100 ● Can incident data be analysed to produce and evidence based understand of the criticality of controls? ● Can the same quantitative approach be extended to track performance of critical controls over time? The remainder of this paper will demonstrate a methodology for analysing incidents to produce quantified evidence of control effectiveness (or probability of failure under demand). It will conclude with commentary about extending the approach to real-time tracking of critical controls performance and the use of artificial intelligence to inform the analysis. 2. METHODOLOGY The following method can be used to analyse incidents in a manner that delivers an understanding of control effectiveness and/or the probability of the control to fail under demand. ● Obtain information on the incident that describes what happened ● If recurring event (rather than novel event), identify the material unwanted event using taxonomy of material unwanted events. Table 1 shows the taxonomy collected by the author over the past 15 years of analysing high-risk industrial accidents. ● If recurring event, obtain the bowtie for the related material unwanted event. If novel event, build a bowtie that identifies all relevant causes and consequences as well as the suite of controls that should be implemented to prevent and mitigate the event to ALARP levels. ● Analyse the incident information to identify which of the controls listed on the bowtie were o Present and effective. o Present but ineffective and if available the reasons for it being ineffective. o Absent and if available the reasons for the absence. Understanding the reasons for ineffective and absent controls can help to define key components of performance specifications. ● Collate results to highlight the percentage of incidents that were attributable to the failure of the control (i.e the control being ineffective or absent). ● If analysis involves high potential incidents (HPIs) – collate results to highlight the percentage of incidents that were arrest with controls that prevented worst outcomes and those where luck prevented the worst outcomes. Also highlight the percentage contributions of those controls that were effective in preventing the incident from being worse. ● Review analysis to identify from the evidence which controls are crucial in preventing fatalities and have this analysis critiqued but local experts and subject matter experts to inform local critical control selection as it is important to not overlook critical controls
RkJQdWJsaXNoZXIy MTM0Mzk2