Of the changes that managed care has brought to family practice, one of the most potentially annoying is the requirement to prescribe according to each managed care plan's drug formulary. Formularies place yet another burden on family physicians who are already strapped for time, and these restrictions can make it difficult to implement the most current treatment protocols. Studies have shown that following restricted formularies (those that stipulate, rather than suggest, what physicians should prescribe) does indeed reduce drug costs1,2 and that physicians' financial incentives to follow formularies don't adversely affect the quality of care.3 But this knowledge doesn't make the implementation any easier.
For practices that contract with several managed care plans, each with its own formulary, the frustration only increases. A physician must know each patient's insurance plan, find the appropriate formulary and be careful to follow its restrictions. Mistakenly prescribing nonformulary medications can trigger numerous phone calls from the patient, the pharmacist and perhaps the plan's pharmacy benefits manager. This creates significant hassles for patients, staff and physicians. And if the practice is at risk for the cost of prescriptions, those mistakes don't just increase hassles; they decrease revenue.
These frustrations — and our desire to practice evidence-based, cost-effective medicine — prompted us to develop our own “preferred practice formulary.” Although some doctors might see a formulary as a weapon of the enemy, we saw it as a weapon that could help us fight back — a way to cope with the hassles of multiple formularies while cutting costs in what continues to be the fastest growing area of health care spending.4 Here's how we did it.
Developing our formulary
Family Health Services is a primary care group in the ambulatory-care department of an academic medical center serving an urban population. Our group, part of a large, multispecialty faculty practice, consists of 12 family physicians, two internists, a pediatrician and three nurse practitioners.
About 65 percent of our patients are covered by managed care plans, two-thirds of which have restricted formularies. Before 1996, when we implemented our formulary, most of our physicians wrote their prescriptions based on personal preference without considering the drugs' cost or their formulary inclusion. But once we were put at risk for the costs of our prescriptions, following the plans' formularies became a higher priority.
We analyzed the obstacles to using insurers' formularies and identified four areas we needed to address: the hassle of having to locate and use several different formularies, the physicians' lack of understanding of our financial risk related to prescriptions, their reluctance to prescribe medications they hadn't prescribed before and the variation in their knowledge about using evidence-based medicine to optimize prescribing. Our solution to these problems involved developing and using a preferred practice formulary — one that would integrate the common requirements of our multiple formularies into one document that our physicians could follow easily. This way, we could increase their compliance and ease their frustration.
To make our formulary easy to use and still beneficial to the practice, we decided to include only our most common drug classes: anxiolytics, antidepressants, diabetic treatments, antihistamines, antibiotics, asthma agents, cardiac medications, hypolipidemics, acne treatments, topical steroids, gastrointestinal therapies, hormonal therapies, muscle relaxants, NSAIDs and otic solutions. For less common drug classes, doctors still must refer to the individual plans' requirements.
To draft our formulary, we first reviewed the health plans' requirements for these 15 classes. If a medication was listed on all the plans' formularies, we included it on ours. For drugs that weren't common to all the formularies, we gave priority to the agents preferred by the plans with which we shared financial risk for drug costs.
It was important that our formulary be based on evidence of efficacy and safety, not simply reflect the least expensive drug options. Fortunately, the MCOs' formularies were broad enough to allow us to base our formulary on commonly accepted references. For example, antibiotic choices were based on data from The Sanford Guide to Antimicrobial Therapy.5 For antidepressants and H-2 antagonists, we referred to guidelines from the Agency for Health Care Policy and Research. We based our choices for anti-hypertensives on the Fifth Report of the Joint National Committee on Prevention, Detection, Evaluation and Treatment of High Blood Pressure (JNC-V).
Then within each drug class we ranked the available agents by their cost according to the MCOs' formularies. Ideally, for any given class, the agent at the top of the list would be accessible to all patients regardless of their health care plans and would be the least expensive of the choices, as well as being medically efficacious. For most classes, a generic agent common to all insurers was available. (The formulary notes the rare cases in which a plan doesn't accept the least-expensive agent in a given class.) In the interest of simplicity, we didn't include information about the numbers of doses or refills allowed. In all, our formulary totals only two pages and includes three to 12 agents in each class. (See “The authors' formulary.”)
