The Implementation of Statistical and Forecasting Techniques in the Assessment of Safety Intervention Effectiveness and Optimization of Resource Allocation

Open Access
Author:
Oyewole, Samuel Adekunle
Graduate Program:
Industrial Engineering
Degree:
Doctor of Philosophy
Document Type:
Dissertation
Date of Defense:
June 05, 2009
Committee Members:
  • Andris Freivalds, Dissertation Advisor
  • Andris Freivalds, Committee Chair
  • Joel B Haight, Committee Chair
  • David J Cannon, Committee Member
  • Ling Rothrock, Committee Member
  • R Larry Grayson, Committee Member
Keywords:
  • incident rate forecasting
  • safety factors
  • intervention effectiveness
  • resource allocation
  • incident rates
  • safety intervention
Abstract:
Most engineering processes involving human and technical sub-systems are designed to achieve a set of objectives. In health and safety, the need to quantify these processes using statistical models cannot be over emphasized, since the high incident rates and ineffective allocation of resources could be costly to several organizations. The objective of this research is to use statistical and forecasting tools to develop an effective resource allocation program, based on the need to reduce incident rates and safety intervention costs. Five main safety intervention factors (Factor A: Leadership and Accountability; Factor B: Qualification Selection and Pre-Job; Factor C: Employee Engagement and Planning; Factor D: Work in Progress; Factor E: Evaluation, Measurement and Verification) were highlighted and investigated to show their effects on incident rate performance. A safety intervention factor is a group of safety and health activities which are implemented in order to reduce incident rates. Analysis of variance test showed that four safety factors (A, C, D, and E) were significant. Factor B was not selected for model development, since it was not significant. A safety model was developed to assist practitioners in making resource allocation decisions, and to better predict incident rates. Statistical techniques such as response surface designs and contour plots were used to determine the resource allocation method. The developed safety model recommended the allocation of 16.66% of the available resources to the significant safety intervention activities in order to achieve the desirable incident rate, and 10.34% of the available resources to achieve the lowest acceptable incident rate. The developed safety model was validated using the comparison between the actual incident rates in a one-year period and the predicted incident rates that was obtained using the double exponential smoothing technique (Holt’s Model). Comparison of the actual and predicted incident rates indicated a forecast accuracy of 71.58%. The analysis of the forecasting error showed an unbiased forecast with a tracking signal of -4.08. This dissertation offers a new dimension into the practice of safety intervention evaluation. For the first time, this research contributes to the body of knowledge through the use of response surface design methodology and contour plots in the determination of an effective method for the allocation of resources, with the aim of reducing incident rates. Safety personnel, supervisors and managers could use the methods proposed and results obtained from this research work to develop an effective resource allocation program which would ultimately reduce safety intervention costs.