Braceurself Limited v NHS England

Introduction

When is a Respondent’s Notice really a cross-appeal in disguise? That is the issue that arises in an ongoing appeal brought by Braceurself Limited (“Braceurself”) against the respondent, NHS England (“the NHS”). The issue matters because, if the NHS’s Respondent’s Notice is, in reality, a cross-appeal, then they need permission in order to bring it, whilst if it is a proper Respondent’s Notice, they do not.

This is the judgment of the court, to which we have all contributed. We are very grateful to both Ms Morris KC and Mr Moser KC for their clear submissions on the underlying issues. We recognise that, although we can and will decide the issue on the particular facts of this case, it may have a wider application. That is particularly so given the paucity of authority directly in point.

The Factual Background

Braceurself made a claim against the NHS for breach of the NHS’s statutory procurement obligations. The tender process in question concerned the provision of orthodontic services in East Hampshire, lot reference PR002368. The claim was based on what were said to be manifest errors in the tender evaluation process. By the time of the hearings before the judge, the claim was for a declaration and damages: the automatic suspension had been lifted some time previously, and the contract had been awarded to the successful tenderer.

On 12 February 2021, at a costs and case management hearing, Fraser J ordered a split trial between liability and quantum. He expressly indicated that the issue of liability, to be heard at the first trial, would include the issue as to whether any manifest breach was “sufficiently serious” to warrant an award of damages. The importance of that latter issue in English law derives from the decision of the Supreme Court in Energy Solutions EU Limited v Nuclear Decommissioning Authority [2017] UKSC 34; [2017] 1 W.L.R. 1373, which itself applied the decision of the European Court of Justice in Francovich v Italian Republic (Joined Cases C-6/90 and C-9/90) EU:C:1991:428; [1995] ICR 722. In short, in a procurement challenge where the remedy sought is for damages only, the claiming party has to demonstrate that the authority’s manifest breach of the procurement obligations was sufficiently serious to justify an award of damages.

There was some debate before us as to the proper categorisation of the Francovich damages issue, and whether it was properly included as part of the liability trial. We think it was. It is a species of causation argument. In a procurement case where damages are the only remedy, it is critical to know whether any breach that may be proved surmounts the “sufficiently serious” hurdle. If it does, the remaining issues are matters of quantification only; if it does not, that is the end of the case. In more old-fashioned language, it is a key element of determining “liability proper”.

The liability trial took place between 28 February 2022 to 4 March 2022 before Alexander Nissen KC, sitting as a Deputy High Court judge (“the judge”). The parties had agreed a list of 23 issues. Issue 19 addressed a schedule of alleged manifest errors in the marks given by the NHS to the two tenderers (Braceurself and the successful bidder). Issue 23 asked whether any of those breaches that were proved were sufficiently serious to justify an award of Francovich damages.

The judge’s judgment on the substantive issues is at [2022] EWHC 1532 (TCC). He found at [127] – [146] that there had been a manifest error in respect of the marking of Braceurself’s responses to question CSD02, concerned with clinical and service delivery, and in particular the issue of accessibility. He found that the NHS had misunderstood Braceurself’s bid in two respects, concerned with: i) access to the first floor premises and the references to a stair-climber; ii) alternative premises. This resulted in Braceurself being awarded a score of 3 for premises and equipment, when they ought instead to have been awarded a 4 [147], [186]. In consequence, the total bid score would have been 2.5% higher and would have meant that Braceurself’s tender would have been the successful bid [188].

At [192], the judge observed that, through no fault of the parties, they had not made adequate submissions as to whether the particular breach which he had found was sufficiently serious to trigger the claim for Francovich damages. This was partly because, although the errors which the judge found were properly before him, they had not been the subject of any particular focus before or at the trial. In consequence, the judge adjourned Issue 23 and all consequential matters until after this question had been decided.

The judge’s order of 21 June 2022 reflected this position. The Recitals to the order included:

By the time the judge made the order of 21 June, the NHS had indicated that they intended to appeal against the findings of manifest error, and Ms Morris had prepared draft Grounds of Appeal. There were four stated grounds. Grounds i) and ii) were concerned with the access issue: i) complained about the judge’s treatment of the mistake about the stair-climber as a legally significant error; ii) was the suggestion that the judge had failed to attach proper weight to the access requirements of parents with buggies. Grounds iii) and iv) were concerned with scoring: iii) criticised the judge for discounting the evidence of the evaluators that, even if they had known that it was a stair-climber, not a stair-lift, it would probably not have changed their mark; iv) was an allegation that, in re-scoring, the judge made various errors in his evaluation and did not take the entirety of CSD02 into account.