Of course, even the best practice formulary will have little effect on prescribing behavior unless busy physicians have the information at hand when they write prescriptions. So we posted the formulary in each exam room and at each physician's desk.
We also knew it was crucial that the physicians understand the need to control cost in our increasingly capitated environment before they would accept and embrace our drug list. So we developed an education program to explain how and why we created the formulary. We also gave each physician detailed information, provided by our largest capitated plan, about his or her pharmaceutical utilization, and we met with each physician individually to review the data. (For our doctors, the most expensive and the most commonly prescribed drug classes were SSRI antidepressants, H-2 antagonists, antibiotics and oral contraceptives.) Giving them their utilization data and noting the preferred agents raised the doctors' awareness without causing undue anger or frustration.
An important part of preventing that anger and frustration was focusing on the formulary's effect on quality of care. We organized a series of monthly journal clubs, in which our physicians presented evidence-based articles about the treatment of patients with common ambulatory problems. The journal clubs helped educate the physicians about the medical validity of the drugs included in our formulary. Their acceptance of it grew as they realized that substitutions for their first choices were equally efficacious but more cost-effective.
The authors' formulary
The list of preferred agents for Family Health Services is two pages long and includes the 15 drug classes most commonly prescribed there. A portion of the practice's formulary appears below.
Family Health Services' preferred formulary choices
The following list of medications is based on efficacy and availability for all our commonly used insurers. Please use these drugs preferentially. Within each category, choice 1 is the least expensive and 5 is the most expensive. If you need to use a drug not listed here, please refer to the individual formulary of the patient's insurer.
1. Ativan (lorazepam) 0.5, 1, 2 mg (1 to 10 mg daily)
1. Librium (chlordiazepoxide) 5, 10, 25 mg (5 to 10 mg qid)
3. Klonopin (clonazepam) 0.5, 1, 2 mg (0.5 mg tid)
1. Elavil (amitriptyline) 10, 25, 50, 75, 100, 150 mg (75 to 150 mg daily)
1. Sinequan (doxepin) 10, 25, 50, 75, 100, 150 mg (75 to 150 mg daily)
2. Desyrel (trazodone) 50, 100, 150, 350 mg (150 to 400 mg daily)
4. Paxil (paroxetine) 20, 40 mg
4. Zoloft (sertraline) 100 mg (1/2 tab q day)
5. Wellbutrin (bupropion) 75, 100 mg (100 to 250 mg daily)
We instituted our formulary late in the second quarter of 1996. A comparison of the average cost per prescription from the first and the fourth quarters of 1996 (again based on data from our largest capitated health plan) shows a 10 percent decrease for our primary care group. In contrast, the average cost per prescription for the multispecialty faculty practice as a whole rose 6 percent over the same period.
Two years after instituting this formulary, we continue to see cost savings. Our group's average cost per prescription remains lower than it was before we introduced the formulary. Unfortunately, the overall average cost per prescription for the multispecialty practice has continued to rise.
We've also seen anecdotal evidence of other important benefits, such as fewer phone calls to our practice to correct prescriptions for nonformulary drugs and reduced frustration for physicians as they care for patients covered by a variety of managed care plans. Having access to formulary information at the time of prescription writing has resulted in most new prescriptions meeting the requirements of the patients' insurance.
It's not just about cost
We've seen significant cost savings as a result of giving physicians concise information about payers' prescription preferences for specific drug classes. That success has come because our physicians were willing to make a difficult adaptation to the world of managed care. Prescribing according to something other than personal preference is a major change. But our doctors have come to see the practice formulary as more than a tool for reducing costs in a capitated environment. Because of its base in medical evidence and its ease of use, it's also a tool for providing quality care.