1. Excess or shock claims, especially their timing, number and amount, are examples of real world disruptions to a health actuary’s IBNR calculations. There are other outside Influences on Health Claim Reserves and Patterns
2. Shock claims have a material impact on completion factors produced by development IBNR calculation methods. Often the adjudication time for these excess claims is longer; thus, when they are paid, they can lower all paid lag month’s completion factors, raising the overall claim reserve produced.
3. By incorporating the excess claim’s impact (e.g., lower completion factors), one is essentially providing an ongoing reserve for a similarly expected excess claim. Alternatively, in the rare case that the large excess claim is paid much faster than other claims, the resulting completion factors will be increased, thus lowering reserves, a likely unwanted result.
CLASSIFYING EXCESS CLAIMS
1. The most common way to account for excess claims’ impact on the development methods’ completion factors is to remove the excess claims completely from the completion factor calculation.
2. If done this way, the actuary then has two choices:
(i) keep the excess claims in the cumulative paid claims to which the completion factors are applied or
(ii) continue to keep the excess claims removed even from the cumulative paid claims and thus need another method to calculate claim reserves for future potential paid excess claims already incurred.
3. A key factor in the choice of whether to remove excess claims from completion factor development is the relative materiality that excess claims have on the block in terms of size, completion factor impact and reserve impact.
SEASONALITY
1. One type of seasonal variation is due simply to the calendar. Not all months have the same number of days, so the total claims incurred in each month will vary. Holidays also affect seasonality as people are much less likely to go through with elective procedures during extended holiday periods, especially late in the year during Thanksgiving and Christmas.
2. It takes analysis of several years of data in order to develop a stable history from which to draw results and further complicated by the fact that benefits and plan provisions often change from year to year.
3. It is usually only feasible to attempt to measure seasonality on mature blocks with several years of historical data that have not had any major changes to benefit plan design.
4. If an actuary is setting IBNR for a health product that exhibits a material and consistent pattern of claim incurrals for certain days of the week, then the actuary could use that information to (a) be a basis for the IBNR and/or (b) influence the judgment for decisions in employing a particular method.
CLAIMS COSTS TRENDS
1. This trend is made up of two pieces: unit cost trends, the increase in per service unit cost for a particular service over time; and utilization trends, the tendency for a people to utilize more services over time. Because these trends can run at more than three times the average inflation rate, they are very important to consider when establishing reserves.
2. it is important to note that other issues can distort the trend values, resulting in understating or overstating the true claims trend. Changes in benefits, provider reimbursement levels and demographics can all affect the cost of services, which can compound the effective trend when analyzing historical data.
3. It is very important to analyze trend when using exposure-based IBNR methods, Historical loss ratios or Per Member Per Month (PMPM) amounts are used to project future claims, and if these amounts are not adjusted correctly for trend, the result could be misestimating the claim reserve. Often graphing and fitting trend lines to monthly PMPM incurred claim levels is a useful tool for estimating PMPM incurred claim levels for recent incurral months where calculated reserve levels are unstable yet large.
GROWING/DIMINISHING BLOCKS
1. The beginning and ending phases, however, where the block is rapidly changing in size, present their own individual sets of challenges when it comes to calculating the claim reserve. When a product is first issued, there is no claim history on which to base a reserve. If a similar product exists in the company’s portfolio, the data from that policy could be used to help estimate a reserve, but the credibility of the data for the new product still creates issues.
2. The primary way to generate reserves early on is to use the Loss Ratio method with the pricing loss ratio. This method is normally used until the block reaches a size at which its own data can be considered credible.At this point, assuming there is enough historical data, other IBNR methods can be used.
3. At the other end of the life cycle, when the block is decreasing in size, a different set of problems is encountered. While sufficient historical data is not a problem, you can again run into the problem concerning the credibility of the data. Without solidly credible data, the lag methods are difficult to use since they will generate volatile results.
4. Exposure-based methods again have to be used. However, this is compounded by the tendency, especially for individual health insurance, for a rate spiral to begin. This results in a block of business that gets less healthy as the healthy insureds lapse and look for better coverage while the ill population persists.
PREMIUM CHANGES
1. Premium changes only directly affect reserve calculations using premiums (e.g., Loss Ratio method). Indirectly, rate increases can affect the growth rate and demographic makeup of a block, creating some of the issues described in the previous section.
2. When using the Loss Ratio method, or modification thereto, it is prudent to be aware of the rate increase percentages, their implementation dates, impact on block size and possible anti-selection. If exposure (e.g., membership or insured count) is available, these can be tracked by developing per capita premium as well as per capita incurred claims on a monthly (or quarterly) basi
BENEFIT CHANGES
1. the Loss Ratio method is the only method that is mostly unaffected by benefit changes, with the exception that over time it is expected that premiums will change relative to claims changes in response to a benefit change.
2. The level of effect benefit changes has on the different lag methods depends upon the type of benefit changes implemented. If the changes are in the form of cost sharing, benefit limits or other monetary type changes, the lag methods should not be affected since the underlying payment pattern will not change.
3. The methods that use membership as the exposure basis (e.g., PMPM and Paid PMPM methods) would be affected the most by a change in benefit structure. Any type of benefit structure change could result in the average claim size moving one direction or another. Historical averages are no longer applicable because of the changes. For the initial months after the benefit change, the claim reserve would have to be calculated by either estimating the average PMPM amount, or by using one of the other methods.
CLAIMS ADMINISTRATION DISRUPTION/BACKLOGS
1. Claims inventory or claims adjudication processing changes are a special example of real world disruptions to a health actuary’s IBNR calculations. Staffing changes, missed work days due to weather or other unexpected events, administration system hardware/software changes and changes in how providers file claims are just a few of the types of situations that cause disruptions and backlogs.
2. If the actuary were tracking the working days in a month and assessing the In Course Of Settlement (ICOS) reserve separately from the true IBNR reserve, he/she could use the days in each month for which the payor’s claim department had available-to-adjudicate claims to assist in the assessment of ICOS and possibly of the reasonability of pending claim levels. For example, it would be reasonable for the month-end ICOS to be relatively high for months in which the number of working days is low.
THOUGHTS
1. Advances in computing power and increasing software sophistication are making new and nontraditional methods more accessible to practicing actuaries. Yet, more sophisticated IBNR calculation methods, while possibly more accurate, often require additional assumptions and supporting data as well as adjustments to derive reasonable results. These methods may also require advanced technical knowledge and the purchase of specialized software as compared to those methods generally in use by practicing health care actuaries.
2. Most health care actuaries use a variety of methods to estimate IBNR, and the preferred method may be a combination of two or more of the tested methods (or other ones). For instance, lag methods tend to be the most common methods used by health actuaries; however, the results consistently indicate these methods have the highest standard deviations.
3. Understanding a health plan’s particular facts and circumstances (current environment) may be the most important assumption to specifically document prior to deciding on the most appropriate method for a specific analysis.
(Source: SoA)
