Keeping the right people on the DNA Database

Published 10 August 2009

1. Consultation process and overall approach

We disagree with the piece-meal approach to legislation being taken, in which the power to set and amend retention times for samples and data in regulations is being given to the Home Secretary via an amendment to the Policing and Crime Bill; and the various governance structures and oversight mechanisms for the Database have not been put on a statutory basis.

The use of regulations and non-statutory bodies does not inspire public confidence because they may be subject to arbitrary future changes and ‘mission creep’, without full scrutiny by parliament or public debate. In our view, the destruction of samples and deletion of records should be enshrined in primary, not secondary, legislation.

There are also some serious omissions from the options outlined in the consultation, including the failure to revisit the need and process for deletion of linked Police National Computer (PNC) records; and failure to lay out in full the possible options for a more restrictive policy on sample collection, particularly for children.

Further, the consultation process has been hampered by the poor quality of the information provided in the Annexes, which suffer from substantial errors and omissions.

Finally, we remain concerned that ministers continue to regularly misinterpret and misrepresent statistics regarding DNA matches as well as the role of the DNA database in individual cases. In particular, the identification of Mark Dixie in the Bowman case occurred when his DNA was taken by police nine months after the murder following his arrest for a drunken brawl. It did not require any individual’s DNA profile to be stored on a database and the investigation would not be affected by the removal of innocent people’s DNA profiles from the database. The Consultation cites the Bowman family’s support for a universal DNA database, which they argue would have identified Dixie more quickly, since the DNA profiles of the whole UK population (and perhaps all visitors) would be retained. However, the Government is rightly not proposing to adopt this option. It is well aware that establishing a universal DNA database would have enormous resource implications, give rise to numerous false matches with crime scene DNA profiles (providing false leads for the police) and be open to abuse by anyone who could infiltrate the system and use it to track victims and their relatives and/or plant evidence at crime scenes. These problems would arise in addition to its adverse implications for human rights and the difficulties of collecting DNA without consent from the general population and from visitors. The use of the Bowman case to support the Government’s position is both misleading and offensive, since it falsely implies that the Government supports the family’s views and that critics of the Government’s position (including the European Court) have not considered the needs of victims and their families.

The database of individuals’ DNA profiles also plays no role in exonerating the innocent, who carry their DNA with them at all times and do not need it to be stored on a database in order for it to be compared with a DNA profile from a crime for which they may have been falsely accused.  The NDNAD provides match reports to the police, which list potential suspects, not lists of the millions of individuals on the database who do not match the crime scene profile.

2. Sample destruction

We strongly welcome the proposals in the consultation to destroy stored DNA samples from individuals for all existing and future cases. We note that samples are not needed for identification purposes once a computerised DNA profile has been obtained; that the storage of samples raises serious privacy concerns since they contain unlimited genetic information; and that stored samples have in the past been accessed in order to undertake controversial genetic research without consent. There is no justification for storing or using samples in this way, whether the individual whose DNA is used has been convicted of an offence or not.  

The six month retention time proposed is more than ample for quality assurance purposes. Any proposals to upgrade the DNA profiling system in the future are best applied prospectively, rather than retrospectively, due to the large number of stored samples involved, which makes retrospective profiling of all individuals’ samples technically and financially unfeasible. Additional samples can always be taken from a suspect who is appearing in court if confirmation or upgrading of their profile is needed for evidential purposes.

We note that the cost calculations provided on page 52 are incorrect, as a ridiculously large cost for sample destruction has been wrongly included. In practice, sample destruction would save millions of pounds a year in refrigeration costs (a saving of 90p per year per sample means a saving of £4.5 million over one year alone). There are also considerable practical advantages in destroying all stored samples, rather than seeking to track back those of unconvicted persons individually.

3. Retention of profiles from arrested persons

We disagree with the Home Office’s proposals for the retention of DNA profiles. The timeframes for retention are disproportionate, both for unconvicted persons and for persons cautioned or convicted for minor offences. Both the ECtHR judgment and the available data on re-offending and DNA detections have been misinterpreted or ignored and there are major errors in the impact assessment included in Annex D of the consultation document. The downsides of expanding the Database have also been neglected.

