Patient and physician demand for antibiotics
Download this section
Chapter 3

David Howard

Interest in understanding patient and physician demand for antibiotics is rooted in the perception that antibiotics are overused, especially in outpatient settings. In the medical literature, “overuse,” or “inappropriate use,” refers to situations where patients receive antibiotics for conditions that are mostly due to viral pathogens (which are unresponsive to antibiotic treatment) or tend to clear up quickly even if left untreated. In both cases, patients’ benefit from antibiotic treatment compared with watchful waiting is minimal or nonexistent. Overuse also occurs if a patient unnecessarily receives a broad-spectrum antibiotic—one that is active against multiple pathogens and typically more effective in treating resistant strains than older, narrow-spectrum drugs (e.g., amoxicillin).

In the economics literature, “overuse” is defined in terms of benefits and costs. The benefit of antibiotic consumption to patients is faster resolution of symptoms and cure. The benefit to society, which is often downplayed in the medical literature on resistance, is reduced transmission of infectious diseases. The cost of consumption, aside from the price of the drug itself, is borne by future generations, who because of resistance will have fewer
options for treating infectious diseases. When costs to society exceed benefits to society, an antibiotic is overused. The economic standard for overuse will tend to be more restr ictive (i.e., less likely to categorize consumption as inappropriate use) when the costs of antibiotic consumption are large or when consumption
reduces transmission. Unfortunately, the empirical relationships between antibiotic use and the benefits and costs to society are not well understood.

Patient demand

The benefit to patients of using antibiotics includes a faster recovery time and possibly avoidance of severe complications and even death. Costs include the financial cost of the prescription, the risk of allergic reaction, and the development of primary resistance (patients who take antibiotics frequently require more powerful drugs, possibly raising future costs). The presumption in the medical literature is that patients are not well informed about costs and benefits. Most do not understand the difference between viral and bacterial infections (Vanden Eng, Marcus et al. 2003); education-based patient interventions have reduced antibiotic consumption, indicating that lack of knowledge contributes to overuse.

Prices

The standard economic prescription for addressing negative consumption externalities (in this case, antibiotic resistance) is to make the consumer bear the costs of the externality by using a tax (or quotas or permits) to reduce consumption. However, given the availability and wide use of health insurance, it is unclear whether an “antibiotic tax” would have the desired effect. In most cases, consumers pay only a fixed cost per prescription (usually between $10 and $20), and so the health plan would be directly responsible for paying the tax. The impact of the tax would depend on health plans’ responses. If plans pass along some of the tax to consumers in the form of higher copayments, or if plans take additional steps to discourage physicians from prescribing antibiotics, then a tax will reduce consumption. In general, though, the case for using a tax to reduce antibiotic consumption is not as clear-cut as it is in settings where consumer and payer are one and the same.

In place of a tax, government might consider encouraging health plans to increase patients’ copayments for antibiotics (Laxminarayan 2003). In recent years there has been a renewed interest in using cost sharing, in the form of “tiered formularies,” to influence prescribing behavior and drug consumption. Antibiotics have not been a focus of these efforts, however. Instead, insurers have targeted medications consumed on an ongoing basis, like antihypertensives, and tried to steer patients from brand-name to generic versions. Studies of tiered copayments find that they affect overall consumption and drug choice (Leibowitz, Manning et al. 1985; Hillman, Pauly et al. 1999; Tamblyn, Laprise et al. 2001). Huskamp, Epstein et al. (2003) found that spending and use of certain classes of drugs changed far more when an employer switched from a one-tier to a three-tier formulary with across-the-board increases in copayments than under a more moderate switch, from a two-tier to a three-tier plan with no increase in payments in the first or second tier. Huskamp, Frank et al. (2005) analyzed the change in use of common medications when a health plan changed from a two-tier to a three-tier formulary and found that overall use did not change but that patients were more likely to purchase drugs from the less expensive tier.

The only published study to have evaluated the impact of cost sharing on antibiotic use is the RAND Health Insurance Experiment, a randomized controlled trial of cost sharing in health care conducted between 1974 and 1982 (Foxman, Valdez et al. 1987). Consumers in the free care plan, where all medical care expenses were covered by insurance, used 85 percent more antibiotics than consumers in plans that required consumers to pay a portion of their medical care bills. Because cost-sharing requirements were applied to all types of services, it was difficult to isolate the impact of cost sharing for antibiotics from the impact of cost sharing for complementary services like physician office visits.

Cost sharing did not appear to differentially reduce antibiotic prescriptions for conditions that were primarily viral, indicating that cost sharing reduced both “appropriate” and “inappropriate” consumption. This finding suggests that cost sharing is a fairly blunt tool for reducing overuse. In theory, health plans could vary cost-sharing amounts based on patients’ diagnoses and the appropriateness of the prescription, but this would probably be very difficult to implement in practice.

