1. Risk Assessment Concepts: Global Perspectives
1.1. Simple risk matrices have significant limitations, and these must be understood through cross-cultural validation.
1.2. Risk perception includes heuristics, biases and cultural variations (Tversky and Kahneman, 1974; Hofstede, 2001).
1.3. Uncertainty, ambiguity and ignorance require distinguishing between knowns and unknowns (Stirling, 2010).
1.4. Low-probability, high-consequence events present particular assessment challenges.
1.5. Black swan theory and dragon-king theory provide frameworks for understanding rare events (Taleb, 2007; Sornette, 2009).
1.6. Case study: The Fukushima Daiichi disaster illustrates the failure to assess low-probability events (Funabashi, 2012; Acton and Hibbs, 2012).
2. Quantitative Risk Assessment Methods
2.1. Fault Tree Analysis includes Boolean logic and minimal cut sets (Vesely et al., 1981).
2.2. Event Tree Analysis includes initiating events and branching probabilities (Andrews and Dunnett, 2000).
2.3. Bow-tie analysis combines fault tree and event tree approaches (CCPS, 2018).
2.4. Layer of Protection Analysis includes independent protection layers (CCPS, 2001).
2.5. Hazard and Operability Study includes guide words and deviation analysis (Kletz, 1999).
2.6. Bayesian methods and Monte Carlo simulation are used in quantitative risk assessment.
2.7. Case study: The Piper Alpha disaster demonstrates the application of Fault Tree Analysis and Event Tree Analysis (Cullen, 1990; Paté-Cornell, 1993).
3. Qualitative Risk Assessment Methods: Adaptable to Local Contexts
3.1. Hazard identification workshops must be facilitated appropriately in different cultural settings.
3.2. What-if analysis and structured what-if analysis are valuable qualitative methods.
3.3. Checklist methods have limitations, including the need to adapt checklists to local hazards.
3.4. Human Reliability Assessment techniques provide structured approaches (Kirwan, 1994; Bell and Holroyd, 2009).
3.5. Appropriate methods must be selected for different risk contexts.
4. Predictive Analytics and Artificial Intelligence in Risk Management
4.1. Leading and lagging indicators have distinct characteristics and selection criteria.
4.2. Near-miss data analysis uses clustering and pattern recognition techniques.
4.3. Statistical Process Control is applied in health and safety contexts (Shewhart, 1931; Oakland, 2003).
4.4. Predictive modelling includes regression, classification and time series analysis.
4.5. Machine learning applications include incident prediction and anomaly detection (Kotsiantis, 2007; Sarkar et al., 2018).
4.6. Digital twins enable safety scenario modelling.
4.7. Wearables and the Internet of Things enable real-time risk monitoring.
4.8. Ethical considerations include privacy concerns addressed by the General Data Protection Regulation, algorithmic bias, consent and the digital divide (European Parliament, 2016; Mittelstadt et al., 2016).
4.9. Case study: Artificial intelligence-based risk prediction in construction demonstrates practical applications (Choi et al., 2020; Kim et al., 2021).
5. Systemic Risk and Control Strategies
5.1. Systemic risks include supply chain disruptions, climate change and pandemics (Helbing, 2013; World Economic Forum, 2023).
5.2. As low as reasonably practicable requires practical application and understanding of tolerability of risk criteria (HSE, 2001; Health and Safety Executive, 2010).
5.3. Variations in "reasonably practicable" occur across different economic contexts.
5.4. Risk appetite and risk tolerance statements guide organisational decision-making (ISO, 2018).
5.5. Cost-benefit analysis in risk control involves valuing life and health (Viscusi and Aldy, 2003).
5.6. Monitoring and review of control effectiveness uses key performance indicators, audits and assurance.
5.7. Case study: The COVID-19 pandemic provides lessons in systemic risk management (World Health Organization, 2020; Independent Panel for Pandemic Preparedness and Response, 2021).
6. International and Local Application: Guidance for Learners
6.1. High-resource settings can employ advanced Quantitative Risk Assessment, artificial intelligence and machine learning applications, and ISO 31000 (International Organisation for Standardisation, 2018).
6.2. Low-resource settings should focus on qualitative methods, low-cost predictive techniques and building reporting culture.
6.3. Multinational organisations must harmonise risk methodologies across countries.
6.4. Major hazard industries require advanced methods including Layer of Protection Analysis, Hazard and Operability Study and bow-tie analysis. Risk assessment methodologies must also consider local regulatory frameworks and regional standards, which vary across jurisdictions.
Example Candidate Response Booklet
Example Candidate Response (ECR) Booklets are a source of crucial information for Centres and Candidates as they use real candidate responses. We ask Senior Examiners to comment on five or more responses in terms of why the mark was awarded with commentary about how to improve the answer (if necessary).
