Friday, April 19, 2019

The Latest Guidance on Expert Reports from England's Forensic Science Regulator

This week, the Forensic Science Regulator for England and Wales issued a third version of guidance on the content of expert reports for criminal cases. Forensic Science Regulator, Guidance on the Content of Reports Issued by Expert Witnesses in the Criminal Justice System in England and Wales, Apr. 17, 2019.

 These reports are required "when the witness will, either in the report or in testimony at court, provide evidence of opinion." Some items are rather mundane -- for instance, a report must be signed by its author. But several parts of the guidance are noteworthy:
  • The author must certify that the report is truthful and, if introduced into evidence, wilful falsehoods would be criminal. (If included in a report in the United States, such declarations could be significant on the question of whether the report is "testimonial" hearsay that would trigger Con frontation Clause protections.)
  • The report must state that "the witness will inform all parties and where appropriate the court in the event that his/her opinion changes on any material issues."
  • "The witness must state any limitations to their experience and whether any particular issue, which is discussed in the statement (or on which they had been requested to discuss), falls outside their expertise."
  • Consideration should be given to stating "[w]hether the witness tends to work for the prosecution, defence or both" amd "[w]hether the witness has been the subject of criticism and is aware of that criticism."
  • Qualifications must be impartially listed, and "[t]he membership of an organisation which is obtained solely by applying for membership (and perhaps, payment of a fee) is not a qualification," and presenting such memberships as proof of expertise "is, at the least, irrelevant and may be misleading."
  • "Where the witness is aware that reputable experts in the field would hold a range of opinions on the issue under consideration, the report must set out what that range is and justify why the witness’s opinion falls at a particular point within that range."
  • The report should delineate "[a]ny results obtained which would undermine the opinion expressed"; [a]ny work, or other information, the witness is aware of which would undermine the opinion"; "[a]ny limitations to the analytical approach adopted"; and "[t]he uncertainty of measurement associated with the methods employed and the manner in which this has been addressed."
The document also advises experts "to avoid the term 'consistent'" because it "has been criticised by the Courts. It can sound significant but, in reality, mean little." On the other hand, "'not consistent' is usually clear and does not give rise to the same issues."

The guidance on "statistics" asks the expert to keep "detailed consideration" of "complex statistical analysis" out of the report but to make it available some other way. Moreover, it seems to place all "Bayesian inference" into this category! In full, the guidance is that
The Courts have, on a number of occasions, expressed concern about the evidence of complex statistical analysis (e.g. Bayesian inference) being discussed before the jury.
It is therefore advisable to avoid incorporating detailed consideration of such analysis in reports. The statement should set out the model employed and the conclusions drawn.
The details of the analysis should be available for disclosure.
This is a matter that the Regulator should reconsider in the next revision of the guidance -- along with the suggestion that "reference to statistical significance" is an "appropriate term" for explaining "how safe or unsafe" is "an inference from any findings." The phrase "statistical significance" has not fared well in recent policy pronouncements from statisticians.

