Technology Corner: Evaluating Settlement Using Decision Tree Software

Thomas P. Redick

In a large mass tort case involving tens, hundreds, even thousands of plaintiffs, the settlement process can be particularly daunting. There are software tools being perfected through practice, however, that can provide assistance in formulating a process for settling cases. By breaking down the key issues – steps in the decision-making process that leads the judge or jury to a verdict – and assigning probabilities to those steps in the decision-making process, a decision tree software can lead a lawyer to the settlement amount that best accounts for all the risks and issues in a case.

This article will review a particular brand of decision tree software that was downloaded for a trial from the Web site of the TreeAge company. This company markets its decision tree for many business uses, including a section on “legal” applications. See, TreeAge Web site, http://www.treeage.com/products/legal.htm.

The first step in using a decision tree to evaluate a particular case for settlement is the identification of various key factual and legal uncertainties (the “issues”). The twists and turns that can arise during fact discovery, motion practice, expert depositions, motions to exclude evidence (e.g., Daubert prospects, or Lone Pine orders) all go into the list of “uncertainties” that must be quantified. The final decision is success or failure at trial, when a judge and jury will resolve the legal and factual issues.

A typical litigator’s assessment of risk is based on a vague notion of probability of success or failure at trial (often too optimistic or pessimistic in hindsight). The decision tree process forces the litigator to break down each step in the process of victory or defeat, and assess the probabilities of that particular step turning out as planned. For example, the program TreeAge offers (DATA), may break down a litigation matter’s issue tree into two stages: liability and damages, each with several sub-issues.

Liability will often depend upon: 1) the outcome of a factual inquiry undertaken through discovery, 2) nature and complexity of the issues to be resolved by the factfinder, 3) credibility of witnesses, 4) the counter-balancing value of an effective counterclaim, 5) unresolved issues of statutory interpretation or caselaw, and 6) jury instruction choices made by the judge.

The size of the judgment is the second key factor to determine, since the program will begin discounting (using probabilities of success on the issues above) from the highest possible verdict. (Some users prefer to run a range of verdict and give each a high, medium or low probability). To develop this hypothetical “worst-case” judgment requires consideration of: 1) financial capacity; 2) whether the defendant will be able to appeal (arbitration limits appeals); 3) the probable time lapse to collect on any judgement; and 4) future legal costs of proceeding, including the internal costs of manager distraction, if that can be quantified.

This analysis will increase a litigator’s odds for early settlement before significant costs are incurred. The DATA program or other decision tools can calculate a fair settlement value (based on an assessment of the value of the right to litigate). Even if no settlement offer or demand is made based upon the review of risk, the identification of key issues will help the litigator prioritize the necessary factual inquiries and legal tactics needed to bring the case closer to settlement.

Many litigators use the term “decision point” to signify moments when a case is poised for possible settlement, due to the uncertain outcome of an interim event. For example, a pending motion for summary judgment may have more value before the court reaches the decision point and denies it. A decision tree allows that litigator to take all the pertinent future decision points into account, and try to resolve the case while the uncertainty of those issues is holding the parties closer together. After the motion is lost, the scales tip toward the winning party, and settlement can become increasingly difficult.

In the context of mass tort cases, where there is enormous pressure placed on parties to settle rather than burden the court system, the use of decision trees can help to standardize settlement offers across a large group of plaintiffs. Each plaintiff has identifying characteristics and uncertainties (e.g., statute of limitations problems, credibility problems, lack of exposure to the substance, lack of product identification, plausible alternative medical causation, etc.). When a settlement matrix is created to standardize payments to similar plaintiffs, a national coordinating counsel can use this matrix to persuade reluctant plaintiffs and their attorneys that the settlement being offered is fair.

By way of example, in one billion-dollar mass tort case, a settlement was engineered in a particularly troublesome individual case, despite a $19 million difference between settlement offer and the demand. This case was one of thousands filed nationwide, and it was not settling as readily as other cases.

The logjam was broken by breaking the case down into distinct issues, including a motion for discovery allowing alternative cause of the medical condition (disclosure of third party medical records from an adopted child’s “birth mother”), and another court order granting a defense medical examination. The exam was controversial because it would include genetic testing for a rare genetic disorder that would exculpate the defense (because the disease was not caused by exposure to the substance in question, but solely by genetic factors). Plaintiff was only in her late thirties, a successful professional, and she did not want to learn that she might have a gene for early-onset dementia.

The decision tree identified critical decision points that should be front-loaded in the case, based on medical records hinting at a possible genetic condition. In that case, the value of the case was greatly reduced by fact discovery that detected the genetic condition in the plaintiff’s birth mother (who had given the child up at birth and entered an institution for care of the degenerative genetic condition). With the plaintiff threatening suicide if she were to learn of the particular genetic disease, and resisting the defense medical examination, this uncertainty was used to discount the potential award. While there was no certainty that the genetic test would prove positive, plaintiff was not willing to take the chance of finding out she had the gene. Defendant was willing to take a bird in the hand (some payout), rather than risk a negative finding and attendant risk of a high verdict. A successful compromise was reached before the medical exam was conducted.

The amount of the settlement was then entered into a defendant’s settlement database, where the value was recorded and key factors influencing that value were also recorded. As a result, any defense litigators for that client in other jurisdictions would have access to the “issue” that created the discount. In similar settings, these members of the defense team in other jurisdictions could consider whether to pursue discovery of genetic evidence in the form of birth mother medical records, and seek a genetic test in a defense medical examinations (neither of which would necessarily occur to every attorney at first glance, given the relative novelty of genetic testing).

For any complex litigation matter, multi-plaintiff or multi-issue, the decision tree method of breaking down and quantifying settlement barriers can work to the advantage of a litigator seeking a cost-effective resolution for a client. In the context of mass tort cases, decision trees can work in tandem with evidentiary databases to ensure that networks of local counsel are following consistent standards in settlement.

Thomas P. Redick is a member of Gallop, Johnson & Neuman, L.C. in St. Louis, and a vice-chair of the Toxic Torts and Environmental Litigation Committee.