What Do We Actually Know?

One of the problems we face in claims administration is that many of our decisions are made in the context of uncertainty.  For example, we may know that the plaintiff is credible, but that the mechanism of injury is questionable and the defense has a strong IME report.  The claims and legal professionals must determine (among other things) the plaintiff’s likelihood of succeeding on the question of whether an injury occurred based upon the available information.  The problem is that this judgment is a guess (though hopefully an educated one) based on experience and the available information.   There is no definite or fixed answer.  In order to make such decisions effectively, we need to know what is fact, what is inference, what is loose conjecture, and what information is likely to be discoverable or otherwise available that will make the guess more educated.  Once we have this information, we can determine what aspects of the claim are uncertain or ambiguous and develop a strategy to deal with them.

This brings us back to Brewer’s strategies for combating cognitive biases and making effective decisions.  His second strategy asks us to:

“Be clearly and explicitly aware of gaps in available information.”

  • Recognize when a conclusion is reached or a decision is made in the absence of complete information and be able to tolerate the ambiguity and uncertainty.
  • Recognize when one is taking something on faith without having examined the “how do we know…” questions.

We normally live with and tolerate an enormous amount of ambiguity and uncertainty in our lives without paying much attention to it.  In fact, imperfect knowledge is the general and pervasive condition of human life.  However, when we assess claims, we become acutely aware of ambiguity and uncertainty and recoil from it.  Why?  We recoil because ambiguity and uncertainty foil our attempts to predict the outcome of claims and hence drive us crazy.  Nonetheless, it is critical that we be able to make effective claims decisions against a background of ambiguity and uncertainty.  And the key to making effective decisions in the context of ambiguity and uncertainty is to specifically and accurately identify what is known (and hence certain) and what is not known (and hence uncertain).  Doing so will help us accurately evaluate the strength of our current position, reveal what we can do to obtain more information, and allow us to make rational decisions without ignoring or being paralyzed by ambiguity and uncertainty.

Once we have asked the “how do we know…” questions, we are in a position to organize what we know.  What we know in any claim falls into several categories.

  • Knowledge based on objectively verifiable evidence (factual knowledge) – The employer report of injury notes that the employee reported an injury that occurred on Y date three days later on X date;
  • Inferences based on evidence (which will always be imperfect knowledge with gaps, the imperfection being, after all, the nature of inferential reasoning) – The employee actually reported the injury to the employer on the date stated on the form. This is not a fact – we infer that the reporting occurred on the date stated because we assume, for whatever reasons (or lack thereof), that employers accurately record the date that employees report injuries;
  • Loose conjecture based on limited evidence, what might also be called a “guess” – A gap between the date of injury and the report of injury suggests the claim is suspect.  But why? 
  • That which we know we do not know or that from which no clear causal or consequential effect can be drawn or predicted – A gap between the date of injury and the report of injury, without further information, is ambiguous because equally plausible competing inferences can be drawn from it – employee didn’t report injury because it didn’t seem that bad at the time v. employee didn’t report injury at time because it actually happened outside of work;
  • That which we know would strengthen inferences, prove conjecture,  and remove ambiguity and uncertainty – Employee alleges a work-related rotator cuff tear in his dominant shoulder occurred on Y date but employee pitched a whole game in an adult league baseball tournament over the weekend before reporting the injury on X date (the inference is that the employee could not have pitched in a baseball tournament if he suffered a rotator cuff injury at work before the tournament);
  • That which we know potentially exists and may be relevant but about which we have no knowledge and which eludes our best conjecture (the infamous “unknown unknowns” in Rumsfeldian parlance) – A video tape surfaces showing the employee at the tournament, but shows that he did not play and was holding his shoulder in a protected manner (this is a highly unlikely occurrence but fits in the category of “we know a bombshell can be dropped but, by definition, if it is a ‘bombshell’ the actual content of the bombshell cannot be predicted”).

To accurately judge the claim, it is important to understand the gaps in available information and to understand when our conclusions are not supported by factual knowledge.  Take the dictum that a delay in reporting an injury increases the likelihood that the injury is fraudulent.  To believe this, one must make assumptions that may or may not be supported by actual evidence.   It is important when evaluating a new claim that we understand what these assumptions are before we make a judgment regarding the validity of the claim. 

First, accepting the dictum as true assumes that there is statistical support for it.  If there is not, the dictum is the equivalent of an old wives tale.  This is not to say that it may not be true, but without statistical support for it then it is equally plausible that the dictum is false.  Thus, the dictum should not be taken to demonstrate the strength or weakness of a claim without the existence of additional supporting evidence such as the softball tournament example above.  Despite the lack of statistical support for the dictum that delayed reporting increases the likelihood that a claim is fraudulent, numerous insurance professionals, companies, and even state agencies continue to hold the dictum out as if it had some sort of predictive significance. 

Second, accepting the dictum can actually create a selection bias in which late reported claims receive a higher level of scrutiny and more intense investigation than claims with contemporaneous reporting.  If one believes based on experience that late reported claims are more frequently bogus than timely reported claims, one must actually investigate her claim handling history and measure the level of scrutiny given to the separate claims to determine if there is any truth to the dictum.  In order to determine if there is a probable statistically significant effect in a retrospective investigation, at a minimum you would have to include only those timely reported claims that receive the same or similar level of scrutiny and investigation to late-reported claims for comparison to at least attempt to eliminate selection bias.  Without making this investigation, the dictum that late-reported claims are more likely to be fraudulent has no basis in fact and is likely to skew results in a way that confirms the dictum.

When managing claims, it is important to consider why a decision is being made and whether the decision is based on factual knowledge, an inference, or an assumption that has been “taken on faith.”  Any claim will have ambiguity and uncertainty.  This is normal.  When the ambiguity and uncertainty are identified, they can be factored into the assessment of the claim and will help generate the strategy for developing the claim (which will be the topic of the next post in this series).  When deciding to give a claim heightened scrutiny or making any other tactical decision, the decision will be more effective and will likely yield better results if it is based on factual knowledge than if it is based on an unsupported assumption.  The only way to ensure that the decision is based on factual knowledge is organize what you know.  Once the knowledge in a claim has been organized, it is easy to identify if something is being taken on faith rather than fact.


log in to comment

Back to Blog
Recent Posts