There are many different definitions for what constitutes an incident or an accident, however the focus is always on unintended and often unforeseen events that cause unintended consequences. This article is focused on the process of learning from incidents and accidents. Learning from incidents can been defined as a process through which employees and the organisation as a whole seek to understand any negative safety events that have taken place in order to prevent similar future events. The focus is on making sure that the lessons learned from incident investigations are implemented and lead to an actual improvement in safety. For this article we will use the terms incident for accidents, near misses and other unwanted events.
Many organisations have problems in reducing the number of incidents and this can be partly attributed to the failure to learn the lessons from incidents that had occurred. Incident prevention is strongly based on learning from previous incidents. When incidents occur they raise awareness and understanding of things that went wrong, and perhaps can go wrong again. The challenge is to learn as much as possible about the causes of accidents and near misses that have already happened in order to prevent reoccurrence. When experiences of previous incidents are translated into preventive measures, an organization can prevent incidents in the future and the need for repressive actions at that time.
Investigating and learning from incidents is also a requirement for continuous improvement of an OSH management system. The ISO 45001 standard considers the investigation of the root causes of incidents, the implementation of appropriate measures and their communication throughout the organisation to be essential to the process of continuous improvement.
Learning is acquiring new or modifying existing knowledge, behaviours, skills, values, or preferences and may involve synthesizing different types of information. A learning organisation actively creates, captures, transfers, and mobilises knowledge to enable it to adapt to a changing environment. When learning from incidents the organization uses the information from incidents to change, to actively improve and to prevent future incidents.
A learning organisation is the cornerstone of an OSH culture based on mutual trust and a shared perception of the importance of taking preventive measures and managing OSH. Findings from the Zero Accident Vision’-study support the importance of learning from incidents. The study involved 27 companies in seven different countries and used questionnaires as well as qualitative interviews to identify the factors that contribute to successful implementation of the Zero Accident Vision. The study showed that learning processes play a key role in on-going safety improvement processes. Learning includes not only training workers and managers, but also collective learning from experiences such as incidents.
It is also useful to learn from incidents of others. Knowledge from these incidents allows for comparison with the own situation and systems and enables you to generate creative solutions and take time to prioritize the measures. Immediately after an incident occurs in an organisation, the situation has changed. There is limited time to consider different solutions, sometimes causing the selection of suboptimal measures.
Steps in learning from incidents
The step that has been most frequently described and analysed in literature on learning from incidents is incident analysis. The investigation and analysis of an incident can be made in several ways.
However the need for follow-up steps has been described as well. Schein (1996) for example emphasises the need for dissemination of lessons learnt within an organisation. Lindberg (2010) developed a CHAIN model in which lessons are disseminated and used for preventive actions. Drupsteen et al. (2011) developed an analytical framework to analyse where problems in learning from incidents arise. This framework enables analysis of the learning from incidents process and consists of several steps, divided into four blocks or phases: incident investigation and analysis, planning of interventions, intervening and evaluation. This evaluation includes the implementation of actions, the effectiveness of the action and also an evaluation of the quality of the learning process itself. This is a cyclical process that should be fully completed to establish learning. If an intervention is not effective, the actions or the analysis should be re-evaluated and adjusted. This can to some extent be compared to the evaluation of implementations that are based on recommendations from an audit or a risk assessment. In this way, incidents can be regarded as interesting warning signals and as opportunities to learn. The outcomes of this process are iterative: a plan of action is made, the actions are performed and actions are evaluated. Based on this evaluation new lessons are formulated. This sequence is also known as the plan-do-check-act cycle by Deming. Experiential learning models that are similar to the Deming cycle can also be used to describe and analyse collective or organisational learning processes.
It is important to recognise that there can be a number of barriers to effective learning from incidents and accidents. These can include 'fear, blame and anxieties about reputational loss and legal proceedings' and can be enhanced (or alleviated) by the culture within an organisation.
To learn from incidents, some organisational factors are important. These aspects are related to the structure or to safety management, such as incident registrations and action plans, or to the organisational safety culture, such as transparency, blame free culture, and feedback on unsafe behaviour.
The actual performance – but also the learning – of organisations is determined by the practical activities in organisations, referred to by Argyris as theory-in-use. Managers all too often only learn through ‘talking and thinking’ without actually changing their behaviour. This is what Argyris & Schön (1978) call espoused theory. Also auditors of safety management systems that focus too much on the documentation of procedures, and do not carefully investigate the actual practice, are less effective because they address mainly the espoused theories of the organisation.
