In late 2013, we discussed claims audits as a way to update andimprove claims programs at p&c insurance organizations,beginning with the September 2013 article titled, “WhyAuditing is a Springboard to Cost Reduction.”

After all, one of the most important outcomes of effectiveclaims management is to reduce claims costs and expenses to thelowest reasonable values, while adhering to the legal and ethicalrequirements inherent in good faith claims handling.

We also began a discussion about some of thesteps involved in auditing workers' compensation,liability, and property claims. However, claims audits representjust one of the many components of a quality assurance (QA)program that must be implemented to not only maintain but alsoimprove the claims-handling operations. The size, scope, andcomplexity of QA programs can differ significantly, dependingon:

• Size of the organization
• Lines of business
• Geographic spread of the claims administrator
• Training program for the claims staff
• Experience and expertise of the claims staff
• Structure of the program—for example, claims are managed byinsurers or TPAs versus self-administered.
• Extent to which vendors are used
• Client mix that you target
• Other market-specific and industry issues

Development of a comprehensive QA program, however, is beyondthe scope of this article. Rather, in this article,we will focus on metrics and evaluation methods that can beused on an ongoing basis to augment the claims audit component ofyour QA program.

Activity Based Metrics

The balance of this article will concentrate on two major items:activity-based metrics and results-based metrics. We referfrequently to activity based metrics when we discussclaims-management best practices. It would be fair to describe bestpractices as the requirements and guidelines that many claimsadministrators believe will provide the path to optimaloutcomes. Unfortunately developing and tracking performance bycompliance with best practices focuses on the activities instead ofthe results or outcomes, which are more difficult tomeasure.

Often adjusters concentrate on reaching the milestones or“checking the boxes” rather than concentrating on the best ways tomanage a given claim. Of course, let me emphasize thatbest practices are nevertheless useful in ensuring the properfoundation is set for managing the claim.

Here are couple of examples. Claims administratorshave created their own best practices, some of which differ in thetimelines for the required steps. Some of them include:

1. Make two- or three-point contact within one working day ofclaim receipt/assignment.
2. Input initial reserves within 1 to 5 days of claimreceipt (requirements may vary up to 30 days).
3. Complete the investigation within five working days of claimreceipt.
4. Develop an action plan within 14 days of claim receipt.
5. Create updated action plans every 45 days thereafter.
6. Require the claims supervisor to review the claim at 30 daysafter receipt.

These and other best practices are important to confirm thatclaims adjusters and supervisors understand the steps required inorder to begin the claims process in the right direction. It isalso imperative that the adjuster complete some of the steps beforethe next step can be performed both properly and in a timelyfashion.

Some of the metrics that a claims administrator should follow inthe initial stages of a workers' compensation indemnity (lost time)claim are presented on the next page.

For liability claims, best practices may be the samefor items 1 through 5, with the exception that the number ofcontacts may differ, depending on how many insureds and claimantsare involved in the claim. For property claims, items 1through 5 again apply, although the contact might be limited to theinsured. There will, however, be other best practicesspecific to (and therefore applicable) to those lines of business.This may include, for example, conducting site inspection withintwo days of contact for property claims.

Results-Based Metrics

These are the metrics that are most difficult, especially if themeasurements are for long-tail claims, such as workers'compensation indemnity (lost time) or liability bodily injury(BI) claims. In some cases, the final resolution may not occurfor many months or years, making it difficult to connect earlysteps, activities, action plans, vendor/specialist utilization, andmanagement procedures to the eventual outcomes.

Many companies have tried to use various outcome-based metrics.In some cases, this approach has backfired, because the solutionled to practices that caused issues far worse than the originalproblem. For example, some companies that were appropriatelyconcerned about significant reserve development or reservestairstepping instituted programs in which adjusters were requiredto report to the client or risk manager, should reserves reached aspecific threshold. In such cases, management mayhave reprimanded the adjuster for either allowing the claim toreach that condition, or for not setting thereserve accurately in the first place.

