Allison v London Underground Ltd

This appeal lies from a decision of His Honour Judge Cowell sitting in the Central London County Court on 25 January 2007. It involves consideration of the construction of Regulation 9 of the Provision and Use of Work Equipment Regulations 1998, (the 1998 Regulations) which provides:

“9(1) Every employer shall ensure that all persons who use work equipment have received adequate training for purposes of health and safety, including training in the methods which may be adopted when using the work equipment, any risks which such use may entail and precautions to be taken.”

The first issue in the appeal is whether or not the duty imposed by Regulation 9 is absolute, and does not require the proof of any fault on the part of the employer. Should the adequacy of the training given to an employee be judged by the result in the light of events or should it, as the judge held, be assessed in the light of what the employer knew about the risks at the time, even though, in the light of later events, it could be seen that the training had been inadequate? If neither if those tests is correct, what is the correct test?

The judge held that the duty was not absolute but required the employer to provide such training as was ‘adequate in all the circumstances’ - the circumstances including, as he put it, ‘notions of foreseeability’. The judge recognised that a difficult question of construction arose and gave permission to appeal.

Ms Latona Allison, the claimant/appellant, began work for London Underground Limited, the defendant/respondent, in 1996. Initially she worked as a guard but, by 1998, she had trained as a driver on the Northern Line. After only a few months, in November 1998, she developed a shoulder strain, which was found to be related to her task of handling the traction brake controller (TBC) which drives the train. After treatment, she returned to work in about July 1999 but was transferred to the Jubilee Line where the rolling stock was more modern; it was thought that the TBC would be more suitable for her. Following her return, she was monitored and, after reporting that she was having no problems, she was declared fit for full duties in September 2001. By that time, she had been trained to train other drivers and her work comprised a mixture of ordinary driving and training others. In early 2003, she developed tenosynovitis of the right hand and wrist. It is now accepted that this was due to strain from the prolonged use of the TBC, which the claimant held in a particular manner which I will later explain. She has not recovered fully from this condition and is now unfit for work as a driver.

The basic aspect of the driver’s duty is to operate the TBC. The design of the TBC on the Jubilee Line is slightly different in several respects from that on other lines. On the Jubilee Line, the TBC is operated by a handle, grasped by the driver’s right hand. The handle is made of metal; it is cylindrical, about 5 inches long and just over an inch in diameter. The way in which it is connected to the driving mechanism is not important for present purposes. The handle is situated immediately in front of the arm rest on the right side of the driver’s seat. The armrest is fixed but the seat is adjustable in a number of respects, rather like the driver’s seat in a modern car. When the train is stationary, the handle is in its rest position. In order to make the train move forwards, the driver turns the handle clockwise through 90 degrees in the horizontal plane. That process releases the brakes. The driver then pushes the handle forwards to produce acceleration. The whole mechanism is spring-loaded so that, if the driver reduces the pressure on the TBC, the train will decelerate and if pressure is released completely, the handle will automatically return to its rest position with the brakes applied. This is a safety device to ensure that, if the driver becomes incapacitated, the train is stopped. For that reason, the TBC is colloquially known as the ‘dead man’s handle’.

While the train is in motion between stations, the driver has to maintain pressure on the handle against the spring loading. The pressure is in two directions, pushing and twisting. A shift usually lasts about 8 hours but the time in motion with the driver’s hand on the TBC is variable, usually between 4 and 5 hours. When the train is at rest in a station, the driver has to operate the TBC at intervals, so as to inform the train’s computer that he or she is still awake and at the controls. It appears that some drivers used to prop the handle in position so as to avoid the need for this periodic operation.

It will be apparent that the work of the driver involves the potential for strain injury to the right upper limb. Such strain would come from the static application of pressure rather than a repetitive movement, which is a more common cause of tenosynovitis. However, it appears that this respondent was aware of the potential for upper limb strain as the result of static pressure.

In the early to mid-1990s, the respondent was developing rolling stock for the new Jubilee Line. Problems of strain injury had arisen in the past, particularly on the Central Line. When designing the driver’s seat, armrest and TBC for the new Jubilee Line rolling stock, the respondent consulted some experienced drivers and also experts including at least one ergonomist. The main focus of consultation was the position of the armrest and the adjustability of the driver’s seat, although there was also discussion about the operation of the TBC itself, in particular whether it should be capable of operation in both a clockwise and anticlockwise direction (as was the case on some lines) or whether it should only operate in the clockwise direction. It appears that the more detailed aspects of the design of the TBC handle were not considered.

There is one particular feature of the TBC as used on Jubilee Line trains which I must describe in detail, as it bears directly on the appellant’s injury. At one end of the handle (the end which is towards the driver when the handle is at rest and is on the left when the handle has been turned through 90 degrees), the handle is ‘chamfered’ or, as it was described by Judge Cowell, ‘bevelled’. This feature is clearly demonstrated in photograph 3 which is before the court but I must explain it in words. When the handle is viewed from the driver’s position, the left end slopes inwards at about 45 degrees. This design feature was introduced on Jubilee Line trains at the suggestion of the two experienced drivers who were consulted at the design stage. They suggested that a chamfered end would be more comfortable for them, when grasping the handle. The idea seems to have been accepted by management partly because it would make it more difficult for the driver to prop the handle in position. This modification of the design of the handle was not the subject of any expert advice, either at the design stage or later when it had been put into use.

The appellant is a very small woman, only 5 feet 1 inch in height and of slight build. Her arms are short and her hands small. When driving a Jubilee Line train, she found it natural and comfortable to rest her thumb against the chamfered end of the handle. Indeed, her evidence was that she thought that the end of the handle was chamfered for the purpose of providing a resting place for the thumb. Dr Elizabeth de Mello, the respondent’s expert ergonomist (whose evidence the judge preferred to that of Mr Ridd, the expert called by the appellant), was of the view that it gave that appearance. Most of the male drivers, whose hands were larger, apparently did not rest their thumbs against the chamfered end but tucked them underneath the handle. However, the evidence was that the position of the thumb was a matter for the driver’s choice. Some drivers would change the position of their thumbs from time to time and rest them against the chamfered end. No special instructions were given to the drivers as to how the thumb was to be positioned in relation to the chamfered end while the handle was under pressure, although drivers were trained to keep their wrists straight and to avoid dorsiflexion because it was recognised that applying pressure while holding the wrist in dorsiflexion could give rise to tenosynovitis.

