Aaron S. Kesselheim, MD, JD, MPH

In response to brand-name drug manufacturers developing co-pay assistance programs to help patients pay high out-of-pocket costs for expensive brand-name drugs, insurers have developed co-pay accumulator adjustment programs (CAAPs) that exclude payments made by drug manufacturer assistance programs when calculating whether beneficiaries have met their yearly out-of-pocket maximums. By doing so, these programs are intended to drive beneficiaries to lower-priced generic alternatives. A rule finalized in July 2020 makes such programs permissible even when a brand-name drug does not have a generic or medically appropriate alternative, while not requiring transparency from payers about the existence of such programs. This commentary explains how CAAPs work and how this finalized rule may affect spending on prescription drugs.

This article describes a recently finalized CMS rule addressing the permissibility of co-pay accumulator adjustment programs when no generic is available.

This rule should be narrowed to situations in which a generic is available and medically appropriate, where such programs have the ability to incentivize use of lower-cost alternatives.

If the rule cannot be narrowed, it should be amended to require transparency from payers so that beneficiaries are able to anticipate future co-payments after they reach the manufacturer assistance program limit or, in the alternative, payers should affirmatively notify beneficiaries of the existence of such programs.

The high costs of prescription drugs remain an important issue, with a large majority of Americans saying that lowering drug prices should be a top priority for the new Biden administration.

 In the months before the 2020 election, the outgoing Trump administration issued a number of executive orders and rules aimed at the drug market. Most of the hastily designed measures were intended to lower prices, such as by facilitating drug importation, but are likely to be blocked by litigation.

 Paradoxically, one of the rules likely to persist after 2021 actually threatens to make drug prices more burdensome for patients by affecting how patient co-payments are counted by insurers.

The Trump administration finalized a rule in July 2020 aimed at co-pay accumulator adjustment programs (CAAPs), which are programs instituted by commercial health insurers that allow insurers to ignore co-payments made by drug manufacturer assistance programs on behalf of beneficiaries when determining whether beneficiaries have reached their yearly out-of-pocket maximum. Although CAAPs are designed to incentivize use of available low-cost generic versions of brand-name prescription drugs, the final rule permits CAAPs to be applied even when no generic is available, which will leave some patients unable to afford to fill their prescriptions once manufacturer assistance limits are met.

Many brand-name drug manufacturers have developed patient assistance programs or offer coupons to help patients with their out-of-pocket prescription drug costs. The overall benefits of such programs, however, depend on the context and availability of lower-cost options. These programs can be helpful for patients who require a brand-name drug for which there are no lower-cost alternatives. The out-of-pocket costs associated with these brand-name drugs can be a substantial financial burden, particularly for those filling prescriptions for costly drugs to treat chronic conditions. One study showed that for covered individuals with deductibles, almost 25% of prescriptions are abandoned before the yearly deductible is met because of inability to afford the prescription.

 In these circumstances, by covering all or some of these out-of-pocket costs, manufacturer assistance can improve overall adherence by helping beneficiaries afford their prescriptions, at least until the beneficiary reaches the assistance limit of the program.

By contrast, when lower-cost alternatives are available—either generic versions of the same drug or other lower-cost drugs shown to be useful in treating the condition—eliminating the cost-sharing responsibilities of the patient through drug manufacturer assistance can incentivize use of brand-name drugs over the generic. Although this may benefit the patient in the immediate transaction of filling a particular prescription, it enriches the pharmaceutical manufacturer and creates economic waste within the health care system. This economic waste, in turn, increases overall spending and drives up premiums, to the detriment of all beneficiaries. For example, one study estimated that the availability of brand-name drug manufacturer coupons in the first 5 years following generic approval resulted in as much as $120 million in extra spending for that particular drug. For drug spending overall, the study estimated an increase in retail spending of 4.6% for this 5-year period.

To eliminate the perverse incentives caused by such programs, health insurers and their pharmaceutical benefit managers designed CAAPs. CAAPs prevent the costs covered by drug manufacturer assistance programs from being counted toward an individual’s yearly out-of-pocket maximum—with the goal being to provide incentives for use of lower-cost drugs not covered by the manufacturer assistance programs.

 This out-of-pocket maximum, as set by the Patient Protection and Affordable Care Act (ACA), is an annual limit on how much commercial insurers are permitted to make beneficiaries pay for health services. Normally, each time a beneficiary pays a co-payment or coinsurance for a health service, that amount is counted toward this yearly limit. Once the limit is reached, the beneficiary is not responsible for co-payments or coinsurance for the remainder of the year.

Under a CAAP, however, if individuals have the costs of their prescriptions filled throughout the year covered entirely by an assistance program and incur no other cost sharing for other medical services rendered, they are still responsible for all co-payments and coinsurance—even if the amount paid by the drug manufacturer assistance program would have exceeded their yearly out-of-pocket maximum. In other words, only expenses actually paid by the beneficiary will count toward the yearly out-of-pocket maximum, ignoring those paid by manufacturer assistance. This is particularly problematic when the manufacturer assistance limit is met, leaving the beneficiary solely responsible for coinsurance and co-payments.

In theory, by ignoring payments made by manufacturer assistance programs on behalf of beneficiaries for high-cost drugs, commercial insurers can incentivize use of lower-cost alternatives. The success of CAAPs therefore requires 3 things: (1) Lower-cost alternatives must exist, such as in the form of generic versions of the brand-name drug or generic versions of drugs in the same class that have similar clinical benefits; (2) beneficiaries must know that these restrictions on manufacturer assistance exist and therefore be able to choose the lower-cost alternative; and (3) beneficiaries must be in a position so that, if not for these restrictions, they would have met their yearly out-of-pocket limits and benefited from being excused from coinsurance and co-payments for a portion of the year. If any of these requirements is not met, a beneficiary is likely to accept the immediate benefits of manufacturer assistance for prescription drug co-payments or coinsurance instead.

The Trump administration’s final rule removes the first 2 requirements listed above and gives insurers substantial latitude to institute CAAPs under the ACA. Starting in 2021, it would permit insurers to prevent coupons and other forms of direct assistance from drug manufacturers from being applied to the patient’s yearly out-of-pocket maximum. When it was first proposed, the rule included language limiting insurers’ exclusion of drug manufacturer assistance to situations in which a generic was available and medically appropriate.

 But the final rule from July 2020 removed that language, giving insurers the option to exclude direct drug manufacturer assistance from a beneficiary’s yearly out-of-pocket maximum, regardless of whether a generic is available. This means that for beneficiaries who rely on brand-name drugs for which a generic alternative is not available, drug manufacturer assistance will not help them get to their yearly out-of-pocket maximum if the plan utilizes such a CAAP, which will eventually translate to increased cost sharing on the individual. For those with chronic diseases or who rely on medications for which a generic is not available and depend on direct manufacturer assistance to mitigate costs at the pharmacy counter, the end of manufacturer assistance may mean reduced adherence to their medications and negative health effects.

