PREDiCT: Transforming Large Loss claims handling

PREDiCT: Transforming Large Loss claims handling

Here at Weightmans we have long recognised that historic large loss claims data could be the key to delivering insights, shaping strategies and informing early settlement offers. This realisation has led to the creation of PREDiCT — a predictive data and analytics capability developed by our in-house data scientists utilising data from over a 1100 individual claims harvested over a 12 year period. PREDiCT provides actionable insights to drive improved performance around reserving accuracy, reduced lifecycles and overall indemnity spend. 

Read transcript

Large loss claims present many significant challenges including changes to the Discount Rate, rising claims inflation across portfolios, the need for accurate reserving, often with limited insight into a claimant’s injuries, recovery and prognosis and drawn out lifecycles, resulting in indemnity leakage. These all impact on claims performance and bottom-line profitability. To tackle these challenges head on, we need to transform the conventional way of handling large loss claims. Weightmans has long recognised that historic large loss claims data could be the key to delivering insights, shaping strategies and informing settlement offers. This realisation has led to the creation of PREDiCT - a predictive data and analytics capability developed by our in-house data scientists utilising data from over one thousand one hundred individual claims harvested over a 12 year period and transforming it into actionable insight for our clients. PREDiCT uses a Meta model trained on our market-leading data set to deliver optimum reserving in the context of a particular claim. Performance validation since the model was launched into the market in July 2021 have shown:- Average PREDiCT modelling outputs to be within 13% of ultimate paid outcomes, compared with 208% on client-approved reserves. Cases on which PREDiCT was used to inform pro-active case handling strategies to have delivered a 27% reduction in median claims lifecycle. Using TBI as an example cohort of claims, we have seen reduced lifecycles result in median costs savings of over £24,000 per case – £1.2m for a book carrying 50 claims with this injury classification PREDiCT mitigates the lack of claim information at initial reserving by modelling for different outcomes, enabling handlers to validate their own thought processes. PREDiCT outputs can be used to inform early Part 36 offers without first gathering a body of costly expert evidence. We believe the benefits of PREDiCT are clear - BUT we recognise that there will be concerns about its efficacy and compliance. Built upon best practice machine learning technology, the modelling processes adopted are used extensively in healthcare, insurance and financial services, where data varies and extremes are common. Ethical data handling has underpinned the development of PREDiCT. Specifically we have liaised with clients and completed a rigorous GDPR audit, resulting in a data set that is not only of a statistical size and relevance but also manged to the highest ethical standards. Weightmans has, and continues, to conduct extensive testing which has shown the predictive performance of the model to significantly outperform handlers on unseen test matters. PREDiCT is an Augmented Intelligence approach, it does not and cannot replace claims handlers’ own experience and expertise. PREDiCT is part and parcel of our Large Loss claims handling service, providing modelling outputs delivered on individual cases auditing and stress testing of reserving strategies. It has also been used in a number of other, highly effectives ways; from portfolio analysis and targeted settlement projects to offering added rigour to due diligence and performance benchmarking, all of which are focussed upon reserving accuracy, reducing claims lifecycles and lowering overall indemnity spend. This data driven approach augments handler expertise to drive consistency of reserving and decision making, improving claims performance by accelerating lifecycles and reducing overall indemnity spend. Unlock the potential of PREDiCT for the benefit of your business.

Unlock the potential of PREDiCT for the benefit of your business

 

High value injury claims present significant practical challenges; not only is the reserving process often carried out with limited insight into the severity of a Claimant’s injury, it is also conducted against a backdrop of rising claims inflation, uncertainty surrounding the Discount Rate and wider economic volatility. All of this can lead to protracted claims lifecycles and indemnity spend leakage.

These all impact on claims performance and bottom-line profitability.

To tackle these challenges head on, we are transforming the conventional way of handling large loss claims.

How can PREDiCT help you?

Improved reserving accuracy

PREDiCT uses a Meta model trained on our significant data set to deliver optimum reserving in the context of a particular claim.

In July 2021 we conducted a full market launch of PREDiCT. Over the last 18 months we have analysed its success, focusing on the performance indicators below:

Average PREDiCT modelling outputs were within 13% of ultimate paid outcomes, compared with 208% on client-approved reserves

This improved accuracy has opened the potential for the release of significant capital lock-up for reinvestment purposes.

Reduced claims lifecycle

Cases on which PREDiCT was used to inform pro-active case handling strategies delivered a 27% reduction in median claims lifecycle.

Lower overall indemnity spend

Using TBI as an example, we have seen reduced lifecycles result in median savings of over £24,000 per case – £1.2m for a book carrying 50 claims with this injury classification

Making the complex simple

PREDiCT mitigates the lack of claim information at initial reserving by modelling for different outcomes, enabling handlers to validate their own thought processes.

Accelerating settlement times

PREDiCT outputs can be used to inform early Part 36 offers without first gathering a body of expert evidence.

COMPLiANCE

We believe the benefits of PREDiCT are clear — BUT we recognise that there will be concerns about its efficacy and compliance.