Richard Scaife, Director of Edinburgh-based psychologists The Keil Centre, noted during a webinar promoting the new guidance that the revamp was driven by the fact that ‘best practice has moved on since 2016’ and regulators required a ‘shift towards more of a learning culture’.

International regulators have expressed concerns regarding organisations’ current capacity to learn from past incidents. They have identified a need for more rigorous analyses that disclose underlying systemic causes. Historically, there has been a tendency to focus investigations on immediate human error, disregarding the broader organisational factors that create the conditions for such errors to occur.

Operators are now encouraged to incorporate human and organisational factors directly into their primary investigation processes.

As regulators suggest more rigorous examinations of underlying factors from accidents, prioritisation of incidents and near misses, and better integration of human organisational factors into investigations, the guidance has evolved. The Energy Institute has released updated incident-reporting guidance that address these requirements. The document emphasises that organisational learning goes beyond simple root cause analysis, and represents a significant advancement in industry best practices. Incident investigation is fundamental to business performance and safety in the energy sector, allowing organisations to identify structural flaws before major failures occur.  

A detailed approach 
A high-quality investigation must address not only the ‘what’ but also the ‘why’ underlying each event. The revised EI 3295 guidance supersedes the 2016 first edition with a more detailed approach. Scaife noted that the document is a comprehensive guide that covers everything from ‘getting notified about the fact that there is something that needs to be investigated, all the way through to... learning from it’.

It features a structured process model (see Fig 1) that leads users from initial notification through to the embedding of long-term change. This model contains dedicated sections on triage, team selection and the formulation of SMART (specific, measurable, achievable, relevant and time-bound) recommendations. Adopting this approach enables organisations to transition from reactive responses to systemic solutions. The overarching objective is to develop an active learning culture rather than one that assigns blame.

Fig 1: The EI 3295 guidance is based on a learning from events process model
Source: Energy Institute

Effective organisational learning begins with a strong reporting culture in which staff feel psychologically safe. When employees fear consequences, they are unlikely to report near-misses or minor errors that serve as early warning signs. The guidance affirms that a just and fair culture is essential for collecting accurate data. Investigators must exhibit objectivity and empathy to reinforce this environment. In the absence of psychological safety, critical information may remain undisclosed, increasing the risk of later incidents.

Incident prioritisation 
Incident prioritisation is a key component of the new EI 3295 guidance. Organisations must establish clear criteria to determine which incidents or near-misses warrant investigation. Triage should account for both actual (or potential) severity and learning potential. High-potential events (HiPos) should always be prioritised to prevent major accidents. Employing a risk matrix allows for objective event classification. Without defined criteria, organisations risk inefficient resource assignment or overlooking significant hazards.

The guidance explains that teams should provide diverse perspectives and possess expertise in both technical and human factors. The organisation must appoint a lead investigator capable of managing the process and the team. Maintaining independence is important to prevent bias or conflicts of interest. In tricky or sensitive cases, engaging external experts may be necessary. Investigation teams should be allocated sufficient time away from routine responsibilities to focus on the inquiry.

A just and fair culture is essential for collecting accurate data. Investigators must exhibit objectivity and empathy to reinforce this environment.

Meticulous data collection is also essential. This includes scene preservation, evidence gathering and conducting interviews. The PEACE framework delivers a planned approach for effective interviewing. The guidance explains that PEACE stands for ‘preparation and planning’, ‘engage’, ‘account’, ‘closure’, ‘evaluation’. The ‘engage’, ‘account’ and ‘closure’ phases take place during the interview, and the ‘preparation’ and ‘evaluation’ phases take place before and after, respectively. Investigators should remain aware of unconscious biases, such as hindsight and confirmation bias, which can lead to the overlooking of evidence.

Fig 2: The guidance offers multiple ways to analyse what went wrong 
Note: taken from Table 9: Quick reference guide on tools and techniques
Source: Energy Institute

SMART recommendations 
The guidance states that, to address root causes and specific issues identified during analysis, recommendations must be SMART. The investigation team should collaborate with the line organisation to ensure each proposed action is feasible and realistic. Recommendations should be prioritised based on their potential to achieve the greatest risk reduction. Organisations must employ effective systems to monitor progress on actions. Monitoring action closure times can reveal resource constraints or managerial deficiencies.

To embed organisational learning, leaders should: 

• Integrate lessons into risk assessments, training and management systems.
• Update safety critical task analysis (SCTA) data following incidents.
• Communicate findings through incident alerts and toolbox talks.
• Revise equipment and policies as necessary.
• Consistently review organisational performance to drive continuous improvement.

Key performance indicators (KPIs) are essential for monitoring the effectiveness of organisational learning. These should encompass ‘hard’ indicators, such as reductions in recurring incidents, as well as ‘soft’ indicators that assess report quality and the robustness of the reporting culture.

Scaife suggested looking at the ‘quality of the analysis that’s been performed and the closure of the actions’. Tracking the speed of action closure demonstrates the seriousness with which findings are addressed. Well-designed KPIs promote desired behaviours and prevent stagnation in the process.

The Energy Institute offers platforms to facilitate communication and knowledge-sharing, such Toolbox, a database of incidental alerts shared by industry, its members and others in industry. The wider learning process incorporates both top-down expert reviews and participative team discussions. Scenario-based exercises allow teams to test their insight into complex systems. Such cooperative efforts are vital to improving global safety performance.

Leadership plays a key role throughout the incident investigation and organisational learning cycle. Leaders must provide enough resources, including time, budget and personnel, to support high-quality investigations. They should actively engage with findings and personally approve the closure of key actions. By demonstrating a commitment to learning, leaders lay the foundation for a strong safety culture.

Additionally, leaders should proactively seek to understand the daily challenges faced by employees. During the webinar, David Garner, Director of Safety, Health and Environment at Cadent Gas, said: ‘It’s as important to get your engagement right as it is to get the right tooling and the right mindset.’ He further emphasised that ‘learning is really only any use if you do something with it’, comparing it to a book that only has value if you take it out of the library. Without strong leadership, even the most effective investigation processes will not yield lasting change.

Technological developments, including artificial intelligence (AI), are influencing incident management. AI tools can analyse large data sets and identify trends that may otherwise go undetected. These tools also assist in generating structured reports and benchmarking performance across multiple sites. However, the guidance emphasises that data accuracy and consistency are essential for the effective use of such technologies. Although technology can support incident management, it cannot substitute for human expertise and critical thinking.

Download the updated EI 3295: Reporting, investigating, and learning from incidents, accidents and events, March 2026

Watch the webinar now: 

• Further reading: ‘The right way, and the wrong way, to deal with near misses’. If industrial accidents were a scientific experiment, near misses would be the control case. They can offer useful insights, but also impose challenges of their own. Four health and safety experts provide a guide to near misses, as interpreted by the Tripod Lite accident investigation methodology.
• ‘Buncefield anniversary prompts appreciation of health and safety standard developments’. Twenty years after the Buncefield event, its legacy remains visible in the ensuing changes in health and safety practice, which included significant contributions from the Energy Institute.