There was some argument before this court as to whether those were challenges to the judge’s findings of fact and/or his evaluation of the evidence, or raised discreet points of principle. Although it is unnecessary to reach a concluded view on that matter, we consider that, at first blush, the four grounds are primarily attacks on the judge’s findings of fact and the inferences to be drawn from those facts, and his evaluation of some of the evidence. Inevitably, there is some overlap with principle, particularly on the topic of re-scoring, although we consider that overlap to be relatively modest. Accordingly, these are the sort of points which, on an application for permission to appeal, may give rise to the concern that they strayed into areas which were the sole preserve of the trial judge (see Fage UK Ltd v Chobani UK Ltd [2014] EWCA Civ 5; [2014] F.S.R. 29 at [114]). However, we make clear that we have formed no view as to the merits of those points one way or the other, and therefore cannot say whether, if they did require permission to appeal, such permission would have been granted.

The order of 21 June 2022 expressly reflected the fact that the NHS had indicated their intention to appeal. Paragraph 4 of the order provided that:

In September 2022 the judge heard the Francovich damages issue. His judgment was dated 16 September 2022, [2022] EWHC 2348 (TCC); [2023] 1 C.M.L.R. 41. The judge concluded that, for the reasons set out there, the breach was not sufficiently serious to trigger a claim for Francovich damages. An order was drawn up to reflect that, and to give directions for a hearing to deal with all consequential matters. The Recitals to the order of 16 September 2022 included:

The consequentials hearing took place on 7 December 2022 and a third and final judgment was handed down at [2022] EWHC 3509 (TCC). In consequence of that hearing, an order dated 16 December was drawn up. That recorded the judge’s decision on costs and his grant of permission to appeal to Braceurself on what was called Ground 7, namely the issue as to whether the breach was sufficiently serious to justify Francovich damages. The judge refused Braceurself permission to appeal on any other ground.

At the consequentials hearing, there had been a debate about whether the NHS was entitled, on the one hand, to object (successfully) to the further grounds of appeal put forward by Braceurself, which primarily sought to impugn other factual findings, and on the other, to advance a Respondent’s Notice which, by including grounds 1-4 of the NHS’s original grounds of appeal, could be said to do exactly that, without the need for permission. During the course of the argument, the judge put it to Ms Morris that “for all the objections that you have been making and will make about the appeals on fact that they (Braceurself) are seeking to run, you get away with it…on the Respondent’s Notice.” Although Ms Morris understandably declined to put it in that way, that might be regarded as the effect of her submissions. That too was reflected in the order of 16 December 2022, both in the Recitals, and at paragraphs 10 and 11 of the order itself:

Subsequently, Braceurself was given permission by this court to expand slightly on the ground of appeal for which permission had been granted by the judge, but permission to appeal on the other grounds was again refused. In their statement of objections to Braceurself’s application to this Court, the NHS had objected to those grounds on the basis that they were appeals against the judge’s findings of fact. That brought into stark relief the NHS’s Respondent’s Notice, served at the same time, which raised the four grounds of appeal against the findings of manifest error to which we have previously referred at paragraphs 10-11 above. This court indicated that a short oral hearing would be required in order to determine the issue of whether the Respondent’s Notice was in reality a cross-appeal, for which the NHS required permission.

The Law

CPR 52.13 is in the following form:

That is confirmed by Practice Direction 52C which contains, at paragraph 8, the following provision in respect of a Respondent’s Notice:

The notes in the White Book dealing with these provisions refer to only one authority, Trinity Logistics USA Inc v Wolff [2018] EWCA Civ 2765; [2019] 1WLR 3997 (“Wolff”). However, a trio of earlier decisions of this court should be addressed before coming to Wolff.

In Lake v Lake [1955] 3 W.L.R. 145, a wife successfully defended her husband’s petition for divorce, but sought to appeal against a finding that the commissioner had made that she had committed adultery. This court held that there was nothing in the order against which an appeal could lie: even if the wife succeeded in disproving the allegation of adultery, that would not lead to an amendment of the order, which had upheld the wife’s case and refused the divorce petition.

