London Borough of Islington v University College London Hospital NHS Trust

Mrs J, as I will call her, was injured through the negligence of the defendants [UCH] and as a result of that disability was provided with residential care by the claimants [Islington], acting under their statutory duty to persons in Mrs J’s position. A local authority is under a duty to charge for such accommodation unless, as was the case with Mrs J, the accommodated person is unable to pay. She remained in such care from 18 June 1998 until September 2003, when she moved to live with, and be cared for by, her daughter. In November 2002 UCH reached a settlement with Mrs J for the payment to her of substantial damages in respect of its negligence; which damages necessarily did not include any compensation for Mrs J for care which she had received free of charge. It was Islington that bore the cost of Mrs J’s care, and Islington now seeks to recover that cost from UCH by an action in negligence. Douglas Brown J, hearing a preliminary point, held that as a matter of law Islington has no claim. The appeal from his ruling has raised a number of difficult issues, which it is necessary to expound in some detail.

The history of the case, and an account of the relevant legislation, was very accurately set out by the learned judge, to the extent that it is not necessary for me to do more than quote it verbatim:

“1.

This case concerns the care costs for a lady who it is agreed should be called Mrs J. She was born in April 1923 and in 1995 was diagnosed with rheumatic mitral valve disease and in January 1996 began taking Warfarin, an anticoagulant. The purpose of treating her with Warfarin was to reduce the risk of a cardiac embolus. She was due to be admitted to Middlesex Hospital on 13^th February 1998 for mitral valve replacement and she was told at the hospital to stop taking her Warfarin a week before the operation in readiness for that. The operation was cancelled and she was told of that on the 11^th February 1998 and told by the cardiac surgeon’s secretary not to recommence her Warfarin. Shortly after that she developed pains in her legs and enquired of the hospital by telephone as to whether that could be connected to the lack of Warfarin. She was told that it could not be and she should not start taking it again.

Although Islington has a statutory duty to recover the cost of care where it is possible to do so, it could only recover virtually nominal amounts (related to her payments of income support) from Mrs J. That was the case in any event before an agreement was reached between Mrs J and UCH as to the payment of damages to her. And it remained the case after that agreement. Islington could not count those damages as part of Mrs J’s capital or income because, under the rules set out in sub-paragraph (v) above, the damages were held on trust for her. That was for two reasons: first, she was a patient; and also, because the damages had been agreed by way of structured settlement rather than paid to Mrs J as a lump sum, they were to be taken to be held on trust and not immediately available to her. The effect of these rules is that, where the local authority is unable to charge the injured person, the financial liability of the tortfeasor is reduced; because the injured person cannot claim, as against him, loss that she has not in fact suffered. By contrast, if care is provided by a relative or other private person, the cost to them is recoverable by the injured person, seen as acting in the litigation on behalf of the carer: see Hunt v Severs [1994] 2 AC 350, a case to which I return below. And the inability of the injured person to make the tortfeasor liable in respect of local authority care continues even after the payment of damages to the injured person if the damages are held on trust. That results in a settlement between the tortfeasor and the injured person such as was reached in the present case, where the structure agreement provides for payment by the tortfeasor in respect of care costs, but only for the period after the injured person has left local authority care.

Islington contends that this outcome is plainly unfair. As between itself and UCH, the tortfeasor, and not the innocent local authority, should bear the cost of looking after Mrs J. And the outcome is particularly erratic when the detailed liability of the tortfeasor comes to be determined, because he can preclude the local authority from charging the injured person out of those damages by paying them by way of structured settlement rather than as a lump sum; and thus, because his liability is only to the injured person and in respect of that person’s actual expenditure, save himself the cost that he would incur in the case of a private carer of reimbursing the injured person’s care costs.

In this action, Islington proposes a pioneering method of righting the injustice, and of moving the costs of care of an impecunious injured person from the local authority to the tortfeasor. Islington seeks to use the law of negligence, to give it a direct claim against the tortfeasor. The action raises some novel questions, not only as being a claim for economic loss that falls outside the normal range of cases that arise from a direct relationship between claimant and defendant; but also more generally as taking the law of negligence into areas that it has not previously entered. It is therefore necessary first to remind ourselves of the broad terms of that law.

