Health insurance plans: Cursed with dimensionality?
How health plans are standardized under the ACA
One outcome of the Affordable Care Act (ACA) was the creation of actuarial value categories by which Americans can compare the individual and family health insurance plans that are available to them. The U.S. Department of Health and Human Services (HHS) defines actuarial value as “a measure of the percentage of expected health care costs a health plan will cover” . By law, carriers are required to classify the plans they make available as belonging to one of four categories, or “metal levels”:
- Bronze plans, which are projected to cover 60% of expected costs
- Silver plans, which are projected to cover 70% of expected costs
- Gold plans, which are projected to cover 80% of expected costs
- Platinum plans, which are projected to cover 90% of expected costs
Furthermore, carriers are required to offer three additional variations of each silver plan to individuals and families whose household incomes fall below certain income levels, as specified each year by HHS.
One challenge regulators had in implementing the above requirements was deciding how to ensure that carriers were projecting their expected coverages in some consistent manner. After considering a few different options, HHS settled on what it calls an Actuarial Value (AV) Calculator. The AV Calculator is actually an Excel Spreadsheet that HHS updates each year and makes available to carriers. The Calculator itself, along with a detailed description of the year’s methodology is published by Center for Medicare and Medicaid Services (CMS) Center for Consumer Information and Insurance Oversight (CCIIO) web portal.
How should consumers really compare plans?
Given this background, one very simple way that a consumer might choose a health plan for the coming year would be to compare the premium charged with the AV, as specified by the plan metal level. As an example, we could consider the little Texas town that I live in. According to healthcare.gov, there are 20 Bronze plans available, 27 Silver plans, and 21 Gold plans. Looking at a representative case for a 40 year old individual, the choices are:
- 60% coverage for a $219 per month premium (+ 19 others)
- 70% coverage for a $260 per month premium (+26 others)
- 80% coverage for a $312 per month premium (+ 20 others)
But this isn’t really a very meaningful comparison, because we really want to compare the expected dollar out-of-pocket costs with the premium cost. This is where things get very complicated and confusing for the consumer.
Why? Because although a carrier is required to classify each of their plans according to the ACA metal level tiers, it is left to the carrier’s discretion as to how each particular plan achieves the AV goal corresponding to that tier. The ACA specifies that whatever AV calculator methodology HHS settles on, it must calculate the expected benefits according to 10 Essential Health Benefits (EHB’s) :
- Ambulatory patient services
- Emergency services
- Maternity and newborn care
- Mental and behavioral health
- Prescription drugs (4 tiers)
- Rehabilitative and habilitative services and devices
- Laboratory services
- Preventative and wellness services and chronic disease management
- Pediatric services
For each of these benefits, the carrier must identify whether or not the benefit is subject to the corresponding deductible (medical or drug) and what the patient’s cost-share (co-payment or co-insurance) is for the benefit. One can see that the carrier can design the plan in any number of ways to achieve the tier’s AV target.In engineering-speak, a consumer must solve a 16-dimensional (non-linear) optimization problem to find their best plan. Ceteris paribus, the number of different implementations resulting from either subjecting or excluding each benefit from the deductible results in over 8,000 possible implementations. Even if the carrier were constrained to use just one of two possible deductible values, maximum out-of-pocket limits, and assign only one of two possible values to each co-insurance or co-payment, there would still be over 4 billion different ways for a carrier to design a plan that would satisfy the same AV requirement!
Cost implications for the consumer
To see the practical implications of all this to a typical consumer, we might consider just the impact of how prescription drug benefits are designed.
Januvia is a very common drug for diabetics, belonging to the class of dipeptidyl peptidase-4 inhibitors (DPP-4i). This drug class features prominently in most glycemic control algorithms (e.g. American Diabetes Association, American Association of Clinical Endocrinologists). There are no generic alternatives and as of December 2015, it retailed from between $420 and $460 for a 30-day supply at the most popular pharmacy chains.
The table below shows the difference in annual costs for Januvia for the lowest premium Silver plans from 6 different representative carriers. Although each of these plans satisfies the HHS requirement that it cover 70% of the patient’s expected medical costs, there is a $4,500 difference in cost between the most favorable plan and the least favorable plan. It is worth noting that in some cases this cost difference exceeds the total dollar amount of the required premiums!
Although this is a somewhat shocking result, we should note that the table might look completely different if some other drug were specified. In the above sample, Carrier B did not provide coverage for Januvia, but it it should be noted that it does provide coverage for an alternative drug in the same DPP-4i class — Onglyza — with very favorable coverage under Tier 2.
Better tools are needed
The point I think I would make from the preceding is not that the plan standardization mandated by the ACA is necessarily flawed, but rather that consumers need access to more sophisticated tools to personalize their plan choice. HHS is trying to make progress in this area, as evidenced by their beta testing user entry of specific prescription drugs in selecting a plan. But even here, Marketplace.com only tells the user whether the drug is covered and not what approximate cost they should expect.
For the past year or so, I have been involved in the development of a combination insurance/drug search platform called Savue. As far as I know, Savue is the first service that not only provides dollar cost estimates of prescription drug spending for a given insurance plan, but also estimates the total monthly and annual cost of the plan as a whole — including premiums, copayments subject to and not subject to the plan deductible, and the cost of prescription drugs that are not covered at all.
Presently, the emphasis is on drugs for diabetes, hypertension, and hyperlipidemia. Plan coverage for Open Enrollment 2016 is focused on Texas, but perhaps people from other states will still find the tool useful for hinting at what plans in their areas might be worth looking into. The insurance search tool, as can be seen in the figure, estimates the total cost a plan subscriber will pay over the course of the year — including premiums and out-of-pocket costs. If the household is eligible for additional cost assistance, such as Advanced Premium Tax Credits (APTCs) or cost-sharing assistance that is available for Silver plans, that is also indicated.
We are hoping to help a lot of people in Texas reduced their health care costs this coming year. I welcome any comments about our service (especially highly-critical ones). You can email me at firstname.lastname@example.org.
 U.S. Department of Health and Human Services, “Actuarial Value and Cost-Sharing Reductions Bulletin”, February 24, 2012.
 Section 1302(b)