Drug formularies ideally are meant to protect health plans against exorbitant drug costs, and to provide members with an array of safe and appropriate products with demonstrated cost-effectiveness. In practice, though, formularies often just seem to generate headaches all around. Patients suspect they're being denied access to the drugs they deserve, physicians believe they're being prevented from prescribing the medications their patients need, and pharmacy directors in health plans feel frustrated, misunderstood, and underappreciated.
And that guy sobbing in the corner with his head in his hands? He's the product manager for one of the newest blockbusters. It's a drug he knows will cut overall health care costs by reducing hospitalizations, but he keeps running headfirst into pharmacy "silos" — a budgetary technique that focuses on drugs' acquisition costs without regard to their overall effect on resource utilization and medical costs.
The guy slumped in the other corner? He's a pharmacy director who has grown weary of telling drug manufacturers that their elegant clinical studies lack real-world relevance. Showing that a new agent is generally safe and more efficacious than placebo may be all the FDA needs to grant approval for marketing, but it's not what a cost-conscious MCO requires. Comparing a new drug head-to-head against yesterday's dominant product does not make a strong impression, either.
The Academy of Managed Care Pharmacy thinks it's high time to introduce an element of rationality into the decision-making process of P&T committees. To that end, it has developed a format for formulary submissions. Based on a system developed by Regence BlueShield in Washington and used there since 1998, it's intended to give the pharmaceutical industry a clear understanding of the materials they should submit to P&T committees if they hope to have their products added to a formulary. AMCP will evaluate the format during the next year, according to Richard Fry, senior director of pharmacy affairs, starting with a few pilot projects launched during the second quarter of this year.
Fry has received some informal feedback about the proposed format, primarily from manufacturers. He says the majority support the AMCP's initiative, and some already are gearing up to use the new format. A few large pharmaceutical firms oppose the effort because they're comfortable with the intensive marketing strategies that they have in place.
Historically, health plans have looked at clinical studies and acquisition costs as the primary criteria for deciding which drugs are added to a formulary or which are excluded. It's a no-brainer — provided your brain doesn't encompass more than the narrow confines of a pharmacy silo. AMCP and Regence, however, take a broader view: At the least, they think a drug's overall health outcomes and economic impact on a health plan have to be taken into consideration.
What to submit
Broadly speaking, the proposed format emphasizes improved communication between P&T committees and pharmaceutical companies. The idea is to let a pharmaceutical company know precisely which documents should be submitted for consideration by a P&T committee. The end product is a dossier filled with evidence-based materials that are tailored to the characteristics and needs of the health plan in question. Of course, submission of the requested information is no guarantee that a drug will be added to the formulary. That depends on the strength of the evidence. At Regence, only 53 percent of the products submitted to date under the new guidelines have been recommended for addition to the formulary.
Some documents must be prepared to a health plan's specifications, particularly the pharmacoeconomic models that are based on the plan's demographic and utilization characteristics, and that ultimately demonstrate the value of the product to the health plan.
Determining value means that a drug's effectiveness is more important than its efficacy. Efficacy, readers of this publication will know, is what's measured in a carefully controlled clinical trial, during which patients take their medication as instructed, physicians prescribe medications correctly, and side effects are carefully monitored. The FDA requires efficacy and safety studies, but not effectiveness or outcomes. Effectiveness, or lack of it, is what happens in the real world — the world in which patients fail to take medications as instructed, physicians fail to prescribe the optimal dose, and side effects go unrecognized. So although summaries of clinical trials — for the manufacturer's product and also competitors' products — are required, they are not sufficient in themselves. The manufacturer has to take an extra step and translate the results of clinical trials into effectiveness.
Typical questions asked by health plans using the AMCP format might be the following: Does that new asthma drug really reduce emergency room visits and trips to the doctor? Does that high-priced biotechnology drug actually improve quality of life and health outcomes more than the tried-and-tested (lower-cost) formulary alternatives? How much more "bang" (health outcomes) does the health plan get for the higher cost of the new product?
As an example of how manufacturers can demonstrate value in accordance with the requirements of Regence and the proposed AMCP format, Robert Ohsfeldt, Ph.D., a health outcomes research scientist at Eli Lilly, oversaw the construction of a model depicting the cost per event avoided for raloxifene, a selective estrogen receptor modulator (SERM) that is indicated for prevention and treatment of osteoporosis. The model compared raloxifene with hormone replacement therapy and the bisphosphonate alendronate, both of which are on Regence's formulary. But it went beyond the prevention and treatment of hip and vertebral fractures due to osteoporosis, and also considered rates of myocardial infarction and breast cancer that could result from the drug's side effects. Such an analysis had to account for the fact that Regence has few retirees in its patient population; most members are well below age 65.
The raloxifene model used conservative assumptions, Ohsfeldt says. For example, the evaluators at Regence thought it would have been reasonable for the manufacturer to assume that compliance rates would be better with raloxifene than with alendronate, but the model assumed that all three agents were equally poor in terms of compliance.
