Managing Income Protection Claims the Smart Way
Issue: July 2017 | Disability, Life | Download PDF | English By Rob Frank
The optimal methods for managing Income Protection (IP) claims have been widely discussed. IP insurers may have an eye on the sometimes adverse customer experiences with IP claims that are aired in local media, or may note the fluctuating fortunes of IP in comparable international markets. Ultimately, any IP insurers’ experience is most assuredly affected by unexpectedly extended claims durations. Often, however, the medical model alone is not enough to explain the length of time a claimant remains absent from work; in fact biopsychosocial factors play a key role in claims durations.
Biopsychosocial factors mean individuals are at risk of becoming stuck, even entrenched, in a sick role and remaining unable to return to work. When a claim is first made, many of the biopsychosocial factors that will influence the outcome may be partly hidden. This article discusses how Gen Re believes it is possible to utilise other elements of the policy and claim to provide at least an initial view of potential risk.
Calculating claims risk
Triage in IP claims allows insurers to identify the claimants most at risk of becoming stuck. Of course, where the actual duration of a claim exceeds the expected duration, the risk for the insurer is a financial one. Gen Re Australia has developed a Claims Risk Scoring System to quantify the risks, either intuitively or through detailed analysis of historical claims experience.
The scoring system uses basic data – policy duration, benefit period, waiting period, sum insured, age, occupation, cause of claim, time of claim notification and previous claims history – to generate, perhaps by actuarial calculation or points score. These data elements determine the relative potential risk of each claim. Although developed primarily for use in Australia, the concept underlying the calculated approach to claims risk has clear application in other markets, with only slight modification.
Moral hazard refers to the concept that a party who is protected from risk of a loss will act differently than an individual who does not have that protection. Anti-selection is a sociological phenomenon in which individuals most at risk are also the most likely to buy life or health insurance; anti-selection, or non-disclosure, happens when a person acts deliberately or otherwise to withhold information that may actually be relevant to the coverage they want.
Scoring the data elements
According to a popularly held view, the longer an IP policy has been in force, the less likely moral hazard or anti-selection will occur (see also blue box). This in turn indicates it is less likely that the claim is in some way premeditated. Premeditation is a barrier to returning to work and adversely impacts likely claim duration.
During claim assessment, a points score is assigned according to the policy duration. A higher score is assigned to short-duration policies since the risk is higher. The score can be assigned in a linear manner, based on a presumed or notional level of risk, as shown in Figure 1. Alternatively, a more detailed analysis of the claims experience of the insurer’s book of business could be undertaken to produce greater accuracy. This could be achieved, for example, by comparing the duration of claims with the same cause or same occupation but different policy durations at the date of claim.
The data element “years to policy expiry” clearly shows the longer this time period, the longer the potential claim. It follows that this will be a key indicator of the potential financial risk of the claim. If the insurer’s book of business that is under review has defined benefit periods, the insurer may choose to remove this data element from the scoring system, as it will not distinguish claims from each other.
Claims are often notified at or close to the end of the waiting period, which potentially means the longer the waiting period, the more likely the claimant is to be entrenched in the sick role before the insurer has an opportunity to provide rehabilitation support. It therefore appears to make sense to “weight” the scoring system with higher risk points values for the longer waiting periods. In addition, the mere fact that a claim has been submitted in a case with a longer waiting period suggests the cause of claim is more significant and therefore potentially will take longer to resolve. We know the probability of a successful return to work reduces dramatically when a period of absence extends beyond the first couple of months.
However, if the book of business in question only has short-term benefit periods, or offers accident-only cover, it could be argued that the shorter waiting periods should be weighted higher. As claims in this scenario are more likely to be injury-based, and as the recovery periods tend to be more consistent, shorter waiting periods make it more likely that a claim will be paid for a longer period of time. Therefore, it is possible to devise a scoring system for this particular data element in more than one way, depending on composition of the portfolio.
Without a diagnosis, symptoms are invariably difficult to treat successfully. It therefore follows that these symptoms are less likely to resolve and the duration of a claim in such a case is less likely to be short-term. The scoring system scores claims with subjective symptoms, or where no diagnosis has been made, as a higher potential risk.
It is generally believed that people involved in manual and physical work represent a greater risk of significant injury than people in other fields. While this may not follow for illness or sickness, the overall combination of injury and sickness would still suggest a greater prevalence of claims in the more manual and physical occupations. For this reason, it could be argued that the higher points values should be applied to manual and physical occupations.
However, consideration should also be given to the increasing proportion of mental health claims, and we must question that impact on the points values assigned, as mental health claims have been more prevalent with professional or white-collar workers. Clearly, the scoring of the data element “occupation” is another calculation, similar to that for “waiting period”, could be presented in more than one way.