The appellant developed tenosynovitis of the flexor pollicis longus tendon. This tendon flexes both joints of the thumb and it appears to have been common ground at the trial that the injury was caused by the position in which she held her thumb while applying pressure to the TBC. By January 2004, the respondent had realised this and it introduced a new element into the training of drivers, advising them of the need to keep their thumbs tucked under the handle and not to let them rest on the chamfered end.

The appellant’s claim rested partly on the alleged inadequacy of the risk assessment which the respondent had carried out on the use of the Jubilee Line TBC. Regulation 3 of the Management of Health and Safety at Work Regulations 1999 (the 1999 Regulations) provides:

“3(1) Every employer shall make a suitable and sufficient assessment of –

(a)

the risks to the health and safety of his employees to which they are exposed whilst they are at work; and

(b)

… (not applicable)

for the purpose of identifying the measures he needs to take to comply with the requirements and prohibitions imposed upon him by or under the relevant statutory provisions.”

The judge did not deal expressly with the suitability or sufficiency of the risk assessment which had been carried out in respect of the TBC. Although that failure is not a ground of appeal, I will have to return to the question of risk assessment in due course.

The appellant also alleged breaches of Regulations 4 and 9 of the 1998 regulations. I have already set out Regulation 9. Regulation 4 provides:

“4(1) Every employer shall ensure that work equipment is so constructed or adapted as to be suitable for the purpose for which it is used or provided.

(2)

In selecting work equipment, every employer shall have regard to the working conditions and to the risks to the health and safety of persons which exist in the premises or undertaking in which that work equipment is to be used and any additional risk posed by the use of that work equipment.

(3)

…. (not relevant for present purposes)

(4)

In this regulation, “suitable” means suitable in any respect which it is reasonably foreseeable will affect the health or safety of any person.”

It was common ground that the TBC was work equipment within the regulations. The judge held that there was no breach of Regulation 4. The expert evidence accepted by the judge was to the effect that, if an ergonomist had been consulted at the time when the TBC was being designed, the design of the TBC could have been significantly improved, so as to reduce the kind of strain which occurred in this case. However, the judge held that the TBC was suitable because it had not been reasonably foreseeable, at the design stage, that this kind of problem would arise. Also, until the appellant suffered her injury, there had been no cases of tenosynovitis arising from the use of the chamfered handle so there had been no reason to consider any change to the design.

There is no appeal against the holding under Regulation 4 and, save to say that I am doubtful about the correctness of the judge’s holding that it was not reasonably foreseeable at the design stage that the chamfered end would give rise to a risk of strain injury, I say no more about it. For the purposes of this appeal, Regulation 4 is only of interest because it provides a definition of ‘suitable’ which expressly incorporates a test of reasonable foreseeability of harm, whereas there is no definition of ‘adequate’ in Regulation 9 and no indication that the duty is governed by any element of foreseeability.

Turning to Regulation 9, the judge observed that it was difficult to construe the meaning of training which is ‘adequate for the purposes of health and safety’. He noted the absence of any qualifying words relating to foreseeability such as were present in Regulation 4. He briefly rehearsed the submission advanced by Miss Newbery, counsel for the claimant, that the duty to provide adequate training was strict or absolute and that the test did not depend upon reasonable foreseeability. Her submission was that, in the light of the development of the claimant’s tenosynovitis, it could now be seen that, if the claimant had been trained not to rest her thumb on the chamfered end of the handle, she would not have suffered injury. It mattered not whether the employer could not have reasonably foreseen the need to provide that particular element of training.

The judge was not attracted by that submission. At paragraph 57 of his judgment, he said:

“I am inclined to think that it is right that reasonable foresight is not the test, or at least not the sole test if it plays some part, but I am also inclined to think that absolute liability in the sense that if some different training had been given which would have avoided the injury, however, impossible it may have been to foresee the need for it, may not sit well with the fact that the adjective “adequate” is used. I suppose it must mean adequate in all the circumstances, and it is at this stage that the first instance judge asks, perhaps in some desperation, what that means.”

The judge then embarked on a consideration of ‘all the circumstances’. Miss Newbery’s submission had been that, if her primary submission were rejected, the test for adequacy of training should be based on what the employer knew or ought to have known was required. She had argued that the employer ought to have consulted an ergonomist to advise on the training to be given to drivers on the Jubilee Line. That should have included training on the operation of the TBC. However, the judge’s view was that it would be impossible for an employer to devise training except by reference to the actual risks foreseen and those recognised from past experience. By implication although not expressly, he rejected Miss Newbery’s submission that the test of adequacy should be by reference to what the employer ought to have appreciated or foreseen as to the risks entailed in the use of the TBC. He did not consider whether the risk assessment (by which the risks might have been recognised) had been sufficient and suitable. Instead, the judge focussed on the need for training to give special consideration to the needs of different ‘shapes and sizes’ of drivers. Within paragraph 58, he opined that

“… a substantial burden must rest on the employer to train so that in certain respects particular training is given which is appropriate for the particular categories of somewhat unusual users. It really comes to this that, if the absolute liability duty mentioned by Miss Newbery does apply then because the training did not include training on where to put the thumb, then the defendant is in breach of Regulation 9. If the test were reasonable foresight, then the defendant would not be in breach. None of the trainers themselves thought the position of the thumb to be anything more than a matter of choice for each driver. ”

Pausing there, I cannot myself understand the relevance of what the trainers thought about the position of the thumb. The trainers were employees such as the appellant; drivers who had been trained to train others. Unless they had been instructed to advise trainees about where to put their thumbs, there was no reason why they would be any better informed than any other driver. At paragraph 59, the judge continued:

“If the test is the adequacy of the training in all the circumstances, it seems to me that much must depend upon the employer’s past experience of problems and the information the employer receives about the particular employee to be trained and the problems to be expected to confront employees of a particular category, in this case an employee of five feet one inch with a particularly small arm and a narrow hand.”

