Artificial intelligence, machine learning, and automation are drastically changing the way law firms and Corporate Counsel perform document reviews. However, some industries and practices are slower than others to adopt eDiscovery tools in their projects. At a recent discussion, Page One’s Managing Partner Zeke Alicea and Jill Ragan, Esq., Senior Community Enablement Specialist at Relativity, sat down with a panel of thought leaders to talk about the impact of Relativity’s analytics tools in healthcare litigation, and how those tools can help overcome the industry’s unique challenge to produce better outcomes.
Joining the discussion were Russell Taber, author of Electronic Discovery in Tennessee: Rules, Case Law and Distinctions and adjunct professor at several local law schools including Nashville School of Law; Rich Moore, eDiscovery Manager at Envision Healthcare; and Andrew Milauskas, VP of eDiscovery at Page One, who shared his viewpoint of RelativityOne use cases and benefits from a vendor perspective.
The panel kicked off with a discussion of what makes healthcare matters different when it comes to litigation. Primarily, healthcare matters generally involve incredible amounts of Personal Health Information (PHI) and Personal Identifiable Information (PII), which patients usually don’t wish to have released to other parties. As Rich Moore said, “Healthcare cases have large amounts of PHI associated with them that have to be scrubbed before producing data. You have to be really aware that you’re representing patients.”
In addition to HIPAA compliance and protective orders, there are other things that are unique to the industry such as physician contract negotiations, which bring their own level of complexity to the discovery process.
Here’s a breakdown of the 5 key takeaways that came from the panel’s discussion on the impact of Relativity’s eDiscovery tools in healthcare, and the challenges that are still ahead.
- Productivity and cost-efficiency are top-of-mind
The panel jumped into a discussion of the time and cost-saving benefits of using RelativityOne in healthcare. As Jill Ragan put it, “[Analytics] can streamline and optimize your review process by helping you organize data, prioritize reviews, and work faster and more accurately.”
Essentially, Analytics gets to the key evidence of your case without having to manually review thousands of documents. To illustrate her point, Ragan gave a quick overview of some of the most productive eDiscovery tools for saving time:
Email threading organizes email chains to reduce the amount of data you need to review, allowing users to search and retrieve unique content without having to review entire chains of emails across multiple documents.
Active learning serves as a sort of technology-assistance review (TAR) 2.0 in which the models continually learn based on the reviewers’ decisions, ranking documents and helping prioritize reviews so teams can get through documents faster.
Conceptual searching allows users to perform searches based on related themes, pulling similar documents containing relevant content.
Communication analysis is a visualization tool that helps identify who’s speaking to whom, establishing the context of the documents being reviewed.
Automated and AI-powered tools like these save time by highlighting relevant information and providing contextual insight that manual document review might not catch. Looking ahead, tools like sentiment analysis will become even more useful at detecting the emotional context of documents, allowing teams to dig even deeper into the ideas that are important to their case.
Aside from eDiscovery tools, protective orders can also help expedite the discovery of information. Usually in a case involving healthcare, there are references to HIPAA and High-Tech to comply with, which poses a question to the reviewer—is that enough protection for the case and the client’s comfort, or should we find all the PHI and redact certain portions (or even all of it)? Depending on the data size, the second choice can quickly escalate expenses and protract timelines. Protective orders go a long way when handling sensitive information in review so teams can keep certain information confidential while complying with the applicable laws.
2. Reducing data volume is key
The sheer volume of data and PHI poses a significant challenge to reviewers. “Going through these documents and redacting them can be really time intensive, and it can require a lot of people,” Ragan pointed out.
Analytics tools not only help identify what’s relevant, but they also help weed out the data that doesn’t need to be reviewed. “If you’re using your Analytics tools and reducing the volume of data that you need to look through and redact, those can certainly help speed up the case, save time and money, and get you through your projects much quicker,” Ragan said.
Moore jumped in to add his perspective from the healthcare field. At Envision Healthcare, they encourage in-house and outside counsel to use analytics to reduce data volume. They also rely on Page One and their team to get Envision Healthcare’s data reductions down.
“The data goes into an ECA workspace, we run terms, and then produce. It’s really reduced that initial data drop down to a manageable level so you can find the relevant information you need for outside counsel,” Moore said. “Most of my job is on the left side of the EDRM, the information governance, identification, preservation, and collection of the data the attorneys need. When all of that is done, we kick off our cases with these outside counsels and provide guidelines, and we strongly encourage the use of ECA workspaces, crafting strong terms, date filters, anything to reduce the initial set of data. For now, it’s more of a bullet point list, but we are working on making it a more comprehensive play book.”
3. The eDiscovery landscape has evolved
Taber shared his unique perspective on the changes he’s seen in eDiscovery. In fact, he’d appeared in that same panel discussion room years ago to discuss the same topic, but with a different purpose. Back then, he was involved in a local think tank called the Prometheus Project, whose mission was to improve the way Nashville lawyers practice eDiscovery. They drafted a proposal for the Middle District of Tennessee Federal Court to improve the default standard. “At the time we had a default standard on eDiscovery that was based on a standard out of Delaware that the State had long since thrown aside,” Taber recalled. The old standard was focused on coercing parties to cooperate and work through issues—and it wasn’t working.