Any approach to a case of negligence has to be through the three-stage test of foreseeability of damage; proximity; and fairness justice and reasonableness; that was accepted by the House of Lords in Caparo v Dickman [1990] 2 AC 605; while in the same breath noting the many warnings in that case that the criteria are not definitive, but in most cases merely facets of the same thing (per Lord Oliver of Aylmerton at p 633C); or are not susceptible of any such precise definition as would be necessary to give them utility as practical tests (per Lord Bridge of Harwich at p 618B).

Cases of pure economic loss equally fall in principle within this rubric, as for instance Lord Woolf explained in Spring v Guardian Assurance [1995] 2 AC 296, but they give rise to special considerations. First, in order to fall within the law of negligence at all a claim based upon a misrepresentation must demonstrate the special relationship between the parties that was envisaged in Hedley Byrne v Heller [1964] AC 465. As Lord Goff of Chieveley pointed out in Henderson v Merrett Syndicates [1995] 2 AC at p 181D, a positive answer to that question determines without more the issues of proximity and fairness. But our case is not a representation case. Islington claims for the direct infliction of financial loss arising from UCH’s negligence towards Mrs J. That has been recognised since as long ago as Cattle v Stockton Waterworks (1875) LR 10 QB 453 as an area into which the law will not easily enter: see the treatment in Clerk & Lindsell on Torts (18^th edition), paras 7-126 to 7-128. This diffidence is in truth only a particular example of the inability of a party to build on a wrong to someone else: Lord Wright in Bourhill v Young [1943] AC at p 108. Accordingly, Islington must establish the existence of a duty directly owed to it on the part of UCH.

Islington can only do that by demonstrating that the relationship between UCH and itself satisfies all three of the Caparo criteria. The judge held that Islington failed comprehensively, because the case did not satisfy any one of the criteria: the financial loss suffered by Islington was not foreseeable by UCH; there was no relationship of proximity between Islington and UCH; and in any event it was not fair just or reasonable that UCH should compensate Islington for the cost of Mrs J’s care.

I shall follow the same pattern of addressing each of the Caparo criteria in turn in the light of the judge’s analysis of them.

The judge said this, at paragraphs 48-50 of his judgment:

“It is convenient here to consider the questions posed by Mr Underwood.

i)

If Mrs J stopped taking her Warfarin as a result of advice was there a reasonable possibility that she would suffer a stroke? The answer to that must be yes. Warfarin was intended to reduce the risk of an embolus which if it occurs is likely to lead to a stroke.

ii)

If she suffered a stroke is it foreseeable that she would require care? That is a possibility but strokes are varied and idiosyncratic in their effect. The stroke could result in a complete recovery with no care. There could be partial recovery with some care at home or there could be disability as in Mrs J’s case which called for full time care.

iii)

If she required care would that care have to be provided by the Local Authority? That was a possibility but care at home and care in a private nursing home were equal possibilities.

iv)

If the Local Authority provided care would she be able to pay for it? This seems to me to be an impossible question for the Hospital Trust to answer at the time when the wrong advice was given. As Mr Underwood accepted their knowledge was restricted to her age and her address and postcode. There was insufficient material for them to infer even on a provisional basis that she might be without means.

The judge may not have been assisted by the way in which Islington put its case, which he criticises in his paragraph 50 set out above. However, that said, I cannot agree with his conclusions. Before addressing those conclusions in detail, it is necessary to make two preliminary points.

First, the issue of whether the creation of a trust fund was foreseeable, that was introduced into the foreseeability enquiry by UCH, is irrelevant. Although that aspect of the case attracted a good deal of attention both below and before us, the only effect of Mrs J’s eventual damages being held on trust was that the already existing position, that Islington could only recover a small proportion of the cost of caring for her, was not interrupted by the damages award. If therefore (as the judge found was not the case) the provision of gratuitous care by the local authority was foreseeable, the only significant effect of the creation of the trust was that the award of damages did not in this case relieve Islington of a financial burden foreseeable by UCH that Islington already bore.

Second, the judge was persuaded to set too high a standard in determining what was “foreseeable”. In particular, it does not conclude that enquiry that outcomes of the defendant’s conduct other than that allegedly foreseeable were possible. It has to be remembered that many people every day commit careless acts that threaten foreseeable injury (like driving at 90 mph down a motorway; or giving incorrect legal advice) without in the event injuring anyone. A turn of events that avoids actionable damage is always a possibility in such cases, but that in itself does not prevent the actual damage from having been foreseeable if it in fact occurs.

