Skip to main content

There’s No Need to Redo a Reasonable Technology-Assisted Review Process

Reasonableness and proportionality in technology-assisted review, with a nod to cooperation, won the day in the multi-district products liability dispute In re: Biomet M2a Magnum Hip Implants Products Liability Litigation. In this case, Judge Robert L. Miller, Jr. of the United States District Court for the Northern District of Indiana refused to order the defendant to “virtually start[] over” as the plaintiffs’ Steering Committee had sought, sparing the company from an additional e-discovery spend in the “low seven-figures.”

This opinion will have important implications for anyone seeking to utilize technology-assisted review to manage discovery costs and deadlines. Why? In his opinion, Judge Miller eschewed the technical nitty-gritty that has filled previous rulings on technology-assisted review, instead choosing to focus on questions of reasonableness and equitability.

The ruling has a number of significant implications. First, it sets a precedent for allowing search term culling to be used in conjunction with technology-assisted review. Second, it emphasizes process efficacy over algorithmic technicalities. Third, it shows an openness to cost-shifting on the part of the courts when a plaintiff attempts to push beyond the boundaries of reasonableness using technology-assisted review.

In this medical devices products liability case, the defendant faced the daunting task of managing a responsive review of over 19.5 million documents. To meet this challenge, they employed a range of discovery methods to meet their objectives. First, they applied keyword culling to remove non-responsive documents from the collection, reducing the review set to 3.9 million documents. They then applied deduplication to the culled set. From there, they used technology-assisted review to bring the documents most likely to be responsive to the forefront. Following a contract attorney review for privilege, confidentiality, and responsiveness, the documents were produced.

The plaintiffs in Biomet objected to the defendant’s use of keyword culling, despite the defendant’s offer to let the plaintiffs propose keywords of their own, calling it a “less accurate” approach. Judge Miller, however, cast serious doubts on the plaintiffs’ claims regarding accuracy, ruling that the defendants have so far satisfied their discovery obligations in full compliance with Federal Rules 26(b) and 34(b)(2), the Seventh Circuit Principles Relating to the Discovery of electronically stored information (ESI), and the Sedona Conference principles. Judge Miller’s caveat: the defendants must “remain open to meeting and conferring on additional reasonably-targeted search terms” and conducting additional review on any new document hits added to the review set following that process.

Beyond endorsing this multimodal approach, Judge Miller sympathized with the defendant’s position that throwing out their use of keyword culling and the technology-assisted review process and “virtually starting over would cost it millions [of dollars] more than the millions it already has spent in document production.” Judge Miller highlighted the proportionality standard in Rule 26(b)(2)(C) in reaching that decision, noting that even if it happened to be the case that using only technology-assisted review from the outset in lieu of keyword culling in conjunction with technology-assisted review would have unearthed additional responsive documents, it would cost the defendants “a million, or millions, of dollars to test” that theory.

Based on this precedent, litigants should be able to argue that a reasonable discovery process should not be second-guessed, especially when the costs of an alternative process are unlikely to deliver any additional benefits to the requesting party. And, when a requesting party tries to push the boundaries of reasonableness when technology-assisted review is leveraged, litigants should consider asking the court to shift the costs to that party.

Stuart LaRosa is a senior search consultant with Conduent. He can be reached at

About the Author