Journal of Health Economics 29 (2010) 743–750 Contents lists available at ScienceDirect Journal of Health Economics journal homepage: www.elsevier.com/locate/econbase Is newer always better? Re-evaluating the benefits of newer pharmaceuticals Michael R. Law a,∗ , Karen A. Grépin b a Centre for Health Services and Policy Research, School of Population and Public Health, University of British Columbia, 201-2206 East Mall, Vancouver, BC, Canada V6T 1Z3 b Robert F. Wagner Graduate School of Public Service, New York University, 295 Lafayette Street, New York, NY 10012, United States article info Article history: Received 11 October 2006 Received in revised form 11 March 2010 Accepted 28 June 2010 Available online 24 July 2010 JEL classification: I12 C14 D61 Keywords: Prescription drugs Drug offsets Health care costs abstract Whether newer pharmaceuticals justify their higher costs by reducing other health expenditures has gen- erated significant debate. We replicate a frequently cited paper by Lichtenberg on drug “offsets” and find the results disappear using a more appropriate model or updated dataset. Further, we test the suitability of similar methods using newer hypertension drugs. We find our observational results run counter to well-established clinical evidence on comparative efficacy and conclude that our model, as well as other studies that do not adequately control for unobserved characteristics that jointly determine drug choice and health expenditures, are likely subject to significant bias. © 2010 Elsevier B.V. All rights reserved. 1. Introduction For many years, prescription drug expenditures in the United States have grown at a rate higher than that of overall health care spending (Catlin et al., 2008). This growth has partly resulted from the use of newer drugs that have higher prices relative to older drugs used to treat the same condition (Duggan, 2005). An important question is whether these newer drugs justify their higher costs. There are several ways in which this could be the case. For example, newer drugs could reduce non-drug health expenditures such as hospitalizations, improve a patient’s qual- ity of life, or produce benefits outside the health sector (e.g. increase the productivity of workers). In this paper we inves- tigate the first of these possibilities, the “offset” of non-drug health care expenditures through the use of newer medicines. We do this by first replicating and evaluating the robustness of the most widely cited paper on the subject (Lichtenberg, 2001, herein L01). We then further investigate the suitability of this and similar research approaches to investigate the existence of offsets. We use the same dataset employed by L01 and other authors to study offsets for hypertension, a condition with signif- ∗ Corresponding author. Tel.: +1 604 822 3514; fax: +1 604 822 5690. E-mail addresses: mlaw@chspr.ubc.ca (M.R. Law), karen.grepin@nyu.edu (K.A. Grépin). icant clinical evidence that can inform the interpretation of our results. The interest in offsets has been at least in part motivated by studies detailing reductions in the cost of care for prevalent con- ditions that have had major drug therapy advances. This literature has investigated changes in the cost of care for a number of medi- cal conditions through the construction of price indices. Using data on the treatment of heart attacks from 1983 to 1994, Cutler et al. (1998) concluded that the real cost of treating a heart attack has declined over time. Similarly, Berndt et al. (2002) found that the real cost of treating major depression declined during the early 1990s. Both of these disease conditions had seen the development of new pharmaceutical treatments that may have contributed to these decreases in the treatment cost. However, this evidence is only suggestive of an offset effect as many other non-drug factors likely contributed. To try and isolate the contribution of drugs in “offsetting” other health care costs, two general research approaches have been used in past work. One approach has investigated the savings from individual newer drugs to another older treatment option in a ran- domized clinical trial setting (Neumann et al., 2000). While there are many advantages to this approach, it only allows comparison of two or a few treatment alternatives, which may not reflect the true range of clinical options available in practice. Moreover, many clinical trials are conducted against a placebo, which gives no indi- cation about the specific offset of newer medicines versus older 0167-6296/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.jhealeco.2010.06.007