Although previous research has emphasized the use of safety-related measures to assess and control safety performance, many construction companies continue to rely on reactive indicators for safety control. The reluctance of industry to use safety-related measures for the proactive evaluation of safety performance is a consequence of the unique characteristics of construction projects, which renders the identification and evaluation of safety-related measures difficult in practice. From a theoretical perspective, models developed to proactively assess the safety performance have difficult considering (1) the specific characteristics of an organization due to the limited amount of data points provided and (2) the dynamic nature of construction sites, which can affect measure performance.

The objective of this research was to develop a framework that could proactively assess safety performance using safety-related measures. A framework, which combines Case-Based Reasoning (CBR) with simulation modelling, was proposed for this purpose. CBR was chosen as an assessment method for its ability to assess safety output under conditions of limited data, while simulation was considered for its ability to reliably reproduce project behavior.

Prior to framework implementation, challenges for understanding current Safety Managements System (SMS) practices and identifying safety-related measures were researched. Then, a risk-rating approach designed to investigate the complex relationship between SMS factors and accident precursors from a holistic perspective was developed. This approach allows for not only the identification of high-priority SMS factors and accident precursors but also for the evaluation of associations between them, and a many-to-many relationship was found to exist between these groups. A case-study was then conducted to analyze and review the effective use of departmental data to control and assess safety performance. Several safety-related measures collected by several departments were found to be useful for proactively controlling safety performance. Altogether, the findings presented here strongly support the use of a holistic approach for the evaluation and control of safety performance

The framework was developed and applied in an industrial construction organization. A total of 27 safety-related measures were identified from data collected from departmental databases, documents, and interviews. After reducing the dimensionality of the database through the application of statistical tests and Correlation Feature Based Selection Algorithm, the final approach considered nine measures. From a practical perspective, the model was found to reliably predict trends in safety performance, and can be used to predict how specific decisions made in practice can affect safety performance.

This research has resulted in (1) the development of a CBR approach that can be used to assess the safety performance of construction projects characterized by few data points (i.e., few incidents) while allowing for the consideration of an organization’s unique conditions and (2) the integration of CBR with a simulation model, which allows managers to more easily predict how decisions, both individually and in combination, influence overall safety performance. Furthermore, the results of this research support the use of (3) a holistic approach for the establishment and evaluation of proactive SMS mitigation strategies and (4) the collection and consideration of various departmental data to more reliably evaluate proactive safety performance.