Lord, R (on the application of) v Secretary of State for the Home Department

The claimant was convicted of murder and sentenced to life imprisonment on 2 October 1981. He remains in prison: currently HMP Frankland. He is a Category A prisoner. That means (and the definition, as we shall see, is important) that he falls within the category of those

“whose escape would be highly dangerous to the public or the police or the security of the State, no matter how unlikely that escape might be; and for whom the aim must be to make escape impossible”.

Prison Service Order 1010, in which that definition is to be found in paragraph 1.2, continues in paragraph 1.3:

“The test to be applied when reviewing the security category of a Category A prisoner is the level of dangerousness he/she would present if unlawfully at large regardless of the likelihood of escape.”

Category A prisoners are placed in one of three escape risk classifications (see PSO 1010, paragraph 1.4): “standard escape risk”, “high escape risk” or “exceptional escape risk”. The claimant’s classification has been downgraded from “exceptional” to “high” and is now “standard”. He continues, nonetheless, to be assessed as a highly dangerous prisoner who requires Category A security conditions.

The security classification of Category A prisoners is reviewed annually by the Category A Review Team and the Category A Committee – both based at Prison Service Headquarters – in accordance with the procedures laid down in PSO 1010. Part of that process involves the preparation of reports (which for convenience I shall refer to as “Category A reports”): reports by staff in the prison where the prisoner is currently detained (ordinarily including reports by the prisoner’s Personal Officer, the Wing Manager, the Medical Officer and the Security Department), reports by other involved professionals (for example, a psychologist and a probation officer) and an overall recommendation by the Governor or Deputy Governor.

Ordinarily the prisoner is not shown the Category A reports. The practice is for a document summarising their content – known as the ‘gist’ – to be prepared and copied to the prisoner. Importantly, the gist maintains the anonymity of the report writers: nothing referred to in the gist is attributed to any particular source.

This practice stems from R v Secretary of State for the Home Department ex p Doody [1994] 1 AC 531 (see per Lord Mustill at p 560G). It was in effect endorsed by the Divisional Court in R v Secretary of State for the Home Department ex p Duggan [1994] 3 All ER 277 and was explicitly approved by the Court of Appeal in R v Secretary of State for the Home Department ex p McAvoy [1998] 1 WLR 790.

It suffices for present purposes to set out a short passage from Rose LJ’s judgment in ex p Duggan and a rather longer passage from Lord Woolf MR’s judgment in ex p McAvoy.

In ex p Duggan at p 288f Rose LJ said that “on the first and subsequent annual reviews, fairness, in my view, requires that the gist of the reports be revealed in order to give the opportunity for comment”. He made clear at p 288g that what the prisoner is entitled to is (and I emphasise this):

“to be informed of the gist of any matter of fact and/or opinion relevant to the determination of his security category”.

In ex p McAvoy starting at p 798H Lord Woolf MR said this:

“For my part, I accept that it is desirable, when something has the impact which being placed in category A has on a prisoner, that the approach should be to ensure, so far as practical, that fairness is achieved. However, in considering whether in any particular situation the procedure which is adopted is fair or unfair, one has to reach a decision not only in the light of the situation of the prisoner, but also in the light of the practical considerations which must apply to the proper running of a prison. The very fact that we are talking about prisoners who have been categorised as category A indicates that they are among those who are the most dangerous within the system. There can be considerable difficulty within the prison service in the managing of those prisoners …

… in the end it seems to me that the question this court has to answer is whether the procedure which is in fact adopted on the review of categorisation is one which complies with the requirements of fairness, having regard to the nature of the exercise being carried out. As to that, I have no doubt, having seen the material in this case, that the way the process was carried out in this case was perfectly satisfactory and perfectly fair.

I can see difficulties for the prison services in adopting the approach which the applicant would urge upon them of normally disclosing all the material which is relied upon and, whenever it was appropriate to do so, seeking public interest immunity. A procedure of that nature seems to me to be inconsistent in that it is too formal for the sort of administrative decision which is being reached in relation to categorisation.

The House of Lords in Doody’s case [1994] 1 AC 531 endorsed an approach which involved providing the gist of the material relied upon rather than the actual material itself. It seems to me that in a great many cases the interests of a prisoner will be fully protected if the procedure envisaged by Lord Mustill in Doody’s case is adopted. In my judgment the procedure which is being followed at present by the Prison Service in relation to the review of the category in which a prisoner is placed accords with Doody’s case. That is a perfectly satisfactory procedure, particularly and most importantly because, where appropriate, the Secretary of State or those responsible for the review in practice are prepared to reconsider, in the circumstances of any particular case, whether additional information should be made available.

In my judgment what is done in pursuance of that policy provides sufficient safeguards for a person in the position of the applicant. It does not seem to me that he should receive either the actual information or the names of those providing that information. It is sufficient if the gist of the reports plus any special information is provided to him.”

Category A reviews, as I have said, are carried out in accordance with the procedures laid down in PSO 1010. Consistently with ex p Duggan and ex p McAvoy, PSO 1010 provides in paragraph 4.1 that:

“The Prison Service’s current policy and practice on disclosure follows the principles laid down in Duggan. The practice is to provide prisoners with a gist of the information which will be taken into account so they have the opportunity to make effective representations. The policy is to disclose as much information as fairness requires. In the normal case a gist is all that will be required although there could be particular cases where fairness might require the disclosure of an actual report. The Head of the Category A Review Team at Headquarters will decide whether, exceptionally, a particular case requires disclosure of an actual report. ” (emphasis in original)

The forms of report that are to be used, as set out in PSO 1010, are prominently marked “Category A reports must not be disclosed to prisoners or their representatives”. The Secretary of State points to this as showing that there will be an expectation of confidentiality on the part of those filling in the forms.

R (Williams) v Secretary of State for the Home Department [2002] EWCA Civ 498, [2002] 1 WLR 2264, is an example of the “exceptional case” or “special circumstances” (both phrases used by Judge LJ at para [32]) in which the court ordered full disclosure of the Category A reports to a prisoner.

Why does fairness require the type of disclosure referred to in ex p Duggan and ex p McAvoy? In one sense the answer is obvious, and indeed it is recognised in PSO 1010 at paragraph 4.1, but it bears repeating. It is because, as Lord Mustill put it in Doody at p 560G:

“the person affected usually cannot make worthwhile representations without knowing what factors may weigh against his interests”.

In this context I draw attention also to what Rose LJ said in R v Secretary of State for the Home Department ex p Creamer and Scholey (1992) in a passage “wholly endorse[d]” by Lord Woolf MR in ex p McAvoy at p 798C:

“A prisoner’s right to make representations is largely valueless unless he knows the case against him and secret, unchallengeable reports which may contain damaging inaccuracies and which result in continuing loss of liberty are, or should be, anathema in a civilised, democratic society.”

In the present case the claimant’s annual review with which I am directly concerned took place in Autumn 2001. The relevant ‘gist’ was dated 28 June 2001. The claimant was anxious to obtain full disclosure of all the Category A reports on which the gist was based. He did not seek to bring himself within the “exceptional” or “special” category recognised in Williams. Rather he sought to outflank the restrictions of the system approved in ex p McAvoy by relying upon the provisions of the Date Protection Act 1998.

In October 2001 he made a request for information in accordance with section 7 of the 1998 Act. A substantial quantity of information was disclosed to the claimant under cover of a letter dated 5 March 2002, but not the relevant Category A reports. On 18 March 2002 the claimant made a complaint to the Information Commissioner (see sections 7, 40 and 51 of the Act). The Commissioner responded on 11 April 2002. A further letter from the Commissioner dated 16 August 2002 shows that the matter was at that stage still under investigation by the Commissioner. Apparently it still is. The letter drew the claimant’s attention to an individual’s right to make an application to the courts under section 7(9) of the Act. That is what he did.

On 27 September 2002 the claimant issued a CPR Part 8 claim in the Queen’s Bench Division seeking relief under section 7(9) against the Secretary of State. Specifically he sought an order directing the Secretary of State “to make disclosure of the category A reports in full or subject to such redaction as is necessary for the purpose of section 29(1) of the 1998 Act”. By an order of Master Foster dated 17 October 2002 the action was transferred to the Administrative Court. The claimant’s application for judicial review was issued on 7 November 2002. The relief sought was a quashing order “in respect of all decisions in this case not to disclose Mr Lord’s Category A reports” and a mandatory order “requiring the Prison Service to disclose said reports under the terms of” the Act. Permission was granted on the papers by Maurice Kay J on 15 January 2003. Directions were given by Master McKenzie QC in a consent order dated 31 January 2003. Consistently with section 15(2) of the Act, paragraphs 3 and 4 of that order directed that the Secretary of State was to serve the Category A reports on the court, but not the claimant, and invited the judge to read them in advance of the parties’ skeleton arguments and submissions. This I did, prior to the hearing that commenced before me on 28 March 2003. At that hearing, as subsequently, the claimant was represented by Ms Phillippa Kaufmann and the Secretary of State by Ms Karen Steyn. I heard argument and reserved judgment.

The gist, as I have said, was dated 28 June 2001. For present purposes the relevant part reads as follows:

“Reports towards this review have been prepared by Frankland prison staff.

Reports advise that since your arrival at Frankland in the summer of 1999 you have made enormous progress. You are now participating in the sentence-planning process and have made significant progress of late. This was clearly illustrated at the Sentence Planning Board in January where it was recognised that you had completed your previous targets and agreed further targets for the next 12 months. You have attained the enhanced level of the Incentives and Earned Privileges scheme, having only one proven adjudication levied against you since your last review for abusive words or behaviour. It is noted that you have had negative Mandatory Drug Tests and Voluntary Drug Tests.

Reports state that you have been made an education orderly and maintain an effective, if limited, degree of interaction with staff. It is reported that you are well behaved on the wing, keeping yourself to yourself, however, you do associate with a select few inmates. One report states that you have matured and mellowed with age and that your attitude has changed so that you now look at life in a better view and appear relaxed and settled. It is noted that you are completing a GCSE course in Painting and Drawing and could move on to a higher course next year.

Reports advise that you acknowledge guilt for the offence, but did not intend to kill your victim. However, you still hold the view that you thought you were punching the victim rather than stabbing him.

Reports indicate that you now have a desire to undertake the Enhanced Thinking Skills course and are awaiting assessment. It is noted that you are also willing to undertake the Anger Management course if it is deemed necessary. These are huge steps forward, however they do not in themselves reduce your risk. It is suggested that you need to be totally focused on self-development and tackle offending behaviour issues so that you are able to demonstrate your level of progress and at the same time make progress in risk reduction. It is noted that you have discussed with the education department the life and social skills modules, however, you declined to attend as you felt that you had the appropriate skills already.

Reports draw attention to the very serious nature of the present offences, your offending history and your escape history. Reports state that there is no evidence through offence related work or otherwise that your security category should be downgraded at this time.

Your case will carefully reviewed.”

The Category A reports which underlie the gist were contained in exhibit “ES2” to a witness statement of Eddie Sprunt, the head at that time of the Category A Review Team. Consistently with the order made by Master McKenzie the claimant and his advisers had seen Mr Sprunt’s statement and exhibit “ES1” (extracts from PSO 1010) but not exhibit “ES2”.

