By John C. Peters, PhD, Chief Science Officer, Gain Life, Inc.
One of the first questions asked by claims leadership when deciding whether to make an investment in a new claims “solution” is, what will be the return on investment (ROI)? Will this new solution reduce claim costs? Will it improve claims processing efficiency? Will it reduce litigation? Will it improve the claimant experience? Implementing a new claims solution could affect each of these areas – be it automation, communications, risk monitoring, analytics, etc. Assessing the value of such a product for each area requires using different analytical approaches and different data.
This article briefly outlines different approaches to assessing the ROI for claims costs when implementing a new claims automation solution.
Determining the effect of implementing a new claims solution on claims costs would seem to be a straightforward problem. Just compare claims costs before product introduction to claims costs after implementation to see if costs went down. While this seems like an obvious approach, it does not account for the many variables that influence costs besides just introducing the solution, and these variables may not be constant in a given population of claims over time.
Many Factors Influence Claim Costs
Among the many factors that may influence claim costs are things like the demographics of the claim pool, injury severities, treatments applied, comorbidities, psychosocial factors, differences among claims professionals and their use of the software, the jurisdiction in which the claims occurred which affects the selection of treating physician, etc. In addition, comparing claims from one year to previous years (historical controls) for example, introduces potential effects on outcomes due to other environmental events (e.g., pandemics, inflation, changes in work circumstances/environments, changes in regulations, improvements made to existing claims systems as part of internal continuous improvement plans, etc.). Any of these factors could affect claims costs such that differences between costs before and after implementation may not be caused by the new solution itself, but by other things happening simultaneously. Thus, potential buyers should beware of vendors’ claims of substantial savings on claims costs that seem otherworldly, unless the seller can demonstrate application of a rigorous approach and methodology in analysis of the data.
What constitutes a rigorous approach?
The most rigorous and defensible approach for determining ROI in terms of claims costs would be to conduct a prospective, randomized controlled trial (RCT) comparing cost outcomes for claims randomly assigned to either the existing system process (the control condition) or the new system/solution (the treatment condition).
The RCT Approach
A randomized controlled trial (RCT) is the gold standard for determining causality between a treatment and control conditions. Either this would involve randomizing incoming claims to the existing system or to the new system, so that there would be no selection bias in determining what claims are processed by what system. It would involve randomizing sufficient numbers of claims that the characteristics of the claim pool in each condition (treatment or control) are similar, minimizing any random bias that might have occurred by chance. For example, you would not want to compare two groups that had significantly different types of claims or claimant demographics, etc., as any differences in cost might be due to those differences and not to the new solution. In addition, you might also want to randomize which adjusters use what system so any differences between adjusters in the usage of the different systems would be randomly allocated to minimize potential bias. After some period and accumulation of claims, the costs between the two groups could be compared after statistically adjusting for any residual differences between the two groups in key characteristics like demographics, claim severity/complexity, state of jurisdiction, job type and other measured variables.
While the RCT approach is the best for determining causality of a treatment (in this case the effect of new solution on cost outcomes), it is likely impractical for most companies with going businesses, so other approaches are necessary. One such approach often used in the pharmaceutical industry to compare the outcomes between two or more different treatments (e.g., a new drug and the current standard drug treatment) is a “comparative effectiveness analysis” (CEA). This analysis would compare two treatments (claims processed by the new solution/system vs existing system) to see which process results in lower claims costs.
The CEA Approach
The CEA approach also requires careful selection of the groups to compare to minimize selection bias and to ensure that the groups are as similar as possible in key variables that may influence claim outcomes. One advantage of this approach over a prospective RCT is that it can be done with data already collected. Ideally, you would want to select a “control” group (claims using the existing system/process) from a large population of claims, only some of which are using the new software and not an historical control from previous years. Using an historical control could introduce bias due to the different period as well as due to any differences in claims that occurred during that historical period vs. recent claims. For example, you would not want to compare costs for a population of claims from 2022 using the new solution to historical claims from 2020 during the middle of the pandemic. A good comparator group would be claims that did not use the solution but were collected over the same timeframe as those that used the software. You would still adjust for potential differences in important variables across the two groups (e.g., demographics, claim complexity, state of jurisdiction, etc.) to increase the probability that any differences between the groups are due to use of the software product.
The Importance of Pre-Planning
It is apparent from this brief discussion that pre-planning is important when considering how to approach measuring ROI. It is best to determine before you implement a new solution how you will assess claim cost differences, what comparator group you will use to assess any cost differences and to plan for what variables you will statistically adjust for to try to make the two- comparator groups as similar as possible except for the treatment (the new solution). The goal is to compare “apples to apples” and not apples to oranges. You may have to collect new data that you never collected before or use data from many different sources to create the best set of independent variables as factors for statistical adjustment.
In a later article, we will discuss approaches to measuring claims management and operational efficiency as well as claimant experience which are additional factors contributing to the overall cost effectiveness of investing in new, innovative, claims solutions.
About John Peters
John C. Peters, PhD, is the Co-founder and Chief Science Officer of Gain Life and Retired Professor of Medicine at the University of Colorado, School of Medicine. Previously, John spent 26 years at the Procter & Gamble Company in various R&D leadership roles. He has served on two National Academies of Science, Institute of Medicine committees, and has authored over 160 scientific papers and book chapters and published two books.
About Gain Life
Gain Life (www.gainlife.com) is an InsureTech company born out of Harvard University’s Innovation Lab, and backed by MassMutual, General Catalyst, Unusual Ventures, and insurance industry luminaries. We build software to help people and organizations return to health, work, and productivity. Our claims communication automation platform is utilized by carriers, employers, and TPAs to save claim costs and provide a better claimant experience across multiple insurance lines-e.g., workers’ compensation, general liability, commercial auto, and disability.