A Family Affair: Forensic DNA Databases and Privacy Implications for Biological Relatives

Remarks delivered at Canadian Institute on the Administration of Justice, “Medical Genetics, Law and Ethics”

Toronto, Ontario
March 27, 2015

Address by Patricia Kosseim
Senior General Counsel and Director General, Legal Services, Policy, Research and Technology Analysis Branch

(Check against delivery)


Introduction

It has been said that “DNA Typing is the single greatest advance in the ‘the search for truth’... since the advent of cross-examination”.Footnote 1  ().  But how did we get to where we are today and where do we go from here are some of the questions I would like to explore with you today.

Canada’s National DNA Database: An Overview

Canada’s National DNA Database, maintained under the auspices of the RCMP, is composed of two indices: the Crime Scene Index (CSI) and the Criminal Offenders’ Index (COI).   

The CSI contains the DNA profiles derived from samples taken from crime scene investigations and uploaded by one of three forensic laboratories in Canada through an encrypted data link, using an internationally-renowned software program called Combined DNA Index System (CODIS).  The DNA profiles are a series of two digit numbers (alleles) associated with specific markers at 13 loci (or parts of DNA) and used for identification purposes only.  For all intents and purposes, these DNA profiles do not, as far as we know at this time, code for any personal traits or characteristics.  The biological samples themselves, capable of revealing much more, are kept separately at the local forensic labs.

The COI contains the DNA profiles derived from samples taken lawfully from offenders convicted of designated primary and secondary offences under the DNA Identification Act.  The biological sample cards are kept in the NDDB with highly restricted access, whereas, identifying information linking the offender’s biological sample with the individual offender (such as name, fingerprint, nature of offence, judicial order and the unique identifying number connected with the sample) is kept in a separate, secure database by the Canadian Criminal Records Real Time Identification Services (CCRTIS), also under the auspices of the RCMP.

On a regular basis, the COI is run against the CSI to look for possible matches between the DNA profiles of convicted offenders and crime scene samples and find new leads by identifying potential suspects and/or eliminating suspects.  When a COI-CSI match – or “Offender Hit”- is found, the match information is revealed to CCRTIS who, alone, can re-identify the COI donor and disclose this information to the forensic laboratory that provided the matching CSI profile in question.  Local investigators can then pursue this new lead, and where justified, seek a judicial warrant necessary to obtain a new DNA sample from the suspected individual.  It is only this new DNA sample that can be used as evidence against the accused. 

Also on a regular basis, the NDDB unit looks for CSI to CSI matches revealing possible links between crimes and/or potential serial offenders.  If any CSI-CSI matches – or “Forensic Hits” are found, these are reported directly to the forensic laboratories that provided the DNA profiles through CODIS, for it is the labs that retain the identifying information.  In turn, the labs inform the investigators involved.

Information in the CSI is retained indefinitely, except where the bodily sample from which the DNA profile was derived is found to have belonged to a victim or to a person who has been eliminated as a suspect, in which cases, the information must be removed.  Likewise, information in the COI is kept indefinitely, save when the order or authorization for DNA collection is set aside, where the person is finally acquitted of a designated offence, one year after absolute discharge or three years after conditional discharge.

By way of regulation, a National DNA Data Bank Advisory Committee has been created to advise the RCMP Commissioner on any matter related to the establishment and operation of the national DNA data bank and to report annually on its activities.  As a representative of the Privacy Commissioner of Canada, I sit as an ex officio member of that Committee.

A Few Facts and Figures

According to its latest published statisticsFootnote 2, the National DNA Databank contained, as at March 15, 2015, a total of 411,938 DNA profiles – 306,991 DNA profiles in the COI (54% for primary offences ad 46% for secondary offences) and 104,947 DNA profiles in the CSI (uploaded almost equally across the three contributing forensic laboratories). 

Since its creation in 2000, there have been 34, 204 Offender Hits – the majority of which served to assist break and enter investigations, and 3,873 Forensic Hits. 

Evolution of DNA Identification Act

Five years prior to the establishment of the National DNA Databank, the Criminal Code had already been amended to provide for a judicial DNA warrant regime allowing police to obtain a biological sample from an individual suspect and to enter DNA test results as evidence against him or her.  The DNA warrant regime was upheld constitutionally under section 8 in R. v. S.A.B.Footnote 3 on the grounds that it appropriately balances the individual right to physical integrity and informational privacy with the significant state interest in bringing criminals to justice and avoiding wrongful convictions. 