There are six reports in all: five Category A reports properly so-called and one report containing recommendations from the ‘Governor/Head of Custody’. Their contents when I first read them caused me concern, for they revealed significant differences of opinion between their authors.

Two of the Category A reports were negative or recommended against recategorisation. I quote the key sentences in the two reports:

“ … he has the inclination to escape if a chance arises and I have no doubt he would do the same again if he could.”

“ … Mr Lord has made some progress this sentence … He has yet to complete any offending behaviour work and, until he has done so, he should remain a Category A prisoner.”

Two, however, expressed a markedly different view:

“His attitude has changed and now looks at life in a better view, and I believe that he is not a danger to the public … I do not think that he is an escape risk. I think he could be downgraded to category B and transferred to a prison with less security.”

“Has demonstrated a change in behaviour and attitude over an extended period and a willingness to progress further … I consider that he could be safely downgraded to Category B.”

The author of the other Category A report was “unable to comment” on the claimant’s dangerousness / risk to public or on his security categorisation and escape risk.

The key passage in the report of the ‘Governor/Head of Custody’ read as follows:

“These are huge steps forward, however they do not in themselves reduce his risk. Whilst acknowledging his excellent progress in the last two years, I could not in good faith recommend recategorisation prior to completion of ETS which should serve to reduce that risk”

Given the basis upon which the Category A reports had been disclosed to me, it seemed to me that the seeming discrepancy between the gist and the Category A reports was something that I ought in the first instance to consider in the absence of the claimant and his representatives. Accordingly on 31 March 2001 I wrote to Ms Steyn:

“There is a matter which concerns me in the light of the documents in exhibit “ES2” and which, in the first instance, I wish to consider in camera with the Secretary of State’s representatives but in the absence of Mr Lord’s representatives. For that reason I did not refer to the matter in open court last week.

“ES2” contains six reports from HMP Frankland. Three of these reports recommend against recategorisation but two recommend recategorisation. One expresses no views on the point.

The ‘gist’ document … contains nothing I can see to indicate that there is this seemingly significant difference of opinion. Indeed, arguably, the penultimate sentence – “Reports state that there is no evidence through offence related work or otherwise that your security category should be downgraded at this time” – is if not wrong then significantly misleading.

I am concerned that the ‘gist’ document may not be, to use Lord Woolf’s words in McAvoy, “perfectly satisfactory and perfectly fair”. And if this ‘gist’ document is not, then the question may arise as to whether there can be appropriate confidence in the present system as a whole.”

In response to my letter I was sent a witness statement of Clare Lewis, the acting head of the Category A Review Team since Mr Sprunt’s retirement, and a further skeleton argument from Ms Steyn. On 11 April 2003 I held a hearing in camera. The claimant and his representatives were neither present nor aware of what was happening. By the end of that hearing the Secretary of State had agreed that there was no objection to the claimant being told of the in camera hearing and (subject only to minor redactions) being shown copies of both Ms Lewis’s witness statement and Ms Steyn’s skeleton argument. (The redactions, which I approved, were directed mainly to concealing the identities of those who had written the reports.)

On 14 April 2003 I wrote to Ms Kaufmann informing her of what had been happening since the first hearing on 28 March 2003 but otherwise in much the same terms as I had previously written to Ms Steyn. Following this the claimant filed further evidence and both Ms Kaufmann and Ms Steyn prepared further skeleton arguments. A further hearing took place on 3 June 2003 when I heard additional submissions from both counsel. I now (1 September 2003) give judgment.

It is convenient at this point to set out the key passages in Ms Lewis’s witness statement:

“The most important document for the Category A review – and gist – process is the Governor’s or deputy Governor’s recommendation at pages 1 & 2 of “ES2”. This is not “a Category A report”, but is the crucial document for the gist purposes as it is the assessment of current progress and the recommendation may have the effect of overriding what may otherwise appear in the reports, in so far as there may be any difference of opinion. It represents the prison’s recommendation, which may not coincide with the views of all those who have written Category A reports on the particular prisoner.

I believe that it is because of the primary importance of the prison’s recommendation that the final sentence of the penultimate paragraph of the gist in the Claimant’s case is drafted in negative terms. It was intended to be a reference to the prison’s recommendation. Unfortunately, as drafted, and particularly because of the reference to “reports” in the plural, that sentence would, I accept, be capable of giving the Claimant the impression that none of the report writers had expressed the view that the Claimant could be downgraded to Category B. I believe that it is unlikely that the Claimant was in fact misled because he was aware that at least one of the report writers, his Personal Officer, had expressed the view that he could be downgraded to Category B … ”

I shall return in due course to consider the significance of that last observation. Ms Lewis continued:

“In this case, I believe that the final sentence of the penultimate paragraph of the gist could be amended to read as follows, without the disclosure of that additional information leading to the identification of individuals or prejudicing the purpose of preventing or detecting crime:

“Some reports express the view that you could be downgraded to Category B, whilst others recommend that you should not be downgraded. Overall, in view of the lack of evidence through offence related work or otherwise that your security category should be downgraded, the prison’s recommendation is that your security category should not be downgraded at this time.””

She continued:

“I have carefully considered whether the gist on the Claimant’s 2001 reports is, in all the circumstances, fair. In so doing I have taken on board the Court’s concern as to whether the gist is an accurate reflection of the reports (or, as outlined above, the recommendation) in the Claimant’s case. As I have said above, I accept that the final sentence of the penultimate paragraph of the gist would have been capable of misleading the Claimant. This is an unfortunate error, for which I apologise on behalf of the Prison Service. As I have explained above, I believe that it has arisen because the overall prison recommendation, along with the Governor’s or deputy Governor’s assessment will naturally be given more emphasis in the gist than the views of individuals.

Apart from this slip, I believe that this gist is a fair and accurate summary of the reports, and it accurately reflects the prison’s recommendation. Whilst I would accept that different caseworkers preparing gists may have different styles, the actual content of the gist is dictated primarily by the matters of fact detailed in the reports. In this case, the detail contained in the Governor’s or Deputy Governor’s assessment has been fully set out and the detailed comments in the Category A reports have been incorporated in the gist in so far as it was possible to do so without identifying the individual report writers or prejudicing the purpose of preventing or detecting crime. In particular, the caseworker who summarised these reports was careful to ensure that the positive comments made by … were conveyed in the gist.

I believe that the Court can have confidence in the present gist system as a whole. I believe that in general gists do summarise the Governor’s or deputy Governor’s assessment, the prison’s recommendation and the Category A reports in a “perfectly satisfactory and perfectly fair” way, as Lord Woolf believed to be the case in the case in McAvoy.”

I have evidence filed on behalf of the claimant in the form of various witness statements from Nicki Rensten, a caseworker at the Prisoners’ Advice Service, who acts for the claimant in these proceedings, from Vicky King, a solicitor in the firm of Hodge Jones & Allen who specialises in prison law, and from Simon Creighton, a partner in the firm of Bhatt Murphy who also specialises in prison law. The vast bulk of that evidence was filed after I had written to Ms Kaufmann and in answer to Ms Lewis’s witness statement. The Secretary of State has not sought to file any evidence in response. All three witnesses have considerable experience of acting for Category A prisoners in relation to, inter alia, Category A reviews.

Much of this evidence is directed to establishing two important points. In the first place they assert in the light of their experience that the gists in Category A cases tend to be standardised in format and formulaic in their terms. Mr Creighton opines that “the terminology used in Mr Lord’s case is extremely common.” They draw attention, in particular, to the standard use of phrases such as “reports state” and “reports advise”. As Ms King comments, when such phrases are used “there is no way at all of knowing whether one report said this, whether all reports said this or whether this was the overall recommendation made by the prison”. They produced for my perusal a number of gists in different cases (redacted to conceal the names of the prisoners involved) that illustrate both these points.

Ms King was also able, with the consent of her client in that case, to produce a report dated 9 June 2003 by the Prisons and Probation Ombudsman for England and Wales, Mr Stephen Shaw, following his investigation of a complaint made to him by her client, Mr S. He is a Category A prisoner who complained about the decision–making process which had informed his last Category A review. Paragraph 16 of the report records Ms King’s complaint on behalf of Mr S that “the gist contains sentences that are standard in all gists depending on whether a prisoner is to be downgraded or not and that its use of language is formulaic”. In paragraph 30 of his report Mr Shaw said this:

“I note that Mr [S’s] gist contains numerous phrases common to many gists on other category A complainants that I have read elsewhere. I have some sympathy for his solicitors’ complaint that the document has become standardised. Mr [S’s] gist should have provided him with clear and unambiguous information on which to base his representations. Whilst I can understand the desire for the Review Team to develop its own house style of communicating sensitive information in an acceptable format that does not breach security guidelines, there is a danger that a document which shows over-rigid adherence to a format at the expense of credible content has been produced. The wording of the replies that the Review Team gave to Mr [S’s] solicitors in response to their representations only serves to underline this point. If a gist renders itself so anodyne that it could equally apply to any number of prisoners then its effectiveness in providing a prisoner with credible material to understand the reasons behind their categorisation is open to question”

Paragraph 36 of his report recommended that

“the Category A Review Team itself reviews the structure of the gist provided to category A prisoners so that it is more flexible in language, tone and scope. It should reflect more closely the individual prisoner and relevant information pertinent to that prisoner.”

The second point is this. All three witnesses emphasise the rarity of any gist disclosing differences of opinion. Ms Rensten says that in her experience it is “wholly exceptional” for gists to be worded otherwise than the claimant’s gist was in the present case. Ms King says that in her experience it is “rare for gists to contain anything other than a unanimous view”. Mr Creighton says that “it is quite unusual for the gist to contain differences of opinion and the vast majority of gists give the impression that all of the report writers are unanimous”. He refers to three cases in the last 18 months involving clients of his in which it transpired that the gist failed to mention recommendations that had been made for downgrading. In relation to one of these cases he comments “what is particularly worrying is not just the failure to mention that a report had recommended downgrading, but that the person responsible for preparing the gist thought it was acceptable to omit the complex and important views of the probation service altogether”. Miss King likewise deposes to a Category A review involving one of her clients where a probation officer had expressed a strong view that the prisoner should have his escape risk classification reduced, and had also given information relevant to the determination of his security classification, but where none of this had emerged in the gist. Mr Creighton’s perception is that “gists will routinely fail to accurately reflect the contents of the reports submitted”.

On behalf of the Secretary of State the most significant evidence is contained in the witness statements of Mr Sprunt and Ms Lewis to which I have already referred.

The main justifications for the present system, and for the Secretary of State’s assertion that the system endorsed in ex p McAvoy also meets the requirements of the Data Protection Act 1998, are to be found in the following passages in Mr Sprunt’s witness statement:

“Security categorisation in relation to Category A prisoners is concerned with ensuring that highly dangerous prisoners with a high risk of re-offending if unlawfully at large are held in conditions of security that make escape impossible. Security categorisation is a public protection measure concerned with the prevention of crime. Category A reports, which are produced for the purpose of reviewing a prisoner’s security categorisation, are invariably processed for the purpose of preventing crime. Such reports may have the additional purpose of detecting crime.

