Chiltern Farm Chemicals Ltd, R (On the Application Of) v The Health And Safety Executive

The Respondent, through its Chemicals Regulation Division (“the CRD”), has responsibility for regulating “plant protection products” including molluscicides, and is the competent authority for the United Kingdom under Regulation (EC) No 1107/2009 of the European Parliament and Council concerning the placing of plant protection products on the market (“the Regulation”).

The Appellant (“Chiltern”) has produced and marketed molluscicides in the form of slug pellets for commercial and domestic use for many years, having been authorised to do so under the predecessors to the Regulation. Under the Regulation, which has more stringent safety requirements, Chiltern was required to apply to the CRD for re-authorisation of those products. As part of its dossier in support of that application, it submitted a bird field monitoring study, which it considered provided data necessary to obtain re-authorisation (“the Prosser Study”). In the study, its slug pellets were applied to several fields in different locations in accordance with its authorisation, and in a standard manner, and adverse effects on tagged birds were noted.

Generally, the Regulation acknowledges that data obtained and thus owned by applicants as a result of studies, which often involve heavy expenditure, should be protected. However, it encourages applicants to cooperate and share tests and study reports; and, in respect of vertebrate animals, where agreement over such cooperation cannot be achieved, article 62 allows the competent authority to use test and study reports for the purposes of another applicant’s application, in return for that other applicant paying a fair share of the costs of the tests or study. Although the data are not shared, the benefit of the data is. These are known as the data sharing provisions.

On 29 January 2016, the CRD notified Chiltern that it considered the Prosser Study to be a vertebrate study that was subject to the data sharing provisions.

Chiltern, for which the decision had potentially substantial commercial consequences, commenced judicial review proceedings challenging that determination. Permission to proceed was granted; but following the substantive hearing, on 25 October 2017, His Honour Judge Saffman sitting as a Deputy High Court Judge refused the claim.

On 22 December 2017, Jackson LJ granted permission to appeal. On 12 January 2018, having been informed that several other applications would be considered by the end of February 2018 in respect of which the CRD will or may use the Prosser Study data thereby making the appeal empty, I ordered the appeal be expedited.

Before us, Tim Ward QC and Robert Palmer appeared for the Appellant, and Jonathan Lewis for the Respondent. At the outset, I thank each of them for his contribution to the debate.

The marketing of molluscicides has been regulated in the United Kingdom since the Control of Pesticides Regulations 1986 (SI 1986 No 1510); and, since 1991, by provisions emanating from Europe. EC Council Directive No 91/414/EEC (“the Directive”) and, since 14 June 2011, the Regulation have had the aim of harmonising the arrangements for the marketing authorisation of plant protection products within the EU and of ensuring they are safe to use. Recital (8) of the Regulation sets out its purpose “to ensure a high level of protection of both human and animal health and the environment and at the same time to safeguard the competitiveness of Community agriculture”. Recital (10) states that substances used in such products “are not expected to have any harmful effect on human or animal health…”, with recital (24) saying that, prior to granting authorisation, it should be demonstrated that plant protection products “do not have any harmful effect on human or animal health…”. Therefore, whilst recital (39) stresses the importance of generally protecting the property rights in test and study reports submitted by an applicant for the purposes of the authorisation of a product – so as to stimulate research – recital (40) requires rules to be laid down to avoid duplication of tests and studies on vertebrate animals.

Chapter 5 of the Regulation deals with “Data protection and data sharing”. Reflecting important property and confidentiality rights, article 59 provides that, if they are necessary for authorisation purposes and are certified as compliant with good laboratory or experimental practice, test and study reports concerning a plant protection product and any active ingredients shall have the benefit of data protection for ten years. That is subject to a number of exceptions, but the only relevant exception for the purposes of this appeal is that found in article 62.

Article 60 requires Member States to keep a list of test and study reports submitted and necessary for any authorisation process; and article 61 requires that list to be provided upon request to anyone intending to seek an authorisation of a plant protection product. Article 61(3) requires the relevant authorisation holder and prospective applicant to take all reasonable steps to share relevant data; but without a mechanism to require them to do so.

However, under article 62, where those parties fail to reach an agreement in respect of test and study reports on vertebrate animals, the competent authority can override the lack of consent, and, although it does not allow the data to be disclosed to the new applicant, it allows the authority to use the reports when considering a new application. Where it does so, article 62(6) entitles the authorisation holder, whose data they are, to claim from the new applicant “a fair share of the costs incurred by him” in obtaining the data (which, of course, may fall short of the value that the data may have to the new applicant). In the meantime, article 62(2) proscribes Member States from accepting duplicate vertebrate tests and studies.

