Controlling Off-Label Medication Use FREE

Off-label prescribing is widespread, accounting for about 20% of all prescriptions and more than 50% of use of expensive, potentially toxic drugs, such as chemotherapeutics.

Biotechnology drugs are among the most expensive and potentially risky medicines, and their use has been increasing by 20% per year, accounting for more than $40 billion of a $200 billion pharmaceutical budget.

Biotechnology drugs often cost more than $1000 per month, and third-party payers typically cover their use, even when they are prescribed off-label.

The Centers for Medicare & Medicaid Services (CMS) have a National Coverage Determination process that can be used to decide whether Medicare will cover U.S Food and Drug Administration (FDA)–approved devices, based on whether evidence shows the devices to be reasonable and necessary.

The FDA approval process evaluates whether drugs are safe and effective in specific cases, but once a drug is approved, it can be used off-label for other indications.

The same 2-step process used by CMS for devices (FDA approval followed by a National Coverage Determination to decide on coverage) could be applied to biotechnology drugs that are both expensive and potentially very toxic, specifying which indications will be covered.

Since its establishment in 1906, the U.S. Food and Drug Administration (FDA) has scrutinized new drugs to determine whether they are “safe and effective.” Once a drug has been approved, physicians are free to use it in other populations, for different conditions, and in varying doses from those for which it was approved. Such “off-label” uses are predicated on the assumption that the FDA does not regulate the practice of medicine and that physicians can be relied on to use their own judgment in deciding how to use an approved agent. As long as drugs are safe, well tolerated, and relatively inexpensive, off-label use generates little concern. But with the rise of several new classes of very expensive drugs, third-party payers have balked at covering drugs for non–FDA-approved indications. Now that Medicare offers a prescription drug benefit, the federal government, too, has a greater incentive to consider whether to pay for drugs used for conditions other than those for which they have been approved. We need a process to control the use of such medications that is ethically sound, logistically efficient, and financially prudent.

Off-label prescribing has been justified on several grounds. First, creative and flexible use of medications can lead to the discovery of new and important indications. For example, β-blockers were initially approved for the treatment of hypertension but were subsequently found to be beneficial for the treatment of heart failure (1). Second, off-label prescribing has been viewed as essential in such fields as pediatrics because few drugs have been specifically tested in children (2). Finally, off-label prescribing avoids the lengthy and costly process of modifying the FDA labeling of a medication.

Off-label use of prescription drugs is widespread: The largest, most carefully conducted analysis of outpatient prescribing patterns found that 21% of prescribed medications are used for non–FDA-approved indications. Among medications used off-label, 73% lacked evidence of clinical efficacy and only 27% were supported by strong scientific evidence (3). In oncology, where many of the most expensive medications with the severest side effects are used, a recent study found that the 5 most widely prescribed chemotherapeutic agents were used for off-label indications in 50% of cases (4). Other studies suggest an even greater prevalence of off-label prescribing in oncology. For example, an in-depth study of the monoclonal antibody rituximab at an academic medical center over 3 years found that 75% of the uses of the drug were off-label (5).

Many medications used off-label are inexpensive, but many others, in particular biotechnology drugs, are costly. Biotechnology drugs are defined by the industry as biopharmaceuticals, such as monoclonal antibodies, recombinant enzymes, and cytokines, which are produced using cellular or molecular processes, as well as more traditional small-molecule drugs, which are manufactured chemically but target a biological process (6). These drugs were developed to treat relatively rare diseases but are now being used for a host of other conditions, despite their high cost. Of the 250 distinct biotechnology drugs in use as of 2007, 94 (37.6%) are single-use products, such as imaging agents or vaccines (7). Of the remaining 156 drugs, 65 (41.6%) cost at least $1000 per month according to price information from multiple online domestic pharmacies and reports of new drug releases in major newspapers. Leading the list of expensive biotechnology agents are chemotherapeutic drugs (36.9%), followed by enzyme-deficiency drugs (18.5%); anti-infectious agents, dermatologics, and rheumatologics together make up another 21.6% of the total (Table 1). Whereas the drugs used for orphan diseases, such as congenital enzyme deficiency disorders, have no off-label uses, chemotherapeutic agents are used for types of cancer other than those for which they have been tested and approved. Many of the rheumatologics are used off-label for dermatologic conditions (8), and the pulmonary vasodilators are used for rheumatoid arthritis and inflammatory bowel disease (Table 2).

Providing coverage for unproved, potentially toxic, and expensive drugs both promotes their use and contributes to the escalating cost of medical care. Retail pharmaceuticals account for 10% of total health care spending, or $216.7 billion in 2007, the third largest component of the health care pie after hospitalizations and physicians' services (9). Medications are the fastest-growing segment of the health care budget, and biotechnology drugs are the fastest-growing of the pharmaceuticals: Between 2002 and 2006, retail sales of biologic drugs in the United States increased from $19 billion to more than $40 billion, or 20% per year, compared with growth rates of 6% to 8% in the remainder of the pharmaceutical industry (10). Off-label prescribing accounts for at least one fifth of all medication use (3) and as much as 70% of the most expensive drugs, such as chemotherapeutic agents.

Medications used off-label vary in terms of cost and potential side effects. Rather than attempting to control the use of all medications prescribed off-label, it may be more efficient to focus on drugs that are both risky and expensive. In practice, the most expensive medications are often those with the greatest risk for toxicity. Although there are no standardized indices of toxicity that allow comparison of the relative risks of medications, data on adverse drug events reported to the FDA show that very expensive biotechnology drugs are among the worst offenders, even though they account for a small proportion of all prescriptions written (11). A recent study calculated that biologicals had a 29% chance of triggering a safety-related regulatory action, such as a black box warning, within 10 years of their introduction to the market (12). Medications with few side effects and low cost may not warrant regulation when used off-label unless they are so widely prescribed that the population-level risk is high.