The judge then embarked on a lengthy consideration of the appellant’s previous problems in connection with her shoulder strain while driving on the Northern Line (apparently caused by the spring-loaded mechanism of the TBC being ‘too strong’ for her), her transfer to the Jubilee Line and her subsequent supervision by the Occupational Health Department. By 2002, the appellant had told the employer that driving Jubilee Line trains was not a problem. At paragraph 69 of the judgment, the judge continued:

“That, it seems to me, is an important circumstance indicating that the absence of training about where to put the thumb or the end of the thumb does not make the training inadequate in relation to the claimant. That important circumstance together with the absence of reasonable foresight of the particular injury leads me to conclude, balancing the opposing arguments, that the training was adequate in all the circumstances.”

In short, the judge held that the training had been adequate because it had been adequate to deal with the risks which the employer had actually foreseen.

The judge then turned to consider the risk assessment which the employer had carried out in March 2000. As I have said, he did not make any finding as to whether the risk assessment had been sufficient and suitable so as to comply with Regulation 3 of the 1999 Regulations. He observed that it had identified the need for training and refresher training in the operation of the TBC. He continued, towards the end of paragraph 70:

“The point it seems to me is that training was given on the use of the handle. It was given to the claimant and it was given to her by others – how to adjust the seat so that the armrest was in the right place and so that the bending of the wrist would be avoided. In short, it seems to me that the risk assessment did identify the measures to be taken, namely initial and refresher training, but that simply leads on to the question which has arisen in this case and which I have attempted to answer: was the training adequate within the meaning of Regulation 9 of the other Regulations?”

I confess that I have not found this passage entirely easy to follow. It seems to me that the judge was saying that it was not necessary to decide on the sufficiency and suitability of the risk assessment because, if training was given in accordance with the results of the risk assessment, the sufficiency of the risk assessment equated to the adequacy of the training. That question he had already decided. Maybe that was intended to be a holding that the risk assessment was sufficient and suitable. However, the judge had not applied his mind directly to the issue of whether or not the risk assessment was ‘sufficient and suitable’; in particular he had not dealt with Miss Newbery’s submission that the employer should have sought expert advice about the risks related to the use of the handle with the chamfered end, either at the design stage or in March 2000, when the formal risk assessment was carried out, presumably in response to the 1999 Regulations.

That passage at paragraph 70 disposed of primary liability, against the claimant. For the sake of completeness, the judge dealt with the defendant’s allegation of contributory negligence. There was evidence that the claimant had experienced some minor aches and pains in her wrist some weeks before her tenosynovitis manifested itself properly. It was said that she should have reported these as and when they occurred. The judge rejected that allegation and held that, if primary liability had been established, there would have been no reduction for contributory negligence. There was no effective cross-appeal on that issue.

The appeal is based only on the judge’s finding under Regulation 9. Mr John Foy QC for the appellant advanced two arguments in the alternative.

First, he submitted that the duty under Regulation 9 of the 1998 Regulations is strict or absolute and not dependent on the proof of any fault. The duty is to ensure the provision of adequate training. The word ‘ensure’ makes it plain that the duty is strict and is not subject to any defence of reasonable practicability. In the context of this regulation, ‘adequate’ means ‘sufficient’; that is sufficient to produce the desired effect or result. The judge appeared to think that ‘adequate’ carried a connotation of being ‘less than ideal’ or ‘barely sufficient’ or ‘just about enough’ as it often does in common parlance. In that respect, he was wrong. Further, adequacy is not qualified by any concept of foreseeability. The question for the judge should have been whether the training was in fact adequate, not whether it was adequate in the light of what the employer had (or even ought to have) appreciated beforehand. The judge had applied the wrong test. If the correct test had been applied, the appellant would have succeeded. She had not been trained not to put her thumb on the chamfered end of the handle. She suffered an injury as the result of putting her thumb there. If she had been trained not to, she would not have suffered the injury. Therefore her training had been inadequate for health and safety purposes.

Mr Christopher Purchas QC, for the respondent, submitted that Regulation 9 did not impose strict or absolute liability on the employer - in the sense that it was to be judged by reference to the result and there was liability for acts or omissions to which no fault could attach. He submitted that, although it exists, such liability is a most unusual feature in English jurisprudence. That is as it should be, because criminal liability attaches to all breaches of the regulations made under the Health and Safety at Work Act 1974, as these regulations were. So unusual is absolute liability that clear words must be used to impose it. Regulation 9 does not clearly impose absolute liability. The regulation must be read in the light of the provisions of Council Directive 89/391/EEC (the Framework Directive) and Council Directive 89/655/EEC (the Work Equipment Directive). The 1998 Regulations with which we are concerned had replaced similar regulations issued in 1992 which had been enacted to implement these two Directives. Both the Framework Directive and the Work Equipment Directive contained articles governing the provision of training for employees. Before 1992, there had been no statutory provisions in the UK governing training requirements. Thus, Regulation 9 must have been intended to comply with the Directives. The Directives did not impose any requirement on Member States to impose absolute liability on employers and there was no warrant for the notion that the UK Parliament might have intended to impose a more onerous duty on employers than was required by the Directives.

Mr Foy cited to us the case of Dugmore v Swansea NHS Trust and another [2002] EWCA Civ 1689, a case involving the Control of Substances Hazardous to Health Regulations 1998 (the COSHH regulations), where, he submitted, the Court of Appeal had held that Regulation 7(1) imposed an absolute, no-fault, liability on an employer, where the duty required the employer to ensure that the employee’s exposure to hazardous substances was ‘adequately controlled’. He submitted that this authority supported his proposition that ‘adequacy’ was to be judged by the result and not by any notion of foreseeability. Mr Purchas submitted that the COSHH regulations were very different from the 1998 Regulations under consideration here and that no analogy could safely be drawn. I will consider this case in greater detail below.

