Summary
To get a sense of what drug coverage will look like across plans offered on the health insurance exchanges (HIEs), we analyzed the drug benefits of all Silver[1] plans offered in the 13 states that account for 51.2% of the nation’s uninsured
As compared to typical plans offered in the employer-sponsored (ESI) market, we find that Silver HIE plans:
- are more likely to require that deductibles be met before prescription coverage begins;
- have higher deductibles;
- have higher average co-pays; and,
- are more likely to use co-insurance in lieu of fixed dollar co-pays
This implies that prescription demand from the average beneficiary covered on the HIEs will – all else equal – be less than from the average beneficiary covered under ESI. However even though the non-group market will expand as the HIEs enroll, because the non-group market will still remain relatively small, the effect of the HIEs’ more austere drug benefit designs on total national drug demand is probably less than 2%
What matters here are the indirect effects – specifically that the HIEs become an experiment in drug benefit design whose findings spill over into the much larger ESI market. ESI plan sponsors (more so) and ESI formulary managers (less so) have hesitated to require large deductibles, raise co-pays and/or use co-insurance in lieu of fixed dollar co-pays, for fear that consumers’ use of medically necessary prescriptions would decline as their out-of-pocket (OOP) costs rose. This is equivalent to saying that ESI plan sponsors and formulary managers feared that brand manufacturers’ co-pay card programs might not offer subsidies large enough to keep consumers’ OOP costs at tolerable levels
If brand manufacturers allow co-pay cards to be used by HIE beneficiaries (we believe they will), and offer subsidies large enough to reduce these beneficiaries’ higher OOP costs to tolerable levels (again, we believe they will), the brand manufacturers will effectively prove to plan sponsors and formulary managers in the far larger ESI market that higher co-pays and/or more extensive use of co-insurance is feasible in ESI drug benefits. The potential effects on US brand pricing are enormous; we estimate that shifting ESI beneficiaries to drug benefits having tier 3 co-insurance can reduce brand manufacturers’ net US pricing power by as much as two-thirds, at a time when US price increases account for more than 100% of trailing US revenue growth for all (save one … Roche) of the 12 large-cap drug makers
Detailed findings
HIE-based plans must simultaneously honor relatively low OOP maximums, and meet any of several actuarial value[2] (AV) thresholds. For the lower AV plans (Bronze 60AV, Silver 70AV), this creates a situation in which much of the AV is ‘consumed’ by the claims costs that exceed the lower overall OOP max, which has the effect of limiting how generous any coverage below the OOP max can be
Deductibles
HIE beneficiaries are substantially more likely than their ESI counterparts to face deductibles before their Rx coverage becomes effective. In 2013, depending on plan type, between 5% (HMO) and 26% (HDHP/HRA) of ESI beneficiaries were required to meet deductibles before receiving Rx coverage; this compares to 63% of Silver HIE beneficiaries who in 2014 will have to meet deductibles before receiving effective Rx coverage (Exhibit 1). 22 percent of Silver HIE plans offered have an Rx-specific deductible, which averages $560; and, Rx coverage for 41 percent of Silver HIE plans offered is subject to the plan’s general deductible, which averages $2,392 (Exhibit 2). In the few cases of ESI plans with Rx deductibles, the average Rx-specific deductible is $96, and the average general deductible varies from $729 (HMO) to $2,003 (HDHP)
Thus it’s clear that many more HIE than ESI beneficiaries will face Rx deductibles, and that the Rx deductibles faced will be higher for HIE than for ESI beneficiaries
Co-payments v. co-insurance
40 percent of Silver HIE plans offered in the 13 states sampled use co-insurance instead of fixed dollar co-payments for drugs covered under tier 3, as compared to 25 percent of 2013 ESI plans (enrollment-weighted). And, 64 percent of Silver HIE plans offered use co-insurance for specialty tiers, as compared to 48 percent of ESI plans (enrollment-weighted) in 2013 (Exhibit 3)
Silver HIE plan beneficiaries with incomes below 250% of the federal poverty level (250FPL) benefit from ‘cost-sharing reductions’, or subsidies which serve to lower effective co-insurance rates and/or co-pay amounts. If we include the effect of these subsidies, then the average effective co-insurance levels faced by Silver HIE beneficiaries are on par with those seen in the ESI market. For Silver HIE beneficiaries above 250FPL who do not have the benefit of co-pay subsidies, the average co-insurance rates are slightly higher than those in the ESI market (Exhibit 4)
For Silver HIE enrollees in plans using fixed dollar co-pays instead of co-insurance, average co-pay amounts (and, the all-important tier 2 / tier 3 ‘spread’) are higher than those seen in the ESI market – even if we include the effect of co-pay subsidies for beneficiaries with incomes below 250FPL (Exhibit 5)
Controlling for mix of underwriters
The preceding analyses of Silver HIE plans are on an ‘as offered’ basis – having no enrollment data at the plan level, we can only analyze on the basis of what’s offered, as opposed to what’s enrolled. This raises an obvious risk: that the enrollment-weighted drug benefit characteristics of Silver HIE plans will differ from the ‘as offered’ averages. Specifically, our concern was that smaller, less well-established underwriters would offer drug benefits that were less generous than the larger underwriters, and that the resulting ‘as offered’ average drug benefit would be more austere than what would ultimately come about on an enrollment-weighted basis. In truth, it appears the opposite is true
As a sanity check, we analyzed drug benefit designs for the four large multinational underwriters (AET, CI, HUM, WLP) offering Silver HIE plans in the states we sampled – and were surprised to find that if anything, these underwriters’ offerings are more austere than the ‘as offered’ average. These underwriters’ Silver HIE drug benefits are more likely to require a deductible, require higher average deductibles, and are far more likely to use tier 3 co-insurance than smaller Silver HIE underwriters (Exhibit 6)