The finalized rule explained that CMS expects insurers to “be transparent” in the plan documents as to whether drug manufacturer assistance counts toward a beneficiary’s out-of-pocket maximum when filling generic or brand prescriptions. However, despite receiving public comments raising concerns about what appears to be optional disclosures, the agency declined to implement mandatory disclosures, leaving the insurers to decide how transparent to make their plan documents.

 The discretionary disclosures leave beneficiaries vulnerable. A study assessing consumers’ understanding of terms such as deductible, co-pay, coinsurance, and out-of-pocket maximum revealed that many Americans do not understand these terms well—with only 14% of those surveyed correctly identifying the definitions of all 4 terms.

 When it comes to CAAPs, this understanding is even more limited, with one study reporting important inconsistencies with transparency about CAAPs in plan designs, including plans with no information at all about the existence of a CAAP.

CMS missed an opportunity to finalize aspects of the rule that could incentivize generics and chose not to mandate transparency—2 choices that could lead to increased out-of-pocket costs for beneficiaries. In 2021, CMS should focus primarily on narrowing CAAPs to situations in which a generic alternative is medically appropriate and available, similar to the original proposal in 2019. Narrowing the rule in this way will ensure that the tool is aimed at increasing use of low-cost generics and will protect beneficiaries who benefit from manufacturer assistance for brand-name drugs that do not have lower-cost alternatives.

If CMS cannot narrow the rule effectively, it should, in the alternative, revisit the mandatory disclosures and lay out a framework for informing beneficiaries of the benefits and detriments of accepting direct drug manufacturer assistance. Although it is unclear how effective such disclosures will be, given consumers’ limited understanding of complex plan designs, broad CAAPs coupled with a lack of transparency will likely lead to unnecessary surprises for beneficiaries.

Use of CAAPs should be reserved to incentivizing medically appropriate and available generics in an effort to lower overall prescription drug costs. When using CAAPs in less restricted situations, plan documents must not only explicitly state that drug manufacturer assistance will not be applied to annual out-of-pocket maximums, but also explain that beneficiaries will be subjected to cost sharing once the manufacturer assistance limits are met. Without these disclosures, beneficiaries may be forced to abandon prescriptions because of unforeseen additional co-pays.

 Program On Regulation, Therapeutics, And Law (PORTAL), Division of Pharmacoepidemiology and Pharmacoeconomics, Department of Medicine, Brigham and Women’s Hospital and Harvard Medical School (BSW, ASK), Boston, MA.

 The authors report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.

 Concept and design (BSW, ASK); acquisition of data (BSW); analysis and interpretation of data (BSW); drafting of the manuscript (BSW); critical revision of the manuscript for important intellectual content (BSW, ASK); and obtaining funding (ASK).

 Bryan S. Walsh, JD, Brigham and Women’s Hospital and Harvard Medical School, 1620 Tremont St, Ste 3030, Boston, MA 02120. Email: bswalsh@bwh.harvard.edu.

1. Lopes L, Hamel L, Kearney A, Bordie M. KFF Health Tracking Poll—October 2019: health care in the Democratic debates, Congress, and the courts. Kaiser Family Foundation. October 15, 2019. Accessed November 20, 2020. https://www.kff.org/health-reform/poll-finding/kff-health-tracking-poll-october-2019/

2. Sachs R. Administration finalizes most-favored nation drug pricing rule at the last minute. 

. November 23, 2020. Accessed December 1, 2020. https://www.healthaffairs.org/do/10.1377/hblog20201122.298613/full/

3. CMS, HHS. Patient Protection and Affordable Care Act; HHS notice of benefit and payment parameters for 2021; notice requirement for non-federal governmental plans. 

4. IQVIA Institute for Human Data Science. Medicines use and spending in the U.S.: a review of 2016 and outlook to 2021. IQVIA. May 4, 2017. Accessed November 21, 2020. https://www.iqvia.com/insights/the-iqvia-institute/reports/medicines-use-and-spending-in-the-us-a-review-of-2016

5. Dafny L, Ody C, Schmitt M. When discounts raise costs: the effect of copay coupons on generic utilization. 

. 2017;9(2):91-123. doi:10.1257/pol.20150588

6. Schmid CE II. New accumulator adjustment programs threaten chronically ill patients. 

. August 31, 2018. Accessed December 2, 2020. https://www.healthaffairs.org/do/10.1377/hblog20180824.55133/abs/

7. CMS, HHS. Patient Protection and Affordable Care Act; HHS notice of benefit and payment parameters for 2020. 

8. Loewenstein G, Friedman JY, McGill B, et al. Consumers’ misunderstanding of health insurance. 

. 2013;32(5):850-862. doi:10.1016/j.jhealeco.2013.04.004

9. Copay accumulator adjustment programs: putting insurance company profits over patients. The AIDS Institute. Accessed December 1, 2020. http://www.theaidsinstitute.org/sites/default/files/attachments/AI_CoPay_Accumulator_Adjustment_Brochure_w%20Appendix_FINAL.pdf

Articles

This article describes a recently finalized CMS rule addressing the permissibility of co-pay accumulator adjustment programs (CAAPs) when no generic is available.

CAAP Rule and Prescription Drug Prices

 To better understand patients’ and pharmacists’ preferences for and experiences with changes in pill appearance (size, shape, color, and markings).

 We conducted independent national surveys of patients 50 years and older taking generic drugs for depression, diabetes, epilepsy, HIV, hyperlipidemia, or hypertension and of licensed pharmacists practicing in chain, franchise, or independent pharmacies. Responses were collected between January and April 2016.

 Of 1000 patient respondents (30% response rate), most reported experiencing changes in pill appearance (51%) and preferred to be notified about them (82%), but less than half recalled being notified (verbally: 36%; via sticker: 45%). Among patients who reported experiencing a change, 12% reported stopping their medication or using it less frequently. Of 710 pharmacist respondents (33% response rate), many reported changes in pill appearance occurring frequently in their pharmacies (47% reported that changes occurred 6 or more times per month) and more than three-fourths reported notifying patients about them often (verbally: 88%; via sticker: 77%).

 Our findings reveal opportunities to improve patients’ experiences with pill appearance changes through better notification practices and patient education.

Conducting separate national surveys of patients and pharmacists, we aimed to better understand their preferences for, experiences with, and responses to changes in generic drug appearance (size, shape, color, and markings).

Most patients reported experiencing changes in pill appearance and preferred to be notified about them, but less than half recalled being notified. Among patients who reported experiencing a change, 12% stopped their medication or used it less frequently.

Most pharmacists reported pill appearance changes occurring frequently in their pharmacies and notifying patients about them often.

These findings reveal opportunities to improve patients’ experiences with pill appearance changes through better notification practices and patient education.

Medication nonadherence occurs when patients do not take their medication as prescribed. About 50% of patients with chronic diseases, such as high blood pressure, high cholesterol, and diabetes, are estimated to be nonadherent.

 Nonadherence can result from factors related to the disease (eg, changes in cognition), medication (eg, adverse effects), patient (eg, forgetfulness), or health system (eg, high drug costs),

 and it has been tied to poorer patient outcomes4 and billions of dollars in US health care spending annually on hospitalizations and other costs associated with disease progression.