In Re B (a minor) [2000] 1 W.L.R. 790, following the first part of a split hearing, the judge at first instance made findings of non-accidental injury. Although there was no formal order, those findings were clearly significant to the second part of the proceedings, which were concerned with the appropriate care orders that needed to be made in consequence. The question for this court was whether the parents were entitled to appeal the findings of accidental injury then and there, or whether they had to wait until after the second hearing. Dame Elizabeth Butler-Sloss P said that to make the parents wait would be “to fly in the face of common sense and an approach to the disposal of litigation which would be totally contrary to Lord Woolf’s proposals in the new civil procedure approach”. She was satisfied that the way the case had been presented to the judge “was by way of hearing of a preliminary issue”. She went on to conclude that the appeal court did have jurisdiction to entertain an appeal. She said of the issues decided by the judge:

In Cie Noga SA v Australia & New Zealand Banking Group [2002] EWCA Civ 1142; [2003] 1WLR 307 (“Noga 3”) a curious issue arose as to the precise form of a particular declaration made by the judge, and whether or not it was deliberately designed to require the defendant to seek permission to appeal. The issues at the trial, which took 6 months, were concerned with whether there had been an agreement by the SJ Berwin defendants to settle with Noga for $100 million (an issue of fact) and, if so, whether it was binding (an issue of law). The judge granted a declaration that “the SJ Berwin defendants have not entered into a binding settlement agreement with Noga…although a figure of $100 million was agreed.” The judge gave Noga permission to appeal on aspects of his conclusion that there was no binding settlement, but refused to grant permission to the SJ Berwin defendants to cross-appeal against the finding of fact as to the $100 million. An issue arose as to whether the judge had been entitled to make the declaration in the form he did, and that was referred to this court. Waller LJ said he was not, because it was not a natural formulation of the preliminary issues, but Tuckey and Hale LJJ disagreed. The application for permission to appeal by the SJ Berwin defendants was therefore refused.

Although Waller LJ dissented in the result, it was his analysis of the law with which the other two judges agreed. In summarising the effect of Lake v Lake and In Re B, he said at [27]:

Waller LJ then went on to consider how far this point could be extended. He said:

Waller LJ addressed CPR 52.13, noting at [35] “the special position of a respondent simply defending the decision in his or her favour”, and went on to say at [36] that a Respondent’s Notice could suggest that the judge’s findings of fact were wrong, without it becoming a cross-appeal. Thereafter, Waller LJ focussed on the principal issue which arose in that case as to the purpose of the declaration granted by the judge. He said at [41] that, if the only purpose of a declaration was to put a respondent in the position where he had to persuade the Court of Appeal to grant permission to pursue his ‘defensive’ Respondent’s Notice, that was not a legitimate basis for the grant of the declaration. Neither statute, nor the CPR provided the court with that power, and it would not be a proper exercise of discretion to create that power.

Also in Noga 3, Hale LJ (as she then was) noted at [53] that “it has long been axiomatic that these words [‘judgment’ or ‘order’] refer to the result of the hearing rather than the reasons given by the judge for reaching that result.” She went on:

That then leads back to Wolff, in which all of those authorities were considered. The issue in Wolff arose in this way. The claimant had a number of different claims against the defendant which all gave rise to the same damage. These were (i) procurement of breaches of a contract to which the claimant was a party; (ii) procurement of a third party’s conversion of goods; (iii) deceit and (iv) conspiracy to cause harm by unlawful means. The deputy judge had allowed the claim for procurement of breaches of a contract to which the claimant was a party, but dismissed the remaining claims. The defendant appealed against the findings on liability and remedy, and the claimant filed a Respondent’s Notice, contending that the deputy judge ought to have upheld the other claims on procuring conversion and conspiracy. Deceit was abandoned.

It was argued on behalf of the claimant, TUSA, that permission to appeal was not required for the maintenance of their defensive position to the effect that, if the judge was wrong to give judgment in its favour on one cause of action, he was also wrong not to do so on another cause of action in respect of which the damages would be the same. It was said that TUSA could not appeal at all against the dismissal of the conversion and conspiracy claims because of the nature of the order. Sir Timothy Lloyd said:

Sir Timothy Lloyd accepted at [86] that “the range of defensive positions open to a respondent defendant may be wider than is open to a claimant, because the defendant is less likely to wish to challenge the lower court’s ruling against it on a separate cause of action.” However he went on to say that it did not seem to him that that made a significant difference. He said at [89] that, if a claimant asserted two claims against the defendant, of which one was successful and the other was dismissed, and the defendant appealed against the judgment on the first claim, then if the claimant (as respondent to that appeal) wanted to argue that the court below was wrong to dismiss its other claim against the defendant and that the order below should be upheld on that basis, that assertion amounted to an appeal against the order and required permission to appeal.

