'The Toss of a Coin', Chapter 13 / 1

Managing safety after the Clapham Junction Accident

After approval by the BR Board of the Du Pont report and recommendations, one of the first moves was to gain its acceptance by Regional and Business management and the railway trade unions. I made a number of presentations to various BR management teams and attended a Board conference with the trade union senior officials of the NUR, ASLEF and the TSSA at the Grove, Watford. The trade union officials were broadly supportive of the proposals, although on the issue of staff safety, were inclined to take the attitude that ‘we’ve been telling you this for years’. Whilst there was some truth in this, in my view the trade unions were just as culpable for the issue of staff safety being contentious and subject to dispute and argument, because all too often safety was used as a lever – often without justification – to achieve other trade union objectives.

One of the problems was that staff safety representatives were part of the formal trade union/management consultation and negotiation machinery so that safety was discussed in the same forum and with the same people as other non safety issues. By eating some humble pie and admitting that management had not been as forthcoming on staff safety matters as it should have been in the past, we got the very useful concession that staff safety representatives should be independent of the trade union/management negotiating machinery and Local Departmental Committees and therefore have one objective only – the safety of the staff they represented. Management could then feel more confident that meeting staff requests for improved safety would not be the ‘thin end of the wedge’ for other non-safety related demands.

There were some hard words for management to accept in Bob Webber’s presentation which I found myself delivering to a number of audiences, but to be fair, most recognised the truth of what he was saying. We set in motion a number of the ‘STOP’ training sessions for managers who would then commit themselves to undertaking monthly inspections in their areas looking not only at safe and unsafe conditions – that was relatively easy – but also at the safety or otherwise of staff working methods, harder to spot and even harder for some managers to intervene after years of condoning less safe practices that got the job done.

As well as the Du Pont proposals which homed in particularly on staff safety, although this would create a safety aware culture that would benefit passenger and public safety, David Rayner and I were keen to use the technique of risk assessment to identify the risks to general railway safety. The report by the chairman of the Clapham Junction inquiry, Anthony Hidden QC, and the Fennell report on the Kings Cross fire, had, between them come up with over 200 recommendations relevant to the railway industry and BR operators and engineers were already holding more than 100 schemes of their own devising to address safety.

One of the experts I had received advice from during my investigations on safety management systems was Tony Taig, then with the Atomic Energy Authority, later to be advisor on railways to the Select Committee on Transport. He advocated undertaking comprehensive identification of risks from all BR’s activities to passengers, staff and the general public. With data from the annual Railway Inspectorate reports and the functional reporting of incidents many of which could be considered as precursors to accidents involving injury or fatality, there was no shortage of data to undertake such an exercise. We undertook this together in 1990 and with the help of a couple of consultants from Cooper Lybrand, Patrick McHugh and Andrew Wells, we were able to link the judicial report recommendations and other BR schemes with the risks that they were designed to reduce or eliminate. We then assessed the benefits likely from implementation of the recommendation or project in terms of reduced number of annual fatalities, serious injuries and minor injuries and linked these with the investment and running costs involved, thus producing a cost-benefit analysis for every project.

The prime benefit of this was to put some order of priority into the large number of potential actions most of which required additional resources and funding. The projects ranged from around £64,000 per life saved (we used a formula agreed with the Railway Inspectorate and the Department of Transport to convert major and minor injuries to ‘equivalent’ lives) to almost infinity, as some schemes – especially those stemming from the Fennell Inquiry on the London Underground Kings Cross fire – were addressing risks that had been eliminated or nearly so by other higher value projects. The government, in the wake of the public concern over the two disasters, had ring-fenced £200 million a year for a maximum of three years for BR to make a step change in safety improvement, but costs of all the schemes BR had identified to address the risks from the joint work with AEA and Cooper Lybrand amounted to more than £300 million a year. The Board therefore set up a Safety Panel chaired by Maurice Holmes, Director of Safety, with a senior member of one of the Businesses to scrutinise every proposal put forward, test the assumptions made, the realism of the benefits claimed and the accuracy of the costs.

We therefore finished up with a list of some 300+ schemes in some sort of priority order. The question now was where to draw the line. How much should BR spend to save a ‘theoretical’ life? Should we use the government money available to determine the boundary? Was there any objective work elsewhere on this? What cost criteria, for instance, did the Ministry of Transport use for investing in road safety highway schemes? I drew a graph of the schemes in ascending order of cost benefit and it was interesting to see that the vast majority of schemes fell in the area up to around £1-2 million per equivalent life or over £10 million. Only about five schemes fell in the £2-£10 million range. Most of the ‘best’ schemes seemed to be arising from either ‘soft’ schemes like training or improved supervision, or environmental schemes involving correcting unsafe ground conditions or removing lineside clutter and redundant materials.

Many of the major investment projects such as Automatic Train Protection came in the third quartile with a value of around £14million cost for every equivalent life potentially saved and some of the recommendations from the Kings Cross fire inquiry were over £100 million per life saved or even impossible to quantify because we could identify no add-on value at all. The most important recommendation to address risk of fire was the banning of smoking at underground and other ‘enclosed’ stations such as Birmingham New Street - frankly this halved the fire risks we had identified. By the time we reached recommendation number 118 from the Fennell report, there was no risk left to be controlled as earlier and higher priority schemes had eliminated the risk from everything we could think of! The problem was that these judicial inquiries would put forward any recommendation that addressed risk without any prioritisation or consideration that some projects addressed the same hazard.