One potential contributor to medication nonadherence is lack of consistency in the appearance of a patient’s medication. The FDA requires that generic drugs—which currently comprise about 90% of dispensed prescriptions in the United States

—have the same dosage form, strength, route of administration, intended use, quality, and performance characteristics as their brand-name versions.

 The FDA, however, does not mandate that generic drugs match their brand-name counterparts or each other in terms of physical appearance, including color, shape, size, and markings (text, lines, grooves, or designs). As a result, patients refilling a medication for a chronic disease may experience variations in their pills’ appearance when the generic manufacturer supplying a drug to a pharmacy changes. One study found that 29% of patients initiating treatment with 1 of 4 classes of cardiovascular drugs after hospitalization for a myocardial infarction experienced a change in their pills’ color or shape within 1 year.

 Well-controlled observational analyses have found increased odds of patients stopping or delaying taking essential cardiovascular and antiepileptic drugs when experiencing such changes.

Why patients might respond this way to changes in pill appearance is not known. To explore this question, we conducted national surveys of patients and pharmacists regarding their preferences for, experiences with, and perceptions and responses to changes in generic drug appearance.

We conducted 2 independent national surveys of patients and pharmacists with assistance from Nielsen Healthcare. The Office of Management and Budget approved the study under the Paperwork Reduction Act of 1995, assigning it the control number 0910-0801. The Brigham and Women’s Hospital Institutional Review Board also approved this study. All analyses were performed using SAS version 9.4 (SAS Institute).

Eligibility for the patient survey was restricted to patients 50 years and older taking a generic medication for depression, diabetes, epilepsy, HIV, hyperlipidemia, or hypertension. These conditions were chosen to capture a wide range of chronic diseases. The age criterion was included to increase the likelihood that contacted individuals—recruited through phone sampling—would have 1 of the listed chronic diseases.

A total of 16,243 landline and cell phone numbers were identified from Survey Sampling International’s Directory Listed Household Sample and Wireless LITe database. From this sample, 6 pretests were conducted in January 2016 that resulted in minor modifications to the survey instrument, such as grammatical fixes and adjustments for clarity. Phone interviews took place between February and April 2016, lasted an average of 18 minutes, and concluded once the target of 1000 respondents was reached.

The patient survey was administered using a 56-question, closed-field instrument (

]). We first inquired about reliance on and preference for consistent pill appearance, asking patients whether they used pill appearance to make sure they took the correct medication and whether they preferred their pills to retain the same color, shape, size, or markings upon medication refill. We then asked patients how often they experienced a change in their pills’ appearance, whether their pharmacist notified them about the change, and what their preference was for being notified when their pills’ appearance changes. Other questions addressed patients’ actions in response to changes in their pills’ appearance. Finally, we collected information about patients’ sex, age, race, education, income, and Census region. Reliance, preference, and perception questions were assessed using Likert scales.

The pharmacist survey targeted licensed pharmacists practicing in independent, franchise, and chain settings in the United States. We randomly selected 2161 practicing pharmacists believed to be practicing in such settings from a master file of more than 100,000 pharmacists in more than 50,000 pharmacies maintained by the data vendor SK&A. The number was chosen based on power calculations using simulated response distributions and an assumed 50% response rate.

Each selected pharmacist was sent a postcard alert and a survey mailing between January and April 2016 that included a paper survey, a URL for optional online completion of the survey, a prepaid return envelope, and a $5 cash honorarium. Nonresponders received a reminder letter and a second packet without the honorarium.

The pharmacist survey was self-administered using an instrument with 23 closed-field questions (

). The survey was pretested in a convenience sample of 5 practicing pharmacists, with small adjustments made for question clarity based on their responses. We asked pharmacists about their perceptions of generic drug safety and effectiveness, followed by their preferences for consistent pill appearance. We then asked how frequently pharmacists observed generic pill appearance changes in their pharmacies and what they generally did to alert patients when such changes occurred. Other questions focused on reports that pharmacists received from patients in response to pill appearance changes, such as a belief that the wrong medication had been given. Finally, we collected information on pharmacists’ sex, race, title, business type, number of years in practice, number of prescriptions filled per day, and source(s) of drug information. As with the patient survey, preference and perception questions were assessed using Likert scales.

The response rate for the patient survey was calculated using the American Association of Public Opinion Research’s second standard definition.

 Multivariable logistic regression modeling was used to examine predictors of patient preferences for pill appearance. For this model, a constructed outcome variable was derived based on the question of whether patients preferred the size, shape, color, or markings of their pills to remain the same, grouping patients who responded “definitely” or “probably” to any of these 4 questions vs patients who responded “no preference,” “definitely not,” or “probably not” to all questions. Multivariable logistic modeling was also used to examine predictors of patients who reported using a medication less frequently or stopping a medication following a pill appearance change. For each model, the covariates were sex, race (white/nonwhite), age (50-57, 58-65, ≥ 66 years), medical condition (hypertension, hyperlipidemia, diabetes, depression, epilepsy, HIV, multiple), region (East, Midwest, South, West), education (high school, college, graduate), and income (<$50,000, $50,000-<$100,000, ≥$100,000).

Multivariable ordinal logistic regression modeling was used to examine predictors of pharmacist preferences for generic pill appearance consistency, grouping pharmacists who responded “much prefer” or “somewhat prefer,” “no preference,” and “somewhat not prefer” or “much not prefer.” The frequency of observing generic pill appearance changes, frequency of being notified about generic pill appearance changes, likelihood of notifying patients verbally in person or on the phone about pill appearance changes, and likelihood of notifying patients via sticker on the pill bottle about pill appearance changes were also modeled using ordinal logistic regression but without any grouping of response categories. Covariates in the models were pharmacy type (chain, nonchain), years in practice (continuous), position (owner/chief pharmacist, staff/other), and pharmacy prescription dispensing volume (≤ 200 prescriptions per day, > 200 prescriptions per day).

A total of 1000 patients (30% response rate) and 710 pharmacists (33% response rate) completed the surveys (

). Of the patient respondents, 61% were women, 79% were white, and 57% were 66 years or older. Of the pharmacist respondents, 47% were women, 80% were white, and 56% identified as a staff pharmacist. These percentages are comparable with those nationally.

 About half of respondents filled more than 200 prescriptions per day (52%), and about one-third practiced in an independent pharmacy (38%) and reported practicing for 30 years or longer (34%). Although we targeted pharmacists in independent, franchise, and chain settings, 10% of respondents reported practicing in other settings, most often hospital-associated outpatient pharmacies.

Patients’ and Pharmacists’ Pill Appearance Preferences

Most patients preferred to have a consistent pill appearance. More than half of patients “definitely” or “probably” preferred their prescription refills to have the same color (63%), size (68%), shape (63%), and markings (55%) (

). In multivariable modeling, only race was associated with a preference for consistent pill appearance (

). Nonwhite patients were more likely to prefer consistent pill appearance (odds ratio [OR], 2.58; 95% CI, 1.52-4.38).

