9. OS and H System vs incident learning

Incident reporting system brings safety problems to the attention of management. Incident reporting is an important and necessary part of the incident learning system and, based on the literature survey a relationship can be established as shown in Figure 8.

Figure 8

The model shows that risky behaviour with unsafe conditions leads to incidents/ accidents. These incidents will only be reported when there is willingness of reporting incidents and it is governed by either the existing regulatory requirement and or due to management policy. The reported incidents will then be investigated to know the reasons for corrective actions. Corrective actions will improve the unsafe conditions and also the risky behaviour. These in turn will lead to review of the other resources and requirement of OS and H as discussed in detail as point 3.0. Therefore management should promote for incidents/accidents reporting.

If management demonstrates a commitment to safety through their words and deeds then eventually this will translate into a higher Personal Commitment to Safety on the part of employees. As shown in Figure 9, this role-modelling behaviour helps to reinforce the balance between productivity and safety that management has struck. To see this, consider what happens when Management Commitment to Safety goes up. After a delay, Personal Commitment to Safety also goes up. Via the outer feedback loop, this leads to a greater Willingness to Report Incidents and higher Safety Pressure to reinforce Management Commitment to Safety. Via the inner feedback loop, this leads to less Risky Behaviour, causing lower Losses and less Productivity Pressure. Less productivity pressure allows management to maintain its focus on safety. The same type of positive reinforcement occurs if management commitment to safety is falling. However, research shows that incidents are typically not reported. Reasons for not reporting incidents include a fear of punishment, bureaucratic or confusing reporting requirements, or quite simply a desire not to interrupt the work flow.

Figure 9

Learning from incidents is not an entirely new concept, but it has not been fully explored as a system for long-term continuous improvement to organizational performance from safety improvement angle. A "disaster dynamics" model that provides insight into the role that a stream of events or frequent interruptions can play in causing disaster by "information overload," but they were not concerned with incident learning. The time period for their dynamic simulation was minutes rather than the months and years involved in incident learning. However, their model does provide a relevant warning that an incident learning system will collapse if it becomes overloaded with incidents. To deal with the incident workload, dedicated resources and processes are required to ensure effective learning. As an example, the commercial airlines have these dedicated resources and, as Haunschild and Sullivan (2002) report, learning from incidents is indeed taking place in this industry. To understand how learning can be facilitated, Figure 10 shows the fundamental components of an incident learning system. We will briefly describe each of these to help clarify how the system works in following points:

  • the importance of identification, without which incident learning is not possible. Unless the organization is sensitized to learning from incidents, deviations from normal behaviour will go unnoticed or be accepted as "normal deviation"
  • The next component of incident learning is reporting. As the Centre for Chemical Process Safety (1989) points out, an incident cannot be investigated unless it is reported. Furthermore, the fraction of incidents reported is dependent on the personal commitment to safety of the workers who observe or are involved in the
  • Incidents. As discussed the management creates the safety climate and so personal commitment to safety of the workers is strongly influenced by management's commitment to safety. Management can show their commitment to safety by creating a climate in which incident reporting is rewarded instead of punished.

Incident investigation is the most well-known component of the incident learning system, involving examination of the site, interviewing witnesses, gathering and evaluating all available data to establish the sequence of events and determine exactly what happened. An investigation team will be more effective than a single investigator.

Figure 10

  • The literature suggests that the purpose of incident investigation is to determine the basic or root causes of the incident. However, since there may be no single "root cause," efforts are better directed towards identifying causal structure (a system model of the causal relationships).
  • Next, it is important to implement corrective actions and follow up on all recommendations made by the investigation team. This is particularly true for actions to eliminate systemic causes of incidents, which may span the organization and involve many people in different departments and locations. Processes outside of the incident learning system, such as management of change, audits and inspections, are useful in checking that corrective actions have been successfully implemented without introducing new risks.
  • Finally, it is important to capture and communicate the learning from the incident, including the relative success or effectiveness of the corrective actions that were taken. This can be done by distributing a summary report by e-mail, website posting, or other means, and should be directed both locally and centrally. In this step organisations should not hesitate in gathering information and lesions learned from other disasters and organisations to translate the best practices into business system to reduce the frequency of incidents.

The whole process should be a continual process after each incident.


Locations of visitors to this page