In combination with a sound Safety Management System (SMS), the absence of an accident is sometimes used as a means to describe an organisation’s operation as “safe”. Is this really a valid argument? Do all organisations have a true picture of their safety landscape and have they been able to identify any close calls? Often it is the manifestation of a number of unknown and unrelated events that can result in a serious incident or accident. These events are usually an element of “business as usual” and occur every day as part of the normal operation.
Within any organisation there are probably a number of unseen issues and events taking place that, in isolation, are completely benign. However, on occasion, these issues can occur simultaneously or in a sequence that might result in a more serious incident or even an accident. Safety experts refer to these events as “Latent Issues” and they can be thought of as a bug in a piece of software code. No one knows it exists until circumstances are such that the bug manifests; sometimes with catastrophic results.
Serious incidents and accidents are, thankfully, something of a rarity in today’s airline industry. For this reason, it is often difficult to determine any safety trends from accident data alone due to the relatively small sample size. However, just because an organisation has not suffered a serious incident does not mean that it is necessarily safe. Unknown, latent issues are probably taking place continuously and it is the ability to determine their presence and then deal with them that separates a safe operation from a lucky operation.
Monitoring the mundane
Having established a sound reporting culture as part of their Safety Management System, airlines will start to receive a large number of reports from employees on the “front line” of the flying operation. Some of these reports will result in an investigation that might make various safety recommendations. Many of the reports will be “closed on receipt” (COR).
One could argue that all reports should be investigated and categorised and COR should only be used as a final option. However, in reality, it is often that case that there are simply not enough staff to deal with an increasing number of reports that often refer to the more mundane aspects of the operation. In isolation, a simple report of some minor issue in the operation may well be COR because the safety team are probably aware of the event type and there is no obvious correlation to a more serious trend. It is the fact that a latent issue is probably unknown, and that safety staff will not be looking for it that makes these such insidious events. On the positive side, there is likely to be a large sample size in the more mundane reports which will help to determine any underlying trends providing suitable analysis techniques are utilised. Indeed, careful analysis of these reports may well provide useful clues to help identify latent issues and emerging trends before they result in something more sinister.
It must be recognised that to achieve adequate analysis of all reports is not a free ride. Investment is required in terms of human resource and, possibly, suitable analytic software especially for the larger, more complex operations. This is probably the first stumbling block that a safety manager will face when trying to produce a business case for investment in the safety analytic capability of the organisation. He or she will be faced with questions such as: “If it is not broken why fix it?”, “We already have Flight Data Monitoring”, “We have had no accidents so we are safe anyway” and finally, “We have just passed our six monthly inspection by the Regulator so we must be safe”. It is a hard sell to convince the Board that trying to identify something that may or may not be present is vital to the preservation of safety. However, it is common knowledge that any accident is the result of a series of minor events that have occurred in just the right order and at just the right time. The elimination of any of these issues might well break the chain of events that result in a serious incident and so avoid significant cost in terms of human life, air-frame damage, depleted reputation and lost revenue.
Although most latent issues are unknown, some others may be the result of common practice. For example, a short cut in the engineering processes that is “always” used or a flying technique that is employed at a certain base although it is not part of the Standard Operating Procedures. These type of “known” (but unwritten) localised processes are often benign and do not result in any major issues, however, left unchecked, they have the potential to do so. Again, careful analysis of reported events will help to highlight where these “unwritten” practices are occurring and then allow management to take suitable, corrective action. Correlation with Flight Data Monitoring information will also help in this area of analysis.
Monitoring the mundane is not as boring as it sounds, and it may well help to eliminate emerging trends before they materialise as an accident. However, certain elements need to be put in place to ensure that this can be achieved successfully:
- The organisation must have a sound and open reporting culture. Unless employees are encouraged to submit reports, there is little opportunity for the safety team to determine latent issues. This starts with the provision of intuitive and accessible reporting processes.
- Management buy in is essential. Providing that managers lead by example and are also seen to embrace a “Just Culture”, employees will feel encouraged to report issues through the safety reporting system.
- Investment in appropriate resource is required in the safety team.
- Provision of suitable analytic software (possibly A.I. capable) is required. In the case of more complex operations, analysis of other data sources such as weather, FDM and crewing information should be conducted although this is very labour intensive.
Can Artificial Intelligence Help?
Although a potentially costly investment, the use of AI can enhance the organisation’s ability to monitor all reports and events and then determine any trends that may be emerging in the background. Additionally, AI will allow the safety team to conduct analysis of multiple data types. For example, crewing information, weather data and FDM can all be assessed alongside submitted reports. By completing a thorough examination of all of the disparate data sources available to the airline, AI will be able to detect and report emerging trends; many of which will not be noted by human analysis alone.
Are you Lucky or Safe?
Your safety record is no doubt very good if viewed in terms of the number of serious incidents or accidents that have occurred in your operation in the last 12 months. However, are latent issues lurking in the background? How many close calls have you had (or even know about)? Is your lack of an accident down to good luck or is it because you have established an excellent SMS in your airline? There is little doubt that a sound SMS will enhance your safety of operation but, without suitable analysis and a watchful eye on the mundane, everyday operation, it is possible that unknown issues could be aligning together to cause a more serious event. Use of AI will only serve to enhance this capability.
The very fact that the presence of latent issues is often undetected until a more serious event occurs, highlights the need to “monitor the mundane” and "business as usual" despite the initial, increased costs and draw down on resource. This investment will pay dividends and result in your organisation’s safety record relying on the engagement and skill of its workforce rather than chance.