I firmly believe that disclosure of the reports would be likely to prejudice the purpose of preventing or detecting crime. If the court holds in this case that the Category A reports do not fall within the s 29(1) exemption, in future all Category A prisoners are likely to make subject access requests for their Category A reports. This would effectively oust the gist procedure.

The Prison Service has major concerns as to the impact the replacement of the gisting procedure with full disclosure would have on the efficacy of the system for the categorisation of highly dangerous prisoners. In order to undertake rigorous reviews of prisoners’ security categorisation, the Prison Service requires prison staff to provide frank personal assessments of some of the most dangerous prisoners in custody. At present they do so in the knowledge that these reports are confidential and the views of individuals will only be conveyed to the prisoner in anonymised form.

It is my belief, and that of senior management within the Prison Service and my staff, that if the gist procedure is replaced by a system of routinely disclosing Category A reports, this will result in less frank personal assessments from prison staff. Prison staff, especially those in close contact with prisoners, are vulnerable. Yet obtaining full and candid reports from such staff is a particularly vital part of the process of reaching an informed decision on categorisation, with a view to protecting the public from serious harm. Accordingly, retaining the gist procedure is designed to protect the public and prison staff from the risk of harm presented by highly dangerous prisoners.

Some reports may also contain information disclosure of which could prejudice prison security, eg information on a prisoner’s associates, manipulation of prison staff or of other prisoners, details of surveillance or monitoring which may be in place. Any report from a security officer would obviously fall within this category. Disclosure of security information could clearly in itself undermine prison security by giving prisoners information as to the importance attached by the Prison Service to particular matters and incidents and as to the surveillance or monitoring that is in place. But reports from other staff may also contain such information.

I accept that there may be instances where reports on Category A prisoners, when looked at individually, would appear to be innocuous and unlikely to prejudice the prevention or detection of crime. This may be so particularly in cases where the prison is recommending downgrading, where one can reasonably expect most, if not all of the reports, to be of a positive nature. But I believe that routine disclosure of any Category A reports would be likely to prejudice the purpose of preventing or detecting crime.

First, if disclosure of Category A reports was ordinarily made it would necessarily become apparent to a prisoner when the Prison Service held specific information on him that it was not prepared to disclose, on the grounds that it would be likely to prejudice the prevention or detection of crime. Thus highly dangerous prisoners would be able to assess whether the Prison Service does or does not have particular information about them relating to any criminal activities that they may be planning or have committed. This may be valuable information for prisoners to obtain, enabling them to thwart the efforts of the prison service to detect or prevent crime.

Secondly, if all positive reports about a prisoner were disclosed, but the gist indicated a negative recommendation it might well be readily apparent to the prisoner, by a process of elimination, who had recommended that he should not be downgraded. This might lead the prisoner to harm or put pressure on the one report writer who he may perceive has caused a decision not to downgrade his security categorisation. Further, fear of this occurring may lead to prison staff being less than frank in the reports they provide to the Category A Committee or Review Team.”

He adds:

“The Category A Review Team have considered the feasibility of disclosure of reports in a redacted form (to remove the report writers’ name, or parts of the report itself). We are of the view that redaction is not sufficient to preserve the identity of the writer, which may be revealed by or easily ascertainable from the content of the report. For example, (i) it will be obvious from the contents of any psychologist’s report that it is the report of a psychologist, and the prisoner will be aware which psychologist has had contact with them, (ii) the content of reports tends to reveal the extent of involvement that the writer has with the prisoner, including specific conversations that they may have had, and in what context, which will lead to the report writer being easily identifiable.

To the extent that disclosure can be made without prejudicing the purposes of preventing and detecting crime, it is made in the form of the gist document. The Category A Review Team do not consider that greater disclosure, in redacted form, should be made unless (exceptionally) it is required in the interests of fairness.”

That evidence can usefully be compared with the contents of the certificate filed by the Secretary of State in ex p Duggan in support of his assertion in that case that he was entitled to withhold Category A reports on the basis of a class claim to public interest immunity. The claim failed. The relevant parts of the certificate are set out in extenso at [1994] 3 All ER 277 at pp 282f–283h. I do not repeat it here. Nor is there any need for me to embark upon any detailed collation of the two documents. Suffice it to say that some of the language used in the two documents is strikingly similar and that, perhaps not surprisingly, many themes are common to both documents: the possible prejudice to prison security and to the effective operation of the Prison Service; the fact that reports are written in the expectation of confidentiality; the risk that reports will become less candid; the danger that partial disclosure or disclosure of redacted documents might enable prisoners to gain inappropriate insight into prison security intelligence and counter–measures.

The Secretary of State’s case in ex p Duggan was summarised by Rose LJ as follows (at p 286j): “A class claim is advanced because of the administrative difficulty of sifting and editing documents, the risk that the differing degrees of disclosure might alert prisoners to monitoring, and the danger to candour if only a contents claim were relied on.” The Divisional Court made brisk work of all this. Recognising (at pp 285d and 288g) that “public interest immunity on a contents basis properly attaches to information which would lead to identification of informants, or impinge upon escape risk by revealing ways in which prison security may be vulnerable or by revealing counter–measures to protect securty”, the class claim was dismissed (see at p 287f), the court drawing attention (at p 285f) to the fact that the liberty of the subject is involved, in effect rejecting (at p 285e) the argument that disclosure would result in a lack of candour and likewise rejecting (at p 285g) the arguments based on the possible administrative burden.

Ms Lewis expressly associates herself with Mr Sprunt’s evidence. But she adds two important points. The first is this. Although, as we have seen, she accepts that the gist in the present case could be amended to disclose the fact that different views were expressed in the various Category A reports, she adds:

“I should stress that whilst I believe that this information can be disclosed in this case, there may be cases where it is not possible to refer to any differing recommendations in the gist without risking identifying individual report writers and revealing their views. In such cases the disclosure of the recommendation may need to be limited to stating the prison’s recommendation.”

The other matter is this. She points out that PSO 1010, which came into force on 15 January 2001, introduced a new format for Category A Review reports which no longer asks report writers to give a recommendation within their reports. (She acknowledges that Mr Lord’s 2001 Category A reports should have been in this new format but “can only assume that they were not due to slow implementation of the new format documents introduced by the PSO”.) In these circumstances, she says, “the problem of conveying individual recommendations that has arisen in this case is not one that should arise in future because, as I have said, the Category A reports should contain detailed comments on which the prison’s recommendation will be based, but should not themselves contain a recommendation.”

The claimant’s witnesses emphasise that what Ms Rensten calls “this standard form of gisting” will continue to cause problems even with the new practice of not requiring report writers to express a view as to whether a prisoner should be downgraded. As she comments: “report writers will still be expected and required to express an opinion whether the inmate has made progress as well as the extent of that progress in bringing about a reduction in risk. There is every indication that the gisting process will continue to leave prisoners in the dark as to whether different report writers are reaching different views on these issues, just as it does in relation to recommendations for downgrading.”

So much for the Category A system in general and for the particular context in which the present dispute arises. But before turning to the 1998 Act there is another part of the system to which I have been referred and which it is convenient to describe at this stage. This relates to the activities of the Parole Board (“the Board”) constituted in accordance with section 32 of the Criminal Justice Act 1991. Two different parts of the Board’s activities are relevant for present purposes:

In relation to ‘lifers’ rule 5 of the Parole Board Rules provides as follows:

“(1)

… the Secretary of State shall serve on the Board and, subject to paragraph (2), the prisoner or his representative –

(a)

the information specified in Part A of Schedule 1 to these Rules,

(b)

the reports specified in Part B of that Schedule, and

(c)

such further information that the Secretary of State considers to be relevant to the case.

(2)

Any part of the information or reports referred to in paragraph (1) which, in the opinion of the Secretary of State, should be withheld from the prisoner on the ground that its disclosure would adversely affect the health or welfare of the prisoner or others, shall be recorded in a separate document and served only on the Board together with the reasons for believing that its disclosure would have that effect.

(3)

Where a document is withheld from the prisoner in accordance with paragraph (2), it shall nevertheless be served as soon as practicable on the prisoner’s representative if he is –

(a)

a barrister or solicitor,

(b)

a registered medical practitioner, or

(c)

a person whom the chairman of the panel [as to whom see rule 3] directs is suitable by virtue of his experience or professional qualification;

provided that no information disclosed in accordance with this paragraph shall be disclosed either directly or indirectly to the prisoner or to any other person without the authority of the chairman of the panel.”

Paragraph 9(1)(d) authorises the chairman of the panel (who is required by rule 3(2) to be “a person who holds or who has held judicial office”) to give, vary or revoke directions in respect of:

“as regards any documents which have been received by the Board but which have been withheld from the prisoner in accordance with rule 5(2), whether the disclosure of such documents would adversely affect the health or welfare of the prisoner or others”.

I need not refer to Part A of Schedule 1 to the Rules, but paragraph 3 of Part B includes amongst the reports referred to in rule 5(1)(b):

“Any current reports on the prisoner’s performance and behaviour in prison and, where relevant, on his health including any opinions on his suitability for release on licence … as well as his compliance with any sentence plan.”

It is common ground (though I have to confess that I am not entirely clear why this should be so given the seemingly all–embracing language of paragraph 3) that this description does not embrace Category A reports and that such reports do not go to the Board. But in this connection I should refer to paragraph 3.2 of PSO 1010 which, headed ‘Synchronisation with other review processes’, provides that:

“Establishments may find it administratively convenient to link security category reviews with other processes such as sentence planning reviews, preparation of F75 [that is, Parole Board] reports for Category A Lifers and parole reviews”.

In relation to prisoners serving determinate sentences paragraph 5.15.1 of PSO 6000 provides that “All reports in the dossier must be written with a view to open reporting.” Paragraph 5.16.1 provides that:

“There are four areas where information may be withheld from the prisoner:

i.

in the interests of national security;

ii.

for the prevention of crime or disorder. This includes information relevant to prison security;

iii.

for the protection of information which may put a third party at risk (it is in this category in which information from the victim should be included);

iv.

if, on medical and/or psychiatric grounds, it is felt necessary to withhold information where the mental and/or physical health of the prisoner could be impaired.”

The claimant in the present case has had his case considered by the Board, most recently in December 2002. In accordance with the procedures which I have just described all the reports prepared for that review were disclosed to the claimant in full and unredacted. In each case they showed the name of the writer. He was accordingly able to make comments on them. Ms Rensten, who acted for him on that occasion, has produced a selection of the reports produced on that occasion which, as she points out, include reports covering the same period as the Category A reports I have been considering. She says that it is “inevitable” that some at least of the two different categories of reports will have been written by the same members of prison staff.

As she correctly points out, some of the parole reports are very favourable to the claimant; others are unfavourable. Thus a report by an officer I shall refer to as officer C (his full name appears in the unredacted copy of the report disclosed to the claimant) says that “in my opinion he could be recategorised to Cat B and given a transfer out of the dispersal system”. Officer CC (again his full name appears in the unredacted copy of the report disclosed to the claimant) suggested that the claimant’s category A status should be “seriously reviewed” and that there was now “a more realistic probability” of his status being lowered. On the other hand, officer B (again his full name appears in the unredacted copy of the report disclosed to the claimant) took a very negative view: “This prisoner in my opinion is not ready for release … Lord should not be on an Enhanced Wing, he is not complying with any regimes to the level that he should be.”