Saturday, April 13, 2019

Questions on Kentucky's Rapid DNA Program for Sex Crime Investigations

An Associated Press report being picked up in newspapers such as the Washington Post and USA Today gives the impression that Kentucky is on the verge of replacing conventional DNA analysis of rape kits with a "rapid DNA" instrument. (1) The article states that
The equipment, known as the ANDE Rapid DNA system, can generate DNA identification from forensic samples in less than two hours, Kentucky law enforcement officials said. ...
“If you are a sexual predator in ... Kentucky, we’re going to come after you,” Kentucky State Police Commissioner Richard Sanders said at a press conference. “And we now have new equipment to come after you quicker. We now have another way to identify who you are.” ...
A 2017 federal law authorized the FBI director to issue standards and procedures for rapid DNA analysis, described as a fully automated process, according to an FBI website.The ANDE system received FBI approval last year for use in accredited laboratories, the company’s website says. A cheek swab is inserted into an ANDE device roughly the size of an office photocopier and results appear within hours. By comparison, DNA samples sent to conventional crime labs can take months to analyze.
The FBI has indeed approved the ANDE 6C Rapid DNA System for use in accredited laboratories -- but only with "[k]nown reference buccal DNA sample[s]." (2) In other words, an STR profile from a cheek swab from a known individual can be added to or checked against a database that is part of the Combined DNA Index System (CODIS). The profile from the male fraction of a sample in sexual assault case ascertained via a rapid DNA machine cannot be.
"The analysis of forensic samples by a Rapid DNA system is not ... permitted to be uploaded [or] searched in CODIS at this time." -- FBI (2)
So what does Kentucky propose to do with the rapid profiles? In a press release, Governor Matt Bevin enthused that they "can help us to identify an assailant in a matter of hours – allowing us to focus the investigations of sexual crimes more quickly than ever before." (3) Of course, it does not take months for a "conventional crime lab" to perform the steps that generate an electropherogram in the ordinary way and to interpret the data, but rapid DNA analysis is less labor intensive. That is a good thing, but how will Kentucky "identify an assailant in a matter of hours" from the rapid profile without using any CODIS database?

Even without a database, rapid DNA can help if there already is a suspect. It might further implicate the suspect, or it might exonerate him, propelling the investigation in a new direction. But that seems to be an afterthought in Kentucky's description of "how rapid DNA will be used." The press release (3) states that
By matching the DNA from the crime to an I.D. in a criminal database, this could give immediate data that focuses the investigation and get [sic] a rapist off the streets. By collecting an additional sample from the victim – just one more swab – a perpetrator’s identity can be matched within two hours."
How Rapid DNA will be used for sexual assault investigations:
● If a victim presents at a hospital or police station, and agrees to having DNA specimens taken for testing, a DNA ID can be generated to inform the investigation. This is a separate process from the Sexual Assault Kit but can be taken during the sexual assault exam.
● Once the male DNA is separated from the female DNA, an operator can test the male DNA for a definitive DNA Identification pattern. This process takes less than two hours.
● If a DNA ID is generated, it can be compared to criminal databases. If a match is found, this will provide important information to law enforcement about the attacker. It may exonerate an innocent suspect. It may connect other crimes to this attack.
The release does not explain how this "separate process" can "match[] the DNA from the crime to an I.D. in a criminal database" when the FBI has yet to approve rapid profiles for CODIS database searches. Another Kentucky State Police document (4) indicates that Kentucky somehow has been able to perform database searches with rapid DNA profiles:
The state of Kentucky has been running a pilot program for several months, using the ANDE Rapid DNA system to test samples from rape cases. It has proved invaluable in several cases where there is a match to DNA in a criminal database and to others in which a specific suspect was under consideration. This has convinced Kentucky officials to more broadly implement Rapid DNA testing.
To be clear, I have no objection to the use of a more efficient technology, but I am left puzzled. Has the FBI given Kentucky a special dispensation to use rapid profiles with CODIS databases? Is Kentucky using some database outside of that system?

And what is the basis for Kentucky's claim that inasmuch as "Rapid DNA uses Short Tandem Repeats for DNA ID, which has no coding information,.... there is no genealogy or health information gathered."? (3) The STRs certainly convey information on (close) genetic relationships. (5, 6) As for health-related information, they lie somewhere between none (like a passport number) and not much (like an ABO blood type). (7)