These are the only provisions that allow the non-consensual use of an authorisation holder’s data for the purposes of an application by another person; and they only apply to test and study reports on vertebrates. “Test and study reports on vertebrate animals” is a phrase not defined in the Regulation. However, recital (40) refers to the restriction of tests on vertebrate animals being “in accordance with Council Directive 86/609/EEC of 24 November 1986 on the approximation of laws, regulations and administrative provisions of the Member States regarding the protection of animals used for experimental and other scientific purposes” (“the 1986 Directive”). This is confirmed in the CRD publication “The Applicant Guide: The Protection of Data” (“the CRD Guidance to Applicants”). Furthermore, each of two European Commission Health and Consumer Protection Directorate guidance documents, European Commission Guidance Document on Data Protection (SANCO/12576/2012 rev 1.1) and European Commission Questions and Answers on the Regulation (SANCO/12415/2013 rev 6), state that:

So far as material to this appeal, article 2 of the 1986 Directive defines “experiment” as:

I pause to note that the 1986 Directive has been superseded by EU Directive 2010/63/EUon the protection of animals used for scientific purposes; but (i) the 1986 Directive regime applies in this case and (ii) it is common ground that there would be no material difference in the outcome under the 2010 Directive in any event. I need not consider the 2010 Directive further.

The 1986 Directive was implemented in the United Kingdom by the Animals (Scientific Procedures) Act 1986 (“the 1986 Act”); and it is also common ground that sections 1 and 2 of that Act properly implemented the 1986 Directive by defining “a regulatory procedure” in terms that are directly equivalent to “experiment” in the Directive (and, thus, to “test and studies on vertebrate animals” in the Regulation).

Section 1(1) of the 1986 Act defines “protected animal” as:

Section 2, so far as material to this appeal, provides:

Two points are worthy of note at this stage.

It is clear from section 2(1) that “experimental procedure” is simply a subset of “scientific procedure”.

The exception incorporated into article 2 of the 1986 Directive, in respect of “non-experimental, agricultural or clinical veterinary practices” is not easy to construe on its face, because (a) the comma suggests that “non-experimental” does not govern “agricultural practice” and “animal husbandry practice”, and (b) it otherwise purports to define an exception to “experiment” in terms of itself, i.e. “non-experimental”. However, before us, it is common ground that section 2(8) of the 1986 Act properly implemented the exception, and I can therefore focus exclusively on section 2(8) without referring back to the 1986 Directive.

Section 3 of the 1986 Act prohibits any person from applying a regulated procedure unless linked personal, project and premises licences are in place. Contravention of section 3 is a criminal offence, with a maximum sentence on conviction of two years’ imprisonment (section 22(1)).

The relevant European methodology for the assessment of the risk of morbidity effect of plant protection products on vertebrates is set out in The European Food Safety Authority Guidance on the Risk Assessment for Birds and Mammals (EFSA Journal 2009; 7(12); 1438). It is helpfully, and uncontroversially, set out in the statement of Peter John Chapman dated 14 July 2017, from paragraph 10 onwards. Mr Chapman is the Director of Regulatory Affairs at JSC International Limited, consultants used by Chiltern.

The approach is two-tiered. The first step is a paper-based screening assessment, comprising a calculation based upon various worst-case assumptions, which are unrealistic in relation to normal conditions of use, e.g. that birds spend all their foraging time eating metaldehyde-containing slug pellets and they eat nothing else. If that calculation shows there is no risk, then no further steps are undertaken. However, if it shows a potential risk, then, if the calculation cannot be refined to exclude the risk, “higher tier data” are required from (e.g.) studies designed to assess the effects of the product on birds and mammals under field conditions. The purpose of such data is to show that, under particular conditions of use, the product is acceptably safe.

The factual background is comprehensively set out in the judgment of Judge Saffman. I can be relatively brief.

Chiltern was established in 1980, and since then it has manufactured slug pellets with the active substance metaldehyde in a cereal-based formulation. The pellets are spread over the surface of the soil. When ingested by slugs, they are fatal. However, the toxic properties of metaldehyde mean that, in certain circumstances, the pellets can also be harmful and even fatal to birds and small animals, to which they are a potential food source.

Under the Directive, authorisation was required before placing a plant protection product on the market. There were two stages to the process. First, active substances (such as metaldehyde) had to be approved. Approved substances were added to Annex 1 of the Directive which triggered a requirement for any plant protection products containing that substance already on the market to be re-authorised. During this process, plant protection products containing those ingredients could continue to be marketed.

