Spillover effects of restrictive drug formularies in the statin class: a descriptive study of the Medicaid preferred drug list in Texas.
A drug formulary restricted to a specific covered population may influence how physicians treat other patients (i.e., the "spillover effect"). Previous studies on the proton-pump inhibitor (PPI) class revealed that spillover effects exist for restrictive drug formularies from Medicaid and private health insurance plans. (1-3) Researchers have found that as a trend, the more a physician is exposed to a restrictive drug formulary, the larger its spillover effects will become.
In contrast with other therapeutic drug classes, PPIs are similar in efficacy and side effects. (4) It is unknown whether similar spillover effects exist between products with large variations in efficacy and/or side effects, for example, the statins. One important consideration, therefore, is that the extent of spillover may depend on the interchangeability of products.
Most Medicaid programs throughout the United States use preferred drug lists (PDLs). These are the formularies of covered medications from which physicians in the individual state can prescribe for Medicaid beneficiaries without prior authorization. This study used the implementation of Texas' Medicaid PDL as a natural experiment and examined whether and how the Medicaid PDL affected physicians' prescribing major statin products for non-Medicaid patients.
A medical prescription database (IMS Health's Xponent PlanTrak, Plymouth Meeting, PA) was used to track physicians' prescribing patterns for Medicaid and non-Medicaid patients in Texas. This database covers more than 70% of prescriptions filled in retail pharmacies in the United States and uses additional statistical methods to project the universe of all retail-filled prescription products. It reports the projected number of retail prescriptions by physician, health plan, month, and product type. (1)
Medicaid accounted for 5.1% of all statin prescriptions in Texas. To examine how the Medicaid PDL affects physicians' prescribing of statin products for non-Medicaid patients (i.e., the spillover effects), this study focused on 8,946 high-volume physicians who, on average, prescribed 20 or more statins per month. These physicians were categorized into four groups with increasing exposure to Medicaid (< 0.1%, 0.1%-2.0%, 2.1%-8.0%, and > 8.0%). The number of physicians in each group ranged from 2,164 to 2,313.
Data were analyzed for one year, between October 2003 and September 2004. Spillover effects were measured as simultaneous changes in a product's non-Medicaid marketshare associated with the implementation of the Medicaid PDL in Texas. No statistical method was used in this descriptive analysis.
Figure 1 describes the monthly marketshares of atorvastatin, simvastatin, and pravastatin in the Texas Medicaid program over a one-year period. The Medicaid PDL took effect in March 2004 and its resulting prescription market fluctuations began to stabilize after three months.
[FIGURE 1 OMITTED]
A 67% (35-point) drop in marketshare for atorvastatin (which was not on the PDL) was noted, corresponding to a 25-point increase for simvastatin and a 10-point increase for pravastatin (both on the PDL). The marketshares of other products, such as fluvastatin, lovastatin, rosuvastatin, and ezetimibe/simvastatin, were not reported, because they accounted for less than 15% of prescriptions in this period.
Figure 2 describes non-Medicaid marketshares of atorvastatin in Texas by physician group, with differing exposure to Medicaid beneficiaries. In the high-Medicaid-exposure group, atorvastatin's non-Medicaid marketshare decreased by about two percentage points soon after implementing the Medicaid PDL. Share trends in other physician groups were less obvious.
[FIGURE 2 OMITTED]
Figure 3 describes non-Medicaid marketshares of simvastatin by physician grouping. A noticeably downward monthly trend was observed across physician groups, likely reflecting the loss of marketshare to other statins, including the newly launched cerivastatin. When compared with physicians with very low to medium Medicaid exposure, physicians with higher exposure prescribed more simvastatin for non-Medicaid patients after the implementation of the Medicaid PDL, resulting in an absolute share increase of greater than one-point.
[FIGURE 3 OMITTED]
Figure 4 describes non-Medicaid marketshares of pravastatin by physician group. Similar to simvastatin, a downward trend was observed across physician groups, with an approximately one percentage point relative share increase among physicians with high Medicaid exposure after the implementation of the Medicaid PDL.
[FIGURE 4 OMITTED]
Although spillover effects may enhance therapeutic competition and facilitate the dissemination of cost-effective products, physicians and patients must be kept informed of interproduct differences and formulary design features to optimize individual drug treatment. A promising development is electronic prescribing, which gives real-time access to product and drug-benefit information.
The scale of spillover in statins seems to be smaller than that in PPIs. (1-3) The two possible explanations are as follows: (1) interproduct differences, in terms of efficacy and side effects, are presumably larger among statins, compared with PPIs, and (2) the Medicaid share of total statin prescriptions in Texas is small (only 5.1%), which might limit its influence on physicians' prescribing.
How product interchangeability and health plan characteristics affect spillover effects of restrictive drug formularies are important topics for future research.
The Medicaid PDL in Texas appeared to influence non-Medicaid sectors, with negative spillover for the statin product that was not on the PDL (i.e., atorvastatin) and positive spillover for the two statins that were on the PDL (simvastatin and pravastatin). These findings verify that spillover effects of restrictive drug formularies, previously verified for the PPI class, also exist for the statin class (albeit on a somewhat smaller scale). Health plans or pharmacy benefit managers must consider such spillover effects (from their own formularies onto others and vice versa) in formulary design and rebate contract negotiations, to maximize formulary compliance and deliver cost-effective pharmaceutical care.
(1.) Wang YR, Pauly MV, Lin YA: Impact of Maine's Medicaid drug formulary change on non-Medicaid markets: Spillover effects of a restrictive drug formulary. Am J Manag Care 2003;9:686-696.
(2.) Wang YR, Pauly MV: Spillover effects of restrictive drug formularies: A case study of PacifiCare in California. Am J Manag Care 2005;11:24-26.
(3.) Wang YR, Pauly MV: Spillover effects of restrictive drug formularies on physician prescribing behavior: Evidence from Medicaid. J Econ Manag Strategy 2005;14:755-773.
(4.) Miner PJ, Katz PO, Chen Y, et al: Gastric acid control with esomeprazole, lansoprazole, omeprazole, pantoprazole, and rabeprazole: A five-way crossover study. Am J Gastroenterol 2003;98:2616-2620.
Dr. Wang was an employee of AstraZeneca, but did not receive financial support for this study.
Address for correspondence: Y. Richard Wang, MD, PhD, Department of Medicine, Attn: Medicine Administration, PP 8th Floor, Temple University Hospital, 3401 North Broad Street, Philadelphia, Pennsylvania 19140. E-mail: y.richard.wang@ gmail.com.
To obtain reprints, please contact Kevin Chamberlain at (914) 337-7878, ext. 202 or visit our website at www.medicomint.com. Copyright 2006 by Medicom International. All rights reserved.
Dr. Wang is a resident at the Department of Medicine, Temple University Hospital, and an adjunct senior fellow at the Leonard Davis Institute of Health Economics, University of Pennsylvania, both in Philadelphia.
Part of the data analysis was completed during Dr. Wang's previous employment at AstraZeneca Pharmaceuticals in Wilmington, Delaware. No financial support was obtained for this study.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||PHARMACY practice|
|Author:||Wang, Y. Richard|
|Publication:||Managed Care Interface|
|Date:||Nov 1, 2006|
|Previous Article:||Is managed care education for new nurses adequate?|