Mr Foy also drew to our attention the case of Stark v Post Office [2000] ICR 1013, a case under Regulation 6(1) of the Provision and Use of Work Equipment Regulations 1992. That required an employer to ‘ensure that work equipment was maintained in an efficient state, in efficient working order and in good repair’. The claimant had been injured due to a latent defect in the bicycle provided by his employers. The Court of Appeal held that the duty to ensure that the bicycle was in an efficient state was an absolute one, in which the claimant had only to prove that he had been injured as the result of the defect and did not have to prove that the employer could or should have discovered and remedied the defect. Equally, it was not open to the employer to escape liability by showing that, even by the exercise of all due care, he could not have discovered the defect. This was an example of ‘no-fault’ liability. As I have said, the 1992 regulations were the predecessors to those of 1998. Regulation 6(1) of 1992 is replicated in Regulation 5(1) of the 1998 Regulations. Mr Foy submitted that the provision under consideration in Stark was closely related to Regulation 9. Parliament was plainly willing to impose no-fault liability within these regulations. If it was willing to do so in what is now Regulation 5, why not in Regulation 9? However, it does not seem to me that Stark is of any assistance. True, it is an example of no-fault liability. But merely because one provision within a set of regulations imposes no-fault liability does not mean that another provision will also do so unless there is a real similarity between the words of the two provisions. Here the only similarity is that the words ‘shall ensure’ appear. That, submits Mr Foy, connotes a mandatory requirement. I agree but those words do not help to establish the extent of the mandatory requirement. In Stark the duty was to ensure that work equipment was maintained in an efficient state etc; that is an absolute duty. But in Regulation 9, the duty is to ensure that adequate training is provided. The question is the meaning of adequate and Stark does not help on that. In any event, as between employer and employee, there had been longstanding no-fault liability on employers for the provision of defective equipment and it is entirely understandable that Regulation 5 of 1998 should impose no-fault liability on the employer.

Mr Foy’s alternative submission was that, if the judge was right to hold that liability under Regulation 9 was not absolute and that the test of adequacy of training was what was ‘adequate in all the circumstances’, he had nonetheless erred in his application of that test. Although the judge had said that the test of adequacy should not be dependent upon reasonable foreseeability, that was the test he had in fact applied. He had not dealt with Miss Newbery’s submission that the employer ought to have consulted an expert ergonomist when thinking of modifying the handle in accordance with the suggestion of the two drivers involved in the consultation. Dr de Mello had said that if the employer had sought advice from an ergonomist, it would have been given advice about the design of the handle and its use. Mr Foy submitted that, had the judge considered the point, he would have been bound to conclude that the respondent should have sought professional advice. It had recognised the need to consult an ergonomist about some aspects of the design (such as the position of the arm rest). It had done that because it recognised the risk of strain injuries from the prolonged use of a TBC. It should also have consulted about the design and use of the handle itself. If it had done so, it would have been advised of the risk of thumb strain if the thumb were placed on the chamfered end for long periods. The need for training on that risk would have been recognised. The failure to consult an ergonomist had resulted in the modification being made and no training being provided as to the manner in which the handle was to be held. Liability must follow.

Mr Purchas submitted that the judge had applied the right test and, on applying the test to the facts as found, had reached the correct conclusion.

Before considering the rival submissions, I want to clarify the terminology I shall use. In the course of argument, counsel used the expressions ‘strict liability’, ‘absolute liability’ and ‘no-fault liability’ interchangeably. I do not think that they all mean the same thing. In my experience, the expression ‘strict liability’ is used where liability cannot be excused on the ground that it is not practicable or reasonably practicable to avoid the risk. That is not the same as ‘absolute’ or ‘no-fault’ liability, which is reserved for the much smaller class of obligations which impose on the employer liability for something which he could not have avoided even by the exercise of all possible care. The distinction is well illustrated by the long established law in relation to section 14 of the Factories Act 1961, which imposed a duty to fence dangerous parts of a machine. That section was said to impose strict liability. However, it was not absolute in the sense that liability automatically followed if a worker was injured by a part of a machine which, by reason of the fact that an injury had occurred, could now be seen to have been dangerous. Liability only arose if the part was dangerous and whether it was dangerous depended on whether it was reasonably foreseeable that a person might be injured by it. What was reasonably foreseeable was not limited to what the employer actually foresaw; still less was it limited to what he had learned from past experience. At the trial of an action, the question of whether a part of a machine was dangerous was almost always a matter for expert evidence. If the judge was satisfied by expert evidence that the part of the machine was dangerous, then the employer ought to have known of it and it did not avail him to show that he had not realised that it was dangerous and had not thought it necessary (before the accident) to consult an expert on the subject. Once the claimant had shown that the part in question was dangerous, liability was strict in that it was no defence for the employer to show that it was not practicable for him to fence it. The duty on the employer was heavier than the duty at common law but it was not ‘no-fault’ liability. The employer would be liable where his fault had been a failure to appreciate the danger. Although the language of risk assessment was not used in those days, the employer’s fault might nowadays be said to have been a failure to assess the risks adequately, to have appreciated that the part was dangerous and to have taken appropriate action to fence the dangerous part. I would describe that kind of liability as ‘strict’ and reserve the terms ‘absolute’ or ‘no-fault’ for liability which attaches where the employer could not have avoided the risk even though he took all possible care.

In the present case, the appellant’s contention that Regulation 9 imposes ‘no-fault’ liability is a bold one, as Mr Foy recognised. I accept Mr Purchas’s submission that such no-fault liability is rare in English law. It exists but requires clear words of imposition. I also accept Mr Purchas’s submission that the 1998 Regulations must be read against the background of the two Directives. Indeed, I did not understand Mr Foy to disagree with that.

Both Directives require Member States to impose duties on employers to provide adequate training. Article 12 of the Framework Directive requires employers to ‘ensure that each worker receives adequate safety and health training’ and article 7 of the Work Equipment Directive requires that ‘the employer shall take measures necessary to ensure that workers given the task of using work equipment receive adequate training, including training on any risks which such use may entail.’ Thus Regulation 9 does no more than to echo the words of the relevant articles of the Directives, both of which use the expression ‘adequate training’. No help is given there as to the meaning of that expression.