The level of certainty required for an outcome to be deemed, after the event, to have been foreseeable is to a large extent a matter of impression. As Lord Wright put it in Bourhill v Young [1943] AC at p107:

“This general concept of reasonable foresight as the criterion of negligence or breach of duty (strict or otherwise) may be criticized as too vague, but negligence is a fluid principle, which has to be applied to the most diverse conditions and problems of human life. It is a concrete, not an abstract, idea. It has to be fitted to the facts of the particular case.”

This fluidity or flexibility of the concept of foreseeability is illustrated for instance in the speeches in Jolley v Sutton LBC [2000] 1 WLR 1083 of Lord Steyn at p 1090C, where he warned that the precise manner and extent of the injury did not have to be foreseeable; and of Lord Hoffmann at pp 1091H-1092D, where he pointed out that, in determining the existence of a duty to prevent a given injury, the likelihood of the injury had to be balanced against the reasonableness or burden of the steps necessary to eliminate the risk.

Mr Miller’s first argument in seeking to support the judge was that the actually negligent person, the unqualified secretary or, possibly, the doctor for whom she was a proxy, could not have had any idea of the ramifications of the care possibilities if, as the judge held, Mrs J foreseeably suffered a stroke. That, it will have been seen, is not how the judge approached the issue. He addressed the foreseeability question in terms of the knowledge, and thus the foresight, of the Hospital Trust; and, with respect, he was obviously right to do so. The failure in this case was not that, or simply that, of the secretary. The failure was an institutional one, in that UCH did not make proper arrangements for advising persons whose drug regime had been varied in anticipation of an operation and whose operation was then cancelled. The question is therefore what was foreseeable by the Trust.

First, as to the prospect of the foreseeable stroke resulting in a need for care, the judge was, I think, misled by Islington’s argument that it was foreseeable that Mrs J would require care. As the judge said, a stroke of a severity that required care was one possibility amongst a number of different outcomes. But there was no reason to think, and certainly no medical evidence, that a stroke of the nature suffered by Mrs J was sufficiently unlikely an outcome of her stopping taking the medicine as to carry it outside the range of reasonable foreseeability: bearing in mind in particular that the institutional body whose notional foresight has to be judged possesses a significant degree of medical knowledge.

Once that step is taken, and the need for Mrs J to receive care is seen as foreseeable, then there are a range of possible outcomes, set out by the judge in his sub-paragraphs (iii) and (iv). Again, the judge may have been diverted from the correct enquiry by the emphasis in argument upon UCH’s knowledge of Mrs J. UCH’s negligence, as we have seen, was in the arrangements that it made for dealing with patients, of which Mrs J happened to be the unfortunate victim. UCH, which must inevitably have worked closely with local authorities in arranging post-hospital care, must be taken to have known that the range of patients whom it treated would have a range of care requirements and financial needs. Care by a local authority in a case in which it could not recover the cost of the care cannot be seen as so unusual as to fall outside that range. Just as, in Jolley, the precise manner in which the injury would occur does not have to be foreseeable, so in this case UCH did not have to know that Mrs J would require local authority care and would not be able to pay for it; but only to have institutional knowledge that some patients with Mrs J’s disability would fall into that category.

I would therefore hold, differing from the judge on this point, that the provision of care to Mrs J by a local authority that could not recover the cost of that care from Mrs J was reasonably foreseeable by UCH.

The judge dealt with this element in the Caparo triology very shortly, as had the parties before him. He said, in his paragraph 53:

“Convenient label or not, the proximity test invites the Court to consider the relationship between the parties. Here as Mr Miller said, there is no relationship between the Hospital Trust and the Local Authority at all at the time of the advice. They were not aware of each other in the context of Mrs J and they were not brought into a proximity relationship by reason of the fact that the Local Authority were the statutory care providers for Mrs J if she made an application to them. In these circumstances the Local Authority fails the proximity test.”

