challenge
Our client, a managed healthcare company, had a relatively small number of documents to review, only 18,539. They wanted to accomplish two goals: first, to review the material quickly; and second, to deploy an effective, cost-efficient, and defensible technology tool to prioritize the review and eliminate nonresponsive content.
solution
With the assistance of the Epiq team, Counsel chose Relativity Active Learning for this review. They reviewed an initial control set of 443 documents, which yielded a responsiveness rate of 3.84%. Based on the control set, the Epiq team estimated that there were 711 responsive documents in total. Counsel also reviewed 502 documents within the Relativity Active Learning tool for training purposes. Each day, new documents and their families were selected for first-level review batching based on their Relativity Active Learning score at the time. Documents that were quality control reviewed by Counsel were used to further refine the Relativity Active Learning tool and update the scores.
execution
After the fourth day of review, the responsiveness rate dropped to 3.4%. At that point, the Epiq team had identified 713 responsive documents. Counsel agreed to shut down the review. We then generated a random sample of 400 documents from the set that the analytics engine thought would likely have had a low probability of being responsive (an “Elusion Test”). The Epiq team reviewed this sample. The Elusion Test returned zero responsive documents.
results
In total, we were able to remove 9,659 (52%) documents from the review. This eliminated four days of the anticipated review schedule (assuming the review of all documents) and an estimated cost savings of 52%.
