On 26 July 2011 The Telegraph reported that “Innocent people’s DNA profiles won’t be deleted after all, minister admits”. It said that
“police will retain DNA profiles in anonymised form, leaving open the possibility of connecting them up with people’s names, ministers have admitted”.
In S and Marper v United Kingdom [2008] ECHR 1581 the European Court of Human Rights held that indefinite retention by the police of fingerprints and DNA samples of two people who had been arrested but not convicted of criminal offences was a breach of their rights under Article 8 of the European Convention on Human Rights (overturning a decision upheld at each instance in the English courts).
The Protection of Freedoms Bill proposes, accordingly, to amend the Police and Criminal Evidence Act 1984 (“PACE”) so that – broadly – a lawfully taken DNA sample (and fingerprints) must be destroyed after three (or in some cases five) years if the suspect has not been convicted of an offence to which the sample relates (Genewatch have a helpful detailed explanation of the proposals).
The Telegraph article said that Home Office minister James Brokenshire “had won agreement from the [Information Commissioner’s Office] that the DNA profiles could be retained by forensic science laboratories”. The Information Commissioner’s Office (ICO) has now, following an FOI request for correspondence between his office and the Home Office about this matter, effectively said that, to quote Ben Goldacre, “I think you’ll find it’s a little bit more complicated than that”.
The complicating factor is that a DNA profile is different to a DNA sample, which in turn is different to the raw data derived from the sample. Christopher Graham, the Commissioner, in his evidence to the Public Bill Committee on the Protection of Freedoms Bill said
“Clause 13 [of the Bill] refers to the destruction of DNA profiles and that no copy must be retained by the police except in a form which does not include information which identifies the person to whom the DNA profile relates. It is assumed that this is aimed at addressing issues relating to the raw data, the electro-phoretogram, from which the DNA profile is created”.
Some existing DNA profiling systems process DNA samples in batches of up to 82 (or possibly 96 – I’m unclear which is the correct figure). In these processes it is not possible to isolate and destroy the raw data relating to a single sample without also destroying the whole batch data (which, of course, might contain raw data relating to samples of now-convicted-persons, which need to be retained).
Graham went on to say
“This provision [Clause 13 of the Bill] should be expressed in a way so it cannot be used to perpetuate such batch processing practices in any new systems used to generate DNA profiles and to require deletion of all the DNA profile information as the norm”.
One hopes this proposal is accepted. Even if it is, however, there will still remain a considerable number of batches of raw data derived from the samples of innocent people, and which it will not be possible to destroy. The question then arises as to what measures can be, and are being, taken to ensure that this remaining raw data cannot be linked to identifiable individuals. In response to my enquiries the IC’s office has said
“the Commissioner has stipulated that forensic science providers remove all the names and identifications from their systems to prevent them being able to link an individual with the ‘raw data’.”
But what confidence can we have that this will be sufficient? The IC’s office continues:
“The Commissioner is satisfied that the deletion of the associated records will remove the link between the identity of the individual and the ‘raw data’ which will be retained in the batch. This will effectively put the retained ‘raw data’ beyond practical use as it should be no longer possible to re-link the individual to the ‘raw data’ retained”.
There remains a lingering concern, however:
“given that the ‘raw data’ is used to create a DNA profile, and a DNA profile is unique to an individual, we are relying on the assurances we have been given and cannot say categorically that there is no possibility of the retained ‘raw data’ ever being linked to an individual.”
These assurances have to be balanced against the contents of a letter from James Brokenshire MP (the Home Office minister quoted in the original Telegraph article) to the joint chairs of the Protection of Freedoms Bill Committee . I’m not sure if this letter has been published yet, but was disclosed in response to my request. Brokenshire says
“Members of the Committee will be aware that most existing DNA records…will include the original barcode, which is used by the police and the FSS [Forensic Science Service] to track the sample and resulting profile through the system. It is therefore theoretically possible that a laboratory could identify an individual’s profile from the barcode, but only in conjunction with the force which took the original sample, by giving details of the barcode to the force and asking for the individual’s name”
This doesn’t strike me as a purely theoretical risk, and one might bear in mind that the FSS’s raison d’etre is to work with the police to detect crime by piecing together and analysing evidence.
Brokenshire explains, however
“Such conduct [i.e. trying to re-identify someone in these circumstances from residual DNA evidence], in clear breach of the requirements set out in the [Protection of Freedoms] Bill, would be likely to constitute offences of misconduct in public office and under the Data Protection Act. In addition, new section 63S of PACE (as inserted by clause 16 of the Bill) specifically excludes the use of such material in evidence or as any part of a criminal investigation.”
Given that both misconduct in public office and offences under the Data Protection Act 1998 can be countered in effect by defences of acting in the public interest, it seems to me that clause 16 might be the best assurance we have against any attempts to use any residual information from innocent people’s DNA samples.
informationrightsandwrongs 