The requirement of “proximity” has an uncertain status. There are some cases in which it has been specifically applied as a vehicle for limiting liability on policy grounds for the foreseeable consequences of a single accident, either in terms of the location of the victims who were physically injured, or in terms of the economic loss to secondary victims. An example in the former category is Alcock v Chief Constable of South Yorkshire, especially in the speech of Lord Oliver of Aylmerton [1992] 1 AC at p 411E. An example of the latter approach is to be found in the judgment of Kirby J in the case in the High Court of Australia of Perre v Apand [1999] HCA 36 [292] – [296].

As Mr Underwood felicitously put it, these are cases of propinquity rather than of proximity. Where such specifically physical relations are not in issue, it is difficult to see that “proximity” adds anything to the other requirements of Caparo. For instance, as Nicholls V-C put it in this court in White v Jones [1995] 2 AC 207 at p 221G:

“the two headings (proximity and justice) are no more than two labels under which the court examines the pros and cons of imposing liability in negligence in a particular type of case”

In the present case, once it is accepted (as of course the judge did not accept) that provision of gratuitous care by Islington fell within the range of outcomes foreseeable by UCH, then it is difficult to see how the relationship between UCH and Islington thereafter fails because of a lack of proximity. The judge’s analysis set out in paragraph 19 above depended very largely upon, and did not expand upon, his analysis of foreseeability. If he was right about foreseeability, then proximity adds nothing. By the same token, if Islington’s involvement was foreseeable, then Islington equally was in a relationship of proximity to UCH.

The foregoing discussion is couched in terms of Islington being in a relationship of proximity with UCH, as was the judge’s analysis. The issue was raised in argument before us of the relationship between UCH and some other local authority into whose care Mrs J might eventually come: for instance, if her daughter who lives in Cambridgeshire proved unable to provide the level of care that Mrs J needs, and the local authority there had to take over. And more extreme cases were hypothesised: for instance, if the person injured were a child, who was in and out of local authority care over the whole of his or her life.

I can understand the potential concern to which these possibilities give rise, but I do not think that they affect the present case. First, the issue in relation to foreseeability, and thus to proximity, is whether the claimant local authority falls into the category of potential carers that the hospital trust should foresee. It is not necessary that a particular claimant local authority should be identifiable as the local authority that will necessarily provide the care. If, however, as in the hypothetical cases mentioned above, a particular need for care arises a substantial time after the negligent treatment of the patient, then the first test will continue to be that of foreseeability, which might very well not be met, at least in the case of the life-span of a younger patient. If the case falls at that hurdle, proximity does not enter into it. But if the successor local authority is foreseeable as a claimant, then that in itself puts it into a relationship of proximity with the hospital trust.

This is an incident or sub-rule of the Caparo doctrine, the exact status of which is not easy to discern. This category or analysis originally entered the law as a reaction to the analysis of Lord Wilberforce in Anns v Merton LBC [1978] AC 728 at p 752, when he identified not a three-stage but a two-stage test: first, foreseeability, and

“Secondly, if the first question is answered affirmatively, it is necessary to consider whether there are any considerations which ought to negative, or to reduce or limit the scope of the duty or the class of person to whom it is owed or the damages to which a breach of it may give rise”.

That approach was thought to give too much licence to the extension of the law on “policy” grounds. It was strongly criticised by Brennan J in the High Court of Australia in Sutherland Shire Council v Heyman (1985) 60 ALR 1 at 43-44 in terms much quoted and approved, not least by Lord Bridge of Harwich in Caparo:

“It is preferable, in my view, that the law should develop novel categories of negligence incrementally and by analogy with established categories, rather than by a massive extension of a prima facie duty of care restrained only by indefinable ‘considerations which ought to negative, or to reduce or limit the scope of the duty or the class of person to whom it is owed’”.

The status of that observation is elusive because it plainly is not intended to take the place of, or to override, the three-stage analysis that was equally asserted in Caparo. I think that Mr Underwood was right when he suggested that incrementalism is a check or guide to the application of the Caparo tests; and that, in particular, if a claimant could demonstrate that his case did no more than incrementally extend an already recognised head of liability, that was a good indication that his claim met the Caparo requirements.

Applying that approach, Mr Underwood argued that his claim could be seen as building incrementally on three previous cases: White v Jones [1995] 2 AC 207; Spring v Guardian Assurance [1995] 2 AC 296 and Kirkham v Boughey [1958] 2 QB 238. In all of these cases, he said, a person other than the immediate object of the act of negligence had recovered for the effect of that negligence upon him. While in verbal terms that is correct, more detailed analysis of the cases shows that they are a long way away from Islington’s claim.

