The long run average price of pharmaceuticals in a cost-effectiveness framework
COMMENTARY by Jaume Puig-Junoy
ABSTRACT
This commentary analyzes the limitations of estimates for the long run average price (LRAC) and net LRAC, as well as the implication of considering the LRAC and the net LRAC in healthcare decision making, especially in HTA. The difference between the entry price, or the price before the loss of exclusivity, and the LRAC (net LRAC) depends on the competitive dynamics and regulation of each particular market, as well as the discount rate. The compensation effect to obtain the net LRAC from the LRAC, if it exists, is very variable, depending on indication, patient sub-groups, doses, duration, and treatment adherence, and it does not need to have a linear relationship with the number of prescriptions. The importance of the LRAC for decision making depends on the adoption of a limited short-term perspective of an agent, or a social perspective with a mid to long term horizon. In the cost-effectiveness or cost-utility approach, the compensation effect is already included in the estimation of incremental costs of the compared treatments. Economic evaluation models should consider the price fall after the loss of exclusivity, which tends to be omitted even in simulations over the patient’s lifetime.
What is the cost of a drug for evaluation purposes?
In a recent article at the AJMC, Lakdwalla et al1 conclude that drug prices before loss of exclusivity (LOE) in the United States could overstate the true long run average cost (LRAC) for buyers by a 40-75%, when the reduction in the price of generics (LRAC) and potential reductions in other medical costs (net LRAC) are considered. The only difference between the LRAC and net LRAC comes from the supposed value of the reduction in other medical costs (compensation), being the net LRAC noticeably lower than the LRAC and the entry price, if the compensation effect is positive.
The results of Lakdawalla et al1 allow us to assert that decisions related to the coverage and pricing of innovations should consider the LRAC, and not only price at a given moment in time. The effect of the loss of a patent on the costs of pharmacological treatments is known and predictable, although variable depending on the therapeutic group, level of market competition, channel of distribution, and patient’s and distributor’s incentives, as well as the regulatory framework2,3.
Variability of the LRAC (net LRAC)
In practice, the estimation of the LRAC and the net LRAC faces at least two types of limitations. Firstly, generic competition reduces the price significantly when exclusivity is lost2,3, although the magnitude of this reduction is variable, depending on real price competition and the market share of generics versus brands (competition dynamics and market specific regulation). Similarly, there exists variability in the effective lifetime of a drug; in this sense, a drug could be pushed out of the market by a more effective treatment, even before the loss of exclusivity; or it could give rise to a shorter, or longer, market lifetime after the loss of exclusivity. The LRAC will be very sensitive to the reduction of prices by generics, the drug’s lifetime after the loss of exclusivity and the market share of generics, if the trademark price is higher than the generic price. In the case of biological drugs, the expected reduction in the price of biosimilars in relation to the entry price is much lower compared to generics. The difference between the LRAC and the entry price will be also very sensitive to the discount rate.
Secondly, the estimation of the compensation effect on medical expenditures from an increase in the number of prescriptions of a particular drug is very variable, depending on the changes in indication and the number of patients, doses, treatment duration, and adherence. The hypothesis that the compensation effect on expenditure is associated with any increase in prescriptions (the elasticity of medical expenditures in front of a variation in prescriptions) taken by Lakdawalla et al1 from a CBO4 survey comes from a particular regulatory framework and population group. It would be a long shot to extrapolate from such particular cases to the general population. The studies selected by the CBO4 estimated the compensation effect of a reduction in pharmaceutical consumption due to an increase in out of pocket payments: that is, the effect on overall medical expenditure of decreasing the consumption of some drugs for a population of 65 years of age and older, many of them suffering from chronic diseases. The economic literature on the crossed impacts of copayments on the use of other healthcare resources highlights the fact that most of the compensation (offset) effect comes from a polimedicated population, suffering from one or many chronic diseases5. It is very risky to assume that the compensation effect is symmetrical for reductions, increases or addition of pharmaceutical prescriptions, something especially uncommon when age and comorbidity increases. For example, a European study performed in Catalonia estimated that the compensation effect on hospitalizations is not significant when old age population increases the consumption of drugs after a reduction of the coinsurance rate6. Also, there is no evidence of the fact the marginal compensation effect should exhibit a linear relationship with consumption or that it is similar to the average compensation effect. If one wants to estimate the compensation effect of a particular innovation, it should be kept in mind that it will change depending on the effectiveness and incremental effectiveness of the new treatment. The compensation effect is also variable along the life cycle of the drug, depending on the extension of its value (broadening indications, combinations, etc.). This variability or heterogeneity makes the application of the estimated effect of a “typical” or representative drug in the study of Lakdawalla et al1 to a specific case (treatment/indication/patient subgroup) not too much useful.