REFERENCES
  1. Bruce Schreiner  (AP), Kentucky To Use Rapid DNA Tests for Sex Assault Cases, Wash. Post, Apr. 10, 2019, https://www.washingtonpost.com/business/technology/kentucky-to-use-rapid-dna-tests-for-sex-assault-cases/2019/04/10/a28d484e-5bd4-11e9-98d4-844088d135f2_story.html
  2. Rapid DNA, www.fbi.gov/services/laboratory/biometric-analysis/codis/rapid-dna (viewed Apr. 13, 2019)
  3. Kentucky State Police, Kentucky Rape Reduction with ANDE Rapid DNA News Conference Fact Sheet, available via link at https://www.ande.com/kentucky-case-study/
  4. Kentucky State Police, The Use of Rapid DNA to End the Sexual Assault Epidemic, available via link at https://www.ande.com/kentucky-case-study/
  5. David H. Kaye, The Genealogy Detectives: A Constitutional Analysis of “Familial Searching”, 51 Am. Crim. L. Rev. 109 (2013), preprint available at ssrn.com/abstract=2043091
  6. Henry T. Greely & David H. Kaye, A Brief of Genetics, Genomics and Forensic Science Researchers in Maryland v. King, 53 Jurimetrics J. 43 (2013), available at ssrn.com/abstract=2403063
  7. David H. Kaye, Mopping Up After Coming Clean About "Junk DNA", Nov. 23, 2007, available at http://ssrn.com/abstract=1032094.

Wednesday, April 3, 2019

OMICS Journals and Conferences Slapped with 20-year Injunction and $50 Million Fine

A few days ago, a U.S. District Court held Srinubabu Gedela and his companies, the OMICS Group, Inc., iMedPub, and Conference Series, liable for over $50 million. It also issued a permanent injunction against many deceptive practices on the part of this nest of companies and the various other entities through which they operate (such as Allied Academies, Meetings International, and Pulsus). Excerpts from the opinion that describe the deceptive practices and other information are posted on the Flaky Academic Journals blog. 1/

The opinion is significant for editors of OMICS-related journals and authors who place their papers in them. The court was persuaded by the Federal Trade Commission's  "evidence indicating that Defendants’ peer review practices are a 'sham.'" The journals affected by the order include
  • Journal of Forensic Research
  • Journal of Forensic Biomechanics
  • Journal of Forensic Medicine
  • Journal of Forensic Pathology
  • Journal of Forensic Psychology
  • Journal of Forensic Toxicology and Pharmacology
  • Journal of Medical Toxicology and Clinical Forensic Medicine
  • Global Journal of Nursing and Forensic Studies
The first journal lists Sheila Willis (Forensic Science Laboratory, Ireland) and Jian Tie (Nihon University School of Medicine) as editors-in-chief, and Peter Gill (University of Oslo), Harvey Ho (Alabama State University), and other individuals inside and outside of academia as editors. Special issue editors from the US are Jianye Ge (University of North Texas), Oliver Grundmann (University of Florida), Thomas J Holt (Michigan State University), Harvey Ho, and Jawahar L Mehta (University of Alabama Little Rock Medical School).

The editor-in-chief of the last journal is, again, Jian Tie. Linda Howe (University of Central Florida) is an editor. I have not looked at the editorial boards of the other journals recently, but earlier postings on the Flaky Academic Journals blog list some of the editors of the Journal of Forensic Medicine and the Journal of Forensic Psychology. The Flaky Academic Conferences blog tracks OMICS conferences on forensic science and technology and some pf the speakers there..The conferences are often marketed under other brand names.

These lists likely include people who would be surprised to learn of their alleged roles or who have tried to have OMICS remove their names. 2/ One of the deceptive practices is misrepresenting the membership of editorial boards.

NOTES
  1. The posting also notes an interview before the order in which Dr. Gedela suggested that any lapses in peer review and quality control were the responsibility of the unpaid editors and authors rather than the publisher. An Ars Technica report characterized OMICS' conduct as "egregious enough that a judge doesn't even wait for a trial." If "egregious" meant "uncontroverted," that would have been correct, but courts do not grant summary judgment in lieu of a trial because the alleged conduct is flagrant. The court granted summary judgment because the defendants failed to point to acceptable evidence contradicting the FTC's showing of persistent and egregious misconduct.
  2. See OMICS Journal of Forensic Research, Flaky Academic Journals, Mar. 18, 2017.