Given current price levels for commonly used antibiotics, the potential for increased copayments for antibiotics to reduce demand is limited. Many commonly prescribed antibiotics, such as amoxicillin, cost less than $20 per course, and so at current copayment levels consumers are already paying a large share of the price, if not the entire amount, out of pocket. For some offpatent antibiotics, patients’ copayments may actually exceed the
wholesale price of the drug.

Cost sharing may be an effective tool to induce consumers to switch from newer, more expensive antibiotics to older drugs (if that were a goal), but evidence on the magnitude of the effect is lacking. Clinical guidelines frequently recommend that broadspectrum drugs be held in reserve, though this policy diminishes incentives for research and development of new antibiotics and may even contribute to the development of antibiotic resistance by loading selection pressure on a handful of older drugs (see Chapter 7 for a detailed discussion).

Health plans that require higher cost sharing for both drugs and office visits, such as health care savings accounts, may be more successful in reducing demand for antibiotics. Such plans, which are marketed as “consumer driven,” have gained in popularity in recent years. The RAND study concluded that “the effect of cost sharing on antibiotic use comes principally through a reduction in visits rather than as a result of reduced antibiotic prescribing given a visit” (Foxman, Valdez et al. 1987).

Dispensing restrictions

Cost sharing is not the only way to raise the “price” of antibiotics to consumers. Policies that make it more difficult for patients to obtain antibiotics—for example, by prohibiting physicians from directly dispensing antibiotics—may reduce consumption even among patients who receive prescriptions. A recent study found that antibiotic prescribing for viral illness in Korea declined by eight percentage points after a 2000 law was passed prohibiting physician dispensing (Park, Soumerai et al. 2005). Physician dispensing of antibiotics is not as widespread in the United States as it is in Asia, and so banning U.S. physician dispensing may have only a small impact. Another policy is to give physicians the option of issuing antibiotic prescriptions with waiting periods: basically, the prescription would entitle patients to receive antibiotics a certain number of days or hours following the physician office visit. This approach permits physicians to acknowledge patients’ symptoms and distress via the prescription but imposes an additional barrier between the receipt of the prescription and the dispensing of the antibiotic. It is consistent with sound clinical practice—patients whose symptoms persist will fill the prescription, while patients whose symptoms resolve probably did not need the antibiotic in the first place—and does not impose additional costs on the health care system, since patients are not required to return for a follow-up visit once the waiting period is over.

Promoting substitutes

Promotion of substitutes to antibiotics is another strategy for reducing use, one that may be more politically palatable than policies that make it more difficult or costly for patients to obtain antibiotics. The economic rationale for promoting substitutes is obvious, but it is also worth noting that this approach to reducing antibiotic demand is consistent with anthropological theories of overprescribing. These emphasize the role of prescriptions in signaling the end to a physician encounter. Once patients have gone to the trouble of visiting a physician or emergency room, they may feel entitled to an antibiotic prescription (Macfarlane, Holmes et al. 1997; Kumar, Little et al. 2003).Use of substitutes affords physicians the opportunity to “do something” about patients’ complaints and acknowledge their validity without having to prescribe an antibiotic that they know will be ineffective.

Government and health plans could increase the use of substitutes by decreasing their price to consumers. For example, a health plan could dispense free or heavily discounted “cold kits” to physicians, who can give them to patients with respiratory symptoms as a substitute for an antibiotic prescription. The kits might contain decongestants, nasal sprays, and cough drops. At least one health plan attempted this strategy but found it costly, time-consuming,
and of limited effectiveness in reducing antibiotic use. Many of the kits were taken by persons outside the target group (e.g., receptionists). Currently, Pfizer Consumer Health Care, manufacturer of Sudafed, distributes cold kits to physicians over the Internet (see http://www.aware.md/ed_materials/cold_kits.asp).

Policies that increase consumers’ access to substitutes may help patients avoid a visit to the doctor’s office. For example, the Food and Drug Administration could shift some products from prescription to over-the-counter status, making it possible for consumers to obtain antibiotic substitutes without a physician visit.

States may also want to reconsider policy changes to limit access to over-the-counter decongestants that contain pseudoephedrine, a crucial ingredient in “crystal meth.” In some cases pharmacies are voluntarily removing or are required by state law to remove decongestants with pseudoephedrine from shelves to behind the pharmacy counter. A pending federal law would require purchasers to sign a log book. Though requiring patients to ask pharmacists for decongestants does not seem to present a significant barrier to legitimate users, many manufacturers are nevertheless reformulating products, substituting phenylephrine for pseudoephedrine. Phenylephrine must be taken more frequently than pseudoephedrine and it may not be as effective (Johannes 2005).