The Prometheus Project created a proposal that focused more on proportionality and reasonableness. The Court largely adopted the proposal in 2018. It’s an improvement from the old standard, but there are still challenges, including how to educate judges on complex eDiscovery issues. Many judges are inclined to allow discovery if it’s requested unless there’s a compelling case of burden or irrelevance.
Still, education and adoption of AI and machine learning in the legal community have grown over the last several years. Taber recalled an example from many years ago when he was working as local counsel for a large firm in another city. The case was huge and involved millions of documents—posing an opportunity to leverage TAR and save money. When this was suggested to the other firm, however, the idea was overruled, as the technology was new at the time and the law for its use less settled. In retrospect, this could have been a missed opportunity. “When you’ve got 40 different lawyers reviewing documents, you can get inconsistent decisions, spend a bunch of money, and it’s harder to be nimble if a judge rules on a disputed discovery issue,” he said. “Then all of a sudden you’ve got to change your coding on some categories of documents, and it’s harder to do that without technology-assisted review.” That was a while back. Now, it appears that many large firms use TAR and clustering in almost every large ESI case—but those tools are still somewhat less accepted in more run-of-the-mill cases.
4. Clients need help understanding eDiscovery tools
Those new to eDiscovery aren’t always familiar with the latest technology or how analytics tools can help. They may have some uncertainty or discomfort when it comes to using modern tools like AI and machine learning, which are still novel to many. Using these tools becomes even more problematic when sensitive personal information is involved, as with healthcare.
In these cases, it’s important to educate clients about how eDiscovery tools work and what the benefits are. Milauskas described how Page One helps clients get comfortable using RelativityOne on their cases, saying, “We handle everything as far as building new templates, workflows, and automation. All of that has to be ready at a client’s fingertips—at the drop of a dime—because they’ll throw a question out and you’ll need an immediate response.”
Together, Moore and Milauskas began referring to these culminations of automations and workflows as their “intel reports.” They package these reports and send them to outside counsel so they can better understand the kind of data they’re dealing with, how much data there is, who the custodians are, etc. Building out these playbooks helps kick off cases with outside counsel, giving them an organized framework for approaching document review.
Of course, data needs can change during the document review process. To help teams mitigate against surprises, Milauskas shared a suggestion: “We started running active learning behind the scenes—just sitting in the background—so if a timeline changes on a matter where you need to expedite, you can use the coding you’ve already applied on the remainder of the dataset and push forward from that standpoint.”
Explaining the available features in RelativityOne helps clients clearly comprehend the functionality and benefits of using those tools. But when it comes to people who are still resistant to using eDiscovery technology, Ragan suggested that it helps to put the numbers right in front of them so they can see the time savings and cost benefits. To do that, she also suggests using active learning.
“We strongly recommend running active learning in the background and then showing it to clients so they can see the savings that they could have had,” she explained. When she used to manage projects, Ragan found that legal teams and attorneys sometimes had trouble wrapping their heads around how the technology worked. When she provided the numbers, it went a long way towards moving forward in the technology space.
On the other end of the spectrum, some teams have no problem using technology-assisted review—but they might struggle with using it in a way that benefits the clients. For that reason, Moore explains the importance of setting down “railroad tracks” that helps outside counsel adhere to best practices. As Moore said, “Most of them are very good about it—but sometimes you have to interject if you see them spinning up a 100,000-document review with associates. That might not be the most useful spend of our money.”
5. Analytics provide a significant head start
When it comes to leveraging Analytics, the ultimate goal is to save time, money, and other resources. Anything you can do to cut down the amount of data and documents for review will help drive the cost down.
To illustrate how Analytics can help achieve that goal, Ragan pointed to reusable models, which are featured within Relativity’s active learning tools.
“If you have, say, a lot of healthcare matters where you have a lot of the same issues, you can take an active learning project that you’ve done, export the information that determined relevance into a new active learning project workspace, and it will help rank the documents in your new project,” she said.
As Ragan described the reusable models feature, it’s a way to take the analytical information from an existing project and import it into a new one. Starting the review process with ranked, prioritized documents provides a significant head start that saves more time and money than if the project had started from nothing.
Milauskas added that Page One has developed their own reusable data and work product application called CodeBank to make it faster and more convenient to locate privileged documents when setting up matters for clients. It works by taking the tags in a workspace’s privilege field and pushing them to a master database. When it comes time to spin up a new matter, CodeBank can pull those results back down into the fresh privilege field. By using this app, organizations not only save time, they also don’t have to worry about missing important documents due to less efficient manual searches.
Another approach to consider at the beginning of the project involves search terms. Taber explained that he often likes to attempt to reach agreement on search terms on the front end. But there may be discussions with the requesting party’s counsel who request very broad terms that could return enormous amounts of irrelevant documents. In those cases, you have a choice: do you litigate the issue in front of the judge and run the risk of irritating the judge? Or do you agree to the terms and accept that you’ll be handling huge volumes of data? Either choice can involve considerable expense. Given the expense of litigating and particular case deadlines, it can sometimes be more effective to agree to the search terms and use TAR. While hosting and running more documents from a broad search isn’t ideal, it can sometimes be more advantageous than trying to litigate for narrower search terms that would reduce your dataset.
More about Page One and Relativity. More Powerful Together.™
At Page One, we help clients leverage RelativityOne and Relativity Server to accelerate their document review process and protect confidential information. Learn more about how Page One can upgrade your eDiscovery and Document Review experience.