In contrast, operators are often sent to training sessions where they learn to do something but the attention paid to this improved understanding and knowledge is often too minimal . A study illustrates the differences in espoused theory and theory-in-use when learning from incidents, according to safety professionals. These results indicate that arranging incident analysis and follow up steps are not sufficient for learning.
Learning from incidents might also be hindered by the gap between "those who investigate" and "those who implement actions". Results from incident investigations are often summarised by the investigators into learning points and disseminated throughout the organisation with little or no input from the workers. Companies that organise the follow-up of incidents by encouraging groups of workers to think about the ways in which incident investigation findings fit with their own work contexts, prove to be more successful.
Levels of learning
An important aspect of learning from incidents is that the processes can be thought to address different ‘levels’ of learning(Argyris & Schön, 1978). Different theorists have presented taxonomy’s for qualitatively different ways to learn either as an individual or organisation. Piaget (1969), who focused on learning in schools, distinguished three distinct levels of learning:
- being able to reproduce certain knowledge;
- being able to apply the knowledge in a similar setting as it was first offered;
- being able to apply the knowledge adequately in other (new) settings.
A well-known distinction in organisational learning research is between so called ‘single loop learning’ and ‘double loop learning’ (Argyris et al., 1979). Argyris’ concepts are related to Gregory Bateson’s concepts of first and second order learning. If an organisation exhibits single loop learning, only the specific situation or processes are improved. However when an organisation exhibits double loop learning, improvements are not limited to the specific situation but the values of the theory-in-use is evaluated and changed as well (Argyris & Schon, 1978).
When investigating an incident that was caused by a malfunctioning valve this could for example mean that:
- a defective valve which caused an incident is replaced by a functioning valve of the same model (single loop learning);
- a defective valve which caused an incident is replaced by a new more appropriate type of valve (double loop learning).
An important kind of double loop learning is the learning through which the members of an organisation may discover and modify the learning system. This so called learning to learn process (called Deutero learning by Argyris & Schön) enables an organisation to continuously improve. In this process systems thinking and the mental models of the key actors – who are able to change the system – are key. Senge also demonstrates that many learning processes are actually hindered by unintended counter-balancing processes.
In our example of an incident caused by a malfunctioning valve a measure that could be seen as a result of ‘learning to learn’ could be:
- In response to an incident caused by defective valve regular equipment audits are improved. In addition systems are put in place to regularly re-evaluate both the overall quality of valves and the quality of the valve monitoring audits.
Another theoretical model describing ‘levels of learning’ is presented in the later works of Argyris. He distinguishes two styles of learning, and called them respectively Model I and Model II learning. His empirical works clearly show that Model II learning, which is characterised by the open sharing and discussing of reliable information, decision-making based on free informed choice, and internal commitment on decisions made, complemented with monitoring the implementation, is much more effective for the achievement of meaningful changes and innovations than the preferred learning style of most organisations, i.e. Model I learning. In Model I learning ‘inferences about another person’s behaviour without checking whether they are valid’ are made and views are advocated ‘abstractly without explaining or illustrating one’s reasoning’.
The following overall conclusions can be made:
- Despite efforts to investigate and learn from incidents many companies still have difficulty to sustainably learn the lessons from previous incidents. More effective learning from incidents or accidents could help prevent accidents in the future.
- Theorists have developed models that describe the steps that a company needs to take in order to learn from incidents. These models could help further improve safety management.
- There are many organisational learning theories that can help scientists and practitioners understand learning from incidents in their organisation. Useful distinctions are between: espoused theory and theory-in-use; single loop and double loop learning; learning to learn or deutero learning.
 Margaryan, A., Littlejohn, A., Stanton, N. Research and development agenda for Learning from Incidents. Safety Science, 99(A), 2017, pp. 5–13. Available at: http://libeprints.open.ac.uk/50579/7/50579.pdf
 Kletz, T., Lessons from Disaster- How Organisations Have No Memory and Accidents Recur, Gulf Professional Publishing, 1993.
 Kjellen, U., Prevention of accidents through experience feedback, Taylor & Francis, London and New York, 2000.
 Kletz, T., Learning from accidents, Butterworth-Heinemann, Oxford, 2001.
