The new age of analytics has led to the development of a variety of solutions that leverage internal and external claims-related data to enable claims handling resources to take preventative actions earlier in the life cycle of a claim. From first notice of incident (FNOI), and resource assignment, to the identification of potentially fraudulent claim activity, models are changing the claims handling world. However, one critical component of helping organizations achieve maximum efficiency using predictive models is the upfront investment claim that organizations make to ensure that the models are implemented effectively.

Modeling Basics

Claims predictive modeling combines internal claim characteristics and external third-party data to calculate a mathematical score that allows claims to be segmented at FNOI and throughout the life of the claim. Several hundred variables taken from background information about the employee, the work injury, external public databases, medical data, and other sources are statistically tested to identify the 50 to 100 candidate variables with the greatest predictive power. The final model’s variables are determined by leveraging a number of model development techniques (for example, correlation analysis, principal components analysis, variable prioritization, exploratory data analysis, and so on), iterative training, and testing of candidate predictive models to evaluate the statistical significance/confidence, robustness, and business reasonability of the candidate variables.

Models for workers’ compensation can take many shapes, but separate models are typically developed for lost time indemnity, lost time medical, and medical-only claims. The output from these models can be applied in many ways including:

Initial routing to the most appropriate unit (low-touch straight through processing/high-touch).
Initial assignment to an appropriately skilled adjuster.
Early identification of claims with a high propensity for fraud or that require medical management. Early escalation of claims to more senior resources (supervisors, regional technical specialists).

Throughout the lifecycle of a claim, models can be run on a periodic basis (monthly, quarterly, and so on) in order to consider any additional information that may help to improve the segmentation power of the model. For example, the receipt of medical bills and pharmaceutical information, along with the claimant’s medical history, can be factored into the model in order to provide medical management guidance throughout the claim. To the extent that the data indicates a potential for higher severity claims development, the case can be reassigned to a more senior adjuster or escalated to a specific technician. If the data indicates a potential for lower severity claims development than originally predicted at FNOI, then the skill level of the adjuster and involvement of medical professionals can be appropriately modified.

Bringing Model Scores to Life

In order to bring claim models to life, it is critical for an organization to invest time upfront to help ensure that the models are implemented into a production environment as smoothly and efficiently as possible. With the right amount of proactive planning, the organization can transition from model development, to claim scoring and the delivery of high impact business workflow on the claims adjuster’s desk top. Traditionally, business rules can be developed in a format that is easy to interpret. Instead of simply being called a score of 950, typically three or four descriptive, rationale-oriented sentences can be used to explain roughly 85 to 90 percent of the drivers behind the score.

In order for this to happen, it is critical to address the following six areas while the models are being developed and tested:

1. Process/Workflow Changes Implementing a claims predictive model requires more than just generating mathematical equations and the resultant scores. Assessing the numerous impacts of the model on current business processes helps to integrate the new claim scoring methods in an effective and efficient manner. An initial assessment of potential impacts on the process will help to identify areas where efficiencies can be gained and risks can potentially be mitigated. For example, referrals to key resources can be automated to reduce the manual burden and shorten each phase of the claim lifecycle. However, it is important to ensure that automated referrals are sent to resources with the appropriate skills and capacity to handle the claim. The risk of inappropriate referrals can be mitigated through the development of a skills matrix and online calendar to guide the automated referral workflow.

2. Business Rules and Workflow Management An analysis of the existing business process rules assists the organization in assessing the expected impact of implementing a claims predictive model. The impact assessment focuses on enabling the optimal alignment of current claims routing and handling the new rules that incorporate claims predictive model scores. The new rules, which reflect lessons learned from the predictive modeling exercise, also help the organization to enhance their existing approach with a strong focus on achieving, then exceeding current industry-leading practices. For example, predictive modeling can assist with the automation of referrals to nurse case managers. By revisiting existing referral rules and incorporating model output into those rules, a company can focus their medical resources on the most complex medical claims. At the same time, these valuable resources can be directed away from lost time only claims.

3. Technology Changes A claims predictive model impacts business processes as well as the supporting technology. Assessing the existing technology infrastructure and how modeling outputs such as the score and reason messages, and modeling inputs such as data feeds from third parties, can be incorporated helps to establish an integrated business and technical solution strategy. For instance, enhancing the integration to an existing claims administration system screen by supplementing select hub user screens with model outputs and reason codes could help ensure that new workflows would be implemented more seamlessly. In addition to enhancing systems, improved data capture and storage in existing systems may be required to support the model calculations and scoring engine. As part of the modeling exercise, it is not uncommon for organizations to identify valuable new data and key performance indicators that were not captured before, or if they were, were captured sporadically. To the extent that data can be captured going forward, the impact of the model can be improved.

Given that many claims organizations are already considering upgrading or replacing their core claims management capabilities, it might be the perfect time to integrate advanced analytics while the “hood to the car” is still open.

4. User Testing As with any technology implementation, testing by process owners and day-to-day users is essential in order to fold the new functionality into existing business processes more efficiently. It is not out of the ordinary for this phase of the development lifecycle to command 30 to 40 percent of the time and resources, particularly when modeling is involved. Test strategies, plans, and scripts should be developed in accordance with the expected processes and business rules. Aligning testing resources by key areas (by line of business, authority level, level of expertise, and so on can help ensure the appropriate level of scenario testing before the system goes live to minimize any potential disruptions.

5. Training Any major process change or technology implementation should be accompanied by a training program that takes the end-user environment into account. The most effective training programs include system demonstrations using sample claim scenarios that are realistic and relevant to the users of the system. Appropriate training helps to ensure that claims personnel fully understand, accept, and then embed the new procedures into their day-to-day claim adjusting activities. Without the appropriate training and buy-in from claims personnel, organizations run the risk of inconsistent adherence and use of the models.

6. Change Management Managing the change resulting from model implementation is important to the successful adoption of the enhanced processes, rules, and systems ultimately targeted at achieving improved business results. Having a small, yet dedicated team of resources (ideally supplemented with claims organization change agents) to judge the impact to the existing organization, engage key stakeholders, and communicate strategic messaging aids the successful adoption of the system, and moves the organization towards full productivity throughout implementation.

The combination of developing advanced predictive models, using a variety of traditional and non-traditional information from internal and external sources, and then undertaking a thorough implementation addressing all aspects of change, can create significant measurable benefit across the entire workers’ compensation claims organization and reduce overall claims costs.

With the right amount of planning upfront and the coordination of all parties involved, a claims organization can effectively implement predictive modeling solutions to realize both short and longer-term benefits. It is important to remember that model development should never be conducted in a vacuum, though. By considering changes to processes, user testing, and the open sharing of business rule impacts throughout the model development, the organization will be much more likely to end up with an effective, productive solution, instead of a very expensive business calculator that sits on adjuster’s desks collecting dust.

Special Reports /

Leveraging Analytics In Workers&rsquo; Comp Claims Handling

Related Stories

Recommended For You

Leveraging Analytics In Workers’ Comp Claims Handling