Now apart from the obvious fact that there appears to have been no difficulty in these reports all being disclosed in full to the claimant, there are two other very striking facts that emerge from the evidence. In the first place, the claimant made a complaint to the Prison Governor about officer B’s report. In a formal memorandum, also disclosed to the claimant, the Governor said: “I have to admit that it was a poor report with very little content. I was concerned that comments made in the report conflicted with other reports to the extent that they could not be substantiated.” (It is quite clear from a reading of the memorandum as a whole that the word “poor” was intended by its author to be descriptive of the quality of the report, qua report, and not of the claimant’s conduct as described in the report.) The Governor decided that officer B should no longer be directly involved with the claimant.

The second point is this. Officer C, as the passage from Ms Lewis’s witness statement that I have set out in paragraph [28] above makes explicitly clear, is also the author of one of the claimant’s category A reports: in fact the first of the two reports the key sentences of which I have set out in paragraph [22] above. One is therefore in the extraordinary situation that whilst Mr Sprunt is asserting that the claimant should not be shown a copy of the Category A report written by officer C, his colleague Ms Lewis is relying on the very fact that the claimant knows what officer C’s views are (because he has been shown a copy of the parole report written by officer C) in support of her assertion that the admittedly defective gist would not have misled him. I must return to the matter below, but I confess that at this point I really do begin to wonder whether we are in the real world or, once again (see R (Anufrijeva) v Secretary of State for the Home Department [2003] UKHL 36, [2003] 3 WLR 252, paras [14] and [28]), in the world of Franz Kafka.

I pause at this point to take stock.

The picture painted by all this evidence is very far from reassuring. It is, in fact, extremely disquieting, and disquieting whether one focuses on the particular case or surveys the wider landscape.

Let me spell out what the law requires. It is, as Rose LJ in ex p Duggan and Lord Woolf MR in ex p McAvoy put it in the passages I have already set out in full, that the prisoner should see “the gist of the reports” (emphasis added): not the gist of the Governor or Deputy Governor’s overall assessment or recommendation, but the gist of the reports – all of them. It is, as Rose LJ put it in ex p Duggan, that the prisoner is entitled to see “the gist of any matter of fact and/or opinion” (emphasis added) that is relevant to his categorisation.

Why does the law require this? It is, as Lord Mustill in Doody and Rose LJ in ex p Creamer and Scholey explained, because otherwise the prisoner will not be able to make worthwhile representations. A gist which, like the gist in the present case, entirely omits – conceals and suppresses – what from the perspective of the prisoner is the vital information that the adverse views on him are not unanimous, that some of the views are indeed decidedly favourable to him, in my judgment falls far short of what both the law and elementary principles of fairness require. Worse than that: it goes a long way to depriving him of any meaningful ability to make worthwhile representations. No more than a moment’s reflection is needed to see that the kind of representations someone in the claimant’s situation would make if given a proper gist will differ markedly from those he will make if given only what the claimant had here.

The defects in the gist in the present case are more serious and pervasive than Ms Lewis is grudgingly prepared to acknowledge. The cause of the problem is not, as she tries to suggest, merely some defect in the drafting caused by the use of the word “reports” in the plural, unsatisfactory though the seemingly inveterate and I suspect deliberately obfuscating use of that word undoubtedly is. The defect cannot be remedied merely by re–drafting the final sentence of the penultimate paragraph of the gist in the manner she suggests. It is not enough for Ms Steyn to protest, even if it be the literal truth, that the gist did not convey the views expressed by three individuals whilst ignoring the views of two others, all of equal weight, and that the gist did accurately convey the overall prison recommendation. The essential vice in the gist is that it failed to reveal the divergence of opinion and that it failed to provide the gist of each of those divergent views.

Ms Kaufmann submits that the gist did not accurately reflect the contents of the reports. I agree. She says that it was not apparent from the gist that whilst some report writers did not recommend downgrading others did. In particular, she says, the words “Reports state that there is no evidence through offence related work or otherwise that your security category should be downgraded at this time” gave the clear impression that no report writers supported downgrading. I agree. She complains that the inconsistencies between the reports and the gist were both significantly misleading and wrong. Both complaints, in my judgment, are amply made out. The claimant has been treated unfairly: he has been wronged.

Proper compliance with the principles to be found in ex p Duggan and ex p McAvoy requires as a minimum that a gist

As the Ombudsman said, and I entirely agree, a gist should provide “clear and unambiguous information”. Indeed in the present case I can see no reason why the gist could not have contained – and in my judgment it should have contained, if not by way of direct quotation then at least in paraphrase – the key information as I have set it out in paragraphs [20]–[24] above.

Lest there be any misunderstanding I should make clear that this will not in any way be affected, as Ms Lewis seems to be suggesting, by the fact that PSO 1010 has introduced a new format for Category A reports. I agree with Ms Rensten. The point does not turn on whether such a report contains, as hitherto, a “recommendation” or, as in future, only “comments”. At the risk of repetition I emphasise that what is required is “the gist of any matter of fact and/or opinion”, and that this applies as much to comments as to recommendations. We are after all concerned here with fairness not semantics, substance not form.

I recognise of course that public interest immunity on a contents basis properly attaches to the, albeit limited, categories of information referred to by Rose LJ in ex p Duggan. And I accept that, as Ms Lewis asserts, there can be cases where reference to differing opinions may risk identifying individuals in circumstances where there is some legitimate reason for saying that they should not be identified. But such cases will, I imagine, be rare. Normally the practice should be as I have indicated. If it is to be said in any particular case that disclosure needs to be limited to stating the prison’s recommendation or to stating the bare fact that views are not unanimous then, as it seems to me, the gist should make explicitly clear that it is so limited.

Ms Lewis describes what happened in this case as an “unfortunate error” and a “slip”. I should like to be able to agree but confess to having some difficulty. The evidence (much of which, as I have already observed, the Secretary of State has not sought to controvert) demonstrates that Category A reports have become standardised in format and formulaic in their terms; that they very rarely disclose any differences of opinion; and that the use of phrases such as “reports state” and “reports advise” is standard. Ms Steyn told me on instructions that the drafting of Category A reports is done by a very small group of people in the Category A Review Team: six in all. I find it hard to imagine that anyone in a team as small and expert as this would have made such a “slip” if it really was the recognised and acknowledged practice in a case such as this to make clear in the gist that the views expressed were not unanimous.

I am sorry to have to say this but to put the point bluntly, in the light of all the materials I have seen I cannot share Ms Lewis’s confidence that the present gist system as a whole is operating satisfactorily let alone fairly. It may be that it is. I am not in a position to say that it is not. But the picture revealed in this case does little to encourage the conviction that it is. Nor does the Ombudsman’s report.

There is one final point that perhaps needs to be made in this context. It is not just the judiciary and the public at large who need to have confidence in the system. Prisoners also – prisoners in particular – need to have confidence in the system. How can the claimant have any confidence in the system, particularly in the light of his previous experience of officer B’s parole report, when he now discovers – and that only by means of litigation – that the gist is as defective as it can now be seen to be? How can others in his situation have any confidence? He and they could be forgiven for thinking (to adopt the Ombudsman’s words) that over–rigid adherence to format has come at the expense of credible content. And what does that do for prisoner/staff relations in the already fraught and difficult world of Category A prisoners? There is no room here for complacency.

I appreciate that the claimant has sought relief only under the 1998 Act. But his claim was of course formulated before the defects in the gist had become apparent. In my judgment the claimant is in principle entitled to relief even if his claim under the 1998 Act wholly fails. I shall return to consider the precise form of that relief below.

I turn to the claim under the 1998 Act.

I need not go through the complicated provisions of the Act in any detail. It is common ground that, in each case within the meaning of the various definitions in section 1(1) of the Act, the claimant is a “data subject”, the Secretary of State is a “data controller” and the Category A reports are both “data” and “personal data”. It is also common ground that, subject to any other relevant provisions of the Act, the claimant is entitled by virtue of section 7(1)(c) of the Act:

“to have communicated to him in an intelligible form –

(i)

the information constituting any personal data of which [he] is the data subject, and

(ii)

any information available to the data controller as to the source of those data”.

Section 8(2) provides, as a general principle, that the obligation under section 7(1)(c)(i) is to supply the data subject “with a copy of the information in permanent form”.

Section 27(5) provides that, except as provided by Part IV of the Act, section 7:

“shall have effect notwithstanding any enactment or rule of law prohibiting or restricting the disclosure, or authorising the withholding, of information.”

Section 63(1) provides that the Act binds the Crown.

So, says Ms Kaufmann, and I agree, the Secretary of State cannot simply point to Doody, ex p Duggan, ex p McAvoy or Williams as providing, without more ado, an answer to the claimant’s claim under the Act.

Pursuant to sections 7(9) and 15(1) of the Act, the High Court and County Court are given power, if satisfied that a data controller has failed to comply with a request under section 7 in contravention of the section, to order him to comply with the request. That is the jurisdiction invoked by the claimant in the present case.

Four limitations on the claimant’s rights under section 7 of the Act have been the subject of argument in the present case. Before summarising the arguments it is convenient if I first identify the relevant statutory provisions.

First, there is section 29(1) of the Act which provides in material part that:

“Personal data processed for any of the following purposes—

(a)

the prevention or detection of crime,

(b)

the apprehension or prosecution of offenders, or

(c)

… ,

are exempt from … section 7 in any case to the extent to which the application of those provisions to the data would be likely to prejudice any of the matters mentioned in this subsection.”

The key question here is thus whether compliance with section 7 “would be likely to prejudice” either the prevention or detection of crime or the apprehension or prosecution of offenders.

Next there are sections 7(4)-(6) which provide that:

“(4)

Where a data controller cannot comply with the request without disclosing information relating to another individual who can be identified from that information, he is not obliged to comply with the request unless—

(a)

the other individual has consented to the disclosure of the information to the person making the request, or

(b)

it is reasonable in all the circumstances to comply with the request without the consent of the other individual.

(5)

In subsection (4) the reference to information relating to another individual includes a reference to information identifying that individual as the source of the information sought by the request; and that subsection is not to be construed as excusing a data controller from communicating so much of the information sought by the request as can be communicated without disclosing the identity of the other individual concerned, whether by the omission of names or other identifying particulars or otherwise.

(6)

In determining for the purposes of subsection (4)(b) whether it is reasonable in all the circumstances to comply with the request without the consent of the other individual concerned, regard shall be had, in particular, to—

(a)

any duty of confidentiality owed to the other individual,

(b)

any steps taken by the data controller with a view to seeking the consent of the other individual,

(c)

whether the other individual is capable of giving consent, and

(d)

any express refusal of consent by the other individual.”

The key question here is thus whether “it is reasonable in all the circumstances” for the Secretary of State to “comply with the request without the consent of the” prison officers and other persons who have made the Category A reports.

Thirdly, there is section 8(2)(a), which exonerates the data controller from what would otherwise be the obligation to supply “a copy of the information in permanent form” if:

“the supply of such a copy is not possible or would involve disproportionate effort”.

Finally, section 7(9) merely empowers but does not compel the court to grant relief:

“If a court is satisfied on the application of any person who has made a request under the foregoing provisions of this section that the data controller in question has failed to comply with the request in contravention of those provisions, the court may order him to comply with the request.”