In contrast with patients, most pharmacists (54%) had no preference for consistent pill appearance of generic drugs when dispensing, and 12% preferred not to dispense generic drugs with the same appearance (Table 2). Multivariable analysis revealed that pharmacists who reported practicing for fewer years were more likely to prefer pill appearance consistency (OR per 10 years of practice, 0.78; 95% CI, 0.69-0.88) (

About half of patients (51%) reported receiving a prescription refill in which their pills’ appearance changed in the last year and, of those patients, about half (53%) reported that it happened 2 or more times (Table 2). Patients reported that pill color was the characteristic that changed most frequently (68%), followed by shape (46%), size (42%), and markings (20%).

More than four-fifths of pharmacists reported that their pharmacy changed generic drug manufacturers with resulting changes in pill appearance at least twice per month: never or once per month (15%), 2 to 5 times per month (38%), 6 to 15 times per month (25%), and more than 15 times per month (22%) (Table 2). In multivariable modeling, chain/franchise pharmacies were more likely than independent pharmacies to experience such changes (OR, 1.57; 95% CI, 1.11–2.22) (Table 4).

A majority of pharmacists reported that they were either “never” (28%) or “rarely” (28%) alerted when their pharmacy, distributor, or wholesaler changed suppliers for generic pills with a different appearance (Table 2). In multivariable modeling, only pharmacy type was associated with such notification, with pharmacists working at a chain/franchise pharmacy having a greater likelihood than pharmacists in independent pharmacies of being alerted when pill appearance changes occur (OR, 1.62; 95% CI, 1.15-2.27) (Table 4).

Patient and Pharmacist Notification Preferences, Experiences, and Practices

Patients overwhelmingly reported that they “definitely” (66%) or “probably” (16%) desired to be notified of a change in their pills’ appearance by their pharmacist and that they preferred being notified verbally (45%) over a sticker on the prescription bottle (36%) (Table 2). Among patients reporting experiencing a change in pill appearance during the prior year, 36% reported that pharmacists notified them about the change verbally, whereas 45% reported being notified with a sticker on the bottle.

When pharmacists were asked about their notification practices when pill appearance changed, most pharmacists reported that they “almost always” (47%) or “commonly” (30%) notified patients verbally (Table 2). A large majority of pharmacists also reported that they “almost always” (73%) or “commonly” (15%) notified patients using a sticker on the prescription bottle. Most pharmacists reported “never” (55%) logging an incident report when a patient complained about a change in pill appearance and “rarely” (27%) or “never” (34%) leaving a note in the patient’s pharmacy record indicating that a change occurred. Multivariable analysis revealed that pharmacists working in pharmacies filling more than 200 prescriptions per day were more likely to notify patients about a change in pill appearance verbally (OR, 1.78; 95% CI, 1.34-2.38) and via a sticker on the pill bottle (OR, 1.65; 95% CI, 1.17-2.33) (Table 4).

Patient Responses to and Beliefs About Pill Appearance Changes

Among patients reporting that they experienced a pill appearance change, almost one-third (29%) reported thinking that they received the wrong pill (Table 2). About 1 in 8 patients reported changing their medication use, either using the pills less frequently (8%) or stopping their use completely (4%). Multivariable analysis found that nonwhite patients were more likely than white patients to use their pills less frequently or stop completely when they experienced a change in their pills’ appearance (OR, 2.89; 95% CI, 1.42-5.85) (Table 3).

Other actions that patients reported taking in response to a change in their medication’s appearance included asking their pharmacist (35%) or physician (9%) about the change (Table 3). Of those patients who reported asking their pharmacist or physician, 97% reported that their pharmacists alleviated their concerns, and 84% reported that their physician alleviated their concerns.

In national surveys, we found that most patients preferred the appearance of their generic pills to remain the same, with some patients reporting using their medication less frequently or stopping it completely when such changed occurred. Pharmacists reported that changes in generic pill appearance happened frequently in their practice, but that they were often not alerted when such changes occurred. Far more pharmacists reported notifying patients about generic pill appearance changes, either verbally or with a sticker on the bottle, than patients recalled being notified by pharmacists.

Our patient findings align well with those of a separate survey we conducted of commercially insured patients taking generic medication indicated for 1 of 4 chronic diseases (depression, epilepsy, hyperlipidemia, and hypertension) who had recently experienced a change in the appearance of their pills.16 In this study, most respondents also reported wanting their pills to remain the same color (72%), size (75%), and shape (71%). Similarly, more than three-fourths of respondents (86%) also reported wanting to be notified when their pills changed appearance, and a minority likewise recalled their pharmacist informing them of the change (verbal: 19%; via a sticker on the pill bottle: 30%). Finally, the same percentage of respondents (8%) reported adjusting their medication use following the change. Combined, these findings suggest that many patients rely on pill appearance to make sure they are taking the correct medication and that changes in appearance may cause confusion, prompting some patients to stop taking their medication. This may be particularly important among elderly patients taking multiple medications for chronic diseases. We are unaware of prior published surveys about the preferences of and experiences with pill appearance changes among pharmacists.

Our study reveals opportunities to improve patient and pharmacist experiences when pill appearance changes: through patient education about the potential for pill appearance changes for generic medication, along with better notification practices and knowledge transfer when such changes occur. Because many pharmacists reported rarely or never being alerted of such changes, creation and widescale adoption of an automated point-of-dispensing system that alerts pharmacists when a patient’s generic pill has changed appearance and indicates what that change is (eg, size, shape, color, and/or markings) would be helpful.

The observed disconnect between pharmacists reporting notifying patients about generic pill appearance changes and patients reporting not recalling being notified, meanwhile, suggests that the mode of communication between pharmacists and patients is important. Pharmacists most commonly reported notifying patients using a sticker on the bottle regardless of their practice type. Patients, however, reported preferring a verbal notification and may not have looked closely at their pill bottles.

Finally, some patients—notably, nonwhite patients—were more likely to report using their medication less frequently or stopping completely when a generic pill appearance change occurred. It is unclear why this may be the case, highlighting an avenue for further investigation in addition to populations for targeted intervention.

There are limitations to our study relevant to the disconnect identified between patients and pharmacists. The surveys were conducted independently, meaning that the patients surveyed did not receive prescriptions dispensed by the pharmacists surveyed. Like all surveys, our study was also subject to possible participation, recall, and response biases. Patients who were counseled by their pharmacist may have misremembered the conversation, and patients who had issues with changes in their pills’ appearance may have been more likely to participate. Similarly, patients with HIV may been reluctant to report having the condition. Survey questions for patients and pharmacists, meanwhile, were not phrased identically and may have been interpreted differently. For example, patients may have interpreted the question “When your pill changed in appearance, as best you recall, did your pharmacist talk to you about the change, in person or on the phone?” literally, thinking that if someone other than the pharmacist (eg, pharmacy technician) talked to them, then they were not notified. By contrast, pharmacists may have had a more general interpretation of their question asking about prescription refill practices when pill appearance changes, thinking that if a member of the pharmacy team, such as a pharmacy intern, counseled the patient, then that was sufficient and reported accordingly.