White v Jones was a conferment of benefit case: the beneficiary lost what the solicitor knew that the testator had intended to give him. That creates a closer and more immediate relationship between the three parties than is apparent where the original object of the negligence has herself suffered loss, and the claim of the third person is only in respect of the loss to him in alleviating that loss. Nonetheless, in a case where the relationship of the claimant and the defendant was sufficiently direct to resemble a relationship of contract the House of Lords could only grant recovery by a bare majority. That strongly indicates that the potential for the incremental development of White v Jones is very limited. And although Spring was a creation of detriment case, the relationship between the three parties was very similar to that in White v Jones, to the extent that Lords Goff of Chieveley, Slynn of Hadley and Woolf thought that there was a contractual as well as a tortious liability on the part of the ex-employer who gave the reference to a third party.

Kirkham v Boughey falls into a different category. Mr Underwood drew attention to Lord Oliver’s description of it in Alcock [1992] 1 AC at p 409C as having permitted a husband to recover directly his expenses in providing medical care for his negligently injured wife because he was under a legal duty to make such provision. Equally, said Mr Underwood, was Islington under a legal duty to care for Mrs J. But closer scrutiny of Kirkham significantly reduces the generality of its effect. Although the report is not entirely easy to follow, it would appear that Diplock J started from the position, based on Best v Samuel Fox [1952] AC 716, that in general persons economically affected by injury to a third party cannot recover against the injurer. But he then, at p342, referred to the husband’s right to consortium, as it then was. Although both Diplock J and Lord Goddard, upon whose speech in Best v Samuel Fox at p 733 Diplock J relied, distinguished the husband’s claim from his claim for loss of consortium, they both undoubtedly viewed it in the latter context, and for that reason as something of a special case. Although the case goes some way towards assisting Mr Underwood, it is a fragile basis for a generalisation as wide as that which he needs to achieve.

And, further, on the other side of this coin there are a number of indications in other areas of the law that are distinctly unhelpful to Islington’s case. I have already noted the position that was taken in Best v Samuel Fox as to lack of recourse for one injured economically by physical injury to another, a passage to that effect by Lord Morton of Henryton [1952] AC at p734 being cited with approval by Lord Oliver in Alcock at p 409E. Second, it is equally axiomatic that a voluntary carer has no cause of action of his own against the tortfeasor, as Lord Bridge assumed without question in Hunt v Severs [1994] 2 AC at p 358F. The injured person does have a claim for the carer’s costs, as part of the damages that she recovers from the tortfeasor, and that claim is intended to recompense the carer, as is shown by resultant damages being held in trust for her. But, as described in Hunt v Severs, that has only been achieved by the construction of a notional liability of the injured party, and notional payment in respect of that liability by the injured party to the carer; rather than by the imposition of direct liability between the tortfeasor and the carer. Third, and a matter to which I will have to return, it has never been envisaged that the National Health Service has a cause of action in negligence against the tortfeasor, any such recourse having to be achieved by special statutory provision. And it is of some interest that when the Law Commission as part of its work on damages for personal injury considered ways of imposing liability to the NHS on tortfeasors by operation of the common law, it felt obliged to address itself principally to the conceptually very difficult area of recoupment for unjust enrichment, rather than to a more direct recourse in negligence: see Consultation Paper No 144 at paras 3.2-3.18, and Law Com No 262 at paras 3.22-3.25.

With that somewhat inconclusive outcome, I turn to the third limb of the Caparo trilogy.

Viewing the matter as between the present parties, it indeed seems all of fair, just and reasonable that UCH rather than Islington should bear the cost of Mrs J’s care. UCH has been negligent, Islington has not. It is not only unreasonable but also unfair that a tortfeasor should escape liability for part of the results of his negligence simply through the double accident of his victim being cared for by a public body rather than privately, and the victim not being able to afford to pay for that care. And, in contrast to many claims for pure economic loss, the tortfeasor is not faced with liability for an uncertain and possibly infinite amount. All that Islington seeks from UCH is the very amount, or something very close to it, that UCH would have had to pay to Mrs J if Mrs J when injured had been sufficiently wealthy to be able to buy care for herself.