I need not at this stage refer to any other provisions in the Act.

I shall have to consider each of these provisions in due course but two preliminary observations are perhaps appropriate. The first is that the Act gives effect to, and in significant measure adopts the language of, Council Directive 95/46/EC of 24 October 1995. In this connection I was referred to Campbell v MGN Ltd [2002] EWCA Civ 1373, [2003] 2 WLR 80 at p 105H (para [96]) where Lord Phillips of Worth Matravers MR said:

“In interpreting the Act it is appropriate to look to the Directive for assistance. The Act should, if possible, be interpreted in a manner that is consistent with the Directive. Furthermore, because the Act has, in large measure, adopted the wording of the Directive, it is not appropriate to look for the precision in the use of language that is usually to be expected from the parliamentary draftsman. A purposive approach to making sense of the provisions is called for.”

Counsel referred me to certain passages in the Directive. It contains no fewer than seventy-two recitals. The important recitals for present purposes are (42) and (43):

“(42)

… Member States may, in the interest of the data subject or so as to protect the rights and freedoms of others, restrict rights of access and information; …

(43)

… restrictions on the rights of access and information and on certain obligations of the controller may similarly be imposed by Member States in so far as they are necessary to safeguard, for example, national security, defence, public safety, or important economic or financial interests of a Member State or the Union, as well as criminal investigations and prosecutions … ”

For present purposes the most important provision in the Directive is Article 13(1) (exemptions and restrictions) which provides so far as material that:

“Member States may adopt legislative measures to restrict the scope of the obligations and rights provided for in [the Directive] when such a restriction constitutes a necessary measure to safeguard:

(a)

national security;

(b)

defence;

(c)

public security;

(d)

the prevention, investigation, detection and prosecution of criminal offences, or of breaches of ethics for regulated professions;

(e)

… ;

(f)

… ;

(g)

the protection of the data subject or of the rights and freedoms of others.”

Neither Ms Steyn nor Ms Kaufmann dissent from the proposition that I should construe these references to “the rights and freedoms of others” as embracing, though not of course confined to, those rights and freedoms conferred or recognised by the European Convention for the Protection of Human Rights and Fundamental Freedoms: cf the Directive, recitals (1), (10) and (37).

The other preliminary point is this. Different types of judicial activity are called for by the four provisions in the Act that I have mentioned.

I go first to section 29(1).

The effect of section 29(1), as we have seen, is that section 7 does not apply, and these are the crucial words, “in any case to the extent to which [disclosure] would be likely to prejudice” either the prevention or detection of crime or the apprehension or prosecution of offenders. The meaning of the two key phrases, “in any case” and “likely”, is thus crucial: central to Ms Kaufmann’s argument is the submission that the words “in any case” are to be understood as meaning “in the particular case”.

The meaning of these two phrases was considered in Equifax Europe Ltd v The Data Protection Registrar (Case DA/90 25/49/7), a decision of the Data Protection Tribunal (Deputy Chairman, Mr Aubrey L Diamond) dated 28 June 1991 which is reported in the Encyclopaedia of Data Protection, Vol IV, paras 6–099 et seq.

Equifax was an appeal by a credit reference agency against an enforcement notice issued against it by the Data Protection Registrar. It related to the provision by the agency, to those making searches, of so called “third party information” relating not to the subject of the search, the applicant for credit, but to persons who had lived at the same address as the applicant for credit. The enforcement notice prohibited the supply of such information. One of the grounds of appeal, in the event unsuccessful, was based on section 28(4) of the Data Protection Act 1984 – the predecessor of the 1998 Act – which exempted data from the Registrar’s powers “in any case in which [their exercise] would be likely to prejudice” either the prevention or detection of crime or the apprehension or prosecution of offenders. Section 28(1) of the 1984 Act, the ancestor of section 29(1) of the 1998 Act provided, in terms very similar to its successor, for the exemption from disclosure of data “in any case in which [disclosure] would be likely to prejudice” either the prevention or detection of crime or the apprehension or prosecution of offenders. In other words, the Tribunal was considering provisions – sections 28(1) and 28(4) of the 1984 Act – materially indistinguishable from the provision – section 29(1) of the 1998 Act – with which I am concerned.

In relation to the words “in any case” the Registrar argued, and more to the point the Tribunal accepted, that the words meant “in any particular case”. The agency argued and the Tribunal seemingly accepted (decision, para [30]) that one of the functions of a credit reference agency is the prevention of crime, because persons applying for credit may commit fraud and the searches performed by the credit reference agency may help to prevent or detect fraud. The agency further argued (decision, para [36]) that the exercise by the Registrar of his powers would be likely to prejudice this function because were searches to be restricted in the manner required by the Registrar “some cases of fraud or suspected fraud might not be exposed.” That argument was rejected by the Tribunal.

The Tribunal found (decision, para [34]) that “the vast majority of applicants for credit were not fraudulent” and that criminals formed only “a tiny proportion of applicants for credit.” The Tribunal continued (decision, para [38]):

“… Personal data are exempt from the provisions referred to “in any case in which the application of those provisions to the data would be likely to prejudice” the prevention or detection of crime or the apprehension of offenders. It seems to us that the words taken as a whole – “in any case in which the application of those provisions to the data” – make it plain that the exemption applies only in particular cases where we can talk about “the data,” that is to say the personal data to which we may or may not apply the provisions.”

I respectfully agree with the Tribunal’s analysis. In my judgment the words “in any case” in section 29(1) of the 1998 Act are to be read as meaning “in any particular case”, so that it is for the data controller, if he wishes to rely upon the exemption in section 29(1), to show that one of the statutory objectives is likely to be prejudiced in the particular case in which the question arises.

The Tribunal went on to consider the meaning of “likely” (decision, para [41]):

“It is clear that in any case where fraud is attempted the Registrar’s notice would be likely to prejudice the prevention of crime. In any such case, therefore, the personal data are exempt from the relevant provisions of … the Act and Equifax are free to process data untrammelled by the notice. But in our judgment section 28(4) does not prevent the Registrar from serving such an order in relation to all the cases – the vast majority – where no crime is, or is going to be, committed. The phraseology is “would be likely to prejudice” the matters referred to, not “might conceivably prejudice” those matters.”

The Tribunal’s approach to the meaning of the word “likely” is fully in accord with general principle. It is well known that “likely” has neither a single nor even a prima facie meaning; it may mean ‘more probable than not’, it may mean no more than ‘more than fanciful’, it may have some intermediate meaning: see In re H (Minors) (Sexual Abuse: Standard of Proof) [1996] AC 563 and Three Rivers District Council v Governor and Company of the Bank of England (No 4) [2002] EWCA Civ 1182, [2003] 1 WLR 210. The general principle was explained in the latter case by Chadwick LJ at p 221H (para [22]) when he said that

““likely” does not carry any necessary connotation of “more probable than not”. It is a word which takes its meaning from context. And where the context is a jurisdictional threshold to the exercise of a discretionary power, there may be good reason to suppose that the legislature - or the rule-making body, as the case may be - intended a modest threshold of probability.”

Thus in In re H (where the context was the establishment of the threshold in public law proceedings under Part IV of the Children Act 1989) the House of Lords treated the word “likely” in section 31(2)(a) of that Act as meaning that there had to be a real, a substantial rather than a merely speculative, possibility, a possibility that cannot sensibly be ignored. But in Three Rivers (where the issue in accordance with CPR rule 31.17(3)(a) was whether disclosure was “likely to support the case of the applicant or adversely affect the case of one of the other parties”) it was held that “likely” meant “may well”, a test which was recognised (see per Chadwick LJ at para [33]) as imposing

“a rather higher threshold of probability than merely “more than fanciful” [but] without reaching the threshold of “more probable than not”.”

Ms Steyn submits that, considered in context, it is plain that “likely” in section 29(1) does not mean “more probable than not”. She submits that Parliament’s intention was to exempt disclosure of information which is processed for the purpose of preventing or detecting crime if there is “a real risk” that such disclosure would prejudice those purposes. Parliament, she submits, cannot have intended that everything should turn in this context on the difference between a 51% and a 49% probability. She suggests that the Tribunal in Equifax did not expressly state what it considered the meaning of the word “likely” to be in this context and submits that the risk at issue in that case – namely the risk that fraud might not be exposed – could properly be described as negligible or merely perceptible. The credit reference agency was processing between fifteen and twenty million searches a year, of which only a “tiny proportion” related to fraudulent credit seekers. So, she says, the risk of prejudicing the purpose of detecting or preventing crime in any case where there was no particular reason to suspect fraud was so minute that it would not pass the “real risk” test.

I can go some, but not all, of the way with Ms Steyn. I accept that “likely” in section 29(1) does not mean more probable than not. But on the other hand, it must connote a significantly greater degree of probability than merely “more than fanciful”. A “real risk” is not enough. I cannot accept that the important rights intended to be conferred by section 7 are intended to be set at nought by something which measures up only to the minimal requirement of being real, tangible or identifiable rather than merely fanciful. Something much more significant and weighty than that is required. After all, the Directive, to which I must have regard in interpreting section 29(1), permits restrictions on the data subject’s right of access to information about himself only (to quote the language of recital (43)) “in so far as they are necessary to safeguard” or (to quote the language of Article 13(1)) “constitute a necessary measure to safeguard” the prevention and detection of crime (emphasis added). The test of necessity is a strict one. The interference with the rights conferred on the data subject must be proportionate to the reality as well as to the potential gravity of the public interests involved. It is for those who seek to assert the exemption in section 29(1) to bring themselves within it, and, moreover, to do so convincingly, not by mere assertion but by evidence that establishes the necessity contemplated by the Directive.

In my judgment “likely” in section 29(1) connotes a degree of probability where there is a very significant and weighty chance of prejudice to the identified public interests. The degree of risk must be such that there “may very well” be prejudice to those interests, even if the risk falls short of being more probable than not.

In the outcome the Tribunal in Equifax substituted for the Registrar’s original enforcement notice an amended notice, set out in the Encyclopaedia of Data Protection, Vol IV, paras 6–155 et seq, which in effect prohibited the supply of such third party information unless it was “reasonable to believe” that the third party had been residing concurrently with the subject at the same address and as a member of the same family as the subject in a single household: even that information could not be supplied if there was information in the possession of the agency “from which it is reasonable to believe that there is no financial connection between” the third party and the subject.

The issue in Equifax arose in essence out of data subjects’ claims that information about them should not be disclosed (to third parties), whereas in the present case the dispute arises in essence out of the data subject’s claim that information about him should be disclosed (to him). In other words, what was being asserted in Equifax was an obligation of non–disclosure, whilst what is being asserted in the present case is an obligation of disclosure. But that does not affect the usefulness of the decision.

In the present case, as we shall see, Ms Kaufmann contends for disclosure combined, where appropriate, with what she calls “targeted non–disclosure”. What is interesting for present purposes is that the solution adopted by the Tribunal in Equifax was in essence non–disclosure combined, to the extent that was appropriate, with what one might call “targeted disclosure”.

Ms Kaufmann’s argument in relation to section 29(1) can, I think, be summarised as follows.

