We found far greater use of two products – lidocaine patch and modafinil – for off-label indications than for any orphan indications. Use of cinacalcet was dominated by its FDA-approved non-orphan indication, but showed an increase in off-label use during the study period. By contrast, imatinib use was mostly related to its FDA-approved orphan indications. Thus, our hypothesis was only partially correct – statistically significant non-orphan and off-label use (compared to approved uses) was found in three of the four drugs we studied. Among all four drugs, we estimated that off-label use accounted for nearly $500 million in Medicaid expenditures, with spending increasing substantially during the exposure period.
The Orphan Drug Act has helped incentivize development of new drugs for rare conditions, although this study shows that there can be significant growth in off-label use of certain orphan drugs. The lidocaine patch and modafinil were approved to treat conditions manifested by common symptoms – respectively, post-herpetic neuralgia (pain) and narcolepsy (daytime sleepiness) – which can be intractable and frustrating conditions for both doctor and patient 
. Although their pre-approval studies enrolled narrow populations covered by the Orphan Drug Act, these drugs showed efficacy in managing such symptoms in one context, so physicians may have been quick to prescribe them for other patients with similar symptoms. Similarly, cinacalcet was approved to treat laboratory abnormalities (hypercalcemia and hyperparathyroidism) found outside the populations for which it was originally approved. Unlike the other three drugs studied, imatinib was approved for distinct conditions (CML and GIST). Perhaps as a result, physicians were less likely to consider using these drugs off-label.
The growth of these drugs into top-sellers may be explained by other factors as well. Patients who present with chronic symptoms, such as pain or fatigue, may learn about newly approved products through media accounts and request prescriptions from their physicians. Some use of medications for non-FDA approved conditions has been illegally promoted by manufacturers; in the case of modafinil, for example, the manufacturer settled a lawsuit in 2008 for $425 million regarding alleged active promotion of its product for use outside of narcolepsy 
Growth of total sales of imatinib was not associated with off-label use, so could have been due to other factors, such as increases in the unit cost (approximately $56,000 per course/patient/year for GIST in 2009) 
. Changes in unit costs also contributed to increases in spending on the study drugs. Despite the substantial commitment of resources through the Orphan Drug Act and other government funding to assist in the development of orphan drugs, their cost remains an important policy issue. For example, in 2009, 84% of Medicare Part D beneficiaries were enrolled in plans that put imatinib in a specialty tier with co-insurance rates as high as $1,366 per month 
. Enacting limited waivers from state and federal antitrust laws could allow insurance plans to voluntarily band together in negotiating groups to seek lower drug prices for expensive orphan drugs where no alternative therapies exist 
Another way to reduce rates of unapproved non-evidence-based used of orphan drugs would be to vary the cost of the drug based on the indication. For example, payers could charge low co-pays to patients prescribed the drug to treat their orphan condition, and correspondingly high co-pays for non-evidence-based use. Applying such value-based insurance design to promote more appropriate use of orphan drugs, however, is limited by payers' ability to distinguish among the reasons that drugs are prescribed. As a result, orphan drugs are commonly priced the same for all indications. Some payers have tried to use administrative pre-certification forms for this purpose, although pharmaceutical companies selling orphan drugs have been investigated for allegedly teaching physicians how to fill out these forms to ensure approval of the drug for off-label uses 
This study has certain limitations. We studied top-selling drugs, so our conclusions do not generalize to all orphan-designated products. We also determined FDA-approved and non-approved uses from diagnosis codes submitted with billing claims from health care encounters. It is possible that some of the patients we identified as receiving a drug for unapproved indications may have had the indicated disease, but did not have a recorded diagnosis of it. Finally, the study population used to identify trends in utilization and spending comprised elderly persons with complete drug coverage residing in two states, and low-income patients in the Medicaid program. Our results may differ from other recipients of these agents.
In this analysis, we primarily focused on off-label use, rather than evidence-based use. However, for one of the products, the lidocaine patch, we found that an overwhelming share of the prescriptions were for patients with diagnoses for non-neuropathic pain syndromes, where no rigorous clinical trial evidence supports its use. Off-label use can have solid supporting evidence, and may be appropriate even in circumstances where gathering of supporting evidence can be difficult 
. On the other hand, when non-FDA approved uses are not supported by adequate clinical evidence, patients do not receive the intended benefits from the drug, are less likely to be prescribed more effective treatments, and are exposed to risks of adverse events with no demonstrated countervailing benefits. The substantial costs of orphan products to Medicaid add to the concern about off-label and non-evidence-based uses. For the drugs in our sample, greater attention to implementing value-based insurance design may help limit non-evidence-based off-label use 
These findings documenting off-label use in certain top-selling orphan drugs may have important implications for the Orphan Drug Act and for drug development policy. Currently, generous orphan drug incentives in the US are earned during drug development or at the point of FDA approval, before the product reaches the market. The Orphan Drug Act has proven useful in encouraging the government and pharmaceutical manufacturers to devote resources towards developing new drugs for orphan conditions. Our data suggest that it might be preferable to continue to monitor orphan drug use after approval to identify products that come to be widely used for non-FDA approved indications – especially those for which there is also inadequate evidence of efficacy. Orphan drugs that quickly find widespread use are likely to provide an adequate return on investment to their manufacturers in a short time frame. When such use occurs widely, it may be reasonable to terminate the orphan drug market exclusivity period. The European version of the Orphan Drug Act contains a “clawback” provision that permits reduction of the statutory exclusivity period if the product is deemed sufficiently profitable, although it has never been invoked despite high prices and substantial revenues made by some orphan drug manufacturers in Europe 
. Another alternative would be to seek reimbursement of the considerable government investments made in orphan drug development on the expectation that these drugs would find only limited use. The concept of reimbursing initial public investment in drug development remains controversial, although it has recently been endorsed by Francis Collins, Director of the National Institutes of Health 
Finally, our study suggests that regulators may be able to take a more proactive step and predict certain orphan products that are likely to be widely prescribed for non-FDA approved indications – those initially designed and approved to treat common symptoms or laboratory abnormalities, albeit in the context of a rare disease. For these products, it may be reasonable to withhold orphan drug status, and instead direct the limited resources of the program to encouraging development of novel products (such as imatinib) aimed at treating truly rare diseases. Applying incentives selectively to developing such products would adhere more closely to the original goals of the Orphan Drug Act.