Errors in the impact assessment

The Consultation Annexes contain significant errors. In addition to the major error in sample destruction costs noted above, the main problems appear to be:

1. Failure to distinguish between different types of DNA detection, some of which do not require the database;
2. Confusion between matches and detections;
3. Failure to recognise that subsequent re-arrests of unconvicted persons will reduce any lost DNA detections;
4. The false assumption that all individuals with a record on the Database are equally likely to commit an offence (based on the research commissioned from the Jill Dando Institute);
5. The failure to distinguish between crimes for which DNA evidence is relevant and the majority of crimes (over 99%) for which no crime scene DNA is likely to be available.
6. The failure to validate the model using existing data, which shows that the number of detections is driven primarily by the number of crime scene DNA profiles loaded and is not correlated with the number of individuals’ profiles retained.

The failure to validate the model may be significant. GeneWatch UK’s preliminary estimates (see Annex) suggest that the consultation estimates of ‘lost’ detections could be one or two orders of magnitude in error. GeneWatch’s calculations are preliminary and are intended to highlight errors and omissions in the information provided in the consultation, rather than to provide final definitive estimates. We recommend that these errors and omissions are addressed urgently in order to provide a more robust evidence base for the relevant policy options.

Downsides of expanding the Database

‘Cold hit’ DNA detections, as opposed to detections where the suspect has been identified by other means, raise significant issues about potential errors. In effect, such detections reverse the burden of proof, so that the suspect has to demonstrate they were not the perpetrator of the crime.

Increasing the size of the DNA Database increases the expected numbers of false matches between crime scene DNA profiles and individuals’ DNA profiles and thus increases the likelihood of false accusations (which may waste police time and be traumatic for the person wrongly identified). False matches could also lead to miscarriages of justice.  

Within the UK, false matches between two full DNA profiles are expected to be rare, however the likelihood of false matches increases if the suspect is related to the perpetrator (a situation that has become significantly more likely as the database expands, and may be a particular risk for young black men, a large proportion of whom are on the database). The likelihood of false matches also increases if the DNA profile from the crime scene is not complete, a situation which occurs frequently because crime scene DNA may become degraded. As a result, the National DNA Database Annual Report 2005/06 states that between May 2001 and April 2006, 50,434 matches with crime scene profiles, or 27.6% of the total number of match reports, involved a list of potential suspects, not a single suspect, being given to the police, because matches with multiple records on the NDNAD were made. A reported 83.7 multiple matches a month were made over the 12 months from May 2008 to April 2009.1
 
The number of false matches that occur simply by chance (rather than laboratory errors etc.) is expected to increase significantly when the routine sharing of DNA profile matches across the EU begins in about two years’ time. The large size of the National DNA Database means that a much greater number of false matches is expected than for any other country.2 The number of false matches that occur by chance is proportionate to the size of the Database and thus one mechanism to reduce these would be to restrict profile retention to a smaller number of people who genuinely pose a danger to the public.

The Impact Assessment wrongly fails to include an assessment of the impact of the size of the NDNAD on likely false matches and false leads and hence on wasted police time, unnecessary questioning and potential extradition of suspects and potential miscarriages of justice. Such an assessment should focus on the impacts of DNA match sharing across the EU, in view of the large number of false matches expected to arise when this commences. In addition, the impact of relatedness should be reassessed, particularly in view of the large proportion of black men who now have records retained on the Database.

The Marper judgment

The judgment of the European Court of Human Rights recognises that, although DNA profiles contain less information than samples, they nevertheless contain “substantial amounts of unique personal data” and that their retention interferes with the right to private life. Destruction of samples alone is therefore insufficient to meet the Court’s requirements.

Regarding the retention of DNA profiles, the judgment notes that the position of Scotland is “notably consistent with Committee of Ministers’ Recommendation R(92)1, which stresses the need for an approach which discriminates between different kinds of cases and for the application of strictly defined storage periods for data, even in more serious cases”. The relevant part of this Recommendation3 states:
“Measures should be taken to ensure that the results of DNA analysis and the information so derived is deleted when it is no longer necessary to keep it for the purposes for which it was used. The results of DNA analysis and the information so derived may, however, be retained where the individual concerned has been convicted of serious offences against the life, integrity or security of persons. In such cases strict storage periods should be defined by domestic law…
Where the security of the state is involved, the domestic law of the member state may permit retention of the samples, the results of DNA analysis and the information so derived even though the individual concerned has not been charged or convicted of an offence. In such cases strict storage periods should be defined by domestic law”.

The judgement also notes that the Government is obliged to destroy the data of volunteers at their request, despite the similar value of their data in increasing the size and utility of the database, and states: ”Weighty reasons would have to be put forward by the Government before the Court could regard as justified such a difference in treatment of the applicants’ private data compared to that of other unconvicted people”.

The Court “further considers that the retention of the unconvicted persons’ data may be especially harmful in the case of minors” and notes the Nuffield Council on Bioethics’ concerns “that the policies applied have led to the over-representation in the database of young persons and ethnic minorities, who have not been convicted of any crime”.

Proposals and rationale

We propose that:

• All DNA profiles should be automatically deleted immediately following the decision of no further action, or acquittal by a court, except in special cases, where the Scotland rule or similar should be applied.
• Weeding rules for all police records, including DNA records, from convicted persons should be reinstated, so that only people convicted of serious crimes have records retained indefinitely.
• For children, retention times for DNA profiles should be determined by the court, ranging from zero to maximum retention times equal to those for adults.
• The policy should be applied retrospectively to all stored profiles and stored DNA profiles without an associated PNC record should also be deleted.
• The same retention rules should be applied to all stored data, including DNA profiles, fingerprints and Police National Computer (PNC) records.

Deletion of all unconvicted persons’ profiles automatically

The information provided in the impact assessment is substantially in error (see Annex) and fails to constitute the “weighty reasons” required to justify treating arrested unconvicted persons any differently than other unconvicted persons, such as people who volunteer their DNA during police inquiries.

Serious violent and sexual offences and the Scotland rule

In our view, although the numbers are small, more weight should be given to any potential losses of DNA detection for serious violent or sexual crimes than other types of crime, because of the risks to the public and the needs of victims.

Committee of Ministers’ Recommendation R(92)1 allows retention of unconvicted persons’ DNA profiles only in cases where the security of the state is involved. It also requires strict storage periods to be defined by law. However, the ECtHR judgment indicated that the Court was not opposed to the temporary retention policy adopted in Scotland for persons charged with serious violent or sexual offences. There is evidence that a small number of rape cases may be solved as a result of such a policy (perhaps one every two years, based on cases cited by the Government, see Annex).

This situation arises because of the poor record of the criminal justice system in dealing with violence against women. In particular violence against a partner can continue for many years without the perpetrator being convicted because of reluctance of the victim to press charges; and the conviction rate for rape is low due to the difficulty in proving that the alleged victim did not consent. This issue would be better addressed by improving policies to address violence against women. However, in the meantime it may be necessary to weigh up the importance of tackling this rare subset of rape cases, against the dangers of wrongly stigmatising a group of men based on allegations that may be entirely false. The 2000 ACPO Guidelines attempted to achieve this balance for PNC records where they state: “Details may be retained for a period of five years of offences where a sexual offence is alleged, but the subject is acquitted, or the case is discontinued because of lack of corroboration or allegation of consent by the victim, providing identity is not an issue…”. Scotland’s legislation for DNA records is an extension of this exception.

We are not opposed to a retention policy similar to Scotland’s, provided:

• The relevant offences are tightly drawn in an attempt to address the small number of relevant cases;
• The time period for retention is not greater than five years;
• There is a right of appeal to a court or other independent body after not more than three years;
• Safeguards are put in place to improve governance and prevent the misuse of such records to discriminate against persons who may be the subject of false accusations.

The temporary retention of DNA profiles from a small number of unconvicted persons in Scotland has not as yet given rise to any DNA detections for serious crimes. However, Scotland’s population is approximately a tenth of that in England and Wales, so, given the small number of relevant cases, this is not unexpected.

GeneWatch UK’s preliminary calculations (see Annex) also suggest that adding a small number of additional crime scene DNA profiles to the Database, particularly profiles obtained from semen collected from rape victims, would easily outweigh any potential DNA detections for such crimes lost through removing the DNA profiles of unconvicted persons. We strongly urge the Government to consider this issue in its forthcoming review of the treatment of rape victims.

Children

The special case of children is considered in more detail in Section 4 below.

Records without an associated PNC record

NDNAD records without an associated PNC record consist only of records where the PNC record has been weeded. ACPO weeding rules were in place from the establishment of the NDNAD in 1995 to end 2005. During that time, NDNAD records were supposed to be weeded at the same time as PNC records4 but this policy was not properly implemented, because the necessary link between the PNC and the NDNAD was never made.  Until 2001, police forces informed the NDNAD of records that should be deleted as a result of acquittals or decisions to take no further action: however, the HMIC estimated in 2000 that 50,000 profiles may have been kept illegally on the Database before the law was changed due to failures in this procedure.5 The law was then changed in 2001 to allow retention of NDNAD records from unconvicted persons and applied retrospectively.

In 2005, changes were made to allow all PNC records to be retained until age 100, allowing the indefinite retention of all records of arrest. The new guidelines to allow records to be ‘stepped down’ rather than deleted were adopted in 2006.6 Under the 2000 ACPO guidelines7, convictions were not weeded until at least ten years had elapsed, thus the rule change in 2005/06 probably prevented any weeding of convicted persons’ PNC records established since the NDNAD was set up. However, most cautions are weeded after 5 years, so the NDNAD will probably contain numerous records of persons cautioned between 1995 and 2000 which have been weeded from the PNC. The approximately 500,000 records without an associated PNC record referred to in paragraph 6.27 of the consultation will consist of these records (from individuals who received a caution in 2000 or earlier), plus the records of unconvicted persons that were held illegally pre-2001. If they have not been subsequently convicted of any offence (in which case they will have duplicate entries on the NDNAD linked with a new PNC record) these people form a low risk population, whose NDNAD records have all been retained in contravention of past ACPO policy and/or legislation that was not properly implemented when it was in force.

Retaining any of these records is unjust and implementing a case by case removal policy, as proposed in the consultation, is an unacceptable waste of police resources.

Weeding rules and deletion of other police data
An important omission from the consultation is the failure to consider the need to reinstate the deletion of Police National Computer records on acquittal or when charges are dropped, and also to set time limits for the retention of all data from persons with cautions or convictions for minor offences.

We note that the Committee of Ministers’ Recommendation R(92) 1, cited above, requires strict storage periods for DNA profiles from convicted persons to be defined by domestic law.

As described above, the past failure to link the PNC and the NDNAD databases as originally proposed in 1995 has led to a situation where the law was first changed to retrospectively legalise the retention of innocent people’s DNA records in 2001, and PNC record weeding rules were then abandoned. This happened for purely administrative reasons, and without any parliamentary oversight, as described in the the National DNA Database Annual Report 2005/06 (page 9), which states: “In support of the powers provided by Section 82 of the CJPA and Sections 9 and 10 of the CJA, it has become necessary to retain a nominal record of every person arrested for a recordable offence on the Police National Computer (PNC) to enable a link to be made between the DNA profile held on the NDNAD and fingerprints held on the national automated fingerprints database (IDENT1) to help the police identify and locate an individual following a match being obtained on the NDNAD”. This has led to a situation unprecedented in British history, in which records of arrest for any recordable offence, as well as all convictions, cautions, final reprimands and warnings, are now retained indefinitely (nominally until age 100), except in Scotland (where records are still weeded). PNC records are available to a wide range of agencies and the information in them may lead to an individual being refused a job or a visa (for example, US visas may be refused on the basis of a record of arrest). Even when records are ‘stepped down’ according to the 2006 rules, they are still available for enhanced criminal record checks, as well as directly to the police.

The Home Office will be aware that a decision by the Court of Appeal on whether to uphold the Information Tribunal’s 2008 decision on PNC records is pending, and, in June the Home Secretary announced a review of police record retention. The Tribunal supported the Information Commissioner’s Decision notices requiring the erasure of conviction data of a number of individuals who had suffered as a result of disclosure of old criminal records to their employers: they included an individual convicted of theft from a supermarket display in 1984, when aged sixteen; and another who had received a reprimand for punching and kicking another girl when aged 13. The Tribunal concluded that the 2006 Guidelines for the retention of police records “do not appear to be a suitable approach to the retention of conviction data in order to comply with the DPA” [Data Protection Act].8

We therefore propose that weeding rules for PNC records are reinstated, so that only the records of persons convicted of serious offences are retained indefinitely. All records, including PNC records, DNA records, fingerprints and associated data (such as photographs) taken from individuals should destroyed/deleted at the same time.

