We set out to estimate the potential savings associated with the drug recommendations in a novel consumer-oriented program of prescription drug education. Our analysis points to significant potential savings for patients and the health care system. If the CRBBD accounted for all drug sales in the four classes we studied in 2006, consumers and payers would have saved a maximum of $2.76 billion dollars or 7.83 percent of total expenditures in these classes.
We found that the majority of savings in these four classes would be achieved through therapeutic as opposed to generic substitution. The rate of generic use in some of our classes is substantially higher than what recently published estimates indicate (Haas et al. 2005
). Payers have become much more aggressive in recent years at requiring patients to fill prescriptions for generic equivalents for multisource drugs (O'Malley et al. 2006
). Nearly two-thirds of prescriptions filled in the U.S. in 2006 were for generic drugs (Frank 2007
). However, our findings also suggest substantial variability in use of generic equivalents across classes and products. Early experience with the Medicare drug benefit suggests that financial incentives to use generic drugs may outweigh concerns among some elderly and their physicians about the safety and efficacy of generics (Banahan and Kolassa 1997
; Hellerstein 1998
; Genther and Kreling 2000
; Gaither et al. 2001
; Saul 2007
We found evidence of variation in the relative prices of generic drugs compared with their brand name counterparts. Studies have shown that generic drugs are priced 10–70 percent lower than brand name drugs (Frank and Salkever 1997
; Suh et al. 2000
). It is important to monitor the price differences between generic and brand name drugs in the midst of dramatic changes in pharmaceutical markets. On the one hand, generic manufacturers have more aggressively challenged the patents of brand name drugs, increasing generic entry and price competition (Frank 2007
). On the other hand, manufacturers of brand name drugs are increasingly releasing “authorized generics” (generic versions introduced by the original brand manufacturer) in an attempt to maintain revenues.
There are limitations to our analysis, which may limit generalizability. First, medication switches may not be appropriate in some cases due to medication side effect profiles or potential drug interactions. We were interested, however, in gaining an understanding of the maximum economic impact of the program and providing a range of possible savings estimates. Second, our findings are not necessarily generalizeable to other medication classes. As of December 2007, CRBBD had produced consumer reports on 19 of the 32 classes for which DERP has conducted systematic reviews. We made simplifying assumptions in our analysis, which may be difficult to apply to other classes (i.e., it would be difficult to apply the assumption of no polypharmacy to the antidepressant or antipsychotic classes). Thus, while savings could be achieved in other classes through therapeutic substitution, the magnitude of savings will vary by class. Third, we estimated savings in a dynamic market, in which prices and market shares can change dramatically from year to year. Notably two of the statins (simvastatin and pravastatin) became available as generics mid-way through our study period. Savings for 2007 and beyond will be substantially higher due to the availability of multiple generic alternatives in the class (Consumer Reports Best Buy Drugs 2006
). Fourth, we cannot take into account any rebates or discounts negotiated by third-party payers. Given that the nature and magnitude of these rebates are not publicly reported, it is unclear how our savings estimates would be affected if rebates and discounts were incorporated. Fifth, because we used aggregate data on total sales in these four pharmacologic classes, we cannot estimate how much of the savings would accrue to patients as opposed to payers. Modest savings that accrue to third-party payers may lower overall health insurance premiums, but are not likely to lead to significant behavior changes among patients. Significant savings that result in lower out-of-pocket spending among consumers could reduce economic barriers to medication adherence.
One important question is how the transition costs associated with medication switches might compare with the savings achieved through increased use of the best buy drugs. For instance, when British Columbia instituted a reference pricing system for ACEIs in 1997, visits to physicians increased slightly among those who switched to lower cost medications. Medication switchers had an 11-percent increase in visits to physicians during the first 2 months following implementation of the policy corresponding to a temporary increase in physician expenditures of approximately $11 per patient just before switching and $13 afterward (Schneeweiss et al. 2002
). It is important to note, however, that these were one-time costs as opposed to savings associated with chronic use of medication over a period of several years. A follow-up study estimated that the reference pricing policy resulted in $6.7 million in savings among existing medication users, while expenditures for increased physician claims amounted to only $0.7 million (Schneeweiss et al. 2004
). Nevertheless, it is important for efforts like those of the CRBBD program to take these transition costs into consideration.
While the findings we report are theoretical, they are of practical importance given that we assessed the potential impact of an existing program targeting consumers. Some payers have provided price and quality information on hospitals and physicians to consumers in conjunction with consumer-directed health care initiatives, yet little is known about their effects (Ginsburg 2006
). In some respects, reporting on the comparative effectiveness of prescription drugs from randomized controlled trials is technically more straightforward than reporting on the quality of care delivered by physicians and hospitals to which patients are not randomly assigned. Programs such as CRBBD may inform broader efforts that encourage consumers to make value-based purchases in health care.
The challenge will be to disseminate the CRBBD information in a way that actually leads to behavior change among consumers and their physicians. Knowing that Internet usage remains low among low-income individuals with high out-of-pocket drug costs who could benefit most from CRBBD-type information (Pew Charitable Trust 2005
; Baker, Wagner, and Bundorf 2003
), CRBBD works with a number of intermediaries to disseminate the educational material. For instance, Medco, one of the country's largest PBMs, delivers direct mailings of the CRBBD information to its members. One such mailing to just over 1 million users of brand name statins about the availability of low-cost generics was followed by 49,020 members switching to simvastatin or pravastatin (G. Shearer, personal communication).
Prescribing practices are difficult to change. Simple one-time educational interventions aimed at changing physician's behavior are seldom effective (Grimshaw et al. 2001
). While strategies such as academic detailing have been shown to alter prescribing in some classes (Spinewine et al. 2007
), they require substantial investment and few cost-effectiveness studies of these types of interventions have been conducted (Simon et al. 2007
). Very few programs that aim to improve prescribing patterns through consumer education have been evaluated (Fillit et al. 1999
; Zorowitz et al. 2005
). However, studies of the impact of direct-to-consumer advertising of drugs indicate that physicians are highly responsive to patient requests for medications (Kravitz et al. 2005
). Providing consumers with information on prescription drugs in conjunction with financial incentives to use lower cost drugs could increase consumer demand for drugs of high value. Indeed, the design of the new Medicare drug benefit market is predicated upon the notion that seniors will “shop” for plans that offer the lowest out-of-pocket price for their drugs.
Our analysis can provide insights for the return on investment for programs to increase the use of evidence-based medicines. Disseminating this type of information to consumers and assisting them to act on it will be critical if these savings are to be realized. A program that achieved only 5 percent of the potential savings we estimated would pay for itself if it costs <$138 million to implement. By comparison, the average annual expenditure on direct-to-consumer advertising for the top 20 drugs advertised in 2005 was $115 million (Donohue, Cevasco, and Rosenthal 2007