In designing a strategy for addressing the off-label use of expensive, potentially dangerous medications, we can benefit from looking at the system already in place for determining coverage in the Medicare program for devices. With devices, the FDA evaluates whether a device is “safe and effective” (13), but the Centers for Medicare & Medicaid Services (CMS) will not automatically reimburse for all such devices. In accordance with the 1965 Medicare statute, CMS will cover a device only if it is “reasonable and necessary.” Although the criteria for what is reasonable and necessary have never been spelled out, they are more restrictive than the FDA “safe and effective” criteria. Moreover, CMS has developed a transparent approach that it can invoke to determine whether to cover a device: its National Coverage Determination process (14). As a result, whereas many devices are both FDA approved and covered by Medicare, others, such as the vagal nerve stimulator for treatment of depression, are FDA approved but not Medicare covered, and still others, such as the left ventricular assist device, are FDA approved but covered by CMS only in particular clinical situations.

A similar process could be mandated for any drug whose cost is similar to that of a typical device. The most commonly used device among hospitalized elderly persons is the cardiac pacemaker (15). The average reimbursement by CMS for insertion of a single-chamber pacemaker is $12 000 (16). Using this as a benchmark, I suggest that any pharmaceutical with a cost of $12 000 per year, or a monthly cost of $1000, automatically trigger the same kind of analysis as devices. The CMS would not be using cost as a criterion for reasonableness, an approach that has consistently been rejected by the U.S. Congress; cost would simply determine which drugs were most appropriate to trigger a coverage decision. Just as National Coverage Determinations increasingly specify the clinical circumstances in which a device will be covered, for expensive drugs they would state which, if any, off-label uses are to be covered based on close scrutiny of all available data.

The CMS have already invoked their coverage decision process to address off-label uses of selected drugs paid for under Medicare Part B—drugs given intravenously in the hospital or office setting. A National Coverage Determination was issued for the bone marrow stimulant erythropoietin, several drugs used in the treatment of colorectal cancer, and radiopharmaceuticals used in lymphoma care (17). After these decisions, CMS embarked on a policy of “coverage with evidence development,” in which off-label uses of medications would be covered provided that patients were enrolled in a clinical trial testing those uses (18).

Consumer advocates constitute a major barrier to the routine scrutiny of expensive medications used off-label. Health insurers are so afraid of negative publicity from advocacy groups, as well as possible lawsuits on behalf of patients with cancer, that they rarely restrict access to new chemotherapeutic agents (19).

Chemotherapy drugs pose additional challenges because they are a protected class. Since the passage of the 1993 Omnibus Budget Reconciliation Act, Medicare has been required to cover off-label uses of anticancer drugs if such use is included in an approved drug compendium. An analogous standard was adopted by the U.S. Congress for use by the Medicaid program and again by use of the Part D prescription drug benefit included in the Medicare Modernization Act of 2003. Many state legislatures have followed suit, requiring reimbursement by private insurers for cancer drugs listed in at least 1 compendium.

The compendia are published by private organizations whose staff members search for new drug information and submit their findings to experts for review. Limited policies are in place designed to avoid conflict of interest for reviewers; for example, they cannot own more than $25 000 of stock in a pharmaceutical company and cannot hold drug patents or have an employment relationship with a pharmaceutical company (20). Many off-label indications are listed in these compendia, based on small observational studies that have not been published in peer-reviewed journals.

Cultural beliefs also stand in the way of change. Defining a standard of care based on preliminary data assumes that most drugs that enter clinical trials are destined for approval. The reality is very different: Only 5% of cancer drugs that enter clinical testing are ultimately approved for use (21).

Other strategies for regulating the use of drugs prescribed for off-label indications have been suggested, but they tend to be unwieldy at best and unworkable at worst. For example, some have recommended prior authorization for all off-label indications, with a neutral, federally sponsored organization deciding which off-label indications are justified by data. This would mean that all prescriptions would have to be screened before they could be filled, creating enormous bureaucratic obstacles to access (22). By contrast, in the system suggested here, prescriptions for expensive medications would be accompanied by a form indicating the condition for which the medicine was being prescribed, allowing third-party payers to determine coverage. Another strategy involves developing a system of postmarketing surveillance. Although this approach would identify unexpectedly dangerous drugs, it would not detect ineffective ones (23).

Yet another approach mandates obtaining informed consent from patients before prescribing a drug for an unapproved indication. It is difficult to understand how meaningful informed consent can be obtained when few or no data on benefit are available. Greater regulation by the FDA is also an option, but recently revised guidelines suggest a softening rather than a strengthening of the agency's position on off-label drugs (24). Comprehensive strategies that involve review of all off-label uses of medications by formulary review and input from expert panels are unlikely to be useful because of the number of drugs that would be subject to review (25). Only by limiting review to high-risk, high-cost medications can such an approach succeed.

In conclusion, the use of off-label medications continues to grow, subjecting patients to substantial risk and contributing to increasing medical expenditures. A reasonable first step in reversing this trend is to address the issue of those medical interventions that do not merely confer small benefits for large costs but also have no proven benefit—and are costly both to patients and to society. For off-label medications, a 2-step process (FDA approval followed by a National Coverage Determination) analogous to that which is already available for medical devices could control use in the Medicare program without depriving patients of beneficial technology. Other insurance companies would probably adopt the recommendations of CMS.