The review of metaldehyde started in 2003. It was concluded, and metaldehyde included in Annex 1, on 1 June 2011. However, as I have indicated, the Regulation came into effect on 14 June 2011, with higher safety standards particularly in relation to minimising harm to vertebrate animals. Upon the advent of the Regulation, Chiltern’s products containing metaldehyde had to be re-authorised on the basis of the new standards, including those within article 62 of the Regulation.

Although the products had been marketed for some years, Chiltern was aware that, if its applications were to be successful, it would need further data in respect of their safety in relation to small wildlife. The first-tier safety assessment did not rule out the possibility of the products posing a risk to birds and mammals. Higher tier data were therefore required. Chiltern commissioned six ecotoxicology studies, including two field monitoring studies, one on mammals and one on birds. The collection of the data for the purposes of the applications for re-authorisation cost Chiltern about £3m, the Prosser Study alone costing £174,100.

The two monitoring studies were commissioned by Chiltern from the Centre for Chemical Safety and Stewardship within the Food and Environmental Research Agency (“FERA”), then an executive agency of the Department for Environment, Food and Rural Affairs. When the studies were initiated, FERA sought and obtained the necessary licences for the mammal study (which involved tagging by punching a hole in the ear of the subject animals), but not the bird study. The director for the bird study was Dr Phil Prosser, hence “the Prosser Study”.

The Prosser Study was higher tier. It involved the tagging of wild birds which happened to be present at the five sites that were the subject of the study, prior to seed drilling and a single slug pellet application. That application was at the recommended rate employing standard equipment normally used for the purpose; and then observation of the birds over a fixed period. It was the express aim of the study to adopt “normal agricultural practice in slug pellet application”; and, in Dr Prosser’s opinion, that aim was achieved (see paragraph 3.3 of the Prosser Study). Before Judge Saffman, it was common ground that the slug pellets were distributed and otherwise used both as authorised at the time and in accordance with “standard”, “normal” and (for the purposes of the 1986 Act) “recognised” agricultural practice. The Prosser Study was completed in April 2014. It was certified as being compliant with the relevant national and European standards of good laboratory practice.

Five other manufacturers of plant protection products containing metaldehyde proposed to apply for the re-authorisation of their products, and they asked the CRD for a list of vertebrate studies already submitted in other applications for metaldehyde-containing products. They were provided with a list, which included the Prosser Study.

On 30 October 2015, Chiltern’s solicitors wrote to the CRD asking whether they considered the Prosser Study fell within or outside the data sharing provisions; and, if within, on what basis. The CRD responded on 29 January 2016, indicating that they did consider the Prosser Study was subject to the data sharing provisions, giving two reasons.

First, it said that the study harmed the birds because “it involves the handling trapping and radio tagging of wild birds unfamiliar with that process”. However, before the application was heard by Judge Saffman, the CRD had accepted that the tagging etc involved in this case did not trigger the data sharing provisions, because it fell within section 2(5) of the 1986 Act (see paragraph 17 above). It has not since been suggested that this reason had any force.

The second reason was as follows:

The Prosser Study recognised that Chiltern’s products posed a risk to birds by consuming the slug pellets directly or by consuming slugs contaminated with metaldehyde from the pellets. Before Judge Saffman, as I have indicated, the CRD accepted that the study involved the use of Chiltern’s slug pellets in accordance with “recognised agricultural practice”; but, it was contended, it became a “regulated procedure”, subject to the data sharing provisions, because of the subjective intention or motive underlying the use. It was submitted that there is a material difference between, on the one hand, merely obtaining feedback from farmers who, in the ordinary course of their farming, have used the slug pellets in the field in accordance with the product’s authorisation and recognised agricultural practice, and then simply observed what happens; and applying slug pellets in exactly the same way, but for experimental or other scientific purposes. In the Prosser Study, the purpose of applying the pellets was experimental, namely to measure the risk of harm to birds. The study therefore fell within the definition of “regulated procedure” for the purposes of the 1986 Act.

It was submitted by Mr Ward, appearing for Chiltern before Judge Saffman as he did before us, that neither the product nor its manner of use in the Prosser Study was experimental. He accepted that, had the Prosser Study been conducted with a new, unauthorised product, then it would have been a recognised procedure under the 1986 Act. But the product was authorised, it had been applied to the relevant sites in accordance with standard agricultural practice, and is likely to have been applied in that way at those sites in any event. Recognised agricultural practice did not become experimental simply because there was subsequent monitoring of the bird population.