However, the preambles to both Directives and the tenor of some of their other provisions give some clues to what the intention of the UK Parliament must have been in enacting the 1998 Regulations pursuant to its duty to implement the Directives. It must be admitted that the clues do not always point in the same direction. It seems to me that the following are of assistance.

The preamble to the Framework Directive states that the Treaty (of European Union) provides that ‘the Council shall adopt, by means of Directives, minimum requirements for encouraging improvements, especially in the working environment, to guarantee a better level of protection of the safety and health of workers.’ The preamble expressly recognises that standards of worker protection vary between Member States. It makes plain that the Directive does not justify any reduction in the levels of protection already achieved in individual Member States. However, the preamble also says that the improvement of workers’ health and safety should not be subordinated to ‘purely economic considerations’. It states that employers shall be obliged to keep themselves informed of the ‘latest advances in technology and scientific findings concerning workplace design’….. so as to be able ‘to guarantee a better level of protection of workers’ health and safety’. The general tenor of the preamble is that the Directive seeks to achieve improvement, in line with recent advances in technology but not to achieve perfection. A requirement to encourage improvements does not in any way suggest that Member States will be expected to impose no-fault liability.

Section II of the Framework Directive, which deals with employers’ obligations, begins with what appears to be a pointer in the opposite direction. Paragraph 1 of article 5 states that ‘the employer shall have a duty to ensure the safety and health of workers in every aspect related to the work.’ Taken literally, the use of the word ‘ensure’ sounds like the imposition of an absolute duty to guarantee and achieve health and safety for all workers. Paragraph 4 of article 5 provides:

“This Directive shall not restrict the option of Member States to provide for the exclusion or the limitation of employers’ responsibility where occurrences are due to unusual and unforeseeable circumstances, beyond the employers control, or to exceptional events, the consequences of which could not have been avoided despite the exercise of all due care.”

It seems to me that the thrust of article 5 is that Member States will be expected to require employers to ensure (or guarantee) their workers’ health and safety, save that provision may be made to exclude or limit liability in circumstances of no fault. That suggests that the expected norm will be a guarantee of health and safety and that special words will be required to limit the employers’ responsibility to anything less.

Article 6, paragraph 1 requires employers to take the measures necessary for the safety and health protection of workers, including prevention of occupational risks and provision of information and training, as well as provision of the necessary organisation and means. The apparently absolute nature of this obligation is counteracted or at least mitigated by the detailed provisions of paragraph 2, which requires employers to implement the paragraph 1 requirement in accordance with various ‘general principles’. These include ‘avoiding risks’, ‘evaluating the risks which cannot be avoided’, ‘adapting to technical progress’, ‘replacing the dangerous by the non-dangerous or the less dangerous’ and ‘giving appropriate instructions to the workers’. It is clear from these so-called principles that it is not intended that the article should require Member States to impose on employers the responsibility of guaranteeing the health and safety of employees.

Paragraph 3 of article 6 contains the provision requiring employers to conduct a risk assessment - or evaluation as it is there called. Also, subsequent to this evaluation, the necessary preventive measures implemented by the employer must ‘assure an improvement in the level of protection afforded to workers with regard to safety and health’. Here the emphasis is on improvement rather than ensuring health and safety; improvement rather than perfection.

I have already mentioned article 12 of the Framework Directive which requires that the employer shall ensure that each worker receives adequate safety and health training in particular in the form of information and instructions specific to his workstation or job. I note that the word there used is ‘adequate’, whereas the general principles set out under article 6 refer to giving ‘appropriate’ instructions to workers.

Taking the Framework Directive as a whole, it seems to me that there are some indications that Member States must require employers to ensure their workers’ health and safety, subject to the option of allowing them to exclude or limit responsibility for matters which could not be foreseen or avoided with all due care. On the other hand, the general tenor of the provision seems to be to achieve improvement rather than a guarantee of health and safety.

Turning now to the Work Equipment Directive, the preamble repeats some parts of the preamble to the Framework Directive. The general tenor is the same; that the objective is to achieve improvement in the protection given to the health and safety of workers.

Article 1 (a general provision) states that the Directive seeks to lay down minimum health and safety requirements. Article 3, which begins the section on employers’ obligations, requires employers to take the measures necessary to ensure that the work equipment made available to workers is suitable for the work to be carried out and may be used by workers without impairment to their safety or health. However, where it is not possible fully so to ensure that work equipment can be used by workers without risk to their safety or health, the employer shall take appropriate measures to minimise the risks. This appears to be the provision underlying Regulation 4 of the 1998 regulations. That regulation imposes a duty to provide suitable work equipment but suitability is defined by reference to what is reasonably foreseeable. I observe in passing that it does not appear to me that regulation 4 is fully compliant with article 3 of the Work Equipment Directive. However, I say no more about that because the point does not arise for decision in this appeal.

Article 5a, inserted by Directive 1995/63, is headed ‘Ergonomics and Occupational Health’. It provides:

“The working posture and position of workers while using work equipment and ergonomic principles must be taken fully into account by the employer when applying minimum health and safety requirements.”

I have already mentioned Article 7, which requires Member States to require employers ‘to ensure that workers given the task of using work equipment receive adequate training, including training on any risks which such use may entail’. But no clue is given as to the meaning of ‘adequate training’.

It is plainly impossible to construe these Directives in the way in which we construe UK statutes; they were not intended for that purpose. However, in so far as they throw light on the purposes behind UK legislation, it seems to me that the overall intention was to require Member States to achieve minimum health and safety requirements and to seek improvement in the levels of protection afforded to workers. Where the Directives use words such as ‘shall ensure’, I do not think that it can have been intended that Member States were to be obliged to impose absolute liability on employers, regardless of fault. Of course, that does not mean that it was not open to the UK to impose absolute liability. The question remains as to whether Regulation 9 does so. I have not found this consideration of the Directives of much assistance.