But “fair just and reasonable” is not to be read literally, nor is it to be read solely in the context of the relationship between the instant claimant and defendant. It still assumes what might be called the residue of Lord Wilberforce’s test, that wider issues of policy may have to intervene. As Lord Keith of Kinkel put it in Yuen Kun Yeu v A-G of Hong Kong [1988] AC 175 at p193:

“The second stage of Lord Wilberforce’s test in [Anns] is one which will rarely have to be applied. It can arise only in a limited category of cases where, notwithstanding that a case of negligence is made out on the proximity basis, public policy requires that there should no liability”.

That is how the judge approached this part of the case. He said, at paragraphs 54-60 of his judgment:

“In view of my conclusion on foreseeability and proximity it is not strictly necessary to proceed to consider the fair just and reasonable test. However because I was addressed on this by both Counsel I briefly now give my view.

Mr Miller was right here when he submitted that there was no lacuna in the law which needed to be filled by the Court.

It does not appear to me that there is any unjust situation which calls for remedy by the Court. Mr Miller is right when he says that the inability of the Local Authority to recover care costs in cases such as this is the result of a statutory scheme which has been thoroughly reviewed in the decisions of Courts.

Parliament must be taken to be aware of the situation created by the combination of the provisions of Section 21 of the National Assistance Act 1948, the 1993 Directive, and decisions of the Court, and has chosen in this case not to intervene to give assistance to that authority.

It is a necessary consequence of the interface between Government and Local Authority in the field of payment for care and it is an abuse of language to describe it as a gross injustice calling for remedy.

If there were an injustice in this area of the law it is not for the Court to intervene but for Parliament. Parliament has imposed liability on tortfeasors under the Road Traffic Acts and recently far more broadly under the Health and Social Care (Community Health and Standards) Act 2003. If intervention is needed Parliament can decide to be involved.

There is in my view no identifiable public policy which requires the Court to extend the law of negligence by providing a novel duty of care to Local Authorities in the circumstances of this case.”

I would respectfully agree with the broad thrust of those observations, but because the matter is of importance it is necessary to go into it in somewhat more detail. Put shortly, the concern in this case is that, in order to correct what seems to be an inequitable distribution of liability between two public authorities, the common law of negligence is being asked to do a job for which it is not qualified. There are a number of strands in that contention, some of which overlap considerably with each other.

First, although there do not seem to be any obvious dangers of unreasonable or uncontrolled liability between the two immediate parties to this case, it would be idle to pretend that a decision in Islington’s favour might not have implications, in terms of its reasoning, for other providers of public services. Mr Miller suggested the case of education authorities, obliged to provide gratuitous services to a child who required a statement of special educational needs because of clinical negligence. The same might be true of a great range of social services that are in principle free to the recipient. It is quite impossible for a court to know, within the confines of a particular case and with the benefit only of a sparse amount of evidence and its own commonsense, what are the wider implications of the move that it is being asked to make. This is a task for law reform, which can be undertaken in the light of the full picture and of informed comment on all of the collateral implications; rather than for law making.

Second, although I have referred in general terms to UCH and Islington as both being public authorities, they are public authorities of rather different sorts and with conspicuously different sources of funding. While the court would not pretend to detailed understanding of the differences, it can at least take notice that UCH is part of the “national” health service, in principle subject to control by national government and financed out of national taxation; whereas Islington is at least in theory controlled and financed by those who pay the community charge in that borough; however much it may be that a high proportion of its income in fact comes from national government, and its freedom of action is, as in the present case, controlled by the national government acting through Parliament. As between these different authorities, it may be the case that the obligation that rests on Islington in relation to impecunious persons in need of care has arisen by statutory accident; and it is very doubtful whether it has arisen in pursuit of any desire to aid the tortfeasor. But the sorting out of the present position, accidental or not, would seem to be essentially a matter for Parliament, or at least for political decision, rather than for a court deciding a particular case.