Judge Saffman accepted that the data sharing provisions of article 62 of the Regulation were an exception to, and a derogation from, the general principle of data protection in respect of studies; so that article 62 had to be construed narrowly. However, for the purposes of the scope of “recognised procedure” in the 1986 Act, he did not consider that the use of an authorised product could be distinguished from the use of a new product. He accepted, as was common ground, that the pellets in the study were applied as in recognised agricultural practice; but the unambiguous motive behind the application here (i.e. to obtain data about the risk posed to birds, under the auspices of a scientist) meant that “it was not in pursuance of standard agricultural practice” (see [113]). He concluded, at [119] of his judgment, that he did not see how the Prosser Study was “anything other than a scientific study of the birds”, being “a study of the effect (if any) of the pellets on birds conducted under the supervision of a scientist, which was monitored by a scientist and which contains a Statement of Compliance with Good Laboratory Practice”. In short, he preferred the submissions made on behalf of the CRD.

As will by now be apparent, this appeal turns upon a narrow point of statutory construction concerning the scope of “regulated procedure” for the purposes of section 2 of the 1986 Act, and particularly whether the use to which the plant protection product was put in the Prosser Study fell within that scope.

The primary definition of “regulated procedure” is found in section 2(1), i.e. “any experimental or other scientific procedure applied to a protected animal which may have the effect of causing that animal pain, suffering, distress or lasting harm”. In my view, the use to which the product was put in the Prosser Study clearly fell within that definition. As Mr Lewis submitted, “experiment” includes a scientific procedure undertaken to make a discovery, test a hypothesis or demonstrate a known fact. The premise upon which the Prosser Study was conducted was that “there may be risks to birds both from directly consuming slug pellets, and from consuming contaminated slugs and other invertebrates”, i.e. risks identified by a paper calculation as a first-tier assessment. The study was higher tier, designed to provide data that would assist in the assessment of that risk. Chiltern no doubt considered that, if used in accordance with its authorisation and standard agricultural practice, the product will not harm birds; but, at the very least, the study was designed to test that hypothesis. In determining whether this use to which the product was put in the Prosser Study fell within that primary definition, purpose was therefore relevant.

However, this definition of “regulated procedure” in section 2(1) is expressly “subject to the provision of this section”, including section 2(8). Section 2(8) provides that “references to scientific procedure do not include references to any recognised… agricultural… practice”. Mr Ward for Chiltern submitted that that exclusion from the primary definition does not, expressly or impliedly, incorporate any element of intention or purpose. I agree. Where, as here, the use of the product was a “recognised agricultural practice”, it falls within the exception. An enquiry as to whether the use was part of a scientific procedure, rather than in “the ordinary course of farming”, was neither required nor warranted. Judge Saffman correctly proceeded on the agreed basis that the use of the product was in accordance with recognised agricultural practice; but, in my respectful view, he then erred in proceeding to consider the purpose or motive for that use, and, having found that the purpose or motive was in pursuance of a scientific procedure, in concluding that it was a recognised procedure attracting the data sharing provisions of article 62.

In coming to that conclusion, I have particularly taken into account the following.

It was common ground before Judge Saffman that the use to which the slug pellets was put in the Prosser Study was a recognised agricultural practice. Before us, Mr Lewis fleetingly sought to suggest that, on the facts of this case, it was not; but that was a point not taken before Judge Saffman and, rightly, not ultimately pursued by Mr Lewis before us. The appeal therefore proceeded on the basis of the same consensus as there was below.

Judge Saffman concluded that the article 62 data sharing provisions are a derogation from the general rule of data protection in article 59, and therefore the article 62 should be interpreted strictly. That conclusion is not the subject of any challenge before this court. However, I am unconvinced that it is either helpful or necessary to rely on the principle of derogation in construing article 62. It is unnecessary because, in my view, the true construction of article 62 is unambiguous.

The construction suggested by Mr Lewis, and adopted by Judge Saffman, would define “scientific procedure” in terms that exclude any “recognised agricultural practice” but except from that exclusion any scientific procedure. It would thus define an exception to “scientific purpose” in terms of itself, which would rob section 2(8) of all sensible content. I am unconvinced that that is a legitimate construction, even though, on one view, it is arguable that the 1986 Directive defines “experiment” in terms of itself (see paragraph 18(ii) above).

On the basis of the construction Mr Lewis suggests, section 2(8) would, at best, be merely confirmation for the avoidance of doubt that “scientific procedure” in section 2(1) excludes recognised agricultural etc practices that are not part of any scientific procedure. I am unconvinced that section 2(8) was intended to have no substantive content: indeed, it was clearly intended to have some purpose. Its purpose was to exclude any recognised use of the product that fell within agricultural etc practice from the definition of “regulated procedure”, irrespective of the purpose or motive of that use.

