Abstract
Drug innovations can outpace regulatory drug approval processes designed to control risk, creating heterogeneous risks among approved drugs. This paper estimates a price model for risky narrow therapeutic index (NTI) drugs. The traditional generic approval process has been criticized as insufficient to guarantee therapeutic equivalence in NTI drugs, leading to higher risks of toxicity or ineffectiveness when a patient switches from a brand-name version of an NTI to a generic version, or between generic versions. Using data from the Medical Expenditure Panel Survey, this paper finds evidence of a significant price penalty for NTI drugs. The paper also finds a smaller gap between brand-name and generic prices for NTI drugs than for non-NTI drugs, consistent with costly switching. An analysis of drug consumption bundles also supports this theory. These results show that despite the many information asymmetries and agency issues in the pharmaceutical market, there is evidence of sensitivity to risk in price and consumption behavior.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
21 U.S. Code 355(j).
Notably, this is not the only case in which generic versions of pharmaceuticals are inferior to their brand-name counterparts. Recent research cast doubt on the appropriateness of generic antiobiotics (Gauzit and Lakdhari 2012).
Other studies have focused on the determinants of generic NTI prescription (Gagne et al. 2013; Chao et al. 2002). Chao et al. (2002) find no difference in switch rates between NTI and non-NTI drugs. Similarly, Gagne et al. (2013) analyze predictors of choosing generic NTI drugs for elderly Medicare participants, focusing on demographics and prior generic use.
Insofar as prices are based on cross-subsidization of other drugs, I cannot see this. However, this only produces biased results if cross-subsidization occurs between NTI and non-NTI drugs.
For clarity, in this paper, a drug class (e.g., anticonvulsants) contains many substances (e.g., carbamazepine). Substance categories contain both generic and brand-name versions of the drug (e.g., both carbamazepine and tegretol belong to the substance category carbamazepine), such that they correspond to the nonproprietary name. Versions of the drug involve unique product codes within each drug category. Versions are an even smaller category than brand-name/generic, as there are multiple generic versions generally available. For the sake of consistency, the drug hierarchy is as follows: drug class, substance, and drug version.
These are not the only payers reported by the MEPS, just the most interesting. MEPS includes Veterans Administration payments, workers’ compensation, state and local, Tricare, and a series of other payment categories (including a set of residual payment categories (”other public” and ”other private”) meant to correct inconsistencies in the data). I do not consider these categories in the analysis except insofar as they contribute to the total sum of payments.
NDA authorized generics are coded as brand-name drugs, given that they are approved under an NDA.
The FDA NDC database can be found at http://www.fda.gov/Drugs/InformationOnDrugs/ucm142438.htm. The data were downloaded in June 2015. Around 80% of the observations match to FDA NDC data based on NDC code and are retained.
The North Carolina Board of Pharmacy lists procainamide hydrochloride, but it is not found in my substance field in the data.
Cyclosporine and tacrolimus are only considered as immunosuppressive agents. While cyclosporine and tacrolimus are used as an eye drop (ophthalmic preparations) and topical ointment (dermatological agents), respectively, I do not consider these drug classes, as the drugs are not consumed orally and the uses seem distinct from the rest of the classes.
This analysis does not take into account any differences in marginal costs of production. However, Berndt (2002) notes that drug pricing reflects marginal value, not marginal production cost. Thus, it is unlikely that differences in marginal costs are driving the result.
Each drug version is associated with an application number and a substance category. The substance category encompasses multiple drug versions. Usually multiple application numbers are nested within one substance; in some cases very similar substances are associated with one application number. Since this is relatively rare and usually involves highly related substances, I see the age of the substance as a broader measure of how novel the substance is, separate from how long a particular drug version has been on the market.
For data observations that report the strength in grams, I assume this is an error and impute the active ingredient unit reported by the FDA NDC data. For the same reasons, I drop any observation with reported strength less than .05 mg.
I also run a weighted OLS regression, clustering by individual, which produce similar risk penalties. These results are available upon request.
I calculate this value by predicting the mean ln(price + 1), switching NTI and generic status on and off, and calculating the corresponding price for each NTI-generic status. I then calculate the generic price gap for NTI and non-NTI drugs.
This information is from a conversation with an expert from the CBO, as well as a CBO report. https://www.cbo.gov/sites/default/files/cbofiles/ftpdocs/118xx/doc11838/09-15-prescriptiondrugs.pdf
This information is from a conversation with an expert from the CBO. A CBO report also estimates that manufacturer rebates to plans averaged about 14 percent of brand-name prescription drug spending (https://www.cbo.gov/sites/default/files/cbofiles/ftpdocs/118xx/doc11838/09-15-prescriptiondrugs.pdf).
These results are available upon request.
One exception being that the interaction term NTI × Generic is negative and significant for self-payments.
Because I am only concerned with capturing the incidence of buying a particular drug, not accurately calculating per unit price and strength, I relax a few of the restrictions from the previous section. Namely, I do not drop prescriptions with fewer than 5 pills, medication that is non-tablet form, or medication with strength below .05 mg.
Product code is the first two segments of the NDC code, distinguishing based on manufacturer and product (but not package).
This does mean that a drug purchase may be considered more than once if it is associated with more than one of the chosen therapeutic classes.
References
Berg, M.J., Gross, R.A., Haskins, L.S., Zingaro, W.M., Tomaszewski, K.J. (2008). Generic substitution in the treatment of epilepsy: Patient and physician perceptions. Epilepsy and Behavior, 13, 693–699.
