The Gatekeepers: Improper Exclusion of Expert Evidence and the Pitfalls of Admissibility Analysis in
In the wake of tragedies like the wrongful convictions resulting from the misleading testimony of Dr. Charles Smith, and the unwarranted parent-child separations stemming from unreliable physical evidence testing at the Motherisk lab, the fields of criminal and family law have had a serious reckoning with the imbalance of information between courts and the scientists called to interpret physical evidence. These tensions were displayed in Nova Scotia (Community Services) v J.M.<1> which illuminated the issues judges face when assessing the admissibility of expert testimony and showed how they can struggle when ruling on matters outside of their expertise.
The case concerned a child protection order made by the Nova Scotia Department of Community Services on behalf of the nine-year-old child of R.R. and J.M. due to their drug use. Months later, custody was restored to the mother, J.M., and soon after, the father, R.R., sought unsupervised access. The Nova Scotia Department of Community Services rejected this application on the grounds that the father had tested positive for cocaine several times. The issue in this case concerned the admissibility of R.R.’s positive test results and the corresponding expert testimony explaining them. The expert testimony came from Dr. Bassam Nassar, the Director of Toxicology at the lab where the tests were conducted. At the hearing, Forgeron J of the Nova Scotia Supreme Court found that this expert testimony was inadmissible on the grounds that the Department did not demonstrate that it was sufficiently reliable. She cited issues like the fact that the lab was “clinical” rather than “forensic,” the fact that Dr. Nassar’s expert opinion was “intertwined” with the lab results, and his lack of knowledge of current international standards for testing.
The jurisprudence establishing expert opinion admissibility comes from a post-Mohan, post-Goudge inquiry<2> understanding that courts need to be very careful about admitting expert opinion due to its ability to mislead triers of fact without challenge due to the technical nature of the information. After the judge has determined threshold reliability based on the factors set out in Mohan, analysis moves to the gatekeeper stage, which requires judges to engage in a balancing of the probative value of the evidence against its prejudicial effect,<3> meant to serve as a safeguard against the inherent risks of expert evidence. <4> This involves assessing legal relevance, necessity, reliability, and absence of bias. In this case, the opinion evidence was excluded at the threshold reliability stage of analysis.
At the Nova Scotia Court of Appeal, Bryson J.A. pointed out the flaws with the decision. Multiple references to the Lang Report<5> make it clear that the decision is borne out of a concern for wrongful decisions based on shoddy scientific evidence. Though well-intentioned, this raised multiple problems. First, it is unclear whether that report was even admissible in the circumstances. Secondly, the use of the report is problematic because the hearing judge ultimately uses it to conflate potential problems with the physical evidence with the overall reliability of the expert’s testimony. The hearing held that because Dr Nassar had a vested interest in the physical evidence appearing credible (since the tests were conducted at his lab), his evidence is “self-serving” and thus unreliable.
Bryson cites White Burgess<6>, where the same issue arose. In that case, the accountant who provided expert testimony was accused of not being able to be impartial because her firm discovered the negligence that resulted in the cause of action. It held that the burden to prove a lack of impartiality is ultimately on the party claiming the bias<7>, which is the opposite of what happened here.
Forgeron J instead presumed that Dr Nassar was biased and falsely placed an onus on the Department to show that he was credible. Finally, Bryson held that it is unclear what stage of analysis she places this exclusion at, since she claims this apparent conflict of interest goes to ‘reliability.’ Per White Burgess, this would suggest the gatekeeper stage. However, a gatekeeper analysis never occurs — most of the analysis instead argues that the issue should only be considered under threshold admissibility, <8> without actually coming to a decision one way or the other.
His decision misses that the key difference between the two stages of analysis is not the elements to be analyzed, but rather how the analysis is conducted. He misses that the gatekeeper stage is supposed to serve as a more holistic balancing between potential prejudice and probative value. In this context, this would primarily mean weighing the reliability of the evidence (the extent of the expert’s opinion basis in fact and the strength of his inferences) against the potential for his testimony to mislead the trier of fact. <9> Bryson’s failure to engage in this analysis indicates an important issue with post-White Burgess jurisprudence. While each of the nominate elements are important to gatekeeper analysis, they ultimately distract from the overall purpose of the stage.
However, the appellate decision makes a valuable warning about the dangers of judges dipping their toes into scientific matters. While Forgeron J’s skepticism came from a well-founded place of concern for wrongful decisions, the scientific facts grounding her argument were ultimately misapplied and rebuttable. As Bryson points out, Dr Nassar appeared to be eminently qualified, at the very least able to pass the threshold reliability test for lack of bias.
As long as the information imbalance between judges and experts continues to exist, courts can expect to see more cases where judges who are not educated on basic science make incorrect rulings, assuming that they are well-equipped to make conclusions outside of the scope of their knowledge. The trial decision shows that even a well-intentioned judge who is up on the literature and aware of the inherent dangers of scientific evidence can still make a ruling that results in a faulty outcome. And so long as the White Burgess framework muddles the differences between the stages, analysis may focus more on what is supposed to be analyzed at each stage rather than on coming to a decision about admissibility. The gatekeeper stage is especially vulnerable to this, with the added wrinkle that the ultimate goal of the stage is balancing prejudicial effect against probative value rather than the assessment of any single specific factor in the evidence. In light of this, it is perhaps, unsurprising that a decision was not actually made.
1 2018 NSCA 71
2 The report issued on the wrongful convictions resulting from the misleading expert testimony of Dr. Charles Smith.
3 White Burgess Langille Inman v Abbott and Haliburton Co, 2015 SCC 23 at paras 52-54.
4 R v Abbey, 2009 ONCA 624 at para 92.
5 The report on the “inadequate and unreliable” hair strand analysis conducted by the Motherisk lab that was used in many criminal and family law decisions.
6 Nova Scotia (Community Services) v J.M., 2018 NSCA 71 at para 27
7 White Burgess Langille Inman v Abbott and Haliburton Co, 2015 SCC 23 at para 48.
8 Confusingly, he instead holds that the test results should be admitted despite that not being an issue on appeal.
9 R v Abbey, 2017 ONCA 640 at para 54