Coding improvement for commercial exchange plans: Is it worth the cost?

  • Print
  • Connect
  • Email
  • Facebook
  • Twitter
  • LinkedIn
  • Google+
By Corey Nathan Berger | 17 December 2015

Executive summary

As we end the second year with plans offered in the Patient Protection and Affordable Care Act (ACA) marketplace, health insurance companies (carriers) are trying to understand the value and potential return on investment of ensuring accurate and complete coding of diagnoses in order to maximize risk scores. Many smaller carriers are struggling to find the resources to perform member outreach or chart review in order to capture and submit all appropriate diagnoses to ensure their risk scores are accurate and complete.

In February 2013, Milliman published a briefing paper entitled, “Commercial risk adjustment and transfer payments: Are you ready?” which hypothesized that “carriers will have to be able to submit accurate diagnosis codes to HHS to ensure their risk scores reflect the actual risk of their population. If they cannot do that, then they are likely to end up subsidizing other carriers.”

Based on discussions with executives at carriers about their financial results relative to their expectations for 2014, we believe this hypothesis is still accurate. The key question that many executives are now trying to answer is “How do I determine whether the investment I make is providing a return?” In addition, these executives are also asking “Is there a return on investment if we increase our risk scores and all other carriers increase their risk scores by a similar amount?”

While it may be surprising, the answer for most carriers is “yes.” Regardless of what other carriers are doing, a carrier will have a return on investment unless it has enrolled a significant majority of the membership in the market or unless the cost of identifying and submitting diagnoses is significant relative to the expected increase in risk scores.

The remainder of this paper provides several examples that illustrate that the return on investment for an individual carrier is only minimally impacted by what other carriers do (unless the carrier owns a significant majority of the membership in the market) and instead is primarily tied to the increase in risk score achieved by each carrier relative to the cost of achieving that risk score increase.

Measuring return on investment

The Milliman briefing paper referenced previously provides a summary of the risk transfer payment formula and the different levers that impact the risk transfer amounts. We would recommend reviewing that paper in conjunction with this paper because this paper focuses strictly on the change in the risk transfer payment for an individual carrier based on its coding efforts. In this paper, we will provide several scenarios and illustrate the return on investment (ROI) under each scenario.

Because the scenarios are illustrative, we will make a number of simplifying assumptions. These assumptions will allow us to present concise results while not materially impacting the actual conclusions.

The assumptions are as follows:

  • There are four carriers in the state.
  • The largest carrier has 50% of the enrollment, two other carriers have 20% of the enrollment each, and our “illustrative” carrier has 10% of the enrollment.
  • Induced demand factor and geographic cost factors are 1.00 for all carriers.
  • The average actuarial value (AV) for all carriers is 0.75 (average of silver and gold plans).
  • The state average premium is $350 per member per month (PMPM).
  • The allowable rating factor (ARF) varies by carrier.
  • The plan liability risk score (PLRS) varies by carrier and equals the ARF multiplied by the AV.

Figure 1 summarizes these initial assumptions.

Figure 1: Summary of Initial Assumptions

Carrier Enrollment ARF AV ARF * AV PLRS
A 500 1.850 0.75 1.388 1.388
B 200 1.800 0.75 1.350 1.350
C 200 1.650 0.75 1.238 1.238
D 100 1.950 0.75 1.463 1.463
Total / Statewide Average 1,000     1.358 1.358

In our initial simplified scenario, the statewide average adjustment factor without risk selection (ARF * AV) and statewide average adjustment factor with selection (PLRS) are the same because the individual values for each carrier are also the same. The result would be no risk transfers, because the ARF multiplied by the AV for each carrier is exactly equal to the PLRS for each carrier. In this example, we have assumed that the combination of the ARF and the AV accurately predicts the actual risk of the population as measured by the PLRS. For purposes of this paper, we will assume that the actual PLRS for each carrier should be 10% higher than the values illustrated above, and that the only way to achieve the “actual” value would be to invest resources in coding improvement through member outreach or chart reviews and supplemental diagnosis submission. We will also assume that the cost of these efforts is $3.50 PMPM, or 1% of the statewide average premium.

For our second scenario, we assume that all carriers have been able to capture and report their diagnoses accurately and that those efforts to submit diagnoses result in all carriers increasing their risk scores by 10%, and that there are no other changes. Figure 2a illustrates the revised values with this one change in assumptions.

Figure 2a: All Carriers Increase Risk Score by 10%

Carrier Enrollment ARF AV ARF * AV PLRS
A 500 1.850 0.75 1.388 1.526
B 200 1.800 0.75 1.350 1.485
C 200 1.650 0.75 1.238 1.361
D 100 1.950 0.75 1.463 1.609
Total / Statewide Average 1,000     1.358 1.493

Figure 2b illustrates the transfer payments based on these revised values.