Additional limitations to our study warrant noting. First, our patient findings were limited to older patients with 6 specific chronic conditions and may not be generalizable to younger patients or those with other diseases. Second, given our response rates and sampling frames, it is possible that our results are not generalizable to all segments of the target populations. Third, our surveys were designed to explore experiences with changes in pill appearance but could not assess the clinical appropriateness of the change. Finally, our ascertainment of patient sex based on voice could have been inaccurate.

There were also several strengths to our investigation. Both patients and pharmacists were surveyed—as opposed to one group or the other—about similar topics. The surveys were national in scope, and the number of respondents for both was large.

Our study indicates that most patients report not being notified when they experience changes in pill appearance during a medication refill even though pharmacists report that these changes happen frequently in the pharmacy. A disconnect exists between patients’ reported experiences and pharmacists’ reported practices regarding notification when that change does occur, with some patients altering their medication use in response. Improving patient education about potential pill appearance changes and notification practices around the time such changes occur may improve the experiences of both patients and pharmacists.

 Program on Regulation, Therapeutics, and Law (PORTAL), Division of Pharmacoepidemiology and Pharmacoeconomics, Department of Medicine, Brigham and Women’s Hospital and Harvard Medical School (REB, ASK, JJG, ZL, AS), Boston, MA; Center for Bioethics and Humanities at the University of Colorado (EGC), Aurora, CO; Office of Generic Drugs Surveillance and Epidemiology, Center for Drug Evaluation and Research, FDA (SKD), Silver Spring, MD; Office of Research and Standards, Office of Generic Drugs, Center for Drug Evaluation and Research, FDA (WJ), Silver Spring, MD.

 This work was supported by a contract from the FDA Office of Generic Drugs and Office of Chief Scientist (HHSF223201310232C).

 Dr Barenie was a consultant to Alosa Health for unrelated work. Dr Gagne was a consultant to Optum, Inc, for unrelated work and received salary support from grants from Eli Lilly and Company and Novartis Pharmaceuticals Corporation to the Brigham and Women’s Hospital for unrelated work. Dr Campbell serves as a paid expert on law cases related to conflicts of interest in medicine. Dr Dutcher and Dr Jiang are employees of the FDA, which funded the work described in this manuscript. The remaining authors report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.

 Concept and design (ASK, JJG, EGC, SKD, WJ, AS); acquisition of data (ASK, EGC, AS); analysis and interpretation of data (REB, ASK, JJG, ZL, SKD, WJ, AS); drafting of the manuscript (REB, ZL, AS); critical revision of the manuscript for important intellectual content (REB, ASK, JJG, EGC, SKD, AS); statistical analysis (ZL, AS); obtaining funding (ASK, AS); administrative, technical, or logistic support (WJ, AS); and supervision (JJG, AS).

 Rachel E. Barenie, PharmD, JD, MPH, Program on Regulation, Therapeutics, and Law (PORTAL), Division of Pharmacoepidemiology and Pharmacoeconomics, Department of Medicine, Brigham and Women’s Hospital and Harvard Medical School, 1620 Tremont St, Suite 3030, Boston, MA 02120. Email: rbarenie@bwh.harvard.edu.

1. Jackevicius CA, Mamdani M, Tu JV. Adherence with statin therapy in elderly patients with and without acute coronary syndromes. 

. 2002;288(4):462-467. doi:10.1001/jama.288.4.462

2. Oung AB, Kosirog E, Chaves B, Brunner J, Saseen JJ. Evaluation of medication adherence in chronic disease at a federally qualified health center. 

. 2017;8(8-9):113-120. doi:10.1177/2040622317714966

3. McLeroy KR, Bibeau D, Steckler A, Glanz K. An ecological perspective on health promotion programs. 

. 1988;15(4):351-377. doi:10.1177/109019818801500401

4. Overcoming barriers to medication adherence for chronic diseases. CDC. February 21, 2017. Accessed November 6, 2019. https://www.cdc.gov/grand-rounds/pp/2017/20170221-presentation-medication-adherence-H.pdf

5. Medication adherence in America: a national report. National Community Pharmacists Association. 2013. Accessed November 6, 2019. http://www.ncpa.co/adherence/AdherenceReportCard_Abridged.pdf

6. Avoidable costs in U.S. healthcare: the $200 billion opportunity from using medicines more responsibly. IMS Institute for Healthcare Informatics. June 2013. Accessed November 6, 2019. http://offers.premierinc.com/rs/381-NBB-525/images/Avoidable_Costs_in%20_US_Healthcare-IHII_AvoidableCosts_2013%5B1%5D.pdf

7. Jha AK, Aubert RE, Yao J, Teagarden JR, Epstein RS. Greater adherence to diabetes drugs is linked to less hospital use and could save nearly $5 billion annually. 

. 2012;31(8):1836-1846. doi:10.1377/hlthaff.2011.1198

8. Generic drug access & savings in the U.S.: access in jeopardy. Association for Accessible Medicines. 2018. Accessed November 6, 2019. https://accessiblemeds.org/sites/default/files/2018_aam_generic_drug_access_and_savings_report.pdf

9. Applications for FDA Approval to Market a New Drug. 21 CFR §314.3, 320 (2019).

10. Kesselheim AS, Bykov K, Avorn J, Tong A, Doherty M, Choudhry NK. Burden of changes in pill appearance for patients receiving generic cardiovascular medications after myocardial infarction: cohort and nested case-control studies. 

. 2014;161(2):96-103. doi:10.7326/M13-2381

11. Kesselheim AS, Misono AS, Shrank WH, et al. Variations in pill appearance of antiepileptic drugs and the risk of nonadherence. 

. 2013;173(3):202-208. doi:10.1001/2013.jamainternmed.997

12. Buttorff C, Ruder R, Bauman M. Multiple chronic conditions in the United States. RAND Corporation. 2017. Accessed November 6, 2019. https://www.rand.org/content/dam/rand/pubs/tools/TL200/TL221/RAND_TL221.pdf

13. Standard definitions: final dispositions of case codes and outcome rates for surveys. American Association for Public Opinion Research. 2016. Accessed November 6, 2019. https://www.aapor.org/AAPOR_Main/media/publications/Standard-Definitions20169theditionfinal.pdf

14. 2014 national pharmacist workforce survey. American Association of Colleges of Pharmacy. April 8, 2015. Accessed November 6, 2019. https://www.aacp.org/sites/default/files/
executivesummaryfromthenationalpharmacistworkforcestudy2014.pdf

15. Pharmacists. Data USA. Accessed November 6, 2019. https://datausa.io/profile/soc/291051/#demographics

16. Sarpatwari A, Gagne JJ, Lu Z, et al. A survey of patients’ perceptions of pill appearance and responses to changes in appearance for four chronic disease medications. 

. 2019;34(3):420-428. doi:10.1007/s11606-018-4791-1

Changes in generic drug appearance occur often. Patients’ and pharmacists’ responses to those changes vary, with some patients stopping their medication or using it less.