How and why this matters
Direct effects
We believe the direct effects of these relatively austere drug benefit designs are limited. Before any expansion of non-group insurance that might occur as a result of the Affordable Care Act (ACA), all forms of private health insurance (essentially ESI plus non-group) accounted for approximately 45% of US drug spending. Non-group beneficiaries (+/- 25M) were about 14% of the total group of private beneficiaries (+/- 181M); thus non-group insurance accounts for roughly 6% (45% x 14%) of US drug spending. Thus even if we assume the ACA increases the size of the non-group market by 50% – which we don’t believe it can – then non-group drug benefit designs would directly influence at most 9% of US drug demand. If we further assume the impact of benefit design changes is to reduce per-capita unit demand by 20% among persons in the non-group market, the total direct demand effect would be less than 2% – not terribly consequential. To our way of thinking, indirect effects are far more important
Indirect effects
As compared to ESI beneficiaries, HIE beneficiaries will face larger deductibles before their drug coverage kicks in, and (barring other influences) higher out-of-pocket costs (because of greater prevalence of co-insurance and higher average co-pays) once coverage is effective
Drug consumption is highly elastic in relationship to beneficiaries’ OOP costs, so if brand manufacturers hope to defend current levels of per-capita utilization among non-group beneficiaries, steps will have to be taken that lower these beneficiaries’ OOP costs
Exhibits 7 and 8 provide context to our claim that drug demand is highly elastic. A large percentage of total demand is concentrated among a relatively small percentage of beneficiaries (Exhibit 7), e.g. fully 70% of US drug sales are accounted for by only 10% of beneficiaries. And, these beneficiaries’ drug spends consume large shares of disposable incomes (Exhibit 8). 21 percent of US drug spending is attributable to just 1% of households, whose OOP drug spend is on par with their housing costs, and the 10% of beneficiaries that consume 70% of US drug sales spend nearly as much OOP on drugs as on food
For the last several years, brand manufacturers have offered co-pay cards that reduce consumers’ effective OOP costs to tolerable levels. Generally the cards are used by consumers purchasing a prescription drug that is on tier 3 (non-preferred) or 4 (specialty), and the value of the co-pay subsidy generally is sufficient to get consumers’ OOP costs down to the prevailing tier 2 (preferred brand) level of +/- $29
Because HIE beneficiaries are more likely to face deductibles, have higher average co-pays, and are more likely to have co-insurance instead of co-pays, to get HIE beneficiaries’ OOP levels to $29, brand manufacturers’ co-pay card subsidies will have to be larger than those applied to prescriptions filled by ESI consumers. And, because co-pay card subsidies reduce net pricing, brand manufacturers’ net pricing to HIE beneficiaries will be lower than to ESI beneficiaries
Because the non-group market is relatively small, we’re less concerned with net pricing to these beneficiaries than we are with the precedent that is being set, and how this precedent might unfold in the ESI market. Logically, if ESI plan sponsors and ESI formulary managers were entirely convinced that drug companies would offer co-pay card subsidies to reduce beneficiaries’ OOP costs to $29 irrespective of the fixed dollar co-pays and/or co-insurance rates used in drug benefits, then ESI plan sponsors and formulary managers would be incentivized to set co-pays and/or co-insurance rates as high as possible. If brand manufacturers do as we expect – namely, use co-pay cards to reduce the relatively high co-pay / co-insurance levels in the HIE market to an effective OOP of +/- $29 – they will have given plan sponsors and formulary managers in the ESI market evidence that co-pays and co-insurance rates can be raised in ESI. Crudely, the high co-pays and co-insurance levels in the HIE market are a test case in the far larger game of ESI drug benefit chicken playing out among manufacturers, plan sponsors, and formulary managers
In an earlier research note[3] we measured the potential impact of rising co-pay card subsidies on manufacturers’ net pricing. In that note we assumed co-pay card subsidies were rising because of a shift in ESI drug benefit design from co-pays to co-insurance. Assuming nominal list pricing continues at a 10% pace, and that ESI beneficiaries are steadily shifted into tier 3 co-insurance-based drug benefit designs over a five-year period, we estimated that rising co-pay card subsidies would reduce net pricing growth to 3.3%, assuming a tier 3 co-insurance rate of 50% (equal to what is in fact being used by AET, CI, and HUM for exchange-based drug plans)
The simple point is that co-insurance and/or higher fixed dollar co-pays require manufacturers to give large co-pay subsidies with their co-pay cards, and that translating the HIE co-insurance rates into the ESI market could reduce brand manufacturers’ US drug pricing power by as much as two-thirds
Relative susceptibility
With a single exception (Roche), 2013 branded price increases account for more than 100% of the large-cap drug companies’ 2013ytd sales growth (Exhibit 9) – i.e. barring Roche the industry is wholly reliant on US pricing power for even modest growth
Exhibits 10 thru 21 give a sense of each individual company’s level of price reliance over time. The lines on each exhibit are a monthly index of the company’s brand prices, indexed to 100 in December of 2006. The dashed lines correspond to 2007 thru 2012, and the green line to 2013. By stacking the yearly price indices on top of one another, it’s possible to more quickly determine whether (and when) a company’s rate of pricing is accelerating (the gaps between yearly indices grow wider, e.g. LLY in Exhibit 16, or SNY in Exhibit 21)