Third, the present case may be unusual in that the tortfeasor is itself a public authority, indeed itself part of the caring professions. Foreseeability considerations might be different in the case of a private tortfeasor (most conspicuously, a motorist), who could not have attributed to him, or at least could not have attributed to him with the degree of confidence displayed in paragraphs 16 and 17 above, the range of institutional knowledge with which UCH is burdened. If the private tortfeasor passed the foreseeability test the detailed concerns set out in paragraphs 37 and 38 above would not apply in his case. But their place would be taken by a different range of issues as to the economic implications of extension to the private tortfeasor or to his insurer of liability for costs that are currently seen as a public liability.

Fourth, it is necessary to rationalise the existence of a direct action by the local authority against the tortfeasor in a context where neither a private provider of care nor the National Health Service enjoy such a right. And Mr Miller suggested other cases where recovery by a public body has never been contemplated: for instance, the Fire Brigade’s costs of attendance at a negligently caused fire. Mr Underwood was constrained to accept, contrary to the constant assumption of the law, that on the logic of his argument private providers would in principle have a right of action; but he submitted that actually to grant such a right in their case would be supererogatory, since the tortfeasor was responsible to the injured party for those costs. That argument does not give weight to the considerable difficulty, some would say artificiality, that the court has encountered in achieving recovery in respect of private care costs, as demonstrated in the account in Hunt v Severs. Far easier, one would think, to permit that carer to sue the tortfeasor directly, and thus avoid the whole of the Hunt v Severs difficulties. But, more pressingly, that reply does not account for the different position of the National Health Service, where a right of recovery has only been achieved by specific legislation.

Mr Underwood said that the NHS was different because its constitution, like that of the Fire Service or Police Force, required it to give free service to all at the point of delivery; whereas local authorities were obliged to make such charges as they could recover, albeit for the latter reason not being able to charge Mrs J. That however is a fragile distinction. First, if the NHS indeed were constitutionally obliged to provide a free service to everyone, irrespective of their social merits, it is difficult to see how Parliament could consistently introduce charging legislation against tortfeasors, most conspicuously by the Health and Social Care (Community Health and Standards) Act 2003. Second, that legislation was thought necessary is a very clear indication that Parliament and those who advised it, in particular the Law Commission, did not think that direct claims were available even where the care provided is “first line” rather than, in the case of the local authority, somewhat more remote. And third, the court will again be trespassing into an area for which it is ill-equipped if it seeks to produce for local authorities something like the solution that was produced for the NHS only after a great deal of law reform work, and which even now has not been brought into operation.

Islington’s basic case attracts a good deal of sympathy. It is difficult to see any good reason why Islington rather than UCH should pay for Mrs J’s care. And the way in which Islington presents its case fits into the formal structure of the law of negligence. But there are important and overriding considerations why this court cannot take the step that Islington requires of it. I would dismiss this appeal.

I have had the advantage of reading the judgments of both Buxton LJ and Ouseley J in draft. I agree with them that this appeal should be dismissed.

In so far as there is or may be a difference between them on the issue of proximity I prefer the approach of Ouseley J. In particular, I agree with him that the question whether or not there was a sufficient relationship of proximity between the parties is closely related to the question whether it would be fair just and reasonable to impose a duty of care on UCH. I agree with both Buxton LJ and Ouseley J that, for the reasons they give, it would not be fair just and reasonable to impose such a duty on UCH in the particular circumstances of this case. Given the considerations which have led me to that conclusions, I would not hold that there was a sufficient relationship of proximity to justify that limb of the test.

I wondered in the course of the argument, when Hunt v Severs was being discussed, whether a voluntary carer could require an injured claimant to include his or her claim in the claimant’s claim for damages against a tortfeasor, in circumstances in which the claimant refused to do so. It struck me that, if such a carer could not, the position would be less than satisfactory because whether he or she was fairly compensated would depend upon the goodwill of the claimant. It further struck me that, if a carer could do so, perhaps an authority like Islington could do so also. As I understand it, no-one has sought to take such a step in the past but, whatever the answer to those questions, it is clear on the authorities that the tortfeasor does not owe a duty of care to a carer. Like Ouseley J, I too do not see that Islington can be in any better position than a voluntary carer, so far as the imposition of a duty of care is concerned.

Quite apart from any comparison between a public authority and a voluntary carer, Buxton LJ has to my mind convincingly demonstrated in paragraphs 36 to 41 of his judgment that it would be wrong in principle to hold that UCH owed a duty of care to Islington on the facts of this case. I see nothing to be gained by repeating those reasons here.