Preferences for and Experiences With Pill Appearance Changes: National Surveys of Patients and Pharmacists

 A “chiral switch” occurs in the pharmaceutical market when a drug made up of 2 enantiomer forms is replaced with a purified single-enantiomer version, often in the context of a patent expiration. We studied the prevalence of chiral switching in the United States over the past decade, including trends in use of, and expenditures on, these products in Medicaid.

 We used US Adopted Names prefixes (lev/levo/ar/es/dex/dextro) to identify all single-enantiomer drugs approved from 2001 to 2011. From publicly available US Food and Drug Administration (FDA) approval documents, we extracted the characteristics of the pivotal premarket trials for the single enantiomers. Specifically, we evaluated whether the single enantiomer was directly compared with the precursor racemic drug and whether there was evidence of superior efficacy. We used quarterly drug expenditure data from each state Medicaid program to chart trends in use of, and spending on, the single-enantiomer products and their racemic precursors during the study period.

 From 2001 to 2011, the FDA approved 9 single-enantiomer products: dexlansoprazole, levoleucovorin, levocetirizine, armodafinil, arformoterol, eszopiclone, escitalopram, dexmethylphenidate, and esomeprazole. Of those 9 drugs, 3 had at least 1 pre-approval randomized trial that included the racemic precursor as a direct comparator, but there was no evidence of superiority of the single enantiomer over the racemic at comparable doses. Between 2001 and 2011, US Medicaid programs spent approximately $6.3 billion on these 9 single-enantiomer drugs.

 Recently approved single-enantiomer drugs showed no evidence of superior efficacy over the older racemic precursors in the pivotal trials leading to their approval, and in a majority of cases, they were not directly compared.

From 2001 to 2011, the US Food and Drug Administration approved 9 single-enantiomer drugs with racemic precursors. Only 3 had pre-approval studies that compared the single-enantiomer with its precursor, and none showed evidence of improved patient outcomes.

		US Medicaid programs spent approximately $6.3 billion on these 9 single-enantiomer drugs during this period.

		Well-designed comparative trials are needed to clarify, soon after approval, whether a single-enantiomer drug represents a measurable, cost-effective improvement over current racemic therapy.

		Large healthcare organizations that bear the brunt of these costs should use their position to educate prescribers about appropriate use of single-enantiomer products.

Most physicians, and certainly most patients, have never heard of the “chiral switch.”

 A chiral drug is a single molecule product that exists in 2 mirror image forms, called enantiomers.

 Despite their similar chemical structures, enantiomers can have different biological properties.

 For example, some enzymatic processes can distinguish between the R- (from the Latin rectus for “right”) and S- (from the Latin sinister for “left”) enantiomers, 

 such that 1 enantiomer may be responsible for much of the pharmaceutical benefit while the other is inactive or even harmful.

The phrase “chiral switch” was coined to refer to the substitution in the marketplace of a racemic drug (the name for a 50:50 mixture of 2 enantiomers) with a single-enantiomer version.

One prominent example of a chiral switch occurred in 2001, when the US Food and Drug Administration (FDA) approved AstraZeneca’s esomeprazole (Nexium), a proton pump inhibitor (PPI) used to treat gastroesophageal reflux disease and erosive esophagitis. At that point, the PPI marketplace was dominated by AstraZeneca’s omeprazole (Prilosec), which had been approved in 1989. Marketed as “the purple pill,” omeprazole earned over $6 billion per year in the United States by 2000, although the key patents protecting its market exclusivity ended in 2001, opening the door to generic competition.

 Omeprazole is a racemic mixture of R-omeprazole and S-omeprazole, while esomeprazole, as its name implies, is isolated S-omeprazole. The S-omeprazole enantiomer was responsible for the drug’s clinical properties while the R-omeprazole enantiomer was inactive. After FDA approval, esomeprazole was marketed as “the new purple pill” to be used in place of generic omeprazole.

Chiral switching is a controversial practice. Some claim that singleenantiomer drugs offer little clinical advantage and are used by pharmaceutical manufacturers to perpetuate revenues as the original racemic pill approaches the end of its market exclusivity. The case of esomeprazole, in fact, has been cited as an example of “corporate waste” of healthcare resources.

However, in some cases, single enantiomers may offer benefit to patients, particularly if the inactive enantiomer may also be responsible for unwanted side effects; manufacturers may pursue single-enantiomer development because of this potential for improved safety or efficacy. Indeed,in guidance originally issued in 1992, and most recently updated in 2011, the FDA identified examples of drug toxicity associated with 1 member of a pair of enantiomers.

Regardless of the motivation for development of a single-enantiomer drug, the FDA review and approval process is the same. New medications are generally approved by the FDA on the basis of effectiveness in reaching a particular outcome, and an improvement over placebo is commonly accepted.

 Since the FDA cannot by law require active comparators in clinical trials leading to drug approvals,

 singleenantiomer products can reach the market by demonstrating improvement over placebo, not their closely related racemic precursors.

How prevalent is chiral switching in the United States? To address that question, we identified all single-enantiomer medications approved from 2001 to 2011 with racemic mixtures already on the market. We examined FDA documents from the approval of these medications, focusing on their pivotal clinical trials and identifying whether the single-enantiomer version was formally compared with the racemic drug in the course of FDA approval. Finally, we use national data on prescription use and expenditures in Medicaid to illustrate the costs associated with their entry and trends in their use.

We used the FDA database of drug approvals to identify all new single-enantiomer products approved for use between 2001 and 2011 that originated from an already FDAapproved racemic mixture. We systematically searched for all United States Adopted Names (USAN) prefixes assigned to drugs with a single-enantiomer formulation: lev-/levo- for the levorotatory, S-isomer form; ar- for the levorotatory, Risomer form; es- for the dextrorotatory, S-isomer form; and dex-/dextro- for the dextrorotatory, R-isomer form.

We extracted the manufacturer name and approval year for each identified drug. We then examined the FDA approval letter(s), summary review, and medical review(s) for each of the single-enantiomer drugs, along with the approved label, and collected reports of the pivotal trials leading to FDA approval. From these reports, we extracted whether the single enantiomer was compared with the racemic mixture in those trials and whether there was any evidence of superior efficacy. We also identified whether these comparisons to the racemic mixture were reported on the approved drug label. Since no pre-approval studies were powered to assess comparative safety, we focused on efficacy.

For a more detailed investigation of chiral switching, we focused on esomeprazole (Nexium) and escitalopram (Lexapro), the 2 single-isomer drugs identified in our search with the largest revenue in the past decade. These drugs were also approved early in our study period, giving us adequate time to examine trends in prescription use. Our data source was the Centers for Medicare & Medicaid Services (CMS), which provides aggregated quarterly drug expenditure data from each state Medicaid program categorized by National Drug Code (NDC).

