Frank Zizzamia is director, actuarial, risk & analyticsat Deloitte Consulting LLP. Jim Kinzie is senior manager,actuarial, risk & analytics, Deloitte Consulting LLP. CarolineBuckley is senior manager, actuarial, risk & analytics,Deloitte Consulting LLP. Andreea Brezeanu, Consultant, actuarial,risk & analytics, Deloitte Consulting LLP. For moreinformation, visit www.deloitte.com.

Chances are, by now you have heard a lot of discussionshappening around the implementation of claim predictive models andthe potential costsavings that can be achieved. However, have you considered theimpact on claims handlers as a result of implementing a claimspredictive model? In this article, we will focus on the humancapital impact of leveraging advanced analytics in the claims life cycle.  

A Balanced Approach ToImplementation 
By taking the plunge andimplementing claim predictive models (CPMs), claims organizationshave started the journey towards identifying and solvingorganizational problems that exist today, next month and years downthe road. If implemented and used correctly, models canprovide immediate insights and tangible cost savings for theorganization.  Leveraged effectively at first notice ofinjury or loss (FNOI/L) and throughout the lifetime of the claim,advanced analytics can have an impact on various aspects of theclaims lifecycle: claims assignment, special investigative unit(SIU) referral, medical case management, litigation, subrogation,escalation and, ultimately, claims settlement and outcome.

Much of the discussion around implementing CPMs to date has beenon the technical model building side of the equation, and rightlyso. There is still quite a bit of 'noise' across the CPMlandscape, however. There are great differences in capability inthis market space from very basic data mining and analytics to far more complex andstatistically advanced predictive models; with hundreds ofpotential claims data variables, groupings or correlations in play.In fact, the observations described later in this article are basedon actual insights and experience from our having implemented thismore sophisticated type of CPM solution.

Beyond Potential CostSavings
Claims organizations tend to focusheavily on how implementing claims predictive models will favorablyimpact the bottom line. However, one of the major benefits ofrolling out a CPM comes from creating workflow efficiencies throughproviding timelier loss insights to claims handlers. Byaccelerating critical adjusting decisions and refocusing thecomposition of their pending workload in a way that better usestheir experience and skills, claim handlers can better focus on theright claims at the right time. This ability to better segmentclaims by severity and align adjuster skill set can help toincrease morale, reduce case workload and can improve the overallspeed and quality of claims file handling. The industry'sacceptance and adoption of CPMs has moved well beyond conceptualdebate to a more focused implementation and benefit realizationdiscussion, and this is anticipated to continue to gainmomentum. Has your organization considered both the costand human capital implications of implementing a CPM?

Facilitate Higher Acceptance of ModelInsights
The effectiveness of a CPM—or any largetechnology change program for that matter—is contingent upon claimshandlers adopting the tool and embracing change. As Spencer Johnsonsaid in the book Who Moved my Cheese?, "If you do notchange, you can become extinct!" Ultimately tools don't adjustclaims, people do. Whether your claims organization is thinkingabout adopting a CPM or you are in the midst of implementing yourown models, there are some key success factors to consider. First,the organization needs to communicate the upcoming changes to theclaims professionals. Communication is important, not only toadjusters and claims department, but to the entire organization.Transparency regarding any changes can reassure the claims handlersof the changes ahead. It is critical to communicate that a CPM isanother adjusting tool to help enable them do a better job—forexample, considering fifty or more predictive factors that no humanbeing could process alone. 

The adjusting process, however, is still very much a combinationof art and science, where the insights of claims handlers willalways be needed. Employee engagement in large transformationalprograms can directly affect morale and retention. If claimshandlers understand how CPMs can help them enhance job performance,then they are more likely to feel at ease and understand that theyare part of the post-implementation plan and solution. Their valueis not being questioned. Rather, their insights are beingacknowledged and institutionalized while being better equipped todo their job. When all is said and done, CPMs are merely anothertool for the adjusting toolbox, albeit an effective one. 