Up to this point, the risk associated with each data element has been assessed in isolation in order to produce a Claim Risk Score. However, evidence has emerged that certain data elements or risk factors, when presented together, have a combined risk value greater than the sum of the individual parts.
Co-existing risk factors
When looking to combine specific data elements, the first consideration is for data that are present on each claim and generally easy to quantify. These are termed “standard co-existing risk factors”, the most simple example being the “sum insured” and “time to benefit expiry” that together produce the calculation for potential liability of a claim.
An argument can be made for adding weighting to the Claim Risk Score reflect the potential liability of each claim, because this has an impact on the potential risk to the insurer of that claim and is not necessarily reflected adequately by looking at the data elements in isolation. Figure 2, which assumes no increase in the monthly benefit, shows this anomaly more clearly.
This additional loading can again be calculated by undertaking a detailed analysis of the impact of the potential liability figure on claim duration, or by simply creating a linear scoring table, as shown in Figure 3.
This amended score can be defined as the True Claim Risk Score. Insurers may review the merits of applying this additional weighting where it is certain that the duration of a claim will be short-term, or the benefit payable significantly limited below the sum insured due to offsets or the calculation of benefit entitlement.
Further data elements with specific combinations will lead to a greater chance of a high-risk claim, such as certain occupations and certain causes. These also need to be considered, but again, the issue to be determined is the weighting to apply.
The combinations of data elements could be extended further, but to do so, and to accurately analyse the effect of these on claim durations, while significantly improving the accuracy of the risk scoring charts, would be both hugely time-consuming and extremely complex. Arguably, as the purpose of the Claims Risk Scoring System is to assess potential risk purely for the purposes of triaging and allocating claims effectively, this isn’t entirely necessary.
Whatever manner the insurer uses to calculate these non-standard co-existing weightings, it can create an Adjusted True Claim Risk Score.
The focus at this point shifts from identifying what potential risk a claim presents to the question of which assessor is best placed to manage the claim.
Currently across Australia and New Zealand, it is common practice for claims to be allocated on the basis of set criteria, which may include geographical location, medical cause, current portfolio size, workloads, or simply on the basis of who’s turn it is to receive the next claim. This means that individual claims are not necessarily being allocated to the assessor with the appropriate skill set to assess the claim. Consider this: does it make sense that a trainee assessor is allocated a claim for Complex Regional Pain Syndrome with a high sum insured, while the most experienced assessor is allocated a claim for a simple fracture with a short-term absence expected and limited benefit payable? This is highly likely to have an adverse effect on claims management and, potentially, customer service and experience.
The advantage of a risk-based claims allocation system is that claims are matched to an assessor with the appropriate skill set, as it is preferable that those claims with potentially the longest durations or greatest financial cost be managed by the most experienced assessors. Gen Re believes this offers the best opportunity to increasing the chances of recovery, maximizing return to work prospects and minimizing claims costs. It requires more effort upfront to determine where to allocate the claim, but puts the right people in the right seats on the right bus, so to speak.
Figure 4 outlines how Gen Re believes a claims department would ideally allocate its resources.
The x-axis shows the Adjusted True Claim Risk Score and the skill level of the insurer’s assessors. The model assumes a relatively small number of inexperienced assessors and of highly skilled senior assessors (who may be managing a smaller number of claims so they have capacity to undertake other activity), with most assessors falling within these two extremes, which is ideally where the majority of cases lie.
Historically, a Claims Assessor’s authority level is a monetary figure that reflects the sum insured. This is just one element of a whole range of data elements that make up the potential risk and complexity of a claim. Are there benefits in taking a more holistic view of the claim, reflecting the competence of an assessor by basing his or her authority level on something additional to the sum insured? Could authority levels be attached to a holistic value, such as the Adjusted True Claim Risk Score produced by the Claim Risk Scoring System?
To award authority levels in this way is only possible once a book of business has been analysed to determine the full range of risk scores. An assessor’s authority level will more clearly reflect the level of experience and competence, and it will be further reflected in the cases seen. It may also provide the assessor with a clearer development pathway, in terms of seniority and promotion, if each assessor level is defined by a Risk Score, as shown in Figure 5.
Gen Re supports a claims management model that focuses on facilitating better claims resolutions by seeking to identify the claims that could potentially represent the greatest relative risk to an insurer’s business in either pure financial terms or extended durations.
While biopsychosocial factors play a significant part in claim durations and claim outcomes, a claims assessor can utilise other elements of the case – even before investigating biopsychosocial factors – to provide an initial view of potential risk, and for this purpose Gen Re has created the Claims Risk Scoring System. Once claims are triaged and potential risk determined, the claim is allocated to an assessor of appropriate experience and skill, with an authority level reflecting a more holistic view of the claim than simply the sum insured.