Not so fast: taking time to learn lessons from Sandilands

On 9 November 2016, Croydon Tramlink Tram 2551 reached its maximum permitted speed of 80 km/h as it entered the first of three closely-spaced tunnels.

The tram should have reduced its speed to 20 km/h as it approached the sharp curve round to Sandilands Junction, in South London. However, it didn’t. The tram overturned and slid along the adjacent gravel. Seven people died and 61 were injured.

When major accidents occur, there is a visceral response and a demand for more to be done. But how can we turn this energy into beneficial action? The priority is to understand what went wrong and implement appropriate measures. Following Sandilands, rapid action was undertaken to improve speed warning signs and to fit a monitoring system to detect when drivers were fatigued.

But beyond immediate fixes, confidence is needed in the underlying safety management. Major accidents serve as a warning for all parties to review their arrangements, but there is also potential for a knee-jerk reaction and poor decisions.

The derailed tram at Sandilands Junction on 9 November 2016 caused seven deaths

The answer to this dilemma is to allow decision-makers to reduce risk efficiently and effectively by understanding the risks and acting in a risk-based way.

When decisions need to be made based on future uncertain events, then subjectivity is introduced. And when that subjectivity relates to events with legal implications, confidence in decisions becomes more difficult for those with accountability. After an accident, maintaining a dispassionate, objective view is very difficult; there are certain traps we fall into.

One is ‘hindsight bias’ – the inclination to see past events, such as accidents, as more predictable than they really were. This can drive a blame culture, the very antithesis of good safety management, which is based on open learning.

The other is ‘narrow framing’ – people’s tendency to view problems in isolation, rather than taking a broader view.

Transport systems have to deal with a whole range of issues and a portfolio of risks, and if one particular issue is focused on disproportionately, this results in sub-optimal outcomes. Focusing on single accidents as the justification for strategic decision-making is likely to be flawed.

So how do we deliver a rational, risk-based approach to safety management? The UK railway industry in the late 90s and early 2000s provides a case study.

The period is marked by several fatal rail accidents, including the collision of two passenger trains following a signal being passed at danger that led to 31 deaths and many injuries at Ladbroke Grove in 1999. This was a watershed moment.

'Hindsight bias’ – the inclination to see past events, such as accidents, as more predictable than they really were, is not helpful when learning lessons from Sandilands

There is an inevitable loss of public trust when a major accident happens. In the case of the Ladbroke Grove railway accident, the holding of the independent public inquiry by Lord Cullen was a key action to regain that trust. To support robust learning and ensure that the right questions are asked, independent expert views and strong governance are important, and this is not just the case in the aftermath of an accident. There is always a place for such a role.

This was recognised in The Ladbroke Grove Rail Inquiry (Cullen report), which included recommendations for a suitably authoritative ‘system authority’, central assurance of contractors to the railway and ultimately the setting up of an independent safety and standards body.

The structures were there to provide the assurance that safety management activity was effective and compliant. However, what is ultimately sought is a strengthened safety culture and this requires an environment whereby there is effective challenge, internal to the industry.

So, was the approach that was put in place a success? Figures show that safety performance of the UK rail network has significantly improved.  It is now more than 11 years since a fatality occurred, and the UK has the safest major rail network in Europe. It is false to allege that UK railway companies don’t manage safety effectively. The culture in the industry is an open one that supports learning and, in my experience, is not at all one of complacency.

This culture of continuous improvement has led to significant drives to push harder, not just on safety, but on health too, with the publication of the cross-industry health and safety strategy.

So, what does this mean for the tram sector in Britain, following the Sandilands accident?

Initial work has been undertaken to deal with local issues and, with the support of the sector body UK Tram, there have been efforts to share and learn more widely across the sector. 

I would suggest that the sector needs to work closely with its regulator, funders and others to be clear about their shared approach to risk management. It needs to be effective and stand up to scrutiny. In particular, clear criteria for safety investment are needed.

To support this, good quality risk information is also vital. Initially, a pragmatic approach can be undertaken using the judgement of experts, but the sector needs a long-term strategy for collecting data that can be pooled to support systemic learning and make the case for investment.

Again, the experience of the overground sector suggests the need for the tram sector to continuously seek to improve collaboration on safety issues, and share information and learning. This approach can be uncomfortable but it is a key part of work to drive improvements in safety culture.

George Bearfield is director of System Safety and Health at the Rail Safety and Standards Board and Visiting Professor of Railway System Safety at the University of Huddersfield

This article is taken from a lecture at Huddersfield university, 19 April 2018