The second key factor is engaging andmotivating a strong team of claims system users representing manyoffices, regions and levels, from the onset of theproject.  While there can be a strong tendency for partiesto be skeptical and uncomfortable with transformational change,selecting the appropriate group of CPM "champions" to carry themessage forward can help assist the claims organization with theadoption of the new tool and technology. This can be done byfacilitating focus groups, surveys, roundtable discussions, orsimply water cooler conversations with the organization'schampions.

Oddly enough, it is very important to engage the users that arethe most vocal opponents of model adoption early in theprocess.  Why?  Because when these earlierdoubters "buy-in" to the benefits of leveraging claims predictivemodels, they often times become the strongest supporters ofchange.  From models we have implemented to date, these"converts" have become an ardent force in rallying theircolleagues.

A third factor is the development of education andtraining materials to help end-users with the adoption andacceptance of the output and insights provided by a CPM. When itcomes to education and training, it is more than teaching claimshandlers how to work with the model's output. It entailstranslating model benefits into a language to which claims handlerscan relate: positive impact on case load closure and quality.

Questions to ponder include:

• How will the model's output positively affect the day-to-dayactivities of the claims handler?  
• What information and insight is new, or what information isbeing looked at differently now? 
• Will the claims handler have less or more capacity based on hisor her skills? 
• What is the best way for the claims handler to use the timeearned from the efficiencies afforded by the model? 
• What additional activities will he or she need to undertake tosupport or complement the model's outputs?

With these three key factors "top of mind" and addressed, theclaims organization can be much better positioned to realize thefull financial value and positive claims handling workflow benefitsexpected from the implementation and roll-out of a CPM.

How Models Help Change Perspective
Whatdoes a claims handler need most? Odds are, most professionals wouldanswer "more time in a day." The use of models can help by freeingup time for a claims professional to focus more on examining theright files at the right time. While the impacts to a claimshandler's productivity and workload efficiency will vary byindividual, tenure, and level or position within the claimsorganization, so will his or her readiness and willingness toaccept, adopt, and change. 

The following two examples demonstrate how the change readinessand apprehension of a team and an individual were addressed in thecourse of implementation of a CPM. The needs of teams andindividuals are quite different. Ultimately, both were enlightenedabout the value of a CPM and have since bolstered adoption andexcitement for CPMs in their respective claims organizations. 

In the early communication phase of a business implementation ofa large CPM, it became clear that some of the more senior workers'compensation examiners were skeptical about the value ofimplementing a CPM. Gaining buy-in and acceptance from  agroup of such seasoned claims adjuster can be challenging. In thisparticular situation, the examiners claimed they "knew a bad claimwhen they saw one and could easily identify the highest severityclaims." It was fairly obvious, at least to them, that they "didnot need a CPM telling them how to adjust a claim."  

When faced with this type of situation, a good approach toconsider is performing a closed file review (CFR) compared againstthe prospective claim scores generated from a claims predictivemodel, using information typically available at FNOL. When claimsorganizations encounter this type of challenge, some have employeda CFR in part as a means of "promoting" a CPM's potential bettersegmentation of claims at FNOL. By doing so, they found a moreobjective way the challenge conventional or tribal wisdom whileengaging a group of highly experienced resources. Typically, a teamof senior claims professionals are instructed to review severalhundred closed, very high-severity claim files in order todetermine if they can correctly identify the high severity claimsbased only on the information typically available atFNOL. 

What is most interesting about this challenge is that by the endof the closed file review, the examiners who claimed they could"spot a bad claim when they see one" correctly identified justone-third of the high-severity claims based on the FNOLinformation. Based on our collective experience implementing CPMsand this added 'real world' closed file perspective, there exists asignificant business opportunity to more consistently identifypotentially high-severity claims at FNOL.