It seems to me that that construction is both unambiguous, and entirely consistent with the aims of the Regulation. It is true that the Regulation enhanced the protection of vertebrates, with the recited aim of ensuring that no marketed plant protection product has harmful effects on animal health (see recitals (10) and (24), referred to in paragraph 8 above). However, all plant protection products pose some risk. Metaldehyde-containing products are required to be toxic to slugs and, dependent upon distribution/dose and other circumstances of use, pose some risk to vertebrate animals, including birds. The balance between the benefits of plant protection products and the risk to vertebrates is set out in article 62 read with the 1986 Directive as implemented in the 1986 Act. It is that which defines the margin of safety.

The concept of “recognised agricultural practice” within this part of the scheme is driven by the directions for use as authorised, the appropriate regulator assessing whether that risk is acceptable. Where a plant protection product is used in accordance with recognised agricultural practice, it is being used in circumstances in which the risk is currently considered to be acceptable.

It would be curious if monitoring an authorised plant protection product used in accordance with the authorised directions and usual, recognised agricultural practice either was prohibited (insofar as similar monitoring had been done before), or alternatively required a licence to perform. There would appear to be no justification for discouraging such practice. In my view, such monitoring is allowed without any such requirement.

Mr Lewis submitted that, once a specific product has been assessed as “safe” and consequently authorised, absent a change in standards, it is unlikely to benefit from further studies (paragraph 42 of his skeleton argument). However, I do not agree. Even where, on available material, a product has been assessed as safe when used in accordance with certain directions, when it is marketed it will be the subject of much greater usage in a much wider range of circumstances than can have occurred prior to authorisation. Problems that were not apparent previously may therefore exhibit themselves after marketing. Anecdotal reports may be inadequate to identify such problems. A responsible authorised person may therefore take the view that monitoring in accordance with authorised use will be beneficial. Where safety standards are raised, data from monitoring a product when it was subject to lower standards may still be helpful. For example, if a standard of safety is raised from X to Y, it will be helpful if monitoring against standard X shows that the product meets standard Y in any event. Simply because re-authorisation of a product will be assessed by higher safety standards than currently apply, does not change the scope of current recognised agricultural practice on which monitoring can be conducted. The results of the re-authorisation process may, of course, change recognised agricultural practice; but that is a different matter.

I do not accept Mr Lewis’s submission that, by using a plant protection product in accordance with recognised agricultural procedure but for the purpose of scientific monitoring, the risk to vertebrates is materially increased, because any risk is not counterbalanced by any benefit in terms of plant protection. Even if the application of the product to the fields that were the subject of the study had not been as part of active farming, the increase in risk was insignificant, given the very modest size of the experimental area monitored in the Prosser Study compared with the area spread with authorised metaldehyde-containing products as part of farming (over 700,000 hectares). But, in any event, the sites used in the Prosser Study were particularly selected as being due to be drilled with rape seed or cereal (see paragraph 2.6 of the Prosser Study); and, as part of the study protocol, the sites were in fact seed drilled prior to pellet application and monitoring. For obvious reasons, a field without being seed drilled would have been less attractive to birds, and may have thus biased the results. Mr Lewis’s criticism therefore has no force on the facts of this case.

The Prosser Study did not increase the risk to birds over and above that inherent in the use of the product during the course of authorised and recognised agricultural practice. That risk has been assessed as being acceptable; and is not made unacceptable by the prospect of a raising of that standard by the new provisions of the Regulation. Therefore, the interference with Chiltern’s property and confidentiality rights in the Prosser Study data inherent in any data sharing would necessarily have been disproportionate, and indeed unreasonable in the sense of legally perverse.

The various guidance that has been issued in relation to the Regulation is not of any real assistance. SANCO/12576/2012 rev 1.1 (which, with the other relevant guidance documents, is referred to in paragraph 12 above) unhelpfully says that “in the case of monitoring of birds and mammals in the fields, it is not very clear whether these constitute ‘tests and studies involving vertebrate animals’”. SANCO/12415/2013 rev 6 makes no reference to the issue. The CRD Guidance to Applicants states:

For those reasons, subject to the submissions of the parties as to the precise terms of the order, I would allow this appeal, quash the order of Judge Saffman and quash the determination of the CRD dated 29 January 2016 that the Prosser Study is a vertebrate study to which the data sharing provisions of article 62 of the Regulation apply.

I agree.

I also agree.