The usual ‘pareto’ principle applied. 20% of the cost would eliminate 80% of the risk - well it was not quite as clear cut as that but the third of projects that cost over £10 million per life saved only addressed around 10% of the risk, much less if you left out ATP. The Ministry of Transport used a figure of around £200,000 per life saved as their criteria for highway safety engineering schemes. We undertook later some research on public attitude to transport risk and found that in certain circumstances travellers were willing to pay more - or expected the transport company to pay more. Road vehicle drivers assumed they had control of their own risks - a dubious assumption - and were prepared to accept more risk than when they placed their lives completely in the hands of another - plane or train. Whether the risk-taker benefits from the activity causing the risk is another element - clearly travellers derive benefit from their ‘hazardous’ journey whereas a farmer living next door to a nuclear plant does not appear to directly benefit and is therefore much more inclined to be risk averse. Some types of fatality are more dreaded than others - fire or drowning, for example - and a more contentious issue is whether catastrophic events are to be weighted more heavily that a ‘routine’ accident. Is killing 100 people in one incident worse than killing 100 people one at a time over a year? Logic says one thing, but the media plays a big part here and the 35 people killed in the Clapham accident apparently warranted far more attention than the 35 people killed in road accidents over a 3 day period in the same month. We developed much of our thinking in close consultation with Professor David Ball and his team at the University of East Anglia.

As a result we identified that the ‘value of life’ we would use to determine investment in safety would be much higher than that used for the same Ministry for avoiding road accidents. From comparisons in other countries and by extrapolating data from other industries, it would seem that
a realistic value for investment justification would be around £1 million per equivalent life saved. This linked well with our risk assessed cost benefit studies which showed a marked dropping off of projects valued at over £1.5 million, so we chose this figure as our guideline. In fact we said that any project costing less than £1 million per life would be implemented and anything between £1 and £2 million would be subject to more detailed consideration before a decision was taken. Anything over £2 million would not be in the budget unless an exceptional case was made for it.

There was one further guideline that we had to take into account - what is known as the ‘ALARP’ principle (‘As Low As Reasonably Practicable’). Guidelines issued by the Health & Safety Executive and used in some industries set individual risk rates, above which the risk was said to be intolerable. At the other extreme, some risks were so unlikely that the risk was purported to be ‘acceptable’. In between these risk levels it was the duty of management to reduce the risks ‘ALARP’.

In the railway industry the maximum individual risk for a staff member was deemed to be 1 chance per 10,000 of a fatality per year, and the ‘acceptable’ risk, 1 in 100,000. For passengers and members of the public the risk levels were an order of magnitude tougher - worse than 1 in 100,000 would be intolerable, 1 in a million (roughly your chance of being struck by lightning) ‘acceptable’. Nearly all BR’s risks fell in the ALARP region where we would use the cost formula to denote what was reasonable. Only with a couple of staff activities did we identify intolerable risk and here we either had to reduce the risk to ALARP whatever the cost, or get out of the activity. The two ‘intolerable’ activities were track installation and maintenance where moving trains were killing up to 15 of our staff annually in the 1980s - the risk was assessed as 1 in 1,650 or 1 fatality per 40 staff over a career length, the second most dangerous activity in the UK next to North Sea fishing. The other was train shunting where a much smaller group of staff were being killed crawling under moving wagons or being caught between buffers and crushed during coupling or uncoupling.

The Board deemed the first fatality rate completely unacceptable and two very senior managers (Graham Eccles and William Hill) were given a two year task of improving the safety of these staff to acceptable levels. They devised a threefold strategy - firstly by staff briefing and training, to make men working on the track much more safety conscious; secondly, to introduce stricter rules; and finally to invest in technical means of warning and protecting men on the track. It was of great interest and some surprise that the first element was so successful that within a couple of years no staff were killed on the track and this performance was sustained for 2-3 years until privatisation and a myriad number of contractors took over track work and we had to inculcate a safety culture all over again. With the shunting activity, the answer was easier. The activity had almost finished - parcels and wagonload freight traffic involving significant remarshalling of rail vehicles had gone. Passenger trains were now virtually all fixed formation trains, many ‘multiple units’ only capable of being parted or joined in factory/depot conditions.

The major contentious item was the one project that was very costly but did address a major risk - Automatic Train Protection (ATP) which would minimise the risk from trains passing signals at danger (SPADs). It was estimated that this would save around four lives a year (including the formula for injuries - serious = 0.1 fatality; minor = 0.005). Over a 30 year project lifespan at a cost of a £1 billion, this gave, as previously indicated, a value of £14 million for every life potentially saved. Even if we used a variety of assumptions to test this, we could get no better value than £9 million. As BR had given a commitment at the ‘Hidden’ Inquiry to introduce ATP, it was installed on the Western Region mainline only as a pilot (and even there failed to stop the Southall collision between an ATP fitted HST and a freight crossing the main line). BR was asked by the Ministry to test this project cost benefit at a seminar with academics and other industries and as BR’s methodology was considered robust, the proposal to fit ATP to all trains was replaced by a simpler system TPWS - Train Protection Warning System - although even that costs about £5-6 million per life saved, well outside our criteria. However, because of the prevalence of SPADs (900 a year of which 100 were potentially dangerous) it was agreed that this was an exceptional case and warranted that degree of expenditure.

It is interesting, however, to consider the different criteria used by different government departments and industry authorities to set safety standards. The money we suggested would be spent on ATP and save at most an average of 4 lives a year would, if spent on road safety save 400 lives, or if spent in the NHS on preventative treatment, probably 4,000 lives - or if spent by the Department for International Development on simple health care such as diarrhoea medicines, would save at least 45,000 children’s lives in Sub-Saharan Africa or the Indian sub-continent.