While the Court recognized that “there is undoubtedly the highest level of personal and private information contained in an individual’s DNA”, it found that the degree of interference with bodily integrity under a properly issued DNA warrant (involving a buccal swab, finger prick or hair sample) is relatively modest.  As for informational privacy, the DNA warrant scheme involves only non-coding DNA, collected for a clearly articulated and strictly limited purpose, and prohibits any other purpose.  DNA searches are specific to an accused, carried out only on issuance of a judicial warrant, in respect of a specific designated offence, based on reasonable and probable grounds.   

With the adoption of the DNA Identification Act in 2000, the National DNA Databank was created to allow for the collection and retention of DNA profiles extracted from designated convicted offenders, regardless of any particularized suspicion in relation to any other crime.  The Ontario Court of Appeal in R. v. Briggs Footnote 4 described the state interest in the National DNA Databank as being much broader than simply law enforcement vis a vis an individual.  It deters repeat offenders; promotes safety in the community; detects presence of serial offenders; assists in solving cold crimes; streamlines investigations; and, helps exonerate the innocent through early exclusion from an investigation.

The Supreme Court of Canada upheld the constitutionality of the DNA databank regime in R. v. Rodgers.Footnote 5  Unlike the DNA warrant regime, the Court found that the DNA Databank does not require reasonable and probable grounds before profiles of convicted offenders can be included.  The DNA data bank provisions do not provide for the creation of evidence to be used in a specific prosecution, but serve only to identify reliable leads, and from there, usual investigative methods, including DNA warrants, must be followed in order to gather evidence that will be used to investigate the specific individual.

Like in R. v. S.A.B., the Court found that the intrusion in bodily integrity is relatively modest.  Because DNA profiles can only serve to confirm identity, and cannot reveal any personal traits or medical conditions about the individual, the infringement on informational privacy is likewise limited.  Given the legislative safeguards that separate the DNA profiles from the actual samples themselves, the Court found that the DNA data bank provisions were more akin to fingerprinting, providing just a more modern form of identification technology for purposes already permitted under the Identification of Criminals Act.

Most importantly, the Court found that the convicted offenders designated under the Act had a reduced expectation of privacy, and by reason of the serious crimes they committed, had “lost any reasonable expectation that their identity will remain secret from law enforcement authorities.”  Citing Briggs, “A person convicted of a crime has a lesser expectation of privacy not because that person’s worth as a human being is less, but because the person’s right to make choices about his or her life is curtailed.”

Through a series of subsequent legislative amendments, the National DNA Databank has undergone significant expansion since 2000, growing from an originally limited list of only 37 designated offences, to more than 250 offences today, depending how you count them. 

Un-traversed Frontiers

Despite the incremental expansion of the National DNA Databank, Canada has so far held true to its policy position of including only DNA profiles of convicted offenders. In contrast with other jurisdictions, such as the U.S. and U.K., Canada has resisted further expanding its National DNA Databank to include DNA on arrest, choosing instead to uphold the reasonable expectation of privacy and presumption of innocence of arrestees prior to conviction.  There have, however, been hints to the contrary, most recently in October 2013, when the Minister of Justice affirmed, in a Globe and Mail interview, that the Federal Government was considering a move to collect DNA samples from suspects upon arrest for certain crimes.  Like fingerprints taken on arrest, he said, a “genetic fingerprint is no different and could be used in my view as an investigative tool”.   

Also unlike the U.K., Netherlands, New Zealand and several states in the U.S. (including California, Wisconsin, Colorado, Texas and Virginia), Canada’s DNA Identification Act does not currently allow for familial searching and would have to be amended to do so.  In very simplistic terms, familial searching is the process of intentionally looking for partial matches between the DNA profile of a convicted offender and a crime scene sample when an exact match has not been found through conventional search.  Depending on the numbers of shared alleles, a familial search indicates a potential biological relationship between two different individuals. 

An oft-cited justification for familial searching is a 1996 survey of the U.S. Department of Justice showing that 46% of jail inmates had at least one close relative who had also been incarcerated.Footnote 6  Several high-profile cases that have been solved using these methods, serve as strong incentive for jurisdictions to move in this direction.  