Implications of LRAC for decision making
Lakdawalla et al1 point out that their results are especially relevant for decision makers and health technology assessment (HTA) agencies. They claim that the models employed by these agencies should take into account, from a social perspective, the LAC if they don’t want to overvalue the price of drugs and provide biased conclusions. For illustrative purposes, we will consider the relevance of the LAC and the net LAC in two restrictive models of decision making for HTA or public and private insurance decision-makers: 1) a decision criteria based only on the comparison of pharmacological treatment prices (a commonly adopted silo approach); and 2) a decision criteria which prioritizes the cost-effectiveness incremental ratio (ICER) for year of live gained (LY) or by quality adjusted life years (QALY) from a cost-effectiveness (CEA) or a cost-utility analysis (CUA).
Firstly, from the perspective of an insurer responsible for the health of the whole population for a long period of time, we can observe that (i) the entry price of a new drug overestimates the pharmacological cost of treatment in the long run (LRAC); and (ii) if the compensation effect is positive, the entry price overestimates even more the true pharmacological insurer cost with respect to the net LRAC. Introducing the net LRAC in an environment where judgements are based only on price also implies a change in criteria, as the impact on the rest of health resources is considered.
Secondly, in the short run and with a perspective limited to pharmaceutical expenditures, the opportunity costs for an insurer or private provider of a drug treatment for a patient with an acute health problem could be acceptably reflected by the price in each moment in time rather than by the LRAC. Notwithstanding, even from the insurer’s perspective, when considering the average opportunity cost for a covered population in the mid or long run, and for treatments of long duration such as those related to chronic diseases, it’s more probable that the opportunity cost would be best approached by the LRAC. From a social perspective, there are good reasons to think that the relevant factor to be considered for more efficient decisions is the long run opportunity cost for the population, so the LRAC also takes preference over entry or before of exclusivity loss price in healthcare decision making.
How does the concept of net LRAC contribute to the improvement of models in HTA agencies, with more than two decades of experience in the use of ICERs, for decision making on the funding and pricing of new drugs? The compensation effect contributes little or nothing to decision making if the HTA model starts from a complete economic evaluation (CEA or CUA). The only relevant contribution of LRAC resides in the temporal modeling and discounting of the price, during the previous and posterior period to the loss of exclusivity. In a CEA or CUA that considers the overall impact of treatments on the use of healthcare resources, as the payer perspective should contemplate, the compensation effect has already been considered when estimating incremental costs of treatments.
CONCLUSIONS
Arguably, the most interesting thing about the LRAC concept that economic evaluations should consider in their simulation is the fall of prices after the expiration of the patent, which is often omitted, even in simulations over the patient’s lifetime. The current practice in a CEA or CUA consists of using the observed price of a drug at a given moment in time to approximate opportunity costs, both for the innovative treatment and the comparators. In the case of a long-term treatment, e.g. across the patient’s lifetime, considering the results of Lakdawalla et al1 would imply using the LRAC instead of the observed entry or before loss of exclusivity price at each point in time (treatments lasting no longer than the exclusivity period). Again, it could seem that some doubts appear on the use of the LRAC as a proxy for opportunity cost when one could consider the true opportunity cost from the individual perspective to be the market price at each point in time. Most likely, from the perspective of an individual patient, the best measure for opportunity costs in the short run is the market price, which is not the case for the insurer responsible for the health of the whole population for a long period of time or long term treatments (static versus dynamic evaluations).
TAKEAWAY POINTS
- A dynamic economic evaluation of new drugs should take into account the long-run average cost (LRAC).
- Value-based decisions of coverage and pricing of innovations should consider the LRAC, not only its price at a given moment in time.
- From a social perspective, it seems that there are good reasons for thinking that the relevant indicator for a dynamic economic evaluation is the long run opportunity cost for the whole population. Therefore, we should give preference to the LRAC, instead of the price, for decision-making.
REFERENCES
1. Lakdawalla D, MacEwan P, Dubois R, Westrich T, Berdud M, Towse A. What do pharmaceuticals really cost in the long run? Am J Manag Care. 2017;23(8):488-932.
2. Berndt ER, Dubois P. Impacts of patent expiry on daily cost of pharmaceutical treatments en eight OECD countries, 2004-2010. Int J of the Economics of Business. 2016;23(2):125-473.
3. Olivier JW, Kanavos P, McKee M. Comparing generic drug markets in Europe and the United States: prices, volumes, and spending. The Milbank Quarterly. 2017;95(3):554-601
4. Congressional Budget Office. Offsetting effects of of prescription drug use on Medicare’s spending for medical services. Congressional Budget Office website. https://www.cbo.gov/sites/default/files/cbofiles/attachments/43741-MedicalOffsets-11-29-12.pdf. Published November, 2012. Accessed October 17th.
- Chandra A, Gruber J, McKnight R. Patient cost sharing and hospitalization offsets in the elderly. American Economic Review. 2010;100(1):193-213.
- Puig-Junoy J, García-Gómez P, Casado D. Free Medicines Thanks to Retirement: Impact of Coinsurance Exemption on Pharmaceutical Expenditures and Hospitalization Offsets in a national health service. Health Econ.2016 Jun;25(6):750-67.