I agree that UCH did not owe a duty of care to Islington in its treatment of Mrs J. The asserted duty of care owed to Islington in the medical treatment of Mrs J was the same as the duty of care owed to Mrs J personally as the recipient of that treatment.

I agree that it was reasonably foreseeable by UCH and by a doctor responsible for the treatment of Mrs J that if she stopped taking the anti-coagulant drug, she would have a stroke so disabling as to require care, care which would either be paid for privately or supplied voluntarily by family, friends or charity, or provided at public expense at a national or local level. It is as foreseeable in most cases that this will cost the patient money as it is that voluntary carers will be put to expense or loss or that the public sector will be unable to charge for the care which statute requires it to provide. The precise level of knowledge which a tortfeasor hospital may have of the patient’s means or residence will not usually alter the foreseeability of that loss. It does not matter for the purposes of this case, to my mind, that a doctor or UCH as an institution might not have been able to foresee which of those carers would be involved.

The considerations of proximity and fairness, justice and reasonableness are essentially limiting considerations upon the extent to which the reasonable foreseeability of loss from the act in question gives rise to a duty of care to the person suffering the loss. I see both as reflecting considerations which may broadly be described as policy considerations. And I agree that all the factors identified by Buxton LJ tell strongly in favour of the absence of a duty of care owed to Islington. In that sense it may matter little under which particular heading the particular factors are posted. There is after all only one question: is there a duty of care? The three stages or tests overlap to a greater or lesser extent.

Here it is said on behalf of Islington that a duty of care is owed to it because it was reasonably foreseeable that a breach of the duty of care to Mrs J, through causing her injury and a consequent need for care, would thereby cause it loss. The loss is reasonably foreseeable but occurs as a consequence of Mrs J’s injury. I do not see that as materially different from the loss which may be suffered by voluntary carers who have no cause of action, or the equally foreseeable losses which may be suffered by a business deprived of the services of a negligently treated patient or of a negligently injured road user. I would see the limit which the law has imposed on the existence of a duty of care towards a person who suffers loss as the result of an injury to another as an aspect of proximity.

There is material in all the cases to support either analysis but I found persuasive what Lord Oliver of Aylmerton said in Alcock v Chief Constable ofSouth Yorkshire Police [1992] 1 AC 310 at 410:

“The failure of the law in general to compensate for injuries sustained by persons unconnected with the event precipitated by a defendant’s negligence must necessarily import the lack of any legal duty owed by the defendant to such persons. That cannot, I think, be attributable to some arbitrary but unenunciated rule of ‘policy’ which draws a line as the outer boundary of the area of duty. Nor can it rationally be made to rest upon such injury being without the area of reasonable foreseeability. It must, as it seems to me, be attributable simply to the fact that such persons are not, in contemplation of law, in a relationship of sufficient proximity to or directness with the tortfeasor as to give rise to a duty of care, though no doubt ‘policy’, if that is the right word, or perhaps more properly, the impracticability or unreasonableness of entertaining claims to the ultimate limits of the consequences of human activity, necessarily plays a part in the court’s perception of what is sufficiently proximate.”

I would regard this legal or judicial policy as different from the particular considerations here which arise out of the statutory provisions which regulate the recovery from tortfeasors of losses incurred by the NHS. There are many public bodies which may be called on to perform their statutory duties as a result of the breach of a duty of care owed to an individual whose injuries thus entitle him to invoke the performance of that statutory duty. They may or may not be able to recover their charges from the individual whom they have assisted, whether because the treatment is required to be free or to be free for the impecunious.

Parliament has so far decided which statutory bodies, and in what circumstances, are able to recover the costs of performing their statutory duty from the tortfeasor. There are issues of policy so far as concerns the recovery of the costs of statutory bodies performing their statutory duties, and charging for that performance according to the terms of the relevant statute, which go beyond the more general considerations of proximity. These issues I would regard as arising under the fair just and reasonableness heading. Although I can see that recovery for Islington would be fair just and reasonable in ordinary terminology, that test is the point at which the factors of public policy which Buxton LJ identifies, in paragraph 36 and following, are brought in. They make a duty of care owed to Islington, so as to enable it to recover the costs which it was required by statute to incur and which it was unable by statute to recover from the recipient of those services, a leap too far.