The principles relating to appeals generally which are of some relevance to this appeal can be summarised as follows:

The principles relating to Respondent’s Notices which are of relevance to this appeal can be summarised as follows:

These principles governing the proper content of a Respondent’s Notice have recently been applied by this court in Morton and Another v Morton [2023] EWCA Civ 700.

The Issues Between the Parties

On behalf of the NHS, Ms Morris submitted that they were not seeking to appeal or vary the September order made by the judge dismissing the claim. Instead, she said that, in seeking to uphold his order for other reasons, the NHS was quite entitled to raise, by way of a Respondent’s Notice, other grounds for concluding the claim against the NHS had been rightly dismissed. She said that a successful defendant was in a special position and was entitled to raise in its Respondent’s Notice matters which were unconnected to the claimant’s grounds of appeal, but which did not alter the ultimate result.

In response, Mr Moser submitted that, in the circumstances of this case, the matters raised in the Respondent’s Notice were a cross-appeal. He said that the NHS was, in effect, seeking to appeal the 21 June order. He also relied heavily on Wolff, to argue that Sir Timothy Lloyd’s analysis of where there were two claims, and one was successful and one was refused, was directly applicable to this case: he said that the same reasoning applied where there were two issues which were unconnected, namely the Francovich damages claim, which was the subject of the appeal, and the manifest error claim, which was the subject of what he called the cross-appeal. In this way, he said, permission for the latter was required. He also argued that, even if the position was not as clear-cut as that, this court should exercise its discretion so as to require the NHS to seek permission to run the points it has identified in its Respondent’s Notice.

The Simple Analysis

It seems to us that the simple analysis is this. Paragraph 8 of PD 52C makes clear that it is when a respondent is seeking to appeal against an order, or seeking a variation of the order, that a cross-appeal is brought, and permission to bring that cross-appeal is necessary. That is consistent with Lake v Lake, In Re B, and Noga 3. The NHS are not seeking to appeal the judge’s order of September 2022. They won, because the claim against them was dismissed. Neither are they seeking a variation of that order, for the same reasons. On the face of it, therefore, they do not need permission to raise their Respondent’s Notice.

Braceurself argued that whatever the position in respect of the September and December orders, the NHS were seeking to appeal the order of 21 June 2022. That was the order which recited the judge’s finding that the NHS made a manifest error in the marking of Braceurself’s bid, and that the mark should have been increased from 3 to 4 which would have resulted in Braceurself’s bid being successful (paragraphs 9 and 10 above). This argument gains additional force from the fact that, at that stage, the NHS actually prepared the grounds of appeal which are now reflected at paragraph 1-4 of the Respondent’s Notice.

Ms Morris’ threshold submission was that, because those matters were recorded in the June order by way of Recital rather than as Declarations, they were not appealable at all. We disagree. We are not persuaded that any significance should attach to the fact that these findings were recorded as Recitals rather than as Declarations: that sets too much store in the form of the order rather than its substance, an approach rejected in In Re B and also deprecated by Sir Timothy Lloyd in Wolff (and see above at paragraph 31(c)).

However, we are in no doubt that Ms Morris was right to say that, in this case, no question of an appeal arose on either side until September, when the judge decided the Francovich damages point. He had been asked to deal with liability, and that included the “sufficiently serious” issue. He was obliged to decide that in two stages, for the reasons we have explained. But it was only when he had done so that the parties were able to see who had won and who had lost on liability. Following the September order, the NHS had no need to appeal: they were the winners because there was no “sufficiently serious” breach. It was Braceurself who needed to appeal, and the NHS who, if so advised, could then seek to uphold the judge’s order of September on other grounds. The fact that both the judge and the NHS approached the matter in a sensible and transparent way cannot alter that position in law. The judge made directions which first extended the NHS’s time for appealing and then, once it turned out that the NHS were in fact the winners, made orders which protected their position for the future.

We should also say that it would be contrary to the overriding objective for parties in split trials, or longer hearings which are sensibly divided up into sections, to be worrying endlessly about whether or not, at different stages of their unfolding litigation, they needed to prepare grounds of appeal, and to seek permission to appeal at every possible stage. It would create uncertainty and potentially increase costs. That is not what the CPR is all about: indeed, the practical problems inherent in current hearings, when there are appeals going on elsewhere, can be seen at [10] – [13] of Energy Solutions at first instance ([2016] EWHC 1988 (TCC)).

