Thus, the opportunity for P&C claims handling improvement toimpact insurer profitability is enormous in several key areas:

Claims Adjusting Efficiency--Well above 40percent of claims adjuster's time is spent on activities that donot actively assist in bringing the claim to a prompt andreasonable conclusion. Inefficiencies lead to longer claimsettlement times which can impact customer satisfaction, drive uplitigation rates and negatively impact indemnity payments. Worse,these inefficiencies waste valuable adjuster time--an everdiminishing resource.

Claims Indemnity Management--Medical costinflation, for example--a key driver of both special and generaldamages--averaged two full percentage points above the ConsumerPrice Index between 2000 and 2010.

Loss Adjustment Expense--Inefficient processes,inappropriate use of claims adjusting resources and excessive legalbills, etc. is estimated at one to four percent of Net WrittenPremium (NWP) in leakage each year.

Claims Indemnity Leakage--Estimated at anannual rate between six to 10 percent of NWP.

Customer Satisfaction--Carriers achieving highlevels of satisfaction retain customers and enjoy lower customeracquisition costs. Among customers who indicate high levels ofsatisfaction with their carrier overall, 65 percent said they"definitely would" renew their policy with their auto insuranceprovider. Conversely, only 43 percent of customers who report lowlevels of satisfaction said they would definitely renew theirpolicy. Claim service can dramatically impact customer satisfactiondriving 44 percent of the overall insurer impression by customerswho filed a recent auto claim.

Combined Ratios

As combined ratios remain high and investment portfolioscontinue to yield low returns, the need to drive maximum efficiencyin the claims handling organization becomes much more important. Ashas always been true, the adjuster is the first line ofdefense.

Imagine if you could have your best adjusters handling each ofyour claims and that those adjusters were guaranteed to be with youfor years to come. Each of the key concerns listed above would begreatly reduced. Your best adjusters by definition are highlyefficient, do not make inappropriate claim or expense payments anddeliver the highest levels of customer satisfaction.

Unfortunately it isn't possible to have your best adjustersassigned to every claim. Not only are carriers losing their toptalent each year due to the normal employee transitional flow of afree market society, but the majority of experienced adjusters areheading for retirement--70 percent are over the age of 45 and 33percent are over the age of 55.

To make this situation worse, new talent is not joining theranks of carrier's claims departments--only four percent are underthe age of 35. In fact, Deloitte Consulting predicts a shortage84,000 claims adjusters in the United States by 2014. It becomesnecessary to preserve these highly experienced and talented claimshandling resources for the claims (or pieces of the claim, such asfeatures or individual coverages) that require their extensiveexpertise, where they can have the most impact while having lessexperienced resources handle more routine tasks. Carriers, however,can struggle to consistently separate those claims that can benefitfrom early intervention of expert adjusting resources from the lowrisk claims. In addition, too often certain claims remain under theradar until they morph in to the complex or high cost entities thatthey really are.

Helping Hand

The adjuster needs a hand. What is needed is a fundamentallymore advanced, more sophisticated approach--one that combinesmodern predictive-analytics and data-driven claims workflowsolutions.

Predictive analytics is a branch of advanced statisticalpractice that specializes in data pattern recognition. Predictiveanalytics examines the data elements surrounding a specific knownevent and then examines the historical changes in those dataelements to determine when and how they changed in a statisticallysignificant way.

Models are then built that accurately and consistently identifyunknown events that fit the pattern, allowing prevention and/ormitigation techniques to prevent or lessen the impact of thepredicted event. Statistical predictive modeling can automaticallydetect extremely complex, subtle patterns in the data that humansalone cannot. Predictive models can also adapt over time tochanging conditions.

Two Decades of Service

Predictive analytics has been applied effectively within thefinancial services community for over 20 years. One of the bestknown applications is around credit card fraud identification wherecombined data driven and technology supported efforts in the creditcard industry has shown very impressive results with modelsscreening 85 percent of U.S. credit card transactions for fraud,resulting in a 50 percent reduction in industry losses. There isnot a credit card company operating today that would want to riskdoing business without predictive-modeling-based fraudidentification tools.