I turn now to consider the case of Dugmore, which Mr Foy submitted is of assistance in construing the meaning of the word adequate. In that case, the claimant became allergic to latex at some time between 1993 and 1995 through using latex gloves during her employment as a nurse at the first defendant’s hospital. Before 1993, international medical literature suggested that there was a risk that the use of latex could result in an allergy. However, until 1996, there was no evidence that the use of latex gloves was giving rise to a problem in England and no guidance had been given about such a risk. In June 1996, the claimant suffered a serious reaction to latex gloves and thereafter the first defendant supplied her with vinyl gloves. In January 1997, the claimant went to work for the second defendant, which was made aware of her allergy. In December 1997, as a result of her extreme sensitivity, she suffered an anaphylactic shock when picking up an empty box which had contained latex gloves. She was unable to return to work.

She sued both defendants for common law negligence and a breach of regulation 7(1) of COSHH, which requires an employer to ensure that the exposure of his employees to a substance hazardous to health was either prevented or, where that was not reasonably practicable, adequately controlled.

The judge dismissed the claim against both defendants on both grounds. As against the first defendant, he held that the date by which the first defendant ought reasonably to have known of the risk of latex allergy was January 1997. It did not and could not reasonably have been expected to know of the risk at the time when the claimant developed her allergy. That holding was made primarily in the context of an allegation of common law negligence. However, the judge also held that the employer’s lack of knowledge was fatal to the claim under the COSHH regulations. He considered that it was not reasonably practicable for the employer to avoid all exposure to latex and that the provision of vinyl gloves to the claimant had amounted to adequate control. However, the vinyl gloves had not been provided until after the damage had been done; the judge plainly considered that, until 1997, there had been no duty of adequate control because the need for such control could not have been reasonably foreseen. The Court of Appeal reversed the judge’s decision in respect of the regulations and held the first defendant liable.

Giving the judgment of the court, Hale LJ, as she then was, considered the extent of the employer’s duty under the COSHH regulations and the extent to which the employer’s knowledge of the risk in question was relevant to that duty. The claimant/appellant’s submission was that the duty ‘to ensure’ that exposure was either prevented or controlled imposed strict liability. The duty was not limited by any reference to the reasonable foreseeability of risk; nor was it in any way related to the outcome of a risk assessment. After a lengthy consideration of authority, the Court accepted that submission in principle. The court then considered the defence of reasonable practicability which applied only to the elimination of all exposure and not to the duty of adequate control and appears to have held (although it is not entirely clear) that the judge had been wrong to hold that it had not been reasonably practicable to avoid exposure to latex because the employer could have ‘gone out and discovered’ the risks arising from latex and could have then provided vinyl gloves.

In any event, the Court then went on to consider whether the claimant’s exposure to latex had been adequately controlled. I am not sure whether this part of the judgment is the ratio of the decision or whether it is obiter. In the COSHH regulations, the term ‘adequate’ is defined or at any rate partially so as follows: ‘In this regulation “adequate” means adequate having regard only to the nature of the substance and the nature and degree of exposure to substances hazardous to health and “adequately” shall be construed accordingly’. At paragraph 25, Hale LJ said:

“Here the duty is to ensure that exposure is adequately controlled. ‘Adequately’ is defined …. without any reference to reasonableness or the forseeability of risk: it is a purely practical matter depending upon the nature of the substance and the nature and degree of the exposure and nothing else. It cannot be adequate control to oblige an employee frequently to wear powdered latex gloves when other barriers are available. ”

Hale LJ then considered the relevant European Directives and at paragraph 27 she said:

“This all reinforces the view … that the purpose of the regulations is protective and preventive: they do not rely simply on criminal sanctions or civil liability after the event to induce good practice. They involve positive obligations to seek out the risks and take precautions against them. It is by no means incompatible with their purpose that an employer who fails to discover a risk or rates it so low that he takes no precautions against it should nevertheless be liable to the employee who suffers as a result.”

Of what assistance is this authority? I do not think it can be said to be binding on this Court which is considering a different regulation. Mr Purchas submitted that, because ‘adequate’ is defined in the COSHH regulations and not in Regulation 9 of the 1998 Regulations, for that reason alone, the case could not be of any assistance to this Court. However, I do not agree on that point. The so-called definition in the COSHH regulations does not really define the meaning of the word; it only restricts the factors which are to be taken into account when deciding whether the control is adequate. The court still has to decide whether the control is adequate and that involves giving some meaning to the word.

Mr Foy submitted that, in Dugmore, the Court held that liability under regulation 7 was absolute; the exposure was not adequately controlled because the control did not achieve the desired result. However, I do not think that the Court went so far. It did not hold that exposure was not adequately controlled merely because the claimant developed the allergy. It held that control was not adequate because it would have been quite possible, well before the claimant developed her allergy, for the employer to have discovered the risks of exposure to latex and to have provided vinyl gloves. Liability was not absolute in the sense that Mr Foy contends for in the present case, namely that it was imposed even though the employer could not have discovered the risk by the exercise of all due care and diligence. Thus I do not think that Dugmore provides the support for which Mr Foy contends. The case is, however, of relevance to the issue in the present appeal in that the Court drew a distinction between the common law duty (for the employer to take reasonable care to avoid reasonably foreseeable risks) and the more onerous duty imposed by the regulations, which, the Court held, required the employer to go out and discover the risks and to take the appropriate steps. In other words, the duty was strict in the sense that I said, in paragraph 30, that I would use the term.

I return to Regulation 9 and the attempt to construe the respondent’s duty ‘to ensure’ that the appellant was given ‘adequate training for the purposes of health and safety’. I accept that the words ‘to ensure’ imply a mandatory duty. However, in the context of this regulation, I do not think that they mean anything more than that the duty to provide training is mandatory. The employer cannot say that it was too expensive or too time-consuming or not reasonably practicable to provide training. He must provide training. Also, the training must be adequate. I accept Mr Foy’s submission that, if the judge thought that the use of the word adequate was intended to imply something ‘barely sufficient’ or ‘just about enough’ that would be wrong. But I do not think that the mere fact that the duty to train is mandatory raises the meaning of ‘adequate’ to the high level which Mr Foy contends for. Thus, I conclude that the judge was right to reject Miss Newbery’s primary submission and I would reject Mr Foy’s submission that the regulation imposes ‘no-fault’ liability.

