Faculty View

Off-Label Prescriptions: The Drug Safety Policy Debate
by Sandra H. Johnson
Professor of Law

Professor Johnson has been working on the issue of off-label prescribing for a series of lectures she gave in 2006, including the inaugural Fallon-Friedlander Lecture in Health Law and Social Science at the University of Chicago Law School in May 2006, and for a forthcoming article in the Minnesota Journal of Law, Science, and Technology.

Health policy issues generally involve significant trade-offs in which efforts to achieve one goal usually produce damage to others. This is certainly true in the case of drug safety policy, as current passionate debates about the effectiveness of the FDA illustrate. 

In fact, the majority of prescription drugs that patients take are prescribed for a purpose, in a higher or lower dose, over a longer period of time, or for a population different from that for which the FDA approved the drug. Recent studies reveal that 21% of all prescriptions written in the medical office setting were prescribed for a purpose for which the drug had not been approved. Drugs for some medical conditions reach even higher levels:  75% of prescriptions for anti-depressant drugs, for example, were for unapproved purposes as were 80% of anticonvulsant medications. Prescriptions for cardiac medications and prescription antihistamines for allergies also approach rates of 50% or higher for unapproved purposes. 

This practice, called “off-label” prescribing, has raised significant safety and effectiveness concerns. A study published in the Journal of the American Medical Association in May 2006 sharpened these concerns when it reported that “most” off-label prescriptions studied had “little or no scientific support.”

This is scary stuff. Counterintuitively, however, restricting off-label prescribing categorically will harm individual patients denied medication that may provide the most effective treatment for them though not yet definitively proven to do so. Furthermore, patients generally may suffer if restrictions on off-label prescribing seriously reduce medical innovation and field discovery of important therapeutics. 

Some examples: Oncologists frequently treat cancer patients with drugs that have been effective with similar types of cancer and have shown promise in ongoing, but uncompleted, clinical trials for the particular cancer that their current patient has. In another common situation, a doctor and patient are likely to find that a pain medication approved for the treatment of neuropathic pain generated by shingles provides relief for the patient’s neuropathic pain caused by another medical condition even though the drug is not approved for that specific use. Sick children may require a drug for treatment even though that prescription will be off-label because the drug has not yet, and may never be, tested specifically on children. Similarly, most drugs are not tested on elderly individuals, where changes in metabolism may significantly alter the effect of a standard drug regimen.

Why is off-label prescribing so prevalent? The answer that has grabbed the headlines recently — because of the over $1 billion paid in settlements of government claims against pharmaceutical companies over the last few years and the proliferation of private class action liability suits drawing on those claims — is that doctors prescribe drugs off-label because the manufacturers pay them to do so through marketing techniques that include financial rewards, big (e.g., consulting fees) and small (e.g., the free lunch). While the story told in this litigation provides a great villain and reveals obvious excesses, the phenomenon of off-label prescribing is much more complicated than that.

Once a drug is approved by the FDA as safe and effective, the FDA has no authority to limit or interfere with physicians in prescribing the approved medication within a legitimate physician-patient relationship. More than merely respecting the issues that would arise with regulating medical judgment, the approval process encourages pharmaceutical firms to seek a narrow approved use, at least initially, to minimize the delay and expense required to meet FDA standards. The FDA itself only rarely requires post-approval clinical trials, and their record for follow-up on such trials has been spotty. Furthermore, because the drug’s patent life is ticking away, there is a disincentive for investment in costly trials for broader uses for approved medications.
Of course, the market could provide incentives for continuing research on approved drugs despite weak regulatory mandates. If doctors refused to prescribe a drug for an unapproved use or for a patient in a population in which the drug has not been tested unless rigorous clinical trials have been completed and published to support that prescription, the market incentives for post-approval testing would increase. 

What we know about the way that practicing physicians learn, however, shows that they are skeptical about clinical trials and place a higher trust in clinical experience. Doctors tend to look to the practices of their peers and respected opinion leaders for guidance, and it may be that the malpractice standards encourage that instinct. Furthermore, the current turmoil in the conduct and publication of clinical research may confirm professional skepticism.

Studies of articles published in medical journals have consistently shown a correlation between the sponsorship of a clinical trial and its outcome. Published studies sponsored by industry are significantly more likely to have “pro-industry conclusions.”  It’s not clear, though, why this is so. One might think that this pattern results from bias on the part of the researcher whose research is supported by a grant or contract from the manufacturer. But, bias in the published literature might result instead from the selectivity of pharmaceutical firms in funding only those clinical trials, say on an off-label use, that seem more likely than not to produce positive results. Or, the pattern may be an outcome of research contracts that give the sponsor the sole right to release the results. Or, it is possible that medical journals themselves contribute to bias in the literature by their own selection processes; for example, by rejecting articles reporting that a clinical trial failed to produce clearly positive or clearly negative results?

Simply prohibiting industry support of post-approval clinical trials, however, is not a realistic option for improving the production of clinical research while public funding for post-approval research remains insignificant. Requiring pharmaceutical firms to pay for more clinical trials of off-label uses will confront the unresolved issue of bias in the results. 

Even if funding issues for post-approval clinical trials could be remedied, there is no miracle that will produce research that is both rigorous and instantaneous. Off-label prescribing is essential to the practice of medicine and the care of patients. In fact, it is unavoidable.