Jurisdictions that currently do use familial searching do so either on the basis of explicit legislative permission, or in some cases, more disturbingly, in the absence of any legislation explicitly prohibiting it.  Different governance policies set out varying limitations, safeguards and oversight mechanisms.     

Several factors can affect the number of leads returned by a partial match.  In some cases, the number can potentially be quite large, for no reason other than the number and rarity of shared alleles.  Through the use of additional markers combined with likelihood ratios, “candidate lists” can be reduced and ranked in terms of probable parent-child or sibling relationships, thereby eliminating what are likely to be more remote instances of distant shared ancestry.   

Non-genetic filters such as age, gender, geography and ethnicity can help further eliminate false leads.  And traditional investigative techniques such as silent searches of publicly available information, general surveillance, police questioning of individuals and family members, requests for voluntary samples or even surreptitious collections of abandoned DNA, can help narrow down the lists even more. 

A number of U.S. scholars have analyzed the case both for, and against, familial searching on constitutional grounds of privacy and equality.  A comparable, in-depth Charter analysis should likewise be done before familial searching can be contemplated in Canada.  For today’s purposes, I am particularly interested in what a privacy analysis might look like and the Supreme Court of Canada’s recent decision in R. v. Spencer Footnote 7 could be highly instructive in this regard. 

Reasonable Expectation of Privacy

Determining whether an individual enjoys a reasonable expectation of privacy is a highly contextual analysis which turns on the totality of circumstances, including: the subject matter of the search; the nature of the privacy interest at stake; and, both the subjective and objective factors which go to the reasonableness of the expectation.  Would familial searches and the warrantless follow-up techniques typically used by investigators to narrow down candidate lists amount to a search? 

R. v. Spencer involved a police investigation into child pornography being distributed on a peer-to-peer file sharing site, which led them to an IP address associated with a computer from which they suspected the files were being shared.  The police requested, and obtained, from the relevant internet service provider (ISP), without a warrant, the subscriber information (basically, name, address and telephone number) associated with that IP address.  The accountholder was a woman. Upon further investigation, which included external surveillance of the woman’s house, the police eventually obtained a warrant to search inside her house and found the computer in question, belonging to her brother living with her at the time.  Although the police had a warrant to search the house, the underlying question was whether they should have obtained a production order earlier in the investigative process to get the subscriber information associated with the IP address?    Ultimately, the answer was yes, they should have.

The Crown argued that subscriber information was akin to white pages information, innocuous and inconsequential from a privacy perspective.  However, a unanimous Court recognized that the subject matter of a search must be assessed not merely in terms of the specific information being sought, but also in terms of the information it can reveal.  Cromwell J. reconciled a number of earlier section 8 cases by virtue of the strength of the inferences that could be drawn from the subject of the search, through technological or other means available at the time.

When it comes to familial searching, what is the subject matter of the search?  Arguably something more revealing than a fingerprint, since a DNA profile does not merely confirm or not identity, but can also reveal possible biological relationships between individuals.   Some have suggested that partial DNA matches are more akin to a partial license plate number reported by a witness, which helps identify potential leads.  Partial license plate numbers might reveal ownership and possible make of vehicle.  Partial DNA matches, on the other hand, can, depending on the algorithm, types and numbers of markers used, reveal parent-child or sibling relationships.  And, based on identifying information of the convicted offender who turned up the partial match, further inferences can be made about the relative’s gender, possible age, geography and ethnicity. 

In considering the nature of the privacy interest, the bodily privacy of family members affected by a familial search may not be engaged since the sample was not taken from them, but their informational privacy clearly is.  In describing informational privacy, Cromwell J. referred not only to the right to confidentiality and the right of control over the use of one’s personal information, but went on to articulate a third conception of informational privacy, namely anonymity – the right of individuals to go about their lives freely, outside the spectre of constant state surveillance.  Partial matches bring a number of law-abiding relatives under intensive scrutiny of investigators who may follow them, surveil their home, trail their comings and goings, question their neighbors and/or family members - sometimes disclosing unknown relationships to them in the process - and even pick up abandoned DNA samples they may innocently leave behind in the form of discarded gum, tissue or even pizza crusts (as in one case), to confirm a possible match.  The privacy invasion of this follow up investigation can be significant I would argue, whether or not the individuals know about it, and whether or not the surveillance occurs in public places.  As the Supreme Court recently confirmed in the case of UFCW v. Alberta, “it goes without saying that by appearing in public, an individual does not automatically forfeit his or her interest in retaining control over the personal information which is thereby exposed.  This is especially true given the developments in technology that make it possible for personal information to be recorded with ease, distributed to an almost infinite audience, and stored indefinitely.”