What then is the extent of the liability? The judge said that adequate training meant ‘adequate in all the circumstances’ and imported some element of foreseeability into the test. The problem with that formulation is that, as the judge himself recognised, it is not at all clear. There must be a clearer way of expressing the duty to provide adequate training of the purposes of health and safety. Also, in my view the test formulated is not the correct test. In my judgment, the test for the adequacy of training for the purposes of health and safety is what training was needed in the light of what the employer ought to have known about the risks arising from the activities of his business. To say that the training is adequate if it deals with the risks which the employer knows about is to impose no greater a duty than exists at common law. In my view the statutory duty is higher and imposes on the employer a duty to investigate the risks inherent in his operations, taking professional advice where necessary. That was the thrust of Ms Newbery’s submission as I set it out in paragraphs 17 and 21 above. I think that her submission was correct. This formulation of the test is consistent with the test applied by this Court in Dugmore.

It follows from my conclusion that the judge applied the wrong test that I will not have to consider in detail the way in which the judge applied his test to the facts. Suffice it to say that I consider that there was much force in Mr Foy’s submission that, although the Judge warned himself that the test should not be the same as common law negligence, he did in fact apply the common law test.

How is the court to approach the question of what the employer ought to have known about the risks inherent in his own operations? In my view, what he ought to have known is (or should be) closely linked with the risk assessment which he is obliged to carry out under Regulation 3 of the 1999 Regulations. That requires the employer to carry out a suitable and sufficient risk assessment for the purposes of identifying the measures he needs to take to comply with the requirements and prohibitions imposed upon him by or under the relevant statutory provisions. What the employer ought to have known will be what he would have known if he had carried out a suitable and sufficient risk assessment. Plainly, a suitable and sufficient risk assessment will identify those risks in respect of which the employee needs training. Such a risk assessment will provide the basis not only for the training which the employer must give but also for other aspects of his duty, such as, for example, whether the place of work is safe or whether work equipment is suitable.

Judge Cowell recognised that there was a connection between risk assessment and adequacy of training but thought that, once he had decided that the training had been ‘adequate in all the circumstances’ he did not need to decide whether the risk assessment had been ‘sufficient and suitable’. With respect to the judge, I think he put the cart before the horse. Risk assessments are meant to be an exercise by which the employer examines and evaluates all the risks entailed in his operations and takes steps to remove or minimise those risks. They should be a blueprint for action. I do not think that Judge Cowell was alone in underestimating the importance of risk assessment. It seems to me that insufficient judicial attention has been given to risk assessments in the years since the duty to conduct them was first introduced. I think this is because judges recognise that a failure to carry out a sufficient and suitable risk assessment is never the direct cause of an injury. The inadequacy of a risk assessment can only ever be an indirect cause. Understandably judicial decisions have tended to focus on the breach of duty which has lead directly to the injury.

In the present case the judge failed to decide whether the risk assessment had been sufficient and suitable. It appeared to me that the right approach for this court to take, in deciding whether the appellant’s training had been adequate for health and safety purposes, was to examine whether the respondent’s risk assessment had been sufficient and suitable. That, as it seemed to me, would provide the answer. In post hearing submissions, Mr Purchas argued that this would not be right as the judge’s failure to decide on the sufficiency and suitability of the risk assessment had not been a ground of appeal. The appeal was limited to the correctness of his decision under Regulation 9. I can see the force of that objection. However, it is quite possible to decide the issue of what the employer ought to have known about the risks without reference to the respondent’s risk assessment. I just think that the most logical way to approach the question is through a consideration of the suitability and sufficiency of the risk assessment. Having said that, I will respect Mr Purchas’s objection and will approach the question of what the employer ought to have known about the risks direct from the evidence before the judge.

Before embarking on this analysis, I have considered whether the issue should be remitted to the County Court. I have concluded that the material before us is sufficient to enable this court to decide. Although we do not have transcripts of Dr de Mello’s evidence, we have her very full report. Dr de Mello gave a detailed account of the consultations and investigations which had taken place in the mid-1990s when the Jubilee Line TBC was designed. She also gave a full description of the risk assessment which the respondent carried out in 2000. In his judgment, the judge set out those aspects of her oral evidence which he regarded as important. Mr Purchas contended that, without the full transcript, we could not properly decide this issue. However, he was unable to tell us that there was any particular aspect of her evidence that we ought to be aware of but are not.

From Dr de Mello’s report, it appears that the design of the Jubilee Line TBC and the associated driver’s seat and armrest was extensively considered in the mid-1990s, with a view to eliminating or reducing the risk of strain and musculo-skeletal injuries. The consideration included seeking the advice of ergonomists. However, it is clear from the evidence that the particular feature of the TBC handle with which we are concerned was not the subject of such expert advice. As I have said, the chamfered end was introduced at the suggestion of two experienced drivers who thought that it would make the handle more comfortable for them. It appears from Dr de Mello’s report and examination of the documents disclosed by the respondent that at no stage was professional advice sought about the existence of risks arising from the prolonged use of the TBC with its chamfered end.

Ought the employer to have sought the advice of an ergonomist in respect of the risk of the TBC handle with the chamfered end? It seems to me that the proper assessment of the risks of developing a static strain upper limb disorder from the prolonged use of a TBC are likely to be beyond the capability of anyone other than an ergonomist. I note that Article 5a of the Work Equipment Directive requires that ergonomic principles should be taken into account when providing work equipment. If when providing equipment, why not when deciding upon the need for training? This respondent was aware of the risks of strain injury arising from the prolonged use of TBCs, as the documents show. In my view, the respondent must have been aware of the risk of strain injury from the use of the TBC and should have taken advice about it from a professional, namely an ergonomist.

