Sorting through electronic discovery haystack

During pretrial discovery in civil litigation, businesses can incur substantial legal fees and must allocate significant staff hours to comply with their document production obligations. With the predominance of electronically stored information in daily business practices, responsive documents are no longer typically found in hard copy form within office file cabinets or at off-site warehouses, but rather are stored on back-up tapes, hard drives and digital media. Business and employment lawsuits can necessitate the review of thousands (and at times millions) of pages of electronic information during pretrial proceedings.
The risks are high if a company’s electronically stored information is not reviewed carefully and thoroughly to determine whether it must be provided to the opposing party in litigation. Courts may impose sanctions upon parties that have not preserved, reviewed and produced in a timely fashion electronic documentation that the other side was entitled to receive. Sanctions can include the award of attorney’s fees to the opposing party, the exclusion of evidence at trial, and the most draconian result of a dismissal of claims or defenses.
Furthermore, if a company and its counsel have poorly coordinated the compilation and review of electronically stored information, the following risks are increased:
• The unnecessary production of documents (a so-called “document dump”) that may go beyond the scope of an adversary’s document request (potentially opening the door for opposing counsel to detour the litigation by pursuing in discovery topics that may not be truly material to the facts in dispute).
• The inadvertent disclosure of documents containing privileged communications or proprietary information (making it difficult to undo the resulting strategic damage once the “cat is out of the bag”).
Many businesses have traditionally provided copies of their electronically stored files (often in hard copy printouts) to their legal counsel for a manual review. While this process may be manageable in smaller cases, it has the inherent disadvantage of running up the billable hour clock as the attorneys sort through the stacks. Also, there is always the potential for human error in a tedious, line-by-line manual review of the documents or inconsistent determinations when multiple persons are evaluating whether to place documents in “responsive,” “nonresponsive” or “privileged” categories. Businesses also have utilized keyword searches to identify potentially responsive electronic documents. Keyword searches can be problematic if not properly performed. For example, the selected keywords may miss relevant documents if the search terms are too selective or, alternatively, compile an unnecessarily large number of documents (so-called “false positives”) by using overly broad terms. Also, keyword searches may fail to capture responsive documents that contain word variations (such as different tenses of verbs) or misspelled words and abbreviations (a common aspect of email communications). As a federal district court judge wrote in a criticism of keyword searching, a simple keyword search alone is often inadequate due to the “inherent malleability and ambiguity of spoken and written English.”
Given the costs and risks associated with manual reviews and keyword searches particularly in large and complex lawsuits, litigants are turning to a relatively new process known as “computer-assisted review” or “predictive coding.” This practice entails a technology-assisted classifying process in which the human reviewer codes a selected number of documents – a small, manageable fraction of the entire collection. Analyzing the results of the human review of the sample, the computer codes the remaining documents in the collection for responsiveness, promising a great reduction in the amount of time and expense spent on reviewing entire document collections. Studies have shown that properly implemented predictive coding typically results in higher accuracy rates than manual reviews and keyword searches. In February, a New York federal court issued the first reported opinion authorizing predictive coding in an employment discrimination case that included approximately 3 million electronic documents collected from the employer’s custodians. Since that ruling, federal and state courts are increasingly reviewing and approving predictive coding protocols. Predictive coding is an evolving concept in practice and before the courts. Currently, it may not be cost-effective in many cases, especially given the up-front costs involved in implementing the software. Over time, predictive coding will likely become more economical, widely utilized and appropriate for smaller cases.
During the early stage of a case pending in federal district court, the parties must confer about the scope of anticipated discovery, including electronically stored information. A similar practice should be undertaken in state court cases in which electronic discovery is anticipated to play a significant part of the proceedings. Key custodians of a company’s electronic files should be reasonably identified, not with an unduly narrow focus. If a keyword search will be utilized, the parties should discuss and agree upon appropriate search terms reasonably capturing the issues at stake. If the complexity of a case and its amount in dispute justify the use of predictive coding, the parties should set forth in writing their mutual understanding of how it will be implemented and overseen. Cooperation, coordination and transparency are essential to efficient electronic discovery in litigation. •


Steven M. Richard is a counsel with Nixon Peabody LLP in Providence.

No posts to display