 Gort, J., Zwaard, A.W., Stavast, K.I.J., & Van Alphen, W.J.T. Leren van ongevallen, een overzicht van analysemethodieken, Sdu Uitgevers, Den Haag, 2010.
 ISO 45001:2018 Occupational health and safety management systems — Requirements with guidance for use
 Argyris, C. Overcoming organizational defenses. The Journal for Quality and Participation 15.2 (1992): 26.
 Senge, P. M., The Fifth Discipline; The art and practice of the Learning Organization, Doubleday, New York, 1990.
 Zwetsloot, G., Kines, P., Ruotsala, R., Drupsteen, L., Merivirta, M., Bezemer, R. The importance of commitment, communication, culture and learning for the implementation of the Zero Accident Vision in 27 companies in Europe, Safety Science, vol. 96, 2017, pp. 22-32. Available at: https://www.sciencedirect.com/science/article/pii/S0925753517303922?via%3Dihub#b0090
 Lindberg A-K., Hansson S. O., Rollenhagen C., Learning from Accidents – What More Do We Need to Know?, Safety Science, 48 (6), 2010, pp. 714-721.
 Schein, E.H., Culture; the missing concept in organization studies. Administrative Science Quarterly, 1996.
 Drupsteen, L., Steijger, D.M.J., Groeneweg, J., Zwetsloot, G.I.J.M., What are the bottlenecks in the learning from incidents process? Paper presented at IchemE Hazards XXII conference, Liverpool, United Kingdom, 2011.
 Deming, W. E., Out of the Crisis; Quality, Productivity and Competitive Position, Cambridge University Press, Cambridge, MA, 1982.
 Swieringa, J. & Wierdsma, A., Becoming a Learning Organization. Longman Group, United Kingdom, 1992.
 Zwetsloot, G. I. J. M. & Allegro, J. T., Organisatieverandering door managementsystemen voor voortdurende verbetering, Gedrag en organisatie 7, 1994, pp. 352-65.
 Royal Society for the Prevention of Accidents. Learning how to learn from accidents. Available at: http://www.rospa.com/rospaweb/docs/advice-services/occupational-safety/noshc/learning-how-to-learn-from-accidents.pdf
 Argyris, C., & Schon, D. (1978). Organizational learning: A theory of action perspective. Reading, MA: Addison-Wesley.
 Drupsteen, L., Groeneweg, J., Zwetsloot, G.I.J.M., Identifying the bottlenecks in learning from incidents: From reporting an incident to verifying the effectiveness of the remedial process, 2010. Available at: https://repository.tno.nl/islandora/object/uuid%3A8a698f7a-8568-4422-b752-12ea83c2edba
 Parker, A.. , Ummels, F.. , Wellman, J.. , Whitley, D.. , Groeneweg, J.. , Drupsteen, L.. How to Take Learning from Incidents to the Next Level. Paper presented at the SPE International Conference and Exhibition on Health, Safety, Security, Environment, and Social Responsibility, Abu Dhabi, UAE, April 2018. Available at: https://doi.org/10.2118/190646-MS
 Piaget, J., The Mechanisms of Perception. Rutledge & Kegan Paul , London, 1969.
 Bateson, G., Steps to an Ecology of Mind: Collected Essays in Anthropology, Psychiatry, Evolution, and Epistemology, University Of Chicago Press, 1972
EU OSHA – European Agency for Safety and Health at Work. New trends in accident prevention due to the changing world of work, 2002. Available at: http://osha.europa.eu/en/publications/reports/208
EU OSHA – European Agency for Safety and Health at Work. Occupational Safety and Health culture assessment - A review of main approaches and selected tools, 2011. Available at: https://osha.europa.eu/en/publications/occupational-safety-and-health-culture-assessment-review-main-approaches-and-selected
EU OSHA – European Agency for Safety and Health at Work. Worker participation practices: a review of EU-OSHA case studies, 2012. Available at: https://osha.europa.eu/en/publications/worker-participation-practices-review-eu-osha-case-studies
ILO - International Labour Organization. Learning from work related accident: why and how? Available at: https://www.ilo.org/budapest/what-we-do/projects/declared-work-ukraine/WCMS_769666/lang--en/index.htm
Vision zero platform https://visionzero.global
DGUV - Deutsche Gesetzliche Unfallversicherung. Culture of Prevention https://www.dguv.de/en/prevention/visionzero/culture_of_prev/index.jsp