High deductible health plans (HDHP)s are supposed to incentivize consumers nee patients to shop more effectively for their care and choose only high value care at good prices. As we looked at yesterday, there is a significant problem in that most people in the health insurance market sphere aren’t touched by the phase transition from high actuarial value plans with low cost sharing to lower actuarial value plans with higher deductibles. The amount of money at risk due to the incentive changes of a HDHP regime is not a plurality of total spending.
More importantly, we also need to exclude people from the previous calculation who don’t have resources to actually fund their HDHPs. Academy Health has an interesting poster with the promise of a paper that I want to read as soon as it comes out. It looks at the conversion of people from low deductible to high deductible plans who have a common chronic condition. It then sees if there are savings and changes in utilization patterns. It then splits the sample into “low income” and “high income” groups.
We used a controlled interrupted time series design to examine employer-mandated HDHP transitions, minimizing selection bias. The intervention group comprised 26,674 HDHP members with diabetes age 12-64 included between 2003-2012. HDHP members were enrolled for 1 year in a low deductible (≤$500) plan followed by 1 year in a HDHP (≥$1000) and propensity score matched 1:1 to diabetes patients with low deductibles. Low income HDHP members (n=9641) were a subgroup of interest….
Principal Findings:
HDHP members experienced small pre-to-post reductions in ED visits (-3.1% [-3.9,-2.3]), hospitalizations (- 4.2% [-5.5,-2.8]), and total healthcare expenditures (-3.6% [-4.3,-3.0]) relative to controls, and no changes in measures of adverse outcomes. However, low income HDHP members experienced relative increases in high severity ED visit expenditures (8.1% [3.0,13.2]) and high severity hospitalization days (26.1% [19.7,32.5]) at follow-up compared to baseline.
People with higher incomes seem to have had very good results. They shifted consumption dollars towards out of pocket spending. Here they were price sensitive with no adverse impacts. However people with low incomes but employer sponsored insurance saw dramatic adverse events. My theory and speculation is that they did not have the readily available resources that could easily be shifted. They are resource constrained and the implied income of health insurance covering maintenance medication was an important part of the family budget.
There are a couple of important policy caveats that we have to take from this study as we think through current legislation. The first is the size of the deductible shift. Here the researchers are defining a large deductible as more than $1,000. In ACA terms, a $1,000 deductible plan is a 2018 Platinum plan assuming no other cost sharing. When this study was started, a $1,000 deductible was probably a weak Gold plan. That is not the type of plan under discussion.
A 2018 58% AV plan has a deductible of $7,000 before any claims are paid. This is, in my opinion, a significant difference past that of degree and towards a difference in kind. People who could absorb an additional $500 or $1,000 annual net income shock as they transitioned from a low deductible plan to a $1,000 deductible plan will find it far more difficult to come up with several thousand more dollars for their maintenance medications.
The second policy caveat that we have to draw out is the population. This is a population that received their insurance through work. We know that the employer sponsored insurance (ESI) covered population tends to be healthier than the general population and also tends to have more income. They have fewer complex co-morbidities and a greater ability to absorb a fiscal shock. And yet, even the low income segment of this subgroup is in trouble in a shift from small to still fairly small deductibles in the context of the BCRA.
How does this generalize to the Medicaid population? We know that the Medicaid population is income and asset constrained by eligibility requirements. We know that the Medicaid population tends to have more complex care needs for the same diagnosis for a variety of reasons. So how does this generalize?
My bet is that we would see the same negative impacts of increased adverse events and lower adherence to care plans because adhering to the care plan and avoiding adverse events requires resources that people on Medicaid are far less likely to have compared to people who have ESI insurance.
The BCRA even in its most generous potential implementation with states fully funding their portion of the stability funds and then dedictating the entire pool to enhanced cost sharing subsidies that match the ACA in actuarial value bumps (24% for people under 150% FPL for example) will still move people from Medicaid with de minimas deductibles to plans with $1,850 deductibles or higher. If the Medicaid population is anything like the low income ESI population, we should expect more adverse events from this financing regime change.