESULTS2015-11-10 6Root Cause1 Hazard Identification and Control2 Human Resources/Professional Development (HR/PD)3 Standard Operating Procedures and Practices 4 Leadership and Administration5 Inspection and Audits6 Orientation and Training7 Site Specific Safety Plan8 Communication Systems9 Security/Emergency Response10 Engineering11 Procurement12 Sub/Trade - Contractor Management13 EnvironmentIntroduction Problem Description Objective Methods Results ConclusionOngoing research2015-11-10 10Worker characteristics that influence the identification and control of hazardsHazard identification and controlThe orientation/training ability  to transfer knowledge to the workerOrientation and trainingManagement attitudes that demonstrate commitment to safetyLeadership and administrationRoot cause descriptions5. RESULTSIntroduction Problem Description Objective Methods Results ConclusionOngoing research2015-11-10 11Root cause IndicatorsHazard identification and controlWork shift; worker’s experience on the project;worker’s ageLeadership and administrationManagement site inspection; participation insafety meetingsOrientation and trainingWorker training hours; evaluation of workers’learning of the course content5. RESULTSIntroduction Problem Description Objective Methods Results ConclusionOngoing research2015-11-10 12Two main categoriesSite risk level increases incidentsIncidents increase accident3 Main loops:R1: Site ConditionB1: Worker KnowledgeB2: Worker IntentionConceptual model5. RESULTSIntroduction Problem Description Objective Methods Results ConclusionOngoing researchIncidentSite UnsafeConditionsWorkerPerceptionAccident+WorkerIntentionSiteConditionSite RiskLevel++Worker SafeBehavior+ +-WorkerKnowledgeWorkerIntention R1B1B2SchedulePressure+ -Congestion+++Site RiskLevelSite UnsafeCondition +IncidentAccident++SchedulePressure+CrewSize +Congestion++ErrorRework+++-ProcurementEngineeringEnvironmentSecurity EmergenceResponseStandardOperatingProceduresSite SpecificSafety PlanR1SiteConditionDesign+-Delay indelivery-+Materials notattend the safetyspecifications+Design safetymeasures+-Activitiessafety risk -+Defective toolsand equipments-+Inadequatewarningsystem- +Exposure torisk+2015-11-10 13Accidents increase schedule pressureHan et al. (2013), Mitropoulos et al.  (2005)Congestion increases workers’ exposure to struck-by or struck-against incidents(Fortunato et al 2012)Both factors can increase the schedule pressure and affect the site conditionsAccording to the company safety investigation these factors affect the site conditions5. RESULTSIntroduction Problem Description Objective Methods Results ConclusionOngoing researchSite RiskLevelWorker SafeBehavior-Incident+SafetyInvestigationHazardIdentification andControl+SafetyCommunicationWorkerPerception+WorkerExperienceHR/PD+Orientation andtrainingB1WorkerKnowledgeWorkerknowledge++++2015-11-10 14Sharing incident investigations with workers can increase their knowledge(Fortunato et al 2012)Meetings increase worker knowledgeCII, (2002)Improvement of worker knowledge improves the perception(Jiang et al., 2015)5. RESULTSIntroduction Problem Description Objective Methods Results ConclusionOngoing researchSite RiskLevel WorkerBehavior-IncidentAccident++Safety Pressure+SchedulePressure+WorkOverload+Fatigue+WorkerIntention-+Safety overschedule--+ ManagementCommitment+Leadership andAdministration+Safety Climate++Inspection andAuditsB2.1FatigueB2.2Safety ClimateWorkerobservation++Sub-ContractorManagement3rd partcommitment withsafety++2015-11-10 15Fatigue can make workers take shortcuts(Alvanchi et al., 2009)Schedule pressure decreases management commitment(Mitropoulos et al., 2005)Management commitment affects the safety climate(Chinda and Mohamed 2008)5. RESULTSIntroduction Problem Description Objective Methods Results ConclusionOngoing research2015-11-10 16Complete conceptual model5. RESULTSIntroduction Problem Description Objective Methods Results ConclusionOngoing researchSite Risk LevelSite UnsafeConditionWorker Safebehavior-+IncidentAccident++SafetyInvestigation+SafetyCommunicationWorkerPerception+WorkerExperienceHR/PD+Safety PressureSchedule PressureCrew Size+Congestion++Work Overload+Fatigue+WorkerIntention-ErrorRework++ Safety overSchedule++ManagementCommitment+Leadership andAdministration+Safety Climate+Inspection andAuditsProcurementEngineeringEnvironmentSub-ContractorManagementStandard OperatingProceduresSite SpecificSafety PlanSecurity EmergenceResponseOrientation andtraining+-B2.1FatigueB2.2SafetyClimateB1WorkerKnowledgeR1Site ConditionWorkerKnowledge Hazard Identificationand Control3rd par commitmentwith safetyDesignDelay indeliverMaterials notattend the safetyspecificationsDesign safetymeasuresExposureto riskInadequatewarning systemDefective Tolls andEquipment Activities safetyrisk2015-11-10 174 scenariosChanging one root cause each timeEnvironmentTrainingInspection and auditsChanging all root causes 1 and 0.1Time: 90 daysEmpirical Equation5. RESULTSIntroduction Problem Description Objective Methods Results ConclusionOngoing research2015-11-10 1801020300 20 40 60 80 100Site Risk LevelDaysEnvironmental effects site risk level0 0.5 101020300 20 40 60 80 100Site Risk LevelDaysTraining effects on site risk level1 0.5 05. RESULTSIntroduction Problem Description Objective Methods Results ConclusionOngoing research5 RESULTS2015-11-10 190510152025300 20 40 60 80 100Site Risk LevelDaysInspection and audit effects on site risk level1 0.5 0-40-2002040600 20 40 60 80 100Site Risk LevelDaysInfluence of root causes on site risk levelAll Root causes = 1All root causes = 0.1Introduction Problem Description Objective Methods Results ConclusionOngoing research2015-11-10 20Root cause indicators can affect the site risk levelRoot causes should be improved concurrentlyRoot cause indicators can be further investigated5. RESULTSIntroduction Problem Description Objective Methods Results ConclusionOngoing research6. CONCLUSION2015-11-10 21Can improve the data type collected during the incident investigation and on a daily basisDifferent parts of the company affect the site risk levelConceptual model built to identify relationships, not to predict incidentOther types of simulation can be used Introduction Problem Description Objective Methods Results ConclusionOngoing research7. ONGOING RESEARCH• Identify the main root causes and indicators for industrial construction• Validate the relationship between the indicators and the site risk level• Build simulation models combining discrete and continuous simulation 2015-11-10 22Introduction Problem Description Objective Methods Results Conclusion Ongoing researchTHANK YOU!!QUESTIONS??2015-11-10 23Estacio Pereiraestacio@ualberta.caSangUk Han, PhD: sanguk@ualberta.caSimaan AbouRizk, PhD, P.Eng:abourizk@ualberta.caINDUSTRIAL RESEARCH CHAIR IN CONSTRUCTION ENGINEERING AND MANAGEMENT A CONCEPTUAL CAUSATION MODEL BASED ON THE INCIDENT ROOT CAUSESEstacio Pereira, PhD CandidateHosein Taghaddos, PhD, PEngRick Hermann, PEngSangUk Han, PhDSimaan AbouRizk, PhD, PEngICSC15 – International Construction Specialty ConferenceVancouver – June, 2015

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.52660.1-0076393/manifest

Comment

Related Items