Mr Moser sought to argue that the judge’s findings of manifest error recited in the 21 June order was equivalent to the findings made (without order) in In Re B, and that an appeal – and therefore the need for permission to appeal – was triggered at that point. We disagree. Here, the issue that the judge had to address was liability, including the Francovich damages point. He had not decided that until September, so there was no question of anyone having a right to seek permission to appeal until then.

Mr Moser’s related argument was that, because the NHS served a draft appellant’s notice in June, even though it was rightly held over until after the outcome of the Francovich issue, it must become a cross-appeal, and could never be transposed into a standard Respondent’s Notice. But there can be nothing automatic or immutable about these labels: they depended on the unfolding story. The NHS’s ultimate success in the action cannot be ignored in favour of the position as it had been before the judge rejected the Francovich claim. Furthermore, we should not lose sight of the fact that, if Braceurself were not themselves appealing, the NHS would have no reason to appeal themselves. It is only because Braceurself are raising one aspect of the liability trial on appeal, namely whether the judge was right to say that the breach was not sufficiently serious, that the NHS are seeking to raise another aspect of the same trial, namely whether they were in breach at all.

Accordingly, on the words of paragraph 8 of Practice Direction 52C and r.52.13, we conclude that the Respondent’s Notice was not a cross-appeal, and permission to raise grounds 1-4 was not required. However, that is not the end of the matter. Mr Moser raised two further arguments. The first was that, by analogy with the decision in Wolff these were two separate issues, just as in that case there were (at least) two separate claims, so permission to appeal for the respondent to raise the second issue was required. Secondly he said that, even if he was wrong about that, the court should exercise its inherent discretion so as to require the NHS to seek permission to appeal. We deal with those arguments in turn below.

The ‘Two Issues’ Argument

Mr Moser’s argument was that, just as Sir Timothy Lloyd concluded in Wolff that the respondent’s desire to resurrect a second, dismissed claim was said to necessitate a cross-appeal, regardless of the terms of the order, so here the NHS’s desire to resurrect the issue on which they lost (manifest error), in circumstances where the appeal is concerned with the issue of Francovich damages, which they won, requires permission to appeal. He said that Sir Timothy Lloyd’s judgment applied just as much to separate ‘issues’ as it did to separate ‘claims’.

Attractively though that argument was put, we cannot accept it. It seems to us that it would blur the divide between, on the one hand, the result as reflected in the order or judgment and, on the other, the reasons for that result. For the purposes of appeals and cross-appeals, that distinction is, and has been since Lake v Lake, critical. There is no justification for blurring that divide now.

If Mr Moser was right, and the outcome of separate issues required separate appeals and cross-appeals, it would mean – amongst other things - that Lake v Lake was wrongly decided (because the issue of adultery was different to the issue of whether or not the husband had proved the contents of his petition, which included allegations of cruelty against his wife). That may explain why Mr Moser suggested that Lake v Lake was an old case which would be decided differently now.

We disagree, for the reasons we have explained. Moreover, given the increase in the volume of civil litigation generally, and the significant increase since the 1950’s of cases run by litigants in person, it is our experience that an important bulwark of the Civil Justice system is the courts’ repeated emphasis that any appeal lies against an order of the court, rather than a finding of fact or a stage in the reasoning along the way. Lake v Lake is commonly cited as authority for that proposition, and none of us have ever heard it suggested that it was ripe for reconsideration.

We do not consider that Sir Timothy Lloyd’s judgment in Wolff can be read as if there was no difference between claims or causes of action, and the issues that arose to permit the determination of those claims. On the contrary, there is a very significant difference. Whatever the order said in Wolff, the reality was that, by not finding that they had been established, the deputy judge had rejected the claims for procuring conversion and conspiracy. The claimant respondent wished to resurrect those claims, albeit as defence to the appeal. He was entitled to seek to do so, but he required the court’s permission, because they were separate claims which had not been upheld. Furthermore, although it was not a part of Sir Timothy Lloyd’s decision, it seems to us that the raising of those resurrected claims would have involved at least some variation of the judge’s order.

By contrast, the reasons for a deciding a single claim in one way or another, and the determination of the issues which arose along the way to that final decision, are inextricably bound up together. They cannot be sensibly separated. Take this case as an example. The claim, and the appeal, centred on Braceurself’s claim for Francovich damages due to a manifest error in the scoring, which amounted to a breach of the NHS’s procurement obligations. The NHS wish to argue that the judge should not even have been considering the issue of damages for breach, because he had been wrong to find that there was a manifest error. The issues raised by both sides all relate to the single underlying claim in this claim, namely the claim for damages for breach of statutory duty. The determination of those issues was a part of the judge’s reasoning along the way to his rejection of that single claim. It is quite impossible to disentangle them.

