The three main intended benefits of PREDiCT are improvement of reserving accuracy, shorter claim lifecycles and reduction of overall indemnity spend. This article looks at two case studies. These cases demonstrate how PREDiCT’s data driven insights have been used to achieve favourable outcomes for our clients.
Case study one (early favourable settlement leading to lower overall indemnity spend)
This claim arose out of a head-on collision which occurred in August 2023, and involved a 50-year-old builder. Liability was not in dispute. The claimant alleged that he suffered a traumatic brain injury resulting in cognitive impairment as well as a fractured right wrist, left knee and hip injuries, and exacerbation of pre-existing shoulder injury. The Immediate Needs Assessment undertaken a few months post-accident identified ongoing orthopaedic issues, poor memory, reduced concentration and headaches. He was able to return to work after a few weeks, but in a limited capacity. He claimed significant loss of earnings. The claimant’s legal representatives indicated that a multi-disciplinary medico-legal approach would be pursued, including expert reports from an orthopaedic surgeon and neurologist in the first instance, and potential further reports from an upper limb surgeon, neuropsychologist, and neuropsychiatrist.
Our case handler utilised PREDiCT modelling for the purposes of a reserve prediction, which gave a recommendation of £349,825. In addition, it was considered that an early offer would be desirable. PREDiCT uses a quantile regression method for modelling, which provides outputs with reference to a percentile. Case handlers can use PREDiCT’s percentiles as a guideline for pitching early offers – a helpful resource to use data driven insights to validate their own assessment as to the appropriate level of an early, competitive offer. In this case, an offer at around the 25th percentile of damages was considered to be appropriate, and an offer of £170,000 was made. This offer was accepted, and the case settled approximately two months after Weightmans were instructed. This settlement represented a significant saving against the damages reserve, and early settlement of the claimant’s legal costs. Clear savings were achieved against the potential for significant disbursement costs, and expert fees.
Case study two (superior reserving accuracy)
This matter involved a claimant motorcyclist who was injured when the insured’s vehicle emerged from a minor road into his path. The claimant sustained injuries including a fracture to left tibia and fibula and subsequent development of compartment syndrome affecting the left leg, fractured ribs and pulmonary embolism, and psychological sequelae. He was unable to return to his pre-accident employment as a motorcycle fast food delivery rider. He did not make a complete recovery from his injuries, and had residual functional limitations including a claim for future assistance to look after his disabled son.
The client opted to maintain a damages reserve of £675,000. PREDiCT recommended a damages reserve of £560,681. The matter proceeded to a JSM, in advance of which the claimant’s Schedule of Loss totalled £1.59 million. Ultimately, settlement was achieved at £550,000 gross.
It is impressive that PREDiCT recommended a damages reserve which ultimately fell within approximately £10,000 of damages paid. Of course, PREDiCT does not (and cannot) provide a damages prediction which falls within 2% accuracy compared to damages paid on each and every case. However, the key takeaway is that PREDiCT has a proven track record of improving reserving accuracy and reducing excess capital lock-up for our clients. As demonstrated with the second case study, it is very natural for case handlers in large loss claims to be cautious when approaching reserving. However, this approach can lead to a propensity to over-reserve cases. PREDiCT’s data driven insights are provided based solely on settled case-data, and without ‘human’ emotion or caution factored into its figures.
As part of our rigorous testing and ongoing scrutiny of PREDiCT, we closely examine performance against settled outcomes. We analysed 147 settled cases in which our clients had the benefit of PREDiCT’s insights (following its full-scale launch to the market in June 2021). Our analysis demonstrated a significant reduction in the level of client over-reserving when compared to damages ultimately paid. The level of over-reserving reduced from +138% compared to settlement outcomes before PREDiCT was available to +55%. Across the portfolio of 147 cases, the improved reserving accuracy resulted in £107 million of capital being redeployed by our clients into other parts of their businesses, rather than being unnecessarily locked-up in reserves against those claims.
The dataset upon which PREDiCT is trained is continually growing, and we continue to develop modelling sophistication and accuracy. We expect to see yet further improvements to the benefits of PREDiCT, and its established track record of impressive outcomes.