The lack of effectiveness of alternatives to antibiotics is a major obstacle to promoting substitutes. Evidence that decongestants and other over-the-counter remedies relieve cold symptoms is limited. Decongestants and antihistamines have been shown to be ineffective in relieving symptoms from acute otitis media in children (Flynn, Griffin et al. 2004). A review of more than 35 published studies found no evidence to support the use of antihistamines alone as a remedy for the common cold, and although in combination with decongestants they displayed a beneficial effect in a majority of the studies, it was unclear whether the results
were clinically significant (Sutter, Lemiengre et al. 2003). Arroll (2005) reviewed several studies on nonantibiotic cold remedies—antihistamines, decongestants, Echinacea, humidification, and others—and found that only one, rest, was at all useful; he also found that antibiotics have no effect on the severity of the common cold. Paul, Yoder et al. (2004) also found little evidence that over-the-counter cough medicine improves symptoms or sleep quality among children with coughs.

Symptoms of respiratory infections may also be alleviated with nonpharmacological treatments, such as nasal irrigation. Although clinical trials have indicated that irrigation may be successful in treating chronic sinusitis (Tomooka, Murphy et al. 2000; Rabago, Zgierska et al. 2002), its efficacy in treating the common cold or acute bacterial rhinosinusitis is unclear (Adam, Stiffman et al. 1998; Rabago, Barrett et al. 2005).

Patient education

Previous efforts to reduce patient demand have tended to emphasize the role of information and education. If patients overestimate the benefit of antibiotics, then providing information about the effectiveness of antibiotics for specific conditions and symptoms may reduce demand. Because patient educational interventions are usually coupled with interventions directed at physicians, it is difficult to isolate the impact of patient education. A recent review of the literature found that studies of combined patient and physician interventions did not report larger effect sizes than studies of physician interventions alone (Ranji, Steinman et al. 2006). (We review this literature in more depth in the next section.)

A handful of studies have examined the impact of patient education alone. Macfarlane, Holmes et al. (2002) randomized patients with acute bronchitis to receive usual care or literature explaining the possible negative consequences of taking antibiotics and an oral message that in most cases antibiotics would do little good. This intervention reduced antibiotic use by 25 percent. Gonzales, Steiner et al. (1999) report that the addition of a patient education component to an ongoing physician education campaign was associated with a slight decline in prescribing rates in pediatric practices but a large reduction in prescribing rates for adult patients. Taylor, Kwan-Gett et al. (2003) found that educational materials distributed to parents on the judicious use of antibiotics had no effect on prescribing rates, though these researchers did not measure actual prescriptions filled or consumption.

Prevention and vaccination

In inpatient settings, strategies to address the spread of resistance focus primarily on reducing the incidence and transmission of disease through hygiene and patient isolation. Prevention has been less of a focus in the community—there are so many more conduits for transmission—but it is the most direct method for reducing antibiotic demand, and unlike policies to reduce use of antibiotics, it does not pose the political and ethical challenges of withholding potentially effective medical care from sick patients. Schools and daycare clinics often require that sick children be kept home, and some employers also require that sick persons stay home. The effect on antibiotic use is unknown. On the one hand, it may reduce transmission. On the other, it may increase demand for antibiotics among people eager to return to work or see their children return to school.

Vaccination with pneumococcal conjugate vaccine and influenza vaccine has also been promoted as a policy to reduce disease incidence, antibiotic demand, and resistance. Predicting the impact of long-term mass vaccination is difficult, but findings from clinical trials (Jacobs 2002) and epidemiological surveillance studies (Whitney, Farley et al. 2003) suggest that pneumococcal vaccination is effective in reducing incidence of pneumococcal disease and incidence of infection with resistant strains (Klugman 2004; Talbot, Poehling et al. 2004; Poehling, Talbot et al. 2006); the vaccine is differentially active against drug-resistant serotypes. In addition, there appears to be a spillover benefit from vaccination to nonvaccinated hosts (Whitney, Farley et al. 2003).

Vaccination creates opportunities for transmission of and infection with nonpneumococcal pathogens and nonvaccine serotypes of Streptococcus pneumoniae (Kyaw, Lynfield et al. 2006). Three trials of pneumococcal vaccination found that rates of acute otitis media did not differ or were only slightly lower in the treatment group compared with controls (Jacobs 2002), but it is conceivable that widespread vaccination could lower incidence rates over a longer timeframe. Many studies of the impact of influenza vaccination in children have found that vaccination reduces the risk of otitis media (see, e.g., Belshe, Mendelman et al. 1998), but not all studies have reached this conclusion (Hoberman, Greenberg et al. 2003).