Property/casualty insurers also routinely use predictiveanalytics in their underwriting organizations. One key area is inrate segmentation. P&C carriers have also started to usepredictive analytics in their claims areas with most efforts todate focused on claim fraud with impressive results. DeloitteConsulting reports, "Although these tools need to be used carefullyand fairly, they can have a significant positive impact....(companies) typically achieve reductions in claims costs of 3percent to 5 percent and some attain savings as much as 5 percentto 10 percent."

But the application of predictive modeling in P&C claims hasnot yet realized its full potential to transform the way all claimsare handled. The application of predictive modeling to the claimsprocess more broadly offers tremendous potential to improveoutcomes for an insurer's entire book of claims.

Data and Workflow

However applying predictive models to real world claimsprocessing is not without challenges.

Two key challenges are around data and workflow.

Essential to highly effective predictive models is the scope ofdata included. Predictive models benefit from lots of data. Sincethey can evaluate complex data patterns, generally the moredisparate data that can be brought to bear on a business issue, themore accurate the decisioning can potentially be. A fundamentalquandary in claims handling is that the more data that is madeavailable to the decision making process, the better decisions thatcan be made; but humans become quickly overwhelmed by more and moredata.

The solution to this quandary is not to restrict the amount ofdata used to examine the business issue but rather to automaticallyanalyze these large quantities of data into results that can beacted upon by humans. This is what predictive modeling does.

The real opportunity to maximize the impact of predictivemodeling lies in inclusion of more data. In addition to a carrier'sown data, predictive models that incorporate public records dataand other sources of data external to the insurance company can beresponsive to conditions relevant to insurance claims but externalto the normal data collected by insurance companies. For example,tapping into claims and policy data across companies gives themodels a broader perspective of the claimant or provider's history.Today the opportunity exists to leverage sources of data notpreviously available to predictive models for claims handling.

But as important as the quantity of data presented to predictivemodels is the quality of the data. High quality data is an absolutenecessity to obtain the maximum results from any predictivemodeling effort. Missing, poor or inconsistent data hashistorically hampered the success of predictive modeling efforts inthe P&C claims space, slowed the progress of individualcarrier's internal efforts and greatly increased their costs.

Another difficulty with working with insurance claims is thatmuch of the most valuable data resides in the freeform text of theclaim notes or loss description fields. Traditionally this data hasbeen difficult for automated systems to utilize. Techniques liketext mining and freeform text searching overcome this limitation.Text mining uses statistical methods to extract "meaning" fromfreeform text and present it to predictive models in a way that itcan be processed automatically.

Good Models

But having well-designed predictive models and lots of data onlysolves part of the problem. To obtain the full benefits of apredictive modeling solution, results must be actionable and mustbe delivered to the resource best able to influence the predictedresult in time to exert appropriate influence. The resultsgenerated by the analytic system need to drive a change in businessprocesses to be fully effective, and that means getting the rightinformation to the right person at the right time.

The best way to ensure that the full value of predictivemodeling is realized is to seamlessly integrate the output of themodel into a robust, modern and flexible claims decision supportapplication. The way this works in practice is that model resultstrigger automated alert guidance to the adjuster--or even initiatethe automated transfer of the claim file to more specialized claimshandling resources. It is important to incorporate the analyticresult into the user's workflow as seamlessly as possibly. The lessthe adjuster is forced to deviate from the normal claims handlingprocess the better.

Because claims handling is a dynamic process, with each claimevolving and changing over its life, the analytic process must becontinuous and dynamic as well, with each claim rescored every timedata changes. Any change in score initiates further adjustment andrefinement of the handling of the specific claim. The goal is thateach individual claim would be handled in the optimal manner--withlow risk claims paid quickly (thus achieving best opportunities forreduced indemnity and loss adjustment expenses and increasingcustomer satisfaction) and more complex claims or features handledby the adjusters best equipped and trained to handle them.

One complex and high impact example of where the potential ofsuch a system can be clearly seen is in the application of modern,integrated predictive modeling and workflow solutions to the earlyrecognition and closer management of high severity auto injuryclaims. Though these claims may not be the most frequent type ofauto injury claim, they are very important because of theirdisproportionate cost to the company. And, though BI and PIPfrequency has been falling in recent years, medical severity hasbeen rising dramatically, which makes addressing this problem evenmore critical.
This is an area of value where Mitchell and LexisNexis RiskSolutions have combined efforts to assist insurers. Mitchell hashigh quality data within its DecisionPoint solution, LexisNexis haspredictive modeling expertise and extensive public records andother specialized databases and Mitchell DecisionPoint has theSentry rules workflow engine--the three essential pieces for asuccessful predictive-analytics-driven claims optimizationsolution.