First, she submits that section 29(1) exempts the Secretary of State from his obligation to make disclosure of Category A reports “in any case” only “to the extent to which” disclosure “would be likely to prejudice” either the prevention or detection of crime or the apprehension or prosecution of offenders. She accepts that Category A reports are, as Mr Sprunt asserts, invariably produced for the purpose of preventing crime and that they may have the additional purpose of detecting crime – crime, that is, both inside and outside the prison. But, as she correctly points out, the determinative test under section 29(1) is not whether the data is prepared for the purpose of preventing or detecting crime but rather whether its disclosure would be likely to prejudice those ends, and that, she says, is a matter of fact to be ascertained by the court objectively in the light of all the evidence. Furthermore, as she emphasises, the data controller is exempt from the duty to disclose “only to the extent” necessary to prevent a disclosure that would otherwise be “likely” to have that effect.

Secondly, she characterises the Secretary of State’s case – fairly as it seems to me – as amounting to this: that the policy of non–disclosure of Category A reports is necessary in every case because anything less would be likely to prejudice the detection and prevention of crime in some cases.

But, thirdly, as she points out, relying for this purpose on Equifax, section 29(1) requires that the issue of whether disclosure is likely to prejudice either the prevention or detection of crime or the apprehension or prosecution of offenders has to be determined in relation to the particular and individual case in which disclosure is being sought. So the question of whether the Secretary of State can rely upon section 29(1) as against the claimant has to be determined in relation to the claimant and the claimant’s Category A reports. The exemption applies only where, and to the extent that, disclosure is likely, in the particular case in relation to which disclosure is sought, to prejudice the prevention or detection of crime.

Putting the same point rather differently, she submits that the effect of the language of section 29(1) is to prohibit a general or blanket policy of non–disclosure of a class of documents – in the present case, all Category A reports – unless the class of documents is such that disclosure in each and every case is likely to prejudice one or other of the enumerated objects, that is, either the prevention or detection of crime or the apprehension or prosecution of offenders. And as she points out, even Mr Sprunt is not prepared to put the factual case that high. For, as we have already seen, he accepts in his evidence (see paragraph [37] above) that

“there may be instances where reports on Category A prisoners, when looked at individually, would appear to be innocuous and unlikely to prejudice the prevention or detection of crime.”

Finally, on this limb of her argument, Ms Kaufmann invites me, having examined the Category A reports (which she, of course, has not seen) to find that in this particular case their disclosure to the claimant is not likely to prejudice either the prevention or detection of crime. She speculates that, judging from the content of Mr Sprunt’s statement, there is no reason to anticipate that there is anything in the claimant’s Category A reports disclosure of which would be likely to prejudice either the prevention or detection of crime and no reason to anticipate that there is anything in them which distinguishes them in any relevant respect from the parole reports that have already been disclosed to the claimant in full, unredacted form.

This is a formidable and compelling argument. How does Ms Steyn seek to meet it? In essence her argument proceeds as follows.

She accepts that section 29(1) requires that the issue of whether disclosure is likely to prejudice the prevention or detection of crime has to be determined in relation to the particular and individual case in which disclosure is being sought. But, she submits, this does not mean that one should ignore the consequential effect that disclosure in this case may have in other cases.

Pointing to Mr Sprunt’s evidence (see paragraph [37] above) Ms Steyn submits that there is a real risk that full disclosure of the Category A reports in this case may have the following consequential effects, all of them, she says, likely to prejudice the prevention or detection of crime:

The fact that the Secretary of State has disclosed to the claimant all the information that it is necessary to disclose in the interests of fairness is, she says, a “crucial element” in applying the section 29(1) exemption. (This is not in fact the case, as I have already found, but Ms Steyn’s real point is that the claimant does not have to have recourse to the 1998 Act to obtain a gist which meets the ex p Duggan and ex p McAvoy requirements and therefore, by definition, meets the requirements of fairness. There are other avenues open to him for obtaining redress.)

Finally, she says that, save in respect of the single offending sentence, the Secretary of State has, in the gist, fairly and accurately communicated to the claimant all the information to which the claimant is entitled pursuant to section 7(1) to the extent that disclosure of that information would not be likely to prejudice the purpose of preventing or detecting crime and so does not fall within the section 29(1) exemption.

Ms Kaufmann of course disputes the fairness and adequacy of the gist. Her response to Ms Steyn’s other arguments is fourfold.

First, she says that the test under section 29(1) is not a test of fairness. It is a different and more stringent test, so it is neither here nor there that the disclosure she is seeking goes beyond that required by ex p Duggan and ex p McAvoy. In this connection she understandably points also to section 27(5).

Secondly, she submits that Category A prisoners are not a homogenous group. Whilst all Category A prisoners (assuming that they are correctly categorised) share the characteristic of being highly dangerous if they escape, the nature of the danger they pose to prison service staff is, she says, far from uniform: contrast, for example, on the one hand the sexual offender who poses a risk of an exclusively sexual kind and only to women or children, and who has neither the resources nor the inclination to try to escape, and, on the other hand, the terrorist, drug trafficker or armed robber who may well have confederates willing to assist and prepared to use lethal force in facilitating an escape. All may fall within Category A, irrespective of the likelihood of escape, for as PSO 1010 makes clear (see paragraph [2] above) categorisation as a Category A prisoner depends on the level of dangerousness the prisoner would present if unlawfully at large “regardless of the likelihood of escape”. But some Category A prisoners – for example the sexual offender – whilst highly dangerous if at large may pose no real threat at all to prison staff. The heterogeneous composition of the general class of Category A prisoners is, says Ms Kaufmann, highly significant when considering the validity of Mr Sprunt’s arguments based on frankness and the need not to prejudice the efficacy of the Category A review system, because it is only in respect of one sub–class that disclosure might impact adversely on the detection or prevention of crime: those Category A prisoners whose dangerousness is liable to manifest itself in attacks on, threats to or intimidation of staff. A targeted form of non–disclosure will, she says, properly protect report writers from the risks presented by the sub–class who do pose a threat and equally protect the integrity of the Category A review system.

Next, Ms Kaufmann points to the procedure in parole reviews which, as I have already described it, provides (both in those cases governed by the Parole Board Rules and in those cases governed by PSO 6000) for full disclosure as the general rule but subject in any individual case to what she calls “targeted non–disclosure” on certain specified but limited grounds. These procedures provide in each case, she says, for targeted non–disclosure where there is particular material the disclosure of which in the particular case might, inter alia, prejudice the prevention of crime. Plainly, full and frank disclosure is just as important if the Parole Board is to make a proper assessment as it is for the Category A Review Team and Committee, but there is nothing to show, she says, that the parole review systems are not operating efficaciously in the case of Category A prisoners. Indeed, as she points out, the Secretary of State has not even sought to assert that the procedure in parole cases jeopardises full and frank reporting, even when it is operating, as it is in the case of Category A prisoners, in respect of the very same class of prisoner and substantially the same body of report writers. The Secretary of State has in fact provided no evidence at all to suggest that the parole review process lacks integrity or efficacy because report writers are too fearful to be frank. In fine, she submits, the procedure in parole reviews presents insuperable difficulties for any attempt by the Secretary of State to justify the present Category A review system. The Secretary of State’s policy is over–inclusive in seeking to protect the efficacy of the Category A review system by a general policy of non–disclosure. A targeted measure will, she submits, plainly secure the necessary degree of frankness and candour. There is no reason why the Secretary of State’s concerns as articulated by Mr Sprunt cannot be fully met by a system of targeted non-disclosure, just as it is in the case of parole reviews. On the contrary, she says, the parole review processes show that there is no real risk at all of the kind that alone would bring section 29(1) into play. In the light of the way in which the parole review system can be seen to operate, the court, she says, is well able to evaluate the situation and can be confident that Mr Sprunt’s fears are unfounded insofar as they suggest the need for anything more than targeted non–disclosure.

So far as concerns Mr Sprunt’s other concern – the risk to prison security – Ms Kaufmann accepts of course that public interest immunity on a contents basis attaches to the categories of information referred to by Rose LJ in ex p Duggan (see paragraph [40] above) and that, consistently with section 29(1), it would be proper for the Secretary of State to withhold in any particular case information which would lead to the identification of informants or impinge upon escape risk, for example by revealing ways in which prison security may be vulnerable or by revealing counter–measures to protect security such as surveillance and monitoring techniques. Such information is already properly withheld as part of the gisting process, as also in the context of parole reviews, although the fact that it is being withheld is something which in either of those contexts the prisoner will either know or be able easily to discover. The relevant need, she says, is to ensure that the prisoner does not learn about the methods employed by the authorities. The detection and prevention of crime are prejudiced only by disclosure of the material. These objectives are not prejudiced by telling the prisoner that material is being withheld because it relates to surveillance or monitoring. The prisoner will be none the wiser if he is told this: it will come as no surprise to him that techniques are deployed for the purpose of surveillance and monitoring, but he will not know from that fact alone either what those techniques may be nor necessarily what, if anything, they may have uncovered. The prisoner’s knowledge that certain unspecified information is being withheld on this ground cannot, says Ms Kaufmann, cause any prejudice to the detection or prevention of crime. There is, she says, no reason why what is in effect a system of targeted non–disclosure on security grounds in the context of the gist system cannot equally and effectively be operated in the context of section 29(1). She points again in this context to the parole review procedures, submitting that if a general ban on full disclosure is not necessary to protect security in relation to parole reviews, then it equally cannot be in relation to Category A reviews. Accordingly, the concerns expressed by Mr Sprunt cannot, she says, afford any basis for a more extensive form of non–disclosure than one targeted to the individual case, just as the gisting and parole review procedures are currently targeted.

In response, Ms Steyn points to the differences between the parole review procedure and the Category A review procedure. She asserts, though Ms Kaufmann disputes, that parole review reports are less detailed than Category A reports; that the issue for the Parole Board – liberty or continuing detention – is different from the issue for the Category A Review Team – detention as a Category A prisoner or detention as a Category B prisoner; and that a ‘toned–down’ parole review report has, as Williams shows, different, and less serious, consequences than a toned–down Category A report, so that there is, as she would have it, less need of frankness in a parole review report.

I can go some of the way with Ms Steyn. Thus, as was recognised both in ex p McAvoy and in Williams, there clearly are differences between the decision–making processes as they affect the Parole Board and the Category A Committee. After all, Ms Kaufmann herself accepts that, although Category A reports are similar to parole review reports inasmuch as the issue addressed in both is public safety, and both are concerned to assess the character and personality of the prisoner and the risk to public safety he presents, there is one significant – and, as it seems to me, very important – difference. In the case of the Parole Board the assessment is made against the assumption of a controlled release into the community under supervision whilst the Category A review falls to be determined against the assumption that the prisoner has escaped, is unlawfully at large and is therefore, by definition, unsupervised. As Judge LJ said in Williams ([2002] 1 WLR 2264, paras [25]–[27]):

“[25] The critical difference between these two decision-making processes in cases involving discretionary life prisoners is readily identified. Release on licence is a formal step. It means what it says. The release can be and generally is made subject to supportive measures as well as stringent conditions such as supervision or treatment, or both. The released prisoner is normally subject to a measure of immediate and continuing control. In the event of non-compliance, the licence is revocable.

