Trends in ACA Risk Adjustment: The Rich Get Richer?

Published in

accordionhealth

5 min readJul 4, 2016

The U.S. Department of Health and Human Services (HHS) recently published a summary report on Affordable Care Act’s transitional reinsurance (RI) payments and permanent risk adjustment (RA) transfers for the 2015 benefit year [1]. The report revealed the amounts of RI and RA transfers between health plans and the program. Centers for Medicare and Medicaid Services (CMS) commented that both RI and RA programs are working as intended, compensating plans that enrolled higher-risk individuals.

Since the report came out, many health plans and RA vendors have shared their own findings and perspectives [2,3,4,5]. Among those, a blog post by Sean Creighton from Verisk Health [2] caught our eyes. He pointed out that there exists some degree of year-to-year volatility on the RA transfer amounts. In other words, some health plans who received money last year, may owe money to the program this year (and vice versa).

Degree of Year-to-Year RA Volatility

To see the degree of Year-to-Year RA Volatility, we have extracted RA transfer amounts data from the summary reports for the 2014 and 2015 benefit years [1].

The scatter plot above shows the RA transfer amounts in the individual market in 2014 (x-axis) and 2015 (y-axis). Each point in this chart represents a health plan that participated in the RA program for both years. Health plans in Quadrant A received money in both years, while plans in Quadrant C consecutively lost money. As can be seen, most plans are located in Quadrant A and C. This implies that the RA transfer amounts have been somewhat consistent over the years.

Do you see the Year-to-Year RA Volatility from the chart? Although not many, you can see some plans in Quadrant B and D. Plans in Quadrant B lost money in 2014, but received money in 2015. On the other hand, plans in Quadrant D received money in 2014, but lost money in 2015. Note that the chart is shown in the log-modulus scale for the simplicity of visualization [6], so the actual gain/loss amounts can be in the million dollar range. We think that some plans mastered how to improve their RA strategies: successfully maintaining high risk scores.

Although not presented here, we have observed similar patterns in the small group market. A majority of health plans are on Quadrant A and C, and a few number of plans are on Quadrant B and D. We confirmed some degree of Year-to-Year RA Volatility in both markets.

Finding Plans with Successful RA Strategies

Some plans have achieved higher RA transfers in 2015 than in 2014. To see who they are, we first introduce a new variable, “RA Improvement from 2014”, which is defined as follows:

RA Improvement = RA Transfer in 2015 — RA Transfer in 2014

where the sign of RA Improvement indicates whether a plan improved its RA strategy (positive: increase, negative: decrease). After a quick exploration of the dataset we created, we found a somewhat interesting pattern.

The box-plot above shows two groups of health plans in the individual market: health plans who received less (blue) and more (red) than $60 million in total from the RI and RA program in the 2014 benefit year. As can be seen, the red group’s RA Improvement is significantly (and statistically) higher than the blue group’s.

The Rich Get Richer in RA

Plans who received more than $60 million last year in the individual market improved $27 million in their RA transfers in the individual market on average. On the other hand, the other group of plans lost $1.36 million on average. It is important to note that this is a statistical statement, and not every plan in the red group improved their RA transfers. There are a couple of outliers, such as Kaiser Permanente (HIOS: 40513) whose RA transfer amount was decreased approximately $180 million compared to the last year’s amount.

The red group consists of 26 health plans, where about a half of them were Blues (BCBS, Anthem, etc.). We have noticed that the red group consists of somewhat larger plans who have been in the market for a long time. This finding is fairly consistent with the claims by small health plans that bigger payers have more experience and resources to perform better at the RA program than smaller payers [3,4,5,7].

Now, let’s switch our focus to the small group market. As you may have guessed, the trends in the individual market similarly appears in the small group market.

RA Polarization among ACA Plans?

The box-plot above shows two groups of health plans in the small group market: health plans who received less (blue) and more (red) than $6 million from the RA program. Similarly, the red group’s RA Improvement is significantly higher than the blue group’s. The red group improved $17 million in their RA transfers in the small market on average, while the blue group of plans lost $400,000 on average.

Our findings indicate that there exists some degree of RA polarization: plans with high risk scores in the previous year further improved their RA strategies this year. When we combined two datasets from the 2014 and 2015 benefit years, and focused on the differences between the RA transfers, we could clearly identify the factors that contribute to the RA Improvement.

Large plans have more data, experience, and resources. Implementing a good RA strategy can cost substantial time, money, and joint efforts by payers and providers. We find that this might be one of the many hurdles that small plans have struggled with their RA strategy over the years.

At Accordion Health, we have developed custom machine learning techniques to provide cost-effective RA strategies to small- to medium-sized health plans. We believe that using advanced data analytics can break the endless spiral of the rich get richer, and we hope you join our endeavor to achieve fairness of the market.