Figure 2b: Transfer Payments – Figure 2a Assumptions

Carrier Normalized ARF*AV Normalized PLRS Difference Transfer
A 1.022 1.022 0.000 $0
B 0.994 0.994 0.000 $0
C 0.912 0.912 0.000 $0
D 1.077 1.077 0.000 $0

Under the second scenario illustrated in Figures 2a and 2b, all carriers spent $3.50 PMPM to ensure accurate coding. None of them had any change in their transfer payments from the initial scenario as the transfer payments were $0 under both scenarios. The only result is an increase in administrative costs and a decrease in gross margin. The question a CEO or CFO may ask based on the results of this scenario is “why bother?” if the carrier has a reduction in gross margin. The following scenario answers this question.

In our third scenario, Carrier D has looked at the results of the second scenario where all carriers increase their risk scores. The result is a reduction in gross margin for Carrier D. Hence, Carrier D assumes that expending resources to submit accurate and complete diagnoses results in worse financial results, and Carrier D decides not to expend those resources, while the other three carriers do achieve the 10% increase.

Figure 3a illustrates the same items as Figures 1 and 2a with the assumption that only carriers A, B, and C submit accurate and complete diagnoses and increase their risk scores by 10%.

Figure 3a: Carriers A, B, and C Increase Risk Scores by 10%

Carrier Enrollment ARF AV ARF * AV PLRS
A 500 1.850 0.75 1.388 1.526
B 200 1.800 0.75 1.350 1.485
C 200 1.650 0.75 1.238 1.361
D 100 1.950 0.75 1.463 1.463
Total / Statewide Average 1,000     1.358 1.479

Figure 3b illustrates the transfer payments based on the values in Figure 3a.

Figure 3b: Transfer Payments – Figure 3a Assumptions

Carrier Normalized ARF*AV Normalized PLRS Difference Transfer PMPM
A 1.022 1.032 0.010 $3.54
B 0.994 1.004 0.010 $3.44
C 0.912 0.921 0.009 $3.16
D 1.077 0.989 (0.088) ($30.89)

Figure 3b indicates that by not pursuing any coding initiatives to ensure accurate and complete diagnosis submission, Carrier D will pay $30.89 PMPM to the other carriers, or almost nine times the cost of pursuing the accurate coding efforts (assuming the cost of the coding efforts are $3.50 PMPM). In this instance, the return on investment would have been almost 800%.

For our final scenario, we look at the transfer payments that would result if only Carrier D increased its risk score while all other carriers do not pursue coding efforts.

Figure 4a illustrates the same items as Figures 1, 2a, and 3a with the assumption that only carrier D increases its risk score by 10%.

Figure 4a: Carrier D Increases Risk Scores by 10%

Carrier Enrollment ARF AV ARF * AV PLRS
A 500 1.850 0.75 1.388 1.388
B 200 1.800 0.75 1.350 1.350
C 200 1.650 0.75 1.238 1.238
D 100 1.950 0.75 1.463 1.609
Total / Statewide Average 1,000     1.358 1.372

Figure 4b illustrates the transfer payments based on the values in Figure 4a.

Figure 4b: Transfer Payments – Figure 3a Assumptions

Carrier Normalized ARF*AV Normalized PLRS Difference Transfer PMPM
A 1.022 1.011 (0.011) ($3.81)
B 0.994 0.984 (0.010) ($3.71)
C 0.912 0.902 (0.010) ($3.40)
D 1.077 1.172 0.095 $33.29

Figure 4b indicates that if Carrier D is the only carrier to pursue coding initiatives, it would receive $33.29 PMPM from the other carriers or more than nine times the cost of pursuing the coding efforts (assuming the coding efforts are $3.50 PMPM). In this instance, the return on investment would have been 850%.

Conclusions

The figures included in this article illustrate that the ROI for improving risk scores in the commercial ACA market is dependent upon the following, which a carrier can measure:

  • The ARF and AV of the population enrolled with each carrier (with carriers that have above-average values for these items being more highly leveraged than plans with lower-than-average values for these items).
  • The cost of coding efforts.
  • The actual risk score increase.
  • The share of the membership for a carrier in the market, although it would take significant market share to eliminate any ROI. For example, in Scenario 4, if Carrier A (with 50% market share) was the only carrier to increase risk scores, its positive transfer payment would still be $16.64 PMPM.

One key item that a carrier cannot measure is what other carriers are doing to ensure accurate and complete coding. What this analysis demonstrates is that an individual carrier’s ROI is not materially impacted by what other carriers do, so each carrier’s decision on the level of resources it wants to expend is independent of this unknown.

Hence, carriers need to calculate their expected risk score increases and the cost of obtaining that increase as the primary determination of ROI. While carriers may increase their risk scores and see no reduction in their payments to other carriers or increases in their payments from other carriers, calculating the ROI based on the change in the transfer payment does not capture what the transfer payment would have been absent such coding efforts.

For more on Milliman's healthcare reform perspective

Visit our reform library at www.milliman.com/hcr

Visit our blog at www.healthcaretownhall.com

Or follow us on Twitter at www.twitter.com/millimanhealth