[26] The Category A Committee is concerned with the risks posed to the public by a prisoner who escapes, something which may occur unexpectedly, at any time. If he escapes, and while he remains at large, the prisoner is uncontrolled and unsupervised, temporarily, at least, untraceable, on the run, subject therefore to the inevitable increased stresses on an individual who, by definition, has not yet satisfied the panel that it would be safe for him to be released on licence.

[27] In summary, the panel is concerned with the protection of the public following a supervised conditional release of the prisoner, whereas the Category A Committee or review team concentrate on the risks to the public posed by an escape. This is a difference of substance. They address the same broad issue – public safety – but they are resolving a different problem ... ”

Moreover, I can accept that, although section 29(1) requires that the issue of whether disclosure is likely to prejudice the prevention or detection of crime has to be determined in relation to the particular and individual case in which disclosure is being sought, this does not mean that one can simply ignore the consequential effect that disclosure in the particular case may have in others.

But that said, Ms Steyn’s arguments do not, in my judgment, suffice to meet the substance of Ms Kaufmann’s submissions. On this point I accept the claimant’s case, essentially for all the reasons given by Ms Kaufmann.

The question of whether the Secretary of State can rely upon section 29(1) as against the claimant has to be determined in relation to the claimant and the claimant’s Category A reports and it is not determined against the claimant by the fact, if fact it be, that he has already been given a gist which, meeting the requirements in ex p Duggan and ex p McAvoy, therefore by definition meets the requirements of fairness. Fairness is not here the test.

The Secretary of State does not seek to make good his case by reference to anything peculiar to or specifically referable to the claimant: his claim is based on the asserted need, in order not to prejudice the legitimate section 29(1) objectives, to impose a general policy confining disclosure in effect to what is contained in gists prepared in accordance with ex p Duggan and ex p McAvoy. But in my judgment the reasons put forward by Mr Sprunt (and endorsed by Ms Lewis) in justification of that policy do not, whether weighed individually or taken together, and even when elaborated by Ms Steyn, suffice to support the weight of the argument sought to be erected upon them. Fundamentally, in my judgment, Mr Sprunt’s analysis breaks down for the reasons given by Ms Kaufmann. I have already summarised her key arguments (see paragraphs [117]–[119] above) and need not repeat them. I agree with Ms Kaufmann’s submissions. It is essentially for those reasons, as it seems to me, that in the final analysis the Secretary of State’s case here breaks down. Mr Sprunt’s and Ms Lewis’s evidence convincingly demonstrates the need for targeted non–disclosure in some cases: it does not, in my judgment, demonstrate that the more general or blanket policy of non–disclosure is needed to prevent the likelihood of prejudice to the section 29(1) objectives.

I emphasise that I am not saying that every Category A prisoner will in every case be entitled to see the full contents of his Category A reports. There will be cases – for all I know there may be many cases – in which the Secretary of State will be able to rely upon section 29(1) as justifying less than complete disclosure. All I am saying is that the Secretary of State’s present policy of blanket non–disclosure cannot be justified under section 29(1). What section 29(1) requires, and Ms Kaufmann of course accepts this, is a more selective and targeted approach to non–disclosure, based on the circumstances of the particular case.

Be that as it may, the Secretary of State cannot in this particular case, and having regard to all the evidence filed in this particular case, rely upon section 29(1) as an answer to the claimant’s claim.

I go next to section 7(4).

The effect of sections 7(4)(b) and 7(5) of the Act is not in dispute. Unless it is “reasonable in all the circumstances” for the Secretary of State to comply with the claimant’s request without the consent of the prison officers and other persons who have made the Category A reports, his obligation is to

“communicat[e] so much of the information sought … as can be communicated without disclosing the identity of the other individual[s] concerned, whether by the omission of names or other identifying particulars or otherwise.”

It is common ground that disclosure of the Category A reports as such would necessarily involve disclosure of information about the authors of those reports – namely their identities and the opinions they hold about the claimant – and that the authors of the reports have not consented to disclosure of this information. It is also common ground that, as sections 7(4) and 7(6) explicitly require, I must have regard to “all the circumstances” in determining what is reasonable.

Ms Steyn’s submissions can be summarised as follows.

First, she submits that, save in respect of the single offending sentence the Secretary of State has, in the gist, fairly and accurately communicated to the claimant so much of the information as can be communicated without disclosing the identities of the authors of the reports.

Secondly, she submits that mere redaction of the authors’ names would be wholly insufficient to protect the identities of the report writers, for the content of each report, it is said, clearly has the potential to lead to the identification of the source. It is likely to be obvious to a prisoner from the context, she says, whether a report is written by, for example, a psychologist, a probation officer, his personal officer or the wing manager, or by the person who interviewed him on a particular day. According to Ms Steyn, it is in the nature of Category A reports that disclosure of them even in redacted form will lead to the identification of their authors if the expressed opinions are not redacted.

Next she submits that, in the light of the matters set out in Mr Sprunt’s evidence (see paragraph [37] above), it would not be reasonable to make disclosure without the consent of the authors. In addition to the various points put forward in support of the corresponding argument in relation to section 29(1) (see paragraph [112] above) Ms Steyn draws attention to:

Fourthly, the Act seeks to protect not only the data subject (the subject of the reports) but also the source (the authors of the reports). Where, as here, the interests of the data subject and the source conflict, a balance has to be struck between the right of the data subject to have access to information about himself and the opposing right of the source not to have his identity as the author of the report disclosed. That balance is properly and fairly struck, says Ms Steyn, by disclosing the information in the anonymised form of the gist, which gives the claimant all the information he requires to enable him to make effective representations.

Finally, she submits, the fact that the Secretary of State has disclosed to the claimant all the information that it is necessary to disclose in the interests of fairness is a “crucial element” in considering the reasonableness of his decision not to disclose information that identifies, or is capable of identifying, other individuals. (In this connection I repeat the point I have already made: see paragraph [113] above.)

Ms Kaufmann in response submits that section 7(4)(b) requires the court to hold a proper balance between the “data subject” – here the claimant – and the “other individual[s]” – here the prison officers and other persons who have made the Category A reports. The question, she says, is whether it is reasonable not to disclose the information sought by the claimant in order to preserve confidentiality in the identities of the authors of the reports. The only basis, she says, upon which the Directive contemplates affording such protection is as provided by Article 13(1)(g), which legitimises such a restriction, as we have seen, only where it “constitutes a necessary measure to safeguard … the rights and freedoms of others.” Those rights and freedoms, she accepts, include those rights and freedoms conferred or recognised by, for example, Articles 2, 3 and 8 of the European Convention. But, she says, referring for this purpose to Campbell, the court should adopt a purposive approach and construe section 7(4)(b), consistently with the Directive, as justifying the kind of restriction contended for here only if it can be shown that the rights and freedoms of others will be interfered with, and then only if and to the extent that the restriction is, within the meaning of the Directive, a “necessary measure”. Section 7(1), she says, confers a right to disclosure. So the limits of the exception to disclosure provided for in section 7(4)(b) should be narrowly construed. The question whether it will be “reasonable” not to disclose the identity of another turns on whether the non–disclosure is, taking a reasonable view of all the circumstances and having regard in particular to the matters referred to in section 7(6), necessary in order to protect the rights and freedoms of that other individual. If it is not, then, she says, that is the end of the matter.

Ms Kaufmann points to the wording of both section 7(5) (“so much of the information sought by the request as can be communicated without disclosing the identity”) and, more particularly, of section 8(2) (“a copy of the information in permanent form”) as indicating that at most section 7(4)(b) justifies redaction, and not gisting, for a summary or a gist, she says, is not “a copy of the information”. This, she says, is not just a semantic quibble, for a summary or gist may very well lose some vital nuance or emphasis apparent in the original information. Moreover, impracticability or difficulty in providing the information required by section 7 is, she says, no reason for taking a restricted view of what is “reasonable”, for that would be to confuse the test under section 7(4)(b) with the quite different exemption under section 8(2)(a).

She points out that the “other individuals” here are in a particular class: they are officials in the employ or service of the data controller; the data the disclosure of which might lead to their identification comprises their expressions of opinion about the data subject; and those opinions have been expressed by them in the course of their employment by the data controller and for the purpose of assisting the data controller – the Secretary of State – in the discharge of his duty to classify prisoners under the relevant provisions of the Prison Rules 1999. She submits that the information is therefore not such as to afford the authors of the reports any claim to privacy or to protection on the ground that they are themselves data subjects and that there is such a close connection between the functions of the data controller and these other individuals that there is no basis to distinguish between them for the purpose of section 7(4). Alternatively, and in any event, she says, the authors of the reports have in these circumstances only a very weak balancing claim.

Moreover, she says, no duty of confidence is owed by the Secretary of State to the authors of Category A reports in relation to those reports. (She contrasts them, for example, with informants, where the circumstances would, she accepts, plainly give rise to a duty of confidence which would fall to be protected under Article 13(1)(g) and section 7(4)(b).) Despite the rubric which appears in PSO 1010 (see paragraph [11] above), Category A reports are, she says, written in the knowledge that there may – even if only exceptionally or in special circumstances – be occasions on which they will be disclosed in full: see PSO 1010, paragraph 4.1, and Williams (referred to in paragraphs [10] and [12] above). The practice in relation to parole review reports is inconsistent, she says, with the Secretary of State owing any duty of confidentiality to prison officer X in relation to a Category A report he prepares when there is – and is known by officer X to be – no confidentiality in the corresponding parole review report he writes about the very same prisoner. Finally, she submits that the existence of any duty of confidence in relation to the identities of the authors of Category A reports is negatived by the Divisional Court’s rejection in ex p Duggan of any class claim for non–disclosure on the grounds that disclosure would imperil candour and frankness.

Turning to the facts of the present case Ms Kaufmann points out that the Secretary of State has provided no evidence, and does not suggest that any exists, to support a finding that the claimant is liable to seek revenge by violence or threats in the event that he is able to identify the author of a Category A report who expressed negative views about him. Indeed, the fact that the claimant’s parole review reports were disclosed to him in full surely suggests quite the contrary. So, looking only to the claimant, there is, she says, nothing to suggest that disclosure is going to compromise the rights and freedoms of others.

The question, she says, then becomes whether “in all the circumstances” it is reasonable to apply to a case such as that of the claimant – who himself poses no threat to the rights and freedoms of others – a blanket policy of non–disclosure for the purpose of protecting the rights and freedoms of others in those other cases where they are liable to be threatened. At this point, as we have already seen (see paragraphs [112] and [134] above), the Secretary of State’s argument reflect those put forward in relation to the claim for exemption under section 29(1). Not surprisingly, Ms Kaufmann’s riposte closely follows her arguments in relation to section 29 (see paragraphs [117]–[119] above). I do need to repeat those of her arguments that I have already set out.

She submits that the proper balance between the legitimate interests of the prisoner and of the authors of the reports can be held by a system of targeted non–disclosure. In just the same way, a system of targeted non–disclosure is adequate to protect the rights and freedoms of informants. There is, she says, no need for a blanket prohibition. Not merely is there, she says, no evidence whatever to suggest that targeted non–disclosure will not adequately protect the rights and freedoms of the relevant other individuals – whether the authors of Category A reports or informants. On the contrary, she submits, the evidence overwhelmingly demonstrates that adequate protection can be secured by targeted non–disclosure.