This punitive approach led to the behaviors in which theadjusters initially set unreasonably high reserves so they wouldnot be forced to increase reserves at a later time and bereprimanded for that occurrence. Or, perhaps they did notincrease reserves until the last minute, possibly when a settlementwas payable, figuring they would only have to be reprimandedonce rather than several times. The resultof these improperly administered, outcome-based metrics led toseveral additional problems, including:

1. If reserves were inadequate, then there could have been anunderstatement of liabilities, which were identified at a laterpoint when other financial problems may havearisen—and at the worst time to recognize theincreased liabilities.
2. Changes in reserving patterns may have made it moredifficult for a casualty actuary to project provide reasonableestimates of claims liabilities and expenses.
3. If reserves were overstated, then this may have ledto improper use of the company's financial resources, whichcould have been used elsewhere for more beneficial purposes.

In some cases, these reservingproblems—when perpetuated on a largescale—led to issues that were much morefar-reaching than the original problem of stairstepping reserves orif reserve development was thought to be excessive. In fact,in certain circumstances, the modified behaviors thatoccurred as a result of the “solution” led to the ultimate failureof the company.

So what are appropriate results-based metrics? The metricsshould be based on final outcomes that are tied back to specificsteps, activities, and management activities that occurred manymonths or even years previously. This makes it more difficult torespond to questions from senior management about how structural orprocedural changes and improvements are affecting today's paymentsand liabilities, but these metrics are based on sound logic.

One example of reserving metrics based on comparisons of finalvalues to earlier steps and activities includes:

• Comparison of the relationship between the final closing valueof a claim (at the only time when a claim's ultimate cost is knownand not subject to debate or conjecture) with the adequacy of casereserves at various points in the claim's lifecycle.

For example, if a claim was closed after 15 months, how did thereserves at 60 days, 90 days, 180 days, 90 days before closing, and30 days before closing compare to the ultimate total incurred cost,and what percentage of the final value did the reserve values reachat those points? The following line graph shows thecomparisons for three adjusters. We will assume for purposesof this illustration that they handled the same claim with the samefinancial outcome. Note that they had very different reservingpatterns, even though they started from the same point (possibly atable reserve mandated by senior management), and ending at thesame paid value.

Analyzing this line graph and the data behind it can lead toseveral hypotheses:

• Adjuster 1 either did not proactively managethe claim to gather the necessary information to make a reasonablyaccurate assessment of the case reserves during its life, or didnot increase reserves when information was received that shouldhave created a reserve change.
• Adjuster 2 seems to have obtained informationthroughout the life of the claim, and while a slight increase wasnecessary prior to closing the claim, the case reserves appeared tobe within reasonable ranges at various points in the claim's life,and certainly well before it closed.
• Adjuster 3 might have been subject to thereprimands that we discussed earlier, so s/he set a high reserveinitially, which required no adjustments along the way, but whichresulted in reserves that were overstated when compared to thefinal value.

Other scenarios may also explain these different reservingpatterns, but gathering information such as this will help claimsadministrators determine the range of reasonable reserve values andreasonable patterns, allowing them to spend their time reviewingand fixing the deviations from the normal range. Certainly ifadjusters consistently produce reserve patterns such as those shownfor Adjuster 1 or Adjuster 3, the company should takeaction.

Other examples that compare outcome values to earlier steps ormanagement activities may include comparisons of final values tothe timeliness of initial reporting; timeliness of initial contact(and a sense of the thoroughness of the initial contact);timeliness of liability or compensability decisions; timeliness ofthe use of return-to-work programs; and use of case managementvendors.

There is no question that creating metrics to measure claimsexcellence can be a challenging proposition, especially if acompany is limited by an old claims system, or has failed to orbeen unable to capture and retain data and information on theactivities and management steps it has taken. However, if theorganization is committed to continuous improvement, the companymust make opportunities to use the data it has maintained, initiatea program to capture more data in the future, and make somethingmeaningful out of the information that will not only measure theactivities, but tie the results back to those activities.