Dr de Mello said that, if the respondent had employed an ergonomist ‘to look at the specifics of the handle design and operation, it would have received advice which would have led it to change both the handle design and operation (my emphasis). This would have reduced the risk of injury’. The judge commented on that statement as follows, at paragraph 55:

“I am reluctant to accept that the use of ergonomists when the TBC was designed would necessarily have resulted in their noticing that the bevel might be used by the end of the thumb, which would be a bad thing, as opposed to its being against some other part of the grip of the user’s hand”.

In so far as the judge there appears to be rejecting Dr de Mello’s statement about what an ergonomist would have noticed, I do not think he was entitled so to do. Of course, a judge is entitled to reject part of the evidence of a witness, including an expert witness, even though he accepts most of it. However, in this case the judge did not apply the correct test; nor did he give any reasons for rejecting this aspect of Dr de Mello’s evidence. In considering whether an ergonomist would have noticed the potential problem of the chamfered end of the TBC handle, the judge was making a finding of fact. Such findings should be made on the balance of probabilities. Here the judge appears to be saying that the ergonomist would not necessarily have noticed that problem; in other words, he was not sure that an ergonomist would have noticed it. He did not say why. It seems to me that this finding was perverse. The judge had accepted Dr De Mello’s evidence in general. Why he should think that she should be wrong in describing what an ergonomist would appreciate when examining this TBC, I cannot understand. There does not appear to be any basis for him to reject her evidence on this point. The evidence before him was that an ergonomist would have noticed the potential problem created by the chamfered end of the handle and would have given advice accordingly.

On the evidence before the judge, it was clear that the respondent had introduced the chamfered handle at the suggestion of two experienced drivers and had not applied its mind to the issue of whether that feature would give rise to any risk of strain injury. More importantly, in the context of this case, it had allowed its drivers to use the chamfered handle in any way they chose. No specific instruction was given as to the correct way to hold the handle so as to minimise the risk of injury. As I have said, Dr de Mello’s evidence was that, if professional advice had been taken from an ergonomist, the respondent would (incidentally) have been given advice about the design of the handle and (crucially for the purposes of this case) have been given advice about its use. If such advice had been taken, the risk of upper limb disorder would have been reduced. Dr de Mello was of the view that the drivers should have been given specific training about how to hold the handle. She said that, had she (or, I take her to mean, a similarly qualified ergonomist) been consulted, she would have described the risk of using the handle as ‘medium’ rather than low, precisely because the drivers had been given ‘no information or training on the specifics of handle holding and use’: see paragraph 48 of her report. Also at paragraph 64, Dr de Mello said that the drivers should have been given training on the TBC hand position from the outset. She had initially been under the mistaken impression that this training had been given but it is common ground that it had not.

My conclusion is that this respondent ought not to have put this new TBC (with the handle with the chamfered end) into service without taking advice from a suitably qualified expert, namely an ergonomist. Had it done so, it would have identified the need for the drivers to be trained in the way in which they held the TBC handle in order to minimise the risk of strain injury. The drivers would have been instructed to grasp the handle with the thumb tucked underneath. Because this advice was not taken, the risk arising from the design of the chamfered end was not recognised as it should have been and the training given to the appellant was not adequate for the purposes of health and safety in breach of Regulation 9 of the 1998 regulations. Had adequate training been given, there is no reason to suppose that the appellant would not have heeded it and, in my view, the only inference which can properly be drawn is that she would probably not have developed the strain injury which she has suffered.

For those reasons I would allow the appeal. The claim should be listed before a district judge for direction to be given as to the assessment of quantum unless agreement can be reached.

I agree with both judgments.

I agree that this appeal should be allowed for the reasons given by Smith LJ. I only add these few observations because I thought it might be helpful if I were to summarise the essential steps that led me to that conclusion.

Regulation 9(1) of the 1998 Regulations provides:

“Every employer shall ensure that persons who use work equipment have received adequate training for purposes of health and safety, including training in the methods which may be adopted when using the work equipment, any risk which such use may entail and precautions to be taken.”

The regulation makes it clear that the training which must be adequate must include training in the methods which may be adopted when using the work equipment. As applied to the facts of this case that means that the training must include training in the methods which may be adopted in using the handle which operates the TBC and that the training must include training in any risk which use of the handle may entail and in the precautions to be taken.

As Smith LJ has observed, how to hold the handle was left to the driver to choose. Some drivers would put their thumbs on the chamfered end, at least from time to time. The employer either knew that or should have known that. In these circumstances, as I see it, regulation 9 required the employer to provide adequate training in holding the handle in that way, including training in any risk that using the handle in that way might entail and training in the precautions to be taken in order to avoid any such risk.

The employer could not of course provide training in respect of a risk created by holding the handle with the thumb on the chamfered end unless it ascertained what, if any, risk was involved. It was therefore necessary to carry out an appropriate assessment of the risks involved. I agree with Smith LJ that in order to carry out such a risk assessment it would be necessary to instruct an appropriate expert, namely an ergonomist. I also agree with her that such a risk assessment would have revealed the risk and led to appropriate training which would have led to instructions to drivers not to hold the thumb over the end, with the result that the appellant would, on the balance or probabilities, have held the handle in a different way and have avoided the strain injury which she suffered.

In these circumstances the employer was in breach of regulation 9 because it failed to ensure that the appellant received training in the risks involved in using the handle by holding her thumb over the end and in consequence they failed to ensure that she received adequate training in how to hold the handle and what precautions should be taken to avoid the risk of strain injury, namely by not holding her thumb over the end.

It was submitted by Mr Purchas that it is not open to this court to reach that conclusion without a detailed analysis of the risk assessment in fact carried out and that the appropriate course was to remit the matter to the judge. It was further submitted that the question whether the employer should have consulted an ergonomist was not before the judge. Like Smith LJ, I would not accept those submissions. My reasons are these:

In short, I agree with Smith LJ that under regulation 9 the employer should have ascertained what risks there were in holding the handle in that way and, having done so as a result of taking the advice of an ergonomist, should have instructed drivers not to hold it in that way but differently.

For these reasons, which are I think essentially the same as those of Smith LJ, the employer was in my opinion in breach of regulation 9 in failing to ensure that the appellant received adequate training for the purpose of health and safety. It follows that I agree that the appeal should be allowed.