 We used the NDC code to link the drugs and their racemic counterparts to the National Drug Data File (NDDF) (First Data Bank, National Drug Data File, San Bruno, CA: Hearst Corporation, 2012). For each drug, we identified the total number of prescriptions filled and the amount spent on the drug by state Medicaid programs in each calendar quarter. For 49 states and the District of Columbia (Arizona data were not available), we analyzed data from 1999 through the most recent quarter of Medicaid data available at the time of the analysis.

From 2001 through 2011, the FDA approved 9 singleenantiomer versions of racemic products (

). Eight of the drugs had racemic precursors available in the United States. The lone exception was eszopiclone (Lunesta), whose precursor zopiclone (Imovane, Zimovane) was only commercially available outside the United States. Three of the 9 drugs—levoleucovorin (Fusilev), levocetirizine (Xyzal), and arformoterol (Brovana)&mdash;came from manufacturers other than the one that applied for approval of the racemic drug. The drugs covered a range of indications from depression to chronic obstructive pulmonary disease (COPD).

Comparative Efficacy of Single-Enantiomer and Racemic Drugs 

Three of the 9 single-enantiomer drugs had at least 1 preapproval randomized, controlled trial that included the racemic precursor as a direct comparator (

). Esomeprazole 40 mg demonstrated statistically significant efficacy over omeprazole 20 mg in 2 studies, although it was not superior in 2 other studies. However, pharmacologically equivalent doses of the racemic and enantiomer medications were not compared in these studies, as 40 mg of the single-enantiomer product would contain at least 3 times as much of the active S-isomer as 20 mg of racemic omeprazole on a per milligram basis.

 Dexlansoprazole (Dexilant) and levocetirizine (Xyzal) demonstrated efficacy similar to that of their racemic predecessors in the pre-approval studies. Two additional drugs—escitalopram (Lexapro) and dexmethylphenidate (Focalin)&mdash;included the racemic as an active control in pre-approval studies but compared the 2 arms with placebo, not with each other.

Trends in the Use of Esomeprazole and Escitalopram 

Between 2001 and 2011, Medicaid programs nationally spent approximately $6.3 billion on the 9 single-enantiomer drugs we reviewed (

). The 2 greatest contributors to this total were esomeprazole ($3.5 billion) and escitalopram ($1.8 billion), each representing over 60% of total Medicaid spending on the single-enantiomer and racemic precursor over that time period. Dexmethylphenidate spending was $678 million from 2001 to 2011, while the single enantiomers approved after 2004 generated substantially lower Medicaid spending.

After esomeprazole’s approval in 2001, its use in Medicaid quickly surpassed omeprazole's and peaked at over 1 million prescriptions per quarter in the second half of 2005 (

). Generic omeprazole entered the market in the fourth quarter of 2002, became available as an over-the-counter drug in June 2003, and peaked in the third quarter of that year with almost 400,000 prescriptions. However, generic omeprazole did not surpass esomeprazole in number of prescriptions until early 2010. The data show an increase in the use of omeprazole between 2009 and 2011, potentially signifying enhanced coverage of over-the-counter omeprazole in Medicaid programs, which had largely excluded this version of the product in the mid-2000s.

Esomeprazole represented almost one-third (32.7%) of prescriptions even at the end of the study period in the third quarter of 2011.

Escitalopram followed a similar pattern. After the introduction of escitalopram in 2002, its use in Medicaid rapidly increased, such that by the second quarter of 2004, over 1 million prescriptions were filled per quarter (

). Generic citalopram entered the market in the fourth quarter of 2004, leading to a steep reduction in the use of brand name citalopram, and also a leveling off in the use of escitalopram, although the number of prescriptions for escitalopram continued to outnumber those for generic citalopram until the end of 2009. Prescriptions of all products fell in 2006 due to transition of dual-eligible elderly patients to Medicare Part D.

From 2001 to 2011, the FDA approved 9 single-enantiomer drugs, several of which reached blockbuster status in the United States.

 Despite substantial spending on enantion mers, most of these drugs were not directly compared with their racemic precursor in the course of FDA approval, and when they were, there was no clear evidence of any improvements in patient outcomes.

In the postapproval period, there has also been little evidence showing any substantial comparative efficacy benefits of these single enantiomers over their racemic precursors. For example, a Cochrane systematic review of escitalopram’s effectiveness identified 6 studies of varying methodological quality (both published and unpublished) directly comparing escitalopram with citalopram. Of the 6 studies, 3 showed a statistically significant advantage of escitalopram in the primary outcome of response to treatment, and each was industry funded.

 The review authors concluded that there was a potential for “overestimation of treatment effect due to sponsorship bias.” The review also found in pooled analyses that escitalopram was no more effective than 3 other antidepressants in the same class (fluoxetine, paroxetine, and sertraline). Evidence of superior effectiveness did not emerge after esomeprazole’s approval, either. The studies reviewed by the Drug Effectiveness Review Project found that comparisons between the enantiomer and racemic precursor uniformly compared esomeprazole at a pharmacologically superior dose (40 mg) with omeprazole (usually at a dose of 20 mg), limiting the possibility of making any conclusions about efficacy.

 The Cochrane review of treatment of gastroesophageal reflux concluded that different PPIs do not show statistically significant differences in effect at comparable dosages.

 Notably, the 2 studies that showed superiority of esomeprazole over omeprazole in the pre-approval RCTs were published within 1 year of approval,

 while the 2 studies that showed no superiority were not published until 5 years later,

 again raising the concern about publication bias for the postapproval comparative effectiveness studies. Although there is no requirement for comparison of single-enantiomer drugs with their racemic precursors for FDA approval, we found numerous examples in the FDA reviews of experience with the racemic product being invoked to justify the need (or lack of need) for additional testing or guidance prior to approving the drugs. For example, while the FDA does not convene external advisory committee meetings for every drug application, it elected to not do so for dexlansoprazole specifically because the “product is not a new molecular entity. It is the R-enantiomer of a product that was approved in 1995.”

 In another instance, the Division of Pulmonary and Allergy Products did not request audits by the Division of Scientific Investigation because arformoterol “is the active enantiomer of the well-characterized LABA [long-acting beta-agonist], racemic formoterol, which is already approved for treatment of both asthma and COPD, the efficacy data are robust and as would be expected for the product, and the sponsor is not making any novel claims for the product.”

 Regulatory consideration of the safety of single-enantiomer products also benefitted from prior experience with racemic precursors. In the medical review for esomeprazole, the reviewer wrote, “Because Nexium is the S-enantiomer of omeprazole and omeprazole is a PPI that is perceived as safe, Nexium is also considered safe. Omeprazole has been marketed worldwide since 1988 and is presently available in 106 countries for various acid-related gastrointestinal disorders.”

Although we did not focus on comparative safety or tolerability of the enantiomers, it is conceivable that the benefit of single enantiomers is not in superior efficacy, but superior tolerability or safety. We are not aware, however, of any strong evidence to support that hypothesis for the drugs we studied. For example, in the Cochrane review comparing escitalopram and citalopram, there was no difference between the 2 drugs in study discontinuation due to side effects or number of patients experiencing a side effect.