Whereas convicted offenders may forfeit a certain degree of their reasonable expectation of privacy, have their presumed-innocent relatives likewise “lost any reasonable expectation that their identity will remain secret from law enforcement authorities” or curtailed their right to make choices about their lives by simple virtue of their biological relationship?  What is their status in the eyes of the law, and on what basis can we justify shining the investigative spotlight on them, as distinguished from anyone else?

Three Privacy Paradoxes

It seems to me that a privacy analysis of familial searching raises three interesting paradoxes.   

A first paradox arises in the privacy trade-off that must be made when deciding which and how many DNA markers to use in familial searching.   A restricted number of markers will be less privacy intrusive when running the search, but also less effective by returning a large number of partial matches, casting a wide net of suspicion over many more innocent individuals who may find themselves subject to police surveillance and questioning, opening up sore family wounds and potentially disclosing information about unknown relationships in the process.  Using additional markers in the up-front screening process (such as alleles on the Y chromosome or mitochondrial DNA), might help weed out false leads and thereby reduce such needless disruption and angst among families.  However, increasing the number and differentiability of markers might also be or become more revealing of individuals depending on what we learn about DNA and how the technology evolves.  What’s better (or less bad) from a privacy perspective? 

A second privacy paradox arises when one decides whether to restrict the use of additional markers for confirmatory testing only.  At an initial stage, familial searching could be restricted to a limited number of strictly autosomal markers (not on the X and Y chromosomes) for the purpose of creating a first candidate list.  Subsequent testing with additional markers could be restricted only for purposes of confirming and winnowing down the first candidate list, thereby reducing the number of individuals whose privacy may be potentially intruded in the process.  While this two-tiered approach may be more privacy protective up front, it also creates an incentive to keep original DNA samples indefinitely, should there be a need for subsequent testing, essentially retaining a far greater amount of highly sensitive personal information.

[Note: When a databank contains only DNA profiles from convicted offenders, such as in Canada, indefinite retention of the original samples has been justified and accepted, though their use is highly restricted.  In the case of DNA databanks that collect DNA on arrest, like in the U.K., indefinite retention of arrestees’ DNA was found to be a contravention of the right to privacy under section 8 of the European Charter of Human Rights (S and Marper vs The Chief Constable of South Yorkshire, 2013).]

A third privacy paradox arises if one justifies familial searching as an acceptable expansion of offender databases.  Then, some have provocatively argued that by logical extension, there is no rational reason for objecting to a universal database containing DNA of all citizens.Footnote 8  Unless one can persuasively demonstrate that the relatives of offenders are themselves more likely to be offenders, there is essentially nothing that distinguishes law-abiding relatives of offenders from law-abiding relatives of non-offenders, other than the accident of biological heritage.   Such a distinction, they argue, is arbitrary and would only exacerbate the snare on minority groups who are already disproportionately represented in the criminal databases.  So paradoxically, is the answer to this argument about inequality and arbitrariness, to invade the privacy of all individuals on equal footing so that no one is treated unfairly?  If this logical end-result feels unpalatable to us as a society, then on what basis can the interim step of familial searching be justified other than “just because we can”?

These are just some of the questions that would need to be thoughtfully considered and carefully worked through in a Canadian context.

Conclusion

While it may be tempting to allow the ends of familial searching to justify the means, particularly in high-profile criminal investigations, Cromwell J. reminded us wisely that “the nature of the privacy interest does not depend on whether privacy shelters legal or illegal activity”.  To paraphrase Cromwell J., who himself was paraphrasing Justice Binnie in R. v. Patrick, the issue is not whether the guilty brother has a legitimate privacy interest in concealing his criminal activity, but whether people generally have a privacy interest in their biological relationships and other personal information about them and their families that can be inferred either directly or indirectly therefrom. 

Thank you.

Report a problem or mistake on this page
Please select all that apply (required): Error 1: This field is required.

Note

Date modified: