Fri, 01 Jan 2010

Home Office gets DNA database funding priorities wrong

There's a wide consensus that adding DNA profiles of crime scenes has a direct impact on detecting crimes. However views differ widely as to what is achieved by retaining the DNA profiles of millions individuals, including that of many innocents, when there's a lack of evidence demonstrating this helps detect crimes. Hence, it is surprising to learn that while the Home Office is keen to waste money on retention of DNA profiles of millions of individuals, it is to stop funding and put at risk Operation Stealth, a national operation to review unsolved murders. Detectives have had great successes when loading the DNA profiles of such cold cases. Continued funding of Operation Stealth should remain a priority.

The Times revealed that Police spending cuts may spell the end for unsolved murders unit:

Cold case units are frequently staffed by retired detectives, who can be laid off more easily than serving police officers. With the cutbacks looming, The Times has also been told that the Forensic Science Service (FSS) is reducing its charges for cold case work to try to encourage police forces to continue the work. FSS labs, which are reviewing more than 200 cases, have assisted in most cold case convictions since 1999.

The success of historic inquiries, which has spawned a spate of TV dramas such as New Tricks and Waking The Dead, is due largely to rapid advances in DNA technology. Two decades ago large amounts of DNA material had to be found at a crime scene to obtain a profile; today offenders can be identified from a few microscopic cells.

Detectives have been able to reopen and keep under constant review unsolved cases from the 1970s, 1980s and 1990s. [...]

The bulk of successful cold case convictions have been for rapes and sex assaults but since 2007 there has been a concerted drive to re-examine murders. Operation Stealth provides financial and investigative help to forces to reopen murder files. Some 46 murders are under review.

The police service’s national lead on homicide inquiries, Jon Stoddart, Chief Constable of Durham, has been been pleading with the Home Office to continue the project until 2012. As yet, however, no agreement has been reached to finance it beyond April next year.

Geoff White, project manager for Stealth, said: “It’s about equality of service. For today’s victim, the police respond with the 2009 toolbox — DNA and telephone evidence are available. Why not look at yesterday’s victim and where material still exists allow detectives to use today’s toolbox to try to find the killer? It is a moral argument, a theoretical argument, but it’s a strong one.”

The Human Genetics Commission, in its Nothing to hide, nothing to fear? (November 2009) report, highlighted that money would be better spent by adding more DNA profiles of crime scenes than that of individuals:

Improved composition – Efforts have been made to calculate an optimum size for the NDNAD [National DNA Database]. In its strategic plan Confident Communities in a Secure Britain: The Home Office Strategic Plan 2004-08, the Home Office estimated that half of all crime in England and Wales was committed by a stable pool of 100,000 offenders, with just 5,000 offenders being responsible for 9% of all crimes.7 It went on to say, however, that “most of these [100,000] offenders are known to the police and other agencies” and are therefore, by implication, already recorded on the database, although there is a 20% turnover each year, with 20,000 new offenders estimated to join the pool of prolific offenders to replace a similar number who leave it. (We note that this is a substantial number in relation to the number of new profiles added to the database each year – around 700,000 in the two most recent years for which data is available.8) It is not clear where the data on which the claim about the 100,000 core offenders was sourced (the source is described as ‘Home Office’) nor whether those who commit the remaining half of crimes are ‘occasional’ or ‘one-off ’ offenders. However, we can infer from these figures that half of all crimes are committed by approximately 2.2% of the people who are currently recorded on the database (assuming that those who are ‘known to the police’ have previously been arrested and their profiles stored on the NDNAD). In this connection we note that others, for example GeneWatch UK and the Nuffield Council on Bioethics, have concluded that putting more effort and resources into the recovery of DNA samples from crime scenes could yield significantly better detection rates than the indiscriminate expansion of criminal justice samples taken from arrestees; we share this view. [emphasis added]

The Home Office, along the years, has offered supporting views – when it's not been busy trying to expand the National DNA Database (NDNAD) by any means. Six months after publication of the National Policing Plan 2004-20079 (November 2003), the Home Office claimed in the Police Science and Technology Strategy: 2004 - 2009 (May 2004):

Significant milestones achieved since this strategy was first published include:

• The profiles of the majority of known active criminal population (2.5 million) on the national DNA database.

(The expressions 'criminal population', 'active criminal population' and 'known active criminal population' are used in several Home Office documents and websites without being clearly defined. I am awaiting an overdue response to a Freedom of Information request for definitions sent on 2009-11-09.)

The Home Office was most clear about retention of what kind of DNA profiles achieves detection of crimes, in the DNA Expansion Programme 2000–2005: Reporting achievement (October 2005, no longer online) by its Forensic Science and Pathology Unit:

Evaluation of the Programme has shown that the number of matches obtained from the Database (and the likelihood of identifying the person who committed the crime) is 'driven' primarily by the number of crime scene profiles loaded onto the Database [emphasis in the original]

GeneWatch UK looked at the data over four years in its background web page (2006) to the The Nuffield Council on Bioethics consultation and found that:

Since April 2003 [to 2006], about 1.5 million extra people have been added to the Database, but the chances of detecting a crime using DNA [as a percentage of recorded crimes] has remained constant, at about 0.36%.

In the Annex: DNA detection model: validation issues (Word) to its response to the Home Office consultation (2009), GeneWatch UK found that the percentage of recorded crimes which involve a DNA detection was again 0.36% in 2007/08.

The National Police Improvement Agency (NPIA), in its National DNA Database Annual Report 2007-09, makes the same point, again:

A crime is said to have been detected when a suspect has been identified for that crime and there is sufficient evidence to charge the suspect. DNA matches are a powerful aid to crime investigation and detection: where DNA profiles from crime scenes are added to the Database the rate of detection can be significantly increased. [emphasis added]

In the graphs below, the curves representing crimes with a DNA crime scene to subject DNA profile matches and crimes detected in which a DNA profile match was available show very close correlation with those for the number of DNA profiles from crime scenes loaded each year on the NDNAD. This is consistent with the consensus that loading DNA profiles of crime scenes is what drives detection. Adding millions of individuals' DNA profiles does not make that much of a difference.

Crimes with a scene-subject match Detections of crimes in which a DNA match was available
(Source data for both these graphs: National DNA Database Annual Report 2007-09.)

The annual report explains the correlations by a causal relation. When fewer DNA profiles of crimes scenes are loaded, irrespective of the number of DNA profiles of individuals added, there are fewer matches. The report also makes the point that the vast majority of matches happen when DNA profiles of crime scenes are added, reinforcing the position that core offenders are already on the NDNAD and adding DNA profiles of innocents or those committing minor offences does not help detect crimes.

During 2007/08, one or more subject profiles were matched with 40,406 crime scene profiles. The total represents a decrease of 8.6% of the total number of crime scenes for which one or more suspects were nominated in the previous year. The fall is due to fewer new crime scene profiles being loaded within the period. [...]

A key objective in recent years has been to ensure that the majority of the active criminal population is represented on the NDNAD, and this has led to a steady increase in the likelihood of a crime scene profile matching to a subject profile already held on the NDNAD upon being loaded. This is referred to as the crime scene to subject match rate.

Matches to crime scenes also occur when, upon being loaded, a subject profile matches to a crime scene profile already held on the NDNAD. This subject to crime scene match rate is a much lower figure. As previously explained, this is because the majority of recorded crimes do not have a crime scene (for example, minor assault, drugs offences, theft, fraud etc.) and consequently there is no crime scene examination. [emphasis added]

In time of tight budgetary constraints, the Home Office and the police can afford not to retain DNA of innocents. Their priority must be to load DNA profiles from crime scenes whether new or old.

Vocabulary note (based on definitions in the NDNAD annual report):

‘Matches’ include cases where the individual had an innocent reason for being at the crime scene and cases where it was not possible to take the investigation forward as well as cases where the individual is eventually convicted.

A ‘DNA detection’ means that the crime was cleared up and a DNA match was available.

websiteblogblog archivenews feedfeedback