Berndt, E.R. (2002). Pharmaceuticals in US health care: Determinants of quantity and price. Journal of Economic Perspectives, 16, 45–66.
Blix, H.S., Viktil, K.K., Moger, T.A., Reikvam, A. (2010). Drugs with narrow therapeutic index as indicators in the risk management of hospitalised patients. Pharm Pr Granada, 8, 50–55.
Chao, J., Taylor, S.D., McKercher, P.L., Kirking, D.M. (2002). Generic narrow therapeutic index drug use: 1996–1998. Value in Health, 5, 139.
Cohen, J.P. (2000). PBMs and the medicare prescription drug benefit. Food and Drug Law Journal, 55, 311.
Crawford, P., Feely, M., Guberman, A., Kramer, G. (2006). Are there potential problems with generic substitution of antiepileptic drugs? A review of issues. Seizure, 15, 165–176.
Duh, M.S., Cahill, K.E., Paradis, P.E., Cremieux, P.Y., Greenberg, P.E. (2009). The economic implications of generic substitution of antiepileptic drugs: a review of recent evidence. Expert Opinion on Pharmacotherapy, 10, 2317–2328.
Fischer, M.A., & Avorn, J. (2003). Economic consequences of underuse of generic drugs: Evidence from medicaid and implications for prescription drug benefit plans. Health Services Research, 38, 1051–1064.
Fox, P.D. (2003). Prescription drug benefits: Cost management issues for medicare. Health Care Financing Review, 25, 7–21.
Frank, R.G., & Salkever, D.S. (1997). Generic entry and the pricing of pharmaceuticals. Journal of Economics and Management Strategy, 6, 75–90.
Gagne, J.J., Polinski, J.M., Kesselheim, A.S., Choudhry, N.K., Hutchins, D., Matlin, O.S., Tong, A., Shrank, W.H. (2013). Patterns and predictors of generic narrow therapeutic index drug use among older adults. Journal of the American Geriatrics Society, 61, 1586–1591.
Gauzit, R., & Lakdhari, M. (2012). Generic antibiotic drugs: Is effectiveness guaranteed? Médecine et Maladies Infectieuses, 42(4), 141–148.
Gothe, H., Schall, I., Saverno, K., Mitrovic, M., Luzak, A., Brixner, D., Siebert, U. (2015). The impact of generic substitution on health and economic outcomes: A systematic review. Applied Health Economics and Health Policy, 13, 21–33.
Gottlieb, S. (2014). FDA’s looming decision on a generic to Teva’s Copaxone reveals drug approval woes. Forbes.
Haas, J.S., Phillips, K.A., Gerstenberger, E.P., Seger, A.C. (2005). Potential savings from substituting generic drugs for brand-name drugs: Medical expenditure panel survey, 1997–2000. Annals of Internal Medicine, 142, 891–897.
Kesselheim, A.S., Misono, A.S., Lee, J.L., Stedman, M.R., Brookhart, M.A., Choudhry, N.K., Shrank, W.H. (2008). Clinical equivalence of generic and brand-name drugs used in cardiovascular disease: A systematic review and meta-analysis. Journal of the American Medical Association, 300, 2514– 2526.
Midha, K.K., & McKay, G. (2009). Bioequivalence; its history, practice, and future. The AAPS Journal, 11, 664–670.
Nabin, M.H., Mohan, V., Nicholas, A., Sgro, P.M. (2012). Therapeutic equivalence and the generic competition paradox. The BE Journal of Economic Analysis and Policy, 12(1), 1–26.
Omojasola, A., Hernandez, M., Sansgiry, S., Jones, L. (2012). Perception of generic prescription drugs and utilization of generic drug discount programs. Ethnicity and Disease, 22, 479.
Reiffen, D., & Ward, M.R. (2005). Generic drug industry dynamics. Review of Economics and Statistics, 87, 37–49.
Shrank, W.H., Stedman, M., Ettner, S.L., DeLapp, D., Dirstine, J., Brookhart, M.A., Fischer, M.A., Avorn, J., Asch, S.M. (2007). Patient, physician, pharmacy, and pharmacy benefit design factors related to generic medication use. Journal of General Internal Medicine, 22, 1298–1304.
Singh, A., Maisch, N.M., Saad, M. (2014). A closer look at generic interchangeability in narrow therapeutic index drugs. U.S. Pharmacist Generic suppl., 39, 8–12.
Viscusi, W.K. (2018). Pricing lives: Guideposts for a safer society. Princeton University Press.
Yu, L.X. (2011). Quality and bioequivalence standards for narrow therapeutic index drugs. FDA Center for Drug Evaluation and Research.
Yu, L.X., Jiang, W., Zhang, X., Lionberger, R., Makhlouf, F., Schuirmann, D.J., Muldowney, L., Davit, B., Conner, D., Woodcock, J. (2015). Novel bioequivalence approach for narrow therapeutic index drugs. Clinical Pharmacology and Therapeutics, 97, 286–291.
Acknowledgments
Support from ANR-Labex IAST is gratefully acknowledged. Special thanks to Tom Kniesner and W. Kip Viscusi for helpful comments and feedback.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendix
Appendix
Rights and permissions
About this article
Cite this article
Gentry, E.P. Empirical evidence of risk penalties for NTI Drugs. J Risk Uncertain 58, 219–244 (2019). https://doi.org/10.1007/s11166-019-09304-6
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11166-019-09304-6
Keywords
- Narrow therapeutic index drugs
- NTI drugs
- Generic substitution
- Drug pricing
- Pharmacy benefit managers
- PBMs
- Hedonic price
- Pharmaceuticals