(22b) Driving Performance Improvement Using Progressive Leading Indicators

Recently, there have been more and more discussions about predictive analytics.  In many ways, predictive analytics offer a promising approach to help organizations focus their efforts and investments. Yet, to be effective, predictive analytics need to be conducted in a rigorous manner - from data readiness assessment to proving insight on the operational factors that underlie and enable the predictive relationship.

This paper sheds lights on the facts and fictions of predictive analytics to prevent organizations from exposing themselves to misleading analytical exercises. Using a rigorous process, the study examines over 700 million work hours and more than 18,000 incidents, including loss incidents and near-misses from the oil and gas industry, and offers breakthrough insights on the concept of progressive leading indicators.

The paper shares results from two tracks of analysis - predictive and prescriptive. It begins with data readiness assessment, which examines data using five different readiness dimensions. In the predictive analysis, the study shows that Learning-Mindedness, Risk Sensitivity and Process Execution are statistically significant (p<0.01) leading indicators of stronger operating discipline and better safety performance. The prescriptive analysis further compares and summaries the distributions of root causes of fire and explosion, LOPC, and injury and illness across various levels of potential risks.

The paper challenges conventional class-room training method and demonstrates how organizations use enterprise information system as a constant learning platform to improve effectiveness of learning by leveraging expert validated content and insights. It showcases how organizations use a process-driven information system to measure reporting behavior, quantify timeliness and diligence in process execution, capture the lessons learned, identify high learning-value events, engage subject matter experts in analyzing root causes and seeking solutions, and to allow “learning” to take place systemically. While this paper provides statistical evidence to the strong correlations between “risk sensitivity”, “process execution” and “learning-mindedness” and injury outcomes, it also offers insights on the supporting organizational factors that are necessary to enable and sustain effective risk mitigation efforts.