Similarly, prior reviews have found no evidence of any difference in adverse events (AEs) or withdrawals due to AEs across trials of PPIs.

 To be sure, there are past examples of chiral switching bringing about important safety benefits, as in the case of levodopa, a key treatment for Parkinson’s disease; initially, the racemic drug caused significant AEs, but shifting to the single enantiomer reduced the AEs and allowed for a higher dose of the active drug.

Despite unclear efficacy or safety benefits, approval of enantiomer drugs over the past decade has had substantial cost implications. In the cases of esomeprazole and escitalopram, the enantiomer rapidly absorbed market share from the racemic precursor and diverted use of generic versions of the racemic product. Increasing the cost of medications with no discernible difference in health outcomes negatively impacts the health of patients taking those drugs. Higher prescription drug costs are associated with lower adherence, higher incidence of serious AEs, greater utilization of healthcare services, and worse health outcomes

 These costs must be weighed against any theoretical, or even real, improvement in patient outcomes with use of single-enantiomer formulations.

The case of armodafinil gives further insight into the use of enantiomers to maintain market share in products losing patent protection. Cephalon received approval for modafinil (Provigil) by the FDA in 1998. In 2002, 4 companies filed applications with the FDA for generic modafinil.

 In order to maintain its market exclusivity, in 2006, Cephalon settled with those generic manufacturers for more than $200 million, delaying entry of a generic modafinil until 2012. In 2007, Cephalon’s single-enantiomer armodafinil (Nuvigil) was approved, meaning that the company then had 5 years to “switch” market share away from modafilil.

 In order to encourage transition of modafinil to armodafinil, Cephalon steadily increased the price of modafinil, from $6.80 per pill in 2008 to $13.62 in 2009, by which time armodafinil was launched at $8.98 per pill.

 By 2011, sales of Nuvigil represented 42% of combined armodafinil/modafinil prescriptions in the United States.

Most patients are probably not aware that the singleenantiomer drugs they are prescribed are so closely related to precursor racemic mixtures. Linguistic similarities between generic names of racemic and enantiomer drugs will likely be missed by most physicians and patients, particularly since most drugs are referred to by their brand name, even after they are being produced by generic manufacturers.

 Thus, naming conventions artificially inflate the differences between the single-enantiomer and racemic drugs.

 If patients (and their physicians) were better educated about the clinical similarities between racemic and enantiomer drugs and were given greater insight into the trials that led to the approval of enantiomers, they may be less likely to pay higher prices for the medication.

A number of measures can be taken to facilitate more rational use of enantiomer drugs by physicians and patients. Although the FDA cannot require comparative efficacy testing prior to approval, the FDA should ensure that single-enantiomer drug labels highlight differences, or the lack of differences, in efficacy and safety from the racemic version.

 Three enantiomer drugs in our sample had direct racemic comparators in approval trials, but only 2 of the labels mentioned the study results (dexlansoprazole and esomeprazole). Special attention should be paid to the single enantiomers when formulary tiering decisions are made by health plans, although heightened attention by insurers to the “chiral switch” may explain, in part, why spending on enantiomers, at least in Medicaid, appears to have slowed for single-enantiomer drugs approved in recent years (Table 3). For another solution, policy makers might look to India, where intellectual property laws prevent patents from regularly being issued on secondary formulations of previously discovered pharmaceutical products that lack substantial clinical benefits.

In the United States, such a change would require an amendment to the patent law, which does not distinguish between pharmaceutical formulations and other products.

Rapid uptake after approval of single-enantiomer drugs by patients and physicians represents a high financial burden on both the US healthcare system and individuals, without an obvious concomitant gain in patient health. More well-designed and widely disseminated comparative trials are needed to clarify, soon after approval, whether a new single enantiomer drug represents a measurable, cost-effective improvement over current racemic therapy. Large healthcare organizations that bear the brunt of these costs should use their position to educate prescribers about appropriate use of single-enantiomer and racemic products to slow the adoption of drugs with no real advantage over cheaper alternatives.

This analysis of single-enantiomer drugs approved since 2001 used publicly available FDA reports, some of which included redacted text. However, such redactions are usually not related to approved-drug indications, so the likelihood that we missed important approval information is not great. We only examined labels at the time of approval, and so did not formally address comparisons between single-enantiomer drugs and their racemic counterparts that may have occurred after approval. Finally, we acknowledge that the FDA does not have legal authority to require comparative efficacy studies for drugs prior to approval and we do not conclude that the FDA erred in approving any of these drugs; there could, in fact, be efficacy or tolerability benefit for individual patients from single-enantiomer drugs that cannot be discerned from population-based randomized trial.

 Nonetheless, our goal was to identify the prevalence of chiral switches and examine whether any comparative effectiveness studies were submitted as part of the approval process.

 Center for Health Equity Research and Promotion, VA Pittsburgh Healthcare System, Pittsburgh, PA (WFG, MM, CBG); Department of Medicine, University of Pittsburgh (WFG, PC, CBG); RAND Health, Santa Monica, CA (WFG); Program on Regulation, Therapeutics, and Law (PORTAL), Division of Pharmacoepidemiology and Pharmacoeconomics, Department of Medicine, Brigham and Women’s Hospital, Boston, MA (ASK); Harvard Medical School, Boston, MA (ASK).

 Dr Gellad is supported by a VA Health Services Research & Development (HSR&D) career development award (CDA 09-207). Dr Kesselheim is supported by the Harvard Program in Therapeutic Science, Greenwall Faculty Scholarship in Bioethics, a career development award from the Agency for Healthcare Research & Quality (K08HS18465-01) and a Robert Wood Johnson Foundation Investigator Award in Health Policy Research. None of the funders played any role in the preparation of this article.

 From 2011 to 2013, Dr Kesselheim acted as an expert witness on behalf of a class of individual plaintiffs in a case against AstraZeneca alleging inappropriate marketing of Nexium. The other authors (WFG, PC, MM, CBG) report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article. This work represents the opinions of the authors alone and does not necessarily represent the views of the Department of Veterans Affairs or the United States government.

 Concept and design (WFG, MM, CBG, ASK); acquisition of data (WFG, PC, MM, ASK); analysis and interpretation of data (WFG, PC, CBG, ASK); drafting of the manuscript (WFG, PC, MM, ASK); critical revision of the manuscript for important intellectual content (WFG, PC, MM, CBG, ASK); statistical analysis (WFG); provision of study materials or patients (WFG); obtaining funding (WFG, ASK); administrative, technical, or logistic support (WFG); supervision (WFG, CBG, ASK).

From 2001 to 2011, the US Food and Drug Administration approved 9 single-enantiomer drugs with racemic precursors. None showed pre-approval evidence of improved efficacy over the racemic precursor.

Aaron S. Kesselheim, MD, JD, MPH

Articles

CAAP Rule and Prescription Drug Prices

Preferences for and Experiences With Pill Appearance Changes: National Surveys of Patients and Pharmacists

Assessing the Chiral Switch: Approval and Use of Single-Enantiomer Drugs, 2001 to 2011