The solution provides insurers:

o Early identification within Mitchell DecisionPoint of highseverity medical claims

o Immediate routing of identified claims to specialized claimsresources via Mitchell Sentry rules engine

o Ability to make maximum efficient and cost effective use ofmajor injury units; nurse case management, medical specialtynetworks and other severe injury care management skill sets

o Increased opportunities for straight through processing of lowseverity cases

o Improved opportunity to achieve early ultimate loss reserveaccuracy as well as early loss adjustment expense and indemnitycontrol via immediate engagement of appropriate claims handlingresources

Catastrophic Injuries

Catastrophically severe injuries, such as significant burns orspinal cord injury are easy to identify as having high medicalexposure from the start and so reserves for such claims are usuallyinitially set appropriately high along with a highly detailedclaims handling plan executed by top claims handling experts.Conversely there are injury claims where the vehicle damage doesnot support high medical exposure, and so these claims are alsogenerally reserved and handled correctly.

But in some cases the nature and extent of the injury is moreambiguous. Sometimes these claims progress for some time before thetrue extent of the medical exposure is recognized and accountedfor. Generally there are two reasons for unrecognized excessivemedical exposure. In some cases the indications that the claim ismore severe than initially apparent are very subtle and easy tomiss until later in the claim. In other cases evidence ofincreasing exposure may be evident but not recognized leading theclaim to not be handled appropriately.

For example, the fact that a simple claim has no activity for aperiod of time is often not a good sign. It may mean that theclaimant has engaged an attorney and is accumulating bills. Howeverin the press of day to day claims adjusting the adjuster may have atendency to not notice this silence and ignore such a claim for toolong. In both cases applying predictive modeling to the data canproperly identify these claims and help to make sure they get theattention (and reserves) that they need.

In work with several clients Mitchell and LexisNexis RiskSolutions have demonstrated potential gains in applying apredictive-modeling and data-driven approach to the identificationof unidentified high medical exposure claims. Models were firstbuilt on several years of historic data. This data included thecarriers' own claim and policy data together with medical bill dataand other external data such as public records data. The modelswere designed to identify the trajectory of a claim from the datapresent early in the claim. When tested against actual data, thesemodels showed the ability to identify 60 percent of the totalmedical indemnity for the group of test claims 12 months laterwithin the top 10 percent of highest scoring claims. These resultswere achieved within 10 days of FNOL. In other words, these modelswere able to predict with uncanny accuracy what the exposure of theclaims would be 12 months later.

In an operational situation, the predictions generated by themodel are compared to the current reserve and claim handling planearly in the life of the claim. If the claim appears to beadequately reserved and handled at the appropriate claims resourcelevel, alerting is suppressed. However if there seemed to be adiscrepancy between the existing reserve and the projected exposureand/or the claims resources assigned, the claim is brought to theattention of the adjuster or manager (or both). The claims continueto be rescored throughout the life of the claim so that asconditions change, the models are able to alert the adjuster at theearliest possible moment that the data suggests an unexpectedtrend.

Summary Judgment
Addressing medical exposure issues is only one example of newopportunities to apply modern predictive-analytics and data-drivenclaims workflow solutions. Application to claims fraud still offerstremendous opportunity as does the recovery area and, of course,further striation of the book of claims by claims special needs (orlack of special needs) to obtain maximum efficiency andeffectiveness and value from your claims professional team.

Greater application of advanced predictive analytics, expandedaccess to robust data sources and tight integration with workflowoffer the opportunity to revolutionize the claims handling process.By generating precise results that allow for each claim to bedirected throughout its lifecycle in the most optimal manner, apredictive modeling and data-driven approach to claims handlingleverages the claim handler and specialist skills in the mostefficient manner, realizing the potential for significantreductions in loss and loss adjustment expense.

(Mike Mahoney is senior director of product marketing forMitchell International, Inc. and

John Lorimer is vice president product development forLexisNexis Insurance Claims Solutions.)