For those reasons, we do not accept Mr Moser’s first way round the simple analysis, namely that the fact that the appeal and the Respondent’s Notice raise separate issues was enough to make this a cross-appeal on which the NHS are required to seek permission to appeal. In our experience, Respondent’s Notices often raise entirely separate matters from those raised in the Appeal Notice.

Mr Moser’s second argument, to defeat what we have called the simple analysis, was advanced particularly by reference to what Hale LJ said in Noga 3. He submitted that, in all the circumstances, the court should exercise its discretion and require the NHS to seek permission to appeal for grounds 1-4 of its Respondent’s Notice. There are three reasons why, although this argument has given us real pause for thought, we decline to do so.

First, we consider that we must follow the CPR, and are bound by the authorities to which we have referred. That leads inexorably to the conclusion that the NHS are not required to seek permission to appeal in this case. Since, for the reasons that we have also explained, this case is not analogous to Wolff, there is no principled basis on which we could ignore both the terms of the CPR and the authorities and require the NHS to seek permission in any event.

Secondly, it would make for considerable uncertainty if this court was entitled, by reference to the overriding objective, routinely to require permission to appeal to be sought in cases where, under the CPR or by reference to the authorities, it is not otherwise required. If that became the practice, then parties would not be certain where they stood and would, for the avoidance of doubt, be drafting and serving notices of appeal at every stage of a split hearing. There would be uncertainties as to the applicable time limits. Indeed, we note here that Braceurself sought to argue that, if the NHS did require permission to appeal, they were out of time and therefore could not make the application at all. Those sorts of procedural disputes would proliferate if this court concluded that these matters should be the subject of the exercise of a broad discretion on a case-by-case basis.

Thirdly, we consider that Hale LJ’s remarks were made in the context of a very extreme case. If the SJ Berwin defendants were allowed to challenge the findings of fact about the agreement for $100 million, leading counsel had suggested (with a nonchalance that he would not get away with today) that the cross-appeal would last between 12 and 28 days. In circumstances where, even if it was entertained, Noga’s appeal was a short matter of law, it is easy to see why the overriding objective might be invoked to prevent such a waste of court resources and costs. Whilst we would not want to say that the overriding objective can never be invoked in these cases, we consider that it would be very rare for such an application to succeed, and that in an appropriate case, an appellant would be better advised to strike out a Respondent’s Notice as being frivolous and vexatious.

That is all a long way from this case. Here, we are confident that the appeal will last no more than a day, with or without the matters raised in the Respondent’s Notice. This is therefore not a case where recourse to the overriding objective is necessary to prevent a potentially significant injustice.

We acknowledge that this case has raised an issue which not uncommonly arises in the Court of Appeal Civil Division, namely the scope and range of the matters raised in a Respondent’s Notice. It seems to us that this problem arises from time to time because the test - namely whether the respondent is seeking to appeal or vary the order, or merely upholding it on other grounds - is a blunt one. It means that, in theory, there is nothing to stop a successful party facing an appeal from raising, by way of a Respondent’s Notice, all and any of the points on which it was unsuccessful before the judge. That would make a mockery of Fage. It also means that, in a particular case, the court may end up treating the successful party more generously than the losing party which, notwithstanding the ‘special position’ of a successful party, is never an entirely comfortable outcome.

But if the blunt test is removed in favour of something more nuanced, the issue immediately arises: what should the new test be? That is not straightforward. We have already said that it could not be by reference to separate issues, because on one view that is too general, and on another, especially in the wrong hands, it could become too granular. No other test or dividing line immediately suggested itself. If it was suggested that the test should be scrapped altogether, and it became the rule that all Respondent’s Notices should be the subject of a permission application, then that would require wide consultation and the involvement of the Civil Procedure Rules Committee. Beyond drawing this issue to their attention, which we will, those are not matters for us.

Conclusions

For the reasons that we have explained, we consider that, on the facts of this case, the NHS does not require permission to appeal to raise grounds 1-4 of the Respondent’s Notice. We invite the parties’ short submissions on the costs of this issue, if they cannot be agreed. Out preliminary view is that the costs of this issue, originally raised by the court itself, should be costs in the appeal.