In this respect Ms Kaufmann points to three things:

Finally, and in answer to the argument based on the approval in ex p Duggan and ex p McAvoy of the gist system as being one that produces fairness, Ms Kaufmann points out that the relevant question under section 7(4)(b), and the only relevant question, is whether non–disclosure is reasonably necessary to protect the rights and freedoms of others, and that question, she submits, is not concluded merely by demonstrating that what has been disclosed is “fair” to the data subject. She points again in this context to section 27(5), which, she says, precludes such an approach. And as she correctly points out, the court’s decision on ex p McAvoy was based in part quite explicitly on balancing the claims of the prisoner against the administrative convenience of the prison service; but that, she says, is not relevant to section 7(4)(b) and, to the extent that it is relevant at all, falls to be considered under section 8(2)(a).

Again, I can go some of the way with Ms Steyn. I do not necessarily accept that the authors of the Category A reports here are not themselves data subjects. I do not accept that they can simply be equiparated with the Secretary of State in the way Ms Kaufmann suggests. On the contrary, I accept that they have important privacy interests that have to be brought into the balance. Likewise, I accept that there will, historically, have been some expectation on their part of confidentiality, albeit coupled with a recognition that the confidentiality would not be absolute. Nor do I accept the distinction that Ms Kaufmann seeks to draw in the light of sections 7(5) and 8(2) between redaction and gisting.

But none of this really goes to the heart of Ms Kaufmann’s case, which otherwise I accept. In particular, I accept her approach to the construction of section 7(4)(b) in the light of the Directive, her analysis of the interrelationship between sections 7(1) and 7(4)(b) and her analysis of the interrelationship between the test under section 7(4)(b) and the wholly different test under ex p Duggan and ex p McAvoy.

Furthermore, and for the reasons she gives, I accept her argument that the proper balance called for by section 7(4)(b) between the legitimate interests of the prisoner and of the authors of the reports can be held by a system of targeted non–disclosure. The blanket policy of non–disclosure of anything which is not contained in the gist is not, in my judgment, a proportionate response to the undoubted problems and concerns identified by Mr Sprunt and Ms Lewis. The Secretary of State’s present blanket policy is not, in my judgment, a “necessary measure” to safeguard – in my judgment, and for all the reasons given by Ms Kaufmann, it goes significantly further than is necessary to safeguard – the interests of the authors of the reports. Their interests are undoubted and, I accept, very important. But so too are the interests of prisoners, for in the final analysis the issue here so far as concerns the prisoner is liberty. So far as concerns the author of a Category A report the issue in some cases may, in the final analysis, I readily accept, be bodily safety or even life itself. But those interests, on my analysis of the evidence, are engaged only in some cases – and not in the present case – whereas the prisoner’s liberty interests are in principle engaged in every case. Ms Kaufmann’s arguments as I have summarised them in paragraphs [141]–[144] above demonstrate powerfully, in my judgment, why adequate protection can be secured by targeted non–disclosure and why the present blanket policy of non–disclosure goes too far.

Again, I emphasise that I am not saying that every Category A prisoner will in every case be entitled to see the full contents of his Category A reports. There will be cases – there may be many cases – in which the Secretary of State will be able to rely upon section 7(4), just as there will be cases in which he will be able to rely upon section 29(1), as justifying less than complete disclosure. All I am saying is that the Secretary of State’s present policy of blanket non–disclosure cannot be justified under section 7(4), any more than it can be justified under section 29(1). What section 7(4)(b) requires, like section 29(1), is in this context a more selective and targeted approach to non–disclosure, based on the circumstances of the particular case.

The issues which arise under section 7(4)(b) are not the same as those which arise under section 29(1), and there is no necessary relationship between the outcomes which may emerge when applying the two different tests. It just so happens that in the present case, where the materials relied on by the Secretary of State in relation to both limbs of the argument are so very similar, the outcome is the same.

Be that as it may, the Secretary of State cannot in this particular case, and having regard to the evidence, rely upon section 7(4) as an answer to the claimant’s claim. The Secretary of State does not seek to make good his case by reference to anything peculiar to or specifically referable to the claimant: his claim is based on the assertion that section 7(4) justifies imposing a general policy confining disclosure in effect to what is contained in gists prepared in accordance with ex p Duggan and ex p McAvoy. Put that way his case fails.

I must next touch on section 8(2)(a).

Although there are references in her skeleton argument to the administrative burden of providing Category A reports in redacted form in those instances where redaction would otherwise be appropriate, I did not understand Ms Steyn to be relying on section 8(2)(a) as such. In this she was, I think, wise. Her comment was that painstakingly checking the whole of a prisoner’s Category A file to ensure that no material is disclosed that would be likely to lead to the identification of sources would involve disproportionate effort in circumstances where the gist of the information has already been communicated to the prisoner.

Ms Kaufmann’s riposte seems to me to hit the nail on the head. As she points out, to prepare a gist it is necessary for the person responsible to read, carefully and closely, all the reports and other material to be placed before the Category A Committee. In the course of that exercise he or she will have to identify any material that should not be disclosed even in gist form, for instance on grounds of public interest immunity. The additional effort involved in identifying any information that ought not to disclosed under the 1998 Act is, she says, and I agree, minimal. Alternatively, as she points out, the Secretary of State could – should, as she would have it – do away with the cumbersome and time consuming gisting procedure. Instead the person responsible for sifting the material to prepare the gist could instead use the time to go through the material and, if and to the extent necessary, redact it for the purposes of disclosure under the 1998 Act. That, as she says, would undoubtedly save time.

I agree with Ms Kaufmann’s submissions on this point. Section 8(2)(a) cannot justify withholding from the claimant information in the form in which he would otherwise be entitled to receive it. The administrative burden – light as it is here – has in any event to be assessed in the context of the significance of the information that is otherwise required to be disclosed. Section 8(2)(a) exonerates a data controller from “disproportionate” effort, but in determining what is proportionate one necessarily, as it seems to me, has to have regard to the intrinsic significance of the information whose disclosure is being sought and its importance for the data subject. Here the information goes in the final analysis to the claimant’s liberty. The claimant’s interests in having disclosed to him the information in the form in which sections 7(1) and 8(2) would otherwise require it to be supplied heavily outweigh any arguments that the Secretary of State can possibly put forward based on section 8(2)(a). In this context the Secretary of State cannot rely upon section 8(2)(a) as a reason for avoiding providing information that neither section 29(1) nor section 7(4) exonerates him from disclosing.

Finally, I must deal with section 7(9).

It is common ground, and rightly so, that section 7(9) confers a discretion on the court. As Wilson J said in P v Wozencroft (Expert Evidence: Data Protection) [2002] EWHC 1724 (Fam), [2002] 2 FLR 1118 at p 1129C:

“I remind myself, however, that under s 7(9) the claimant would have had to establish that the defendant had failed to comply with a request for disclosure in contravention of s 7(1), and, importantly, that, even in that event, the subsection confers upon the court a discretion as to whether to order the disclosure of such documents. I consider it of extreme significance that, even though s 7(1) speaks in terms of entitlement to disclosure on the part of the subject of data, the court is given a discretion, by the use of the word ‘may’ rather than any word such as ‘must’ or ‘shall’, as to whether to make the order.

It is also important to note that an analogous discretion is reflected in the terminology of s 14. As has been seen, s 14 is engaged only if the court is satisfied that personal data are inaccurate; and, even then, a discretion arises as to whether to order their rectification.”

The facts of that case were about as far removed from the facts of the present case as it is possible to imagine. Moreover, Wilson J exercised his discretion against the claimant on the basis that there was an alternative forum in which the claim could and should have been litigated and that the further proceedings were thus an abuse of the process of the court. Entirely appropriately, it might be thought, Wilson J did not even attempt to articulate any principles by reference to which this wholly unfettered statutory discretion ought to be exercised. His judgment therefore does not assist me at all in deciding how my discretion ought here to be exercised.

I was also referred in this context to an unreported judgment of His Honour Judge Zeidman QC sitting in the Edmonton County Court on 24 October 2002 in Durant v Financial Services Authority. That was a case in which the claimant sought disclosure under the Act from the FSA of its file relating to the claimant’s dealings with Barclays Bank. Save in relation to one document (disclosure of which was ordered in redacted form) the claim failed on other grounds, but Judge Zeidman QC nonetheless went on to consider how he would have exercised his discretion under section 7(9). He gave a number of reasons why, even if the claimant had otherwise established his case, he would nonetheless have exercised his discretion against him:

“First, I cannot see that the information could be of any practical value to the appellant. Secondly, the purpose of the legislation it seems to me is to ensure that records of an inaccurate nature are not kept about an individual. A citizen needs to know what the record says in order to have an opportunity of remedying an error or false information. In this case the appellant seeks disclosure not to correct an error but to fuel a separate collateral argument that he has either with Barclays bank or with the FSA, litigation which is in any event doomed to failure. I am entirely satisfied on the facts of the case that the FSA have acted at all times in good faith, and indeed there has been no suggestion to the contrary from the appellant; his argument is with Barclays bank not with the FSA.”

I do not of course dispute that there is a discretion, and I have no particular quarrel with the manner in which in that case Judge Zeidman would have exercised his discretion had the point arisen. But there is, I think, a danger that too much may be read into Judge Zeidman’s words. They are not to be read as, any more than they were intended to be, a gloss on the statute. The discretion conferred by section 7(9) is, as I have said, general and untrammelled. How it is properly exercised in one case may throw little if any light on how it should be exercised in another, and it may be very different, situation.

Ms Steyn accepts that what she calls the “error” in the gist in this case is a factor that the court may take into account in the exercise of its discretion under section 7(9). But she identifies four reasons why, as she would have it, I should exercise my discretion under section 7(9) against the claimant and in favour of the Secretary of State:

Ms Kaufmann’s riposte, in support of her assertion that discretion should be exercised in favour of the claimant, can be summarised as follows:

I agree with Ms Kaufmann. There is nothing here sufficiently weighty to justify the court exercising its discretion against the claimant or in favour of the Secretary of State. On the contrary, the very fact that the interests of the Secretary of State and the various other private and public interests he here represents have, ex hypothesi, been found to be insufficient to justify non–disclsoure under either section 29(1), section 7(4) or section 8(2)(a), points powerfully in favour of the court exercising its discretion in favour of the data subject, at least where, as here, it is the data subject’s liberty interests that are at stake.

I conclude therefore that the claimant has established his prima facie right to disclosure under section 7(1) of all the Category A reports, in full and unredacted; that the Secretary of State has failed to establish any basis for non–disclosure, whether under section 29(1), section 7(4) or section 8(2)(a); and that in all the circumstances I ought to exercise my discretion under section 7(9) in favour of the claimant.

The claimant is, therefore, entitled in principle to the relief he seeks. I will leave it to counsel to draft an appropriate form of order. I should add, though in the event the point no longer arises, that as I have already indicated (see paragraph [67] above) I would in any event have been minded to grant the claimant relief even if his application under the 1998 Act had wholly failed. In that event he would in principle, as it seems to me, have been entitled to a declaration to the effect that the gist did not meet the requirements of fairness mandated by ex p Duggan and ex p McAvoy and also, if necessary, a mandatory order requiring the Secretary of State to provide a gist in proper form.