Ohsfeldt estimates that it took about 50 person-days to create the original model, and that a few days per month have been required since then to keep it up to date as new data become available. Ohsfeldt and his colleagues have presented the model about 30 times since its creation, with mixed results. Some pharmacy directors and medical directors have objected that the model is speculative — an inherent property of a model, from which it derives its strength as a decision-making tool — and therefore not useful. Ohsfeldt says such critics tend to require evidence from a randomized, placebo-controlled trial as the gold standard, especially with respect to assumptions about drug efficacy. They fail, however, to fully appreciate the limitations of trials: that the environment in which they take place is artificial and not commensurate with the milieu in which medicine is practiced, and that the subjects who volunteer for them differ clinically and behaviorally from those who do not.
Some critics of the raloxifene model argue for waiting for the release of new data, such as the results of large multiyear placebo-controlled clinical trials evaluating the effects of raloxifene or hormone replacement therapy on the incidence of coronary heart disease. But deciding to wait for years for more information constitutes a decision in itself, Ohsfeldt observes.
"No one knows what the truth is in these areas," Ohsfeldt says, "but modeling lets health plans determine what areas of uncertainty have a major impact on projected costs and outcomes, and what areas of uncertainty don't really matter much."
And the members of Regence's P&T committee, after dissecting the raloxifene model and rigorously questioning its underlying assumptions, concluded that it indeed constituted a reasonable argument for adding raloxifene to the formulary.
The architect of the Regence BlueShield guidelines is Dwight S. "Pete" Fullerton, Ph.D., R.Ph., vice president for pharmacy services. Coming from a background in academia — he was dean of pharmacy at the University of Utah and dean of faculty at Oregon State before joining Regence (then known as King County Medical Blue Shield) in 1993 as its first pharmacist — he wanted the organization's first formulary to be evidence-based.
Regence's evidence-based system seemed to work well enough at first. Whenever drugs in a therapeutic class appeared to be equivalent, such as the statins for lowering LDL cholesterol or ACE inhibitors for reducing blood pressure, the P&T guidelines called for selecting the most cost-effective agent. A couple of problems emerged: Fullerton couldn't obtain unpublished studies, even though he knew they existed; and he realized that, in an era of tight resources, health plans needed to look at value.
Fullerton enlisted help from Regence's pharmacy director, Debbie Atherly, M.P.H., R.Ph., and one of his former students, Sean D. Sullivan, Ph.D., who had become the director of the pharmaceutical outcomes research and policy program at the University of Washington, where Fullerton also holds an appointment. (Sullivan also serves as the chair of the Regence P&T committee.) Together, they developed a new approach that specified the kinds of data they wanted manufacturers to provide.
From the outset, Fullerton said, manufacturers have been supportive of Regence's initiative. First, they're eager to do business with Regence, because it serves about 1.2 million members and has a high percentage of "closed" and "tiered" pharmacy benefits. Second, drug companies that market their products internationally already were conducting outcomes analyses required by Canada, Great Britain, Australia, and other countries. These companies had established international outcomes departments, but no large health plan in the U.S. market ever had asked for such data before. The AMCP's Fry says, though, that some critics of the new format maintain that it's one thing to create a model that will be used nationwide and quite another to create multiple versions of plan-specific models, especially if the need for many plan-specific models occurs simultaneously.
In addition, Regence was willing to be flexible in its requirements, recognizing that an outcomes analysis for a hypertensive agent would differ from one for a migraine product. Initially, Regence was extremely forgiving. Some manufacturers just submitted cost-minimization analyses, showing merely that their product was cheaper than the competition's. That kind of approach no longer suffices, and a manufacturer is unlikely to try it now that Regence's requirements are well known.
"As Regence has grown, so has the outcomes industry," Fullerton says, noting that the pharmaceutical industry now has 1,000 outcomes analysts on staff.
Outcomes analysis still is a little daunting to some health plans, Fullerton says, but the science is getting stronger. He regards allegiance to the clinical trial as a defense that's raised by people who don't understand outcomes analysis. Now that pharmacoeconomics is part of the curriculum at every pharmacy school, more pharmacy directors have learned to appreciate and understand an outcomes analysis, while the Journal of the American Medical Association and Pharmacoeconomics have published national standards that can help pharmacy directors become more familiar with this complex topic. Part of the appeal of pharmacy silos, after all, is that conceptually they're very simple.
"Some health plans still pick drugs on the basis of rebates and price, but that's short-sighted," Fullerton says.
From Fullerton's perspective, an added bonus of Regence's system is that it reduces, by about 40 percent, the amount of staff time needed to evaluate new pharmaceutical products — once they're up to speed with outcomes. He doesn't think adoption of the AMCP guidelines would be onerous for the pharmaceutical industry, given that it already is generating such data for international markets. Besides, once one dossier has been compiled for a U.S. health plan, it can be easily adapted to conform to the requirements of other plans. In fact, Regence lets pharmaceutical manufacturers give other health plans the same dossiers they're preparing for Regence.
It all makes so much sense, you wonder if that alone might work against it.
How to obtain the AMCP submission format:
Full text of the AMCP's formulary submission format is available at http://www.amcp.org/. Ideas and comments should be sent to Richard Fry, AMCP's senior director for pharmacy affairs, at (703) 683-8416, x309, or email@example.com.