Because the weeding rules were abandoned purely for administrative reasons there has never been any impact assessment of the decision to abandon them. However, the data on hazard rates for convicted persons supplied in Annex D (Section: Research Base for Retention, pages 87-93) strongly supports the re-adoption of weeding rules, as does the evidence supplied to the 2008 Information Tribunal case. The 2000 ACPO guidelines (cited above) provide a suitable starting point for debate about these rules. Any new rules should be put on a statutory basis so that they cannot be abandoned or significantly amended in future without scrutiny by parliament.

4. Taking of samples

The collection of samples other than in connection with the investigation of a specific offence is controversial. Recommendation R(92)1 states: “Where the domestic law admits that samples may be taken without the consent of the suspect, such sampling should only be carried out if the circumstances of the case warrant such action”.

There are particular concerns that sampling DNA on arrest for all recordable offences impacts adversely on ethnic minority communities (who are disproportionately represented on the Database) and on vulnerable persons, including children and the mentally ill, who may be traumatized unnecessarily.

An estimated 37% of black men and 77% of young black men had records on the DNA Database in 2006.9,10 There were 3.8 times more arrests of black people per head of population than of white people in 2007/08.11 The decision to collect DNA routinely on arrest therefore exacerbates racial discrimination in the justice system and erodes the trust of BME communities in policing.
 
We note that a race impact assessment is a statutory requirement but has not been included at the consultation stage.12 There is therefore a failure to consider relevant issues in the consultation document.

Further, we note that large numbers of mentally ill people are arrested on their way into care, often for public order offences, and thus may be sampled unnecessarily. We have anecdotal evidence from phone calls that this can cause great distress and exacerbate existing symptoms, affecting both people against whom no further action is pursued and people who have accepted cautions. Similarly, we have received reports of children who have felt stigmatised and traumatised by the process of arrest and who find it hard to recover from this whilst their DNA, fingerprints and police records are still retained, particularly if their arrest has occurred as the result of a false accusation.  

Taking DNA from children in order to perform speculative searches is also an unnecessary waste of money since, in general, a child or young person is unlikely to reveal a ‘cold hit’ match with a past rape or murder.

For adults, we suggest that DNA is collected routinely on arrest only in relation to a specific recordable offence (to confirm or eliminate the individual as a suspect) and that there is a return to collecting DNA on charge for any recordable offence (for speculative searches against the NDNAD).

For children, we suggest that DNA is collected routinely on arrest only in relation to a specific recordable offence (to confirm or eliminate the individual as a suspect) and that DNA is otherwise not collected or entered on the NDNAD to perform speculative searches unless ordered by a court.

These proposals would limit the adverse impact of DNA testing on vulnerable groups, whilst still allowing speculative searches to be made for adults who have been charged with an offence. They would reduce, although not eliminate, racial bias, partly because charging someone requires somewhat more oversight than arrest by an individual officer, and partly because they would considerably reduce the large number of black children’s DNA profiles entered on the database.

Further we note that, as part of its PACE Review, the outcome of which is not yet public, the Government has proposed taking DNA, fingerprints and photographs in ‘Short Term Holding Facilities’ in shops or town centres, and expanding collection to arrest for any offence (including dropping litter or parking offences). In our view, this proposal carries considerable dangers and would further erode public trust in police use of data. We therefore propose that any new rules for DNA collection are included in primary legislation so that they cannot be eroded without parliamentary scrutiny.

We appreciate that the proposals for post-conviction testing are driven by concerns that some individuals who pose a danger to the public may currently not have records on the database. However, we are concerned that allowing the police to track down past offenders with spent records is contrary to the rehabilitation of offenders, may be open to abuse if drawn too widely (for example, allowing the harassment of particular individuals), and could waste considerable police time. However, there may be merit in sampling persons convicted and imprisoned for violent or sexual offences overseas, on the basis that they could pose a danger to the public.

5. Fingerprints

We support the view that fingerprints should be destroyed at the same time as DNA profiles, however we disagree that such long retention times are needed (see above). We welcome the proposal to delete copies when no longer required for investigative purposes.