To be clear, these exercises were notconducted to discredit these very capable claims examiners, but toinstead illustrate how a mathematical combination of 50 internal,external and synthetic variables could fundamentally change howclaims handlers might approach a new file. The revelationexperienced by the examiners was to understand that a CPM can bringthem insight much sooner in the claims handling process, and insome cases, it could also deliver valuable new insights. Theseclaims handlers quickly realized and appreciated how a CPM could beused to support their claims investigation, evaluation, andsettlement. There remained a real need for their talent and skills,but their time would be better spent on the most severe claims byleveraging a well-educated head start at FNOL.
 

In other situations, change readiness and management mayneed to be considered all the way down to the individual claimshandler level. In one such auto bodily injury (BI) CPMimplementation, a particular claims handler was hesitant to acceptthe CPM outputs. This individual felt it was targeted at"automatically closing claim files and minimizing the need for(his) skills." When faced with this type of situation, an effectiveapproach for management to take could be leading the claims handlerthrough the CPM design and building process. 

More specifically, the claims manager could walk the adjusterthrough what pieces of information led to the creation of thescoring algorithm and what variables carry more weight in terms ofpotentially predicting claims outcomes. 

Talking in a business language that claims handlers can relateto is absolutely key here. It is not necessarily imperative thatclaims handlers know the mechanics of how a CPM is technicallyconstructed. It is, however, quite helpful for them to understandhow some of their day-to-day investigation and evaluationactivities have been incorporated, along with what new insightshave been brought to light. Thus, educating claims handlers abouthow model output provides earlier and better insights isessential. 

This information may have been difficult and time-consuming togather or interpret before, or it may have not have even beenconsidered at all. The information can allow the claims handler totailor the investigation, evaluation, and negotiation by beingarmed with more knowledge to arrive at a quality settlement and/orclosure in a timelier manner. 

The claims handler who was "afraid [his] job was on the line"embraced the insights being provided by a CPM. Only then did herealize that the model being introduced would help move the claimtowards closure, rather than eliminate necessary justing steps orto diminish precious handling skills. This particular claimshandler also became a vocal proponent of the CPM, and helped theorganization accelerate the adoption of the CPMoutput. 

Impact on Claims Handlers
Again, the keytakeaway is this: models don't adjust claims, people do. The betterclaims handlers understand the true motivation and businessobjectives of implementing a CPM, the better the outcomes can befor all of those involved. Many of the business objectives anddesired outcomes of implementing a CPM are squarely focused oncommon human capital challenges. For example, one of the pleasantbyproducts of implementing a CPM is obtaining knowledge of whereskill/resource gaps or excess capacity exist.

Having this insight can help allocate talent to where it may bebetter served.  In turn, this can help the claimsorganization and the individual claims handlers.  Moraleis normally higher when a claims handler feels well-matched for hisor her role and is actively contributing, not caught up inrepetitive, administrative, or lower-value tasks. The insightsgleaned from a CPM can offer claims professionals additional timeto focus their energy on the claims that matter most. With moretime to focus on claims files that truly affect outcomes(delivering value to insureds/claimants as well as facilitatingmore favorable financial outcomes), claims handlers can bettermanage caseload and do more with less. 

This acknowledgement of value in the collective claims handler'sknowledge base and experience, coupled with its scarceavailability, is central to the design and rationale for buildingand implementing CPMs. Having a more thorough understanding of aCPM's design and business intent is helpful. However,consistently and effectively utilizing the insight generated from aCPM and doing so in a balanced manner is often critical to longterm success. While the financial benefits derived from a CPM canbe significant, those who also appreciate the impacts on humancapital and the claims organization can realize the true value of aclaims predictive model.

This publication contains general information only and isbased on the experiences and research of Deloitte practitioners.Deloitte is not, by means of this publication, rendering business,financial, investment, or other professional advice or services.This publication is not a substitute for such professional adviceor services, nor should it be used as a basis for any decision oraction that may affect your business. Before making any decision ortaking any action that may affect your business, you should consulta qualified professional advisor. Deloitte, its affiliates, andrelated entities shall not be responsible for any loss sustained byany person who relies on this publication.

Copyright © 2012 Deloitte Development LLC. All rightsreserved.