We recognise that, if large numbers of fingerprints currently held are to be destroyed, it may be unfeasible to allow people to witness their destruction. However, removal of this right will undermine public trust in the destruction/removals process unless adequate independent scrutiny of this process is guaranteed. To enhance and maintain public trust in police use of data, we recommend retaining the right to witness destruction of data/samples in the case of individual applications or appeals (‘exceptional cases’), combined with an independent scrutiny and oversight process for the much larger numbers of automatic data and sample destructions and removals (see below). ‘Exceptional cases’ should include cases where the individual has identified a failure in the automatic removals process following a subject access request.

6. Volunteer samples and profiles

We welcome the proposals that future volunteer profiles will only be searched against crime scene DNA profiles relating to the specific offence under investigation; and that they will be destroyed when no longer required for investigative purposes.

The proposed approach is consistent with that applied to the DNA profiles of serving police officers held on the Police Elimination Database (PED), which is not speculatively searched and where records are destroyed when the officer leaves the force and his/her profile is no longer needed for elimination purposes.

However, we note that this approach is inconsistent with the proposals for arrested persons. Similar principles should be adopted for all unconvicted persons in order to meet the requirements of the judgment of the European Court of Human Rights, unless there are “weighty reasons” to do otherwise.

7. Governance and accountability

We support the past decisions to improve governance by including members of the Human Genetics Commission on the National DNA Strategy Board; setting up an independent Ethics Committee; and establishing a Forensics Regulator. However, we remain concerned that governance remains piece-meal and has not been put on a statutory basis.

Whilst we welcome the proposal that an independent body monitors the removals process (indeed we regard this as essential) we disagree that an appeals procedure is unnecessary – fairness and independence is vital to restore trust that the database is being managed properly. We strongly support the recommendation for quarterly and annual publication of key statistics: we have repeatedly identified conflicting reports of different figures obtained from different sources and publication of information has been erratic. We agree in principle with independent oversight, but this should be put on a statutory basis and take particular care to ensure full representation of the views of vulnerable groups and those disproportionately affected (including children, BME groups, and the mentally ill).  

Issues that have not yet been properly addressed include:

• The purposes for which retained DNA data may be used are drawn too broadly.
• The rules for the use of familial searching have not been published.
• There is still an inadequate audit trail and security for samples and data (including poor practices in police stations, and the sending of sensitive data to laboratories alongside samples).
• The impact of sharing DNA profile matches across the EU has not been properly debated or assessed.

Summary of recommendations

We support the Home Office’s proposals to destroy all DNA samples and not to include DNA profiles from volunteers on the National DNA Database.

In contrast, the proposed retention times for DNA profiles are disproportionate to the need to tackle crime. In addition, the options to restrict sample collection – particularly for children – have not been properly considered. We make the following alternative proposals.

For adults:

1. Collect DNA routinely on arrest only in relation to a specific recordable offence (to confirm or eliminate the individual as a suspect). Collect on charge for any recordable offence (for speculative searches against the NDNAD).
2. Delete profiles automatically if not convicted by a court, except in special cases (applying the Scotland rule or similar).
3. Reinstate weeding rules for all police records, including DNA records, from convicted persons, so that only people convicted of serious crimes have records retained indefinitely.

1. For children and young people (aged 10-18):
2. Collect DNA routinely only in relation to a specific recordable offence (to confirm or eliminate the child as a suspect)
3. Delete profiles automatically if not convicted by a court (including all reprimands and final warnings).
4. Give the court powers to order DNA profiles to be entered and/or retained on the NDNAD if: a) for entering, they have reason to believe a speculative search of the NDNAD could reveal past offences; b) for retaining, they deem the child at risk of committing future offences for which DNA evidence might be relevant. If the court does not order it, DNA should not be taken for speculative searches or retained.
5. Retention times for DNA profiles should be determined by the court, ranging from zero to maximum retention times equal to those for adults.

The same retention periods should be applied to DNA profiles, fingerprints and Police National Computer (PNC) records and any other data.

The proposals for retrospective sampling of persons not imprisoned should be abandoned. However, there may be merit in sampling persons convicted and imprisoned for violent or sexual offences overseas, on the basis that they may pose a danger to the public.

The level of independent oversight of the Database and the use of DNA evidence also needs significant improvement and all safeguards should be put on a proper statutory basis, with full parliamentary scrutiny