If vaccination is shown to be an effective strategy for reducing the incidence of infection (and, consequently, use of antibiotics), then it will be important to consider new policies for increasing vaccination uptake. The government can require that all school-age children receive pneumococcal vaccinations. In the absence of a mandate, the government may need to offer subsidies or impose coverage mandates on insurers to significantly increase uptake. Currently, pneumococcal vaccination is relatively expensive compared with other common vaccines, and not all health plans cover pneumococcal vaccination.

Physician demand

In the short term, physicians’ incentives to prescribe antibiotics will depend on the reimbursement system and the internal reward structure of the physician group. Most primary-care physicians practice in groups, and so the financial incentives for the group as a whole may differ substantially from those for individual physicians, though one would expect group managers to align group and
individual incentives as closely as possible.

Most groups are compensated using either fee-for-service reimbursement or capitation. Under fee-for-service reimbursement, groups are paid a per-visit fee. Under capitation, groups are paid a per-member per-month fee. In general, fee-for-service reimbursement is associated with stronger incentives to treat patients aggressively, while capitation presents incentives to treat patients conservatively, though the opposite may be true in situations where aggressive treatment means referring patients to specialists not covered under the capitation contract.

The impact of these alternative reimbursement arrangements on physicians to prescribe antibiotics depends on whether otics are a substitute for, or complement to, followup visits. Most likely, antibiotics and follow-up visits are substitutes; physicians instruct patients who receive antibiotics to take the full course of the drug before making another visit. Patients who do not receive antibiotics may be advised to return for a follow-up visit in a few days if symptoms persist. When antibiotics and follow-up visits are substitutes, fee-forservice reimbursement will be associated with lower rates of antibiotic prescribing, and capitation with higher rates. Evidence on the degree to which antibiotics and follow-up visits are substitutes or complements is limited. One study found that a large decrease in antibiotic prescribing rates following a physician and patient educational intervention was not associated with an increase in return visits (Gonzales, Steiner et al. 1999), but another study, in which patients were randomized to treatment (immediate antibiotics, delayed antibiotics, no antibiotics), found that 19 percent of patients in the no-antibiotic group returned for a follow-up visit, compared with 11 percent of patients in the immediateantibiotic group (Little, Rumsby et al. 2005).

Within physician groups, a portion of physicians’ pay may be tied to “productivity,” which is often measured by the number of patients seen. The medical and sociological literature on antibiotic prescribing and inappropriate prescribing in general emphasizes that physicians can use prescriptions as a tool for reducing the length of office visits. The prescription signals to the patient that the visit is concluded, and it indicates that the physician understands and recognizes the legitimacy of the patient’s complaints (Steinke, MacDonald et al. 1999). If a practice operates at full capacity, a physician who sees more patients during the day may earn a productivity-based bonus. If a practice operates at less than full capacity, physicians who finish their appointments ahead of schedule can spend the rest of their workday on administrative tasks or go home early. Studies have found that physicians with greater workloads are more likely to prescribe antibiotics (Gonzales, Steiner et al. 1997; Hutchinson and Foley 1997; Arnold, Allen et al. 1999). Explaining to patients why they do not need antibiotics can take as long as 10 to 15 minutes—valuable time to a busy physician.

Although reimbursement and bonus incentives have been found to play an important role in influencing practice patterns in a variety of settings, the medical literature on overprescribing tends to downplay short-term financial considerations. Instead, survey and focus group results emphasize the role of patient demand (Schwartz, Soumerai et al. 1989; Barden, Dowell et al. 1998; APUA 1999; Bauchner, Pelton et al. 1999; Metlay, Shea et al. 2002). Physicians fear that if they do not meet patients’ expectations, they will lose patients to other practices. Survey results must be interpreted cautiously. Responses reflect survey structure, and so a survey that asks physicians about the role of patient demand but not reimbursement may mischaracterize the nonclinical factors responsible for overprescribing. The survey that included the most comprehensive list of factors influencing prescribing (APUA 1999) found that 59 percent of respondents cited patient demand, but only 26 percent cited time pressure and 20 percent cited the potential for a return visit (respondents were allowed multiple responses). Also, such studies measure physicians’ perceptions of patient behavior, not actual behavior. We are not aware of any studies showing that patients switch physicians if they are refused antibiotics. A study that analyzed actual patient-physician interactions found that explicit demands for antibiotics by patients are rare (Stivers 2002a). Other studies have found that physicians perceive patient demand for antibiotics when there is none (Mangione-Smith, McGlynn et al. 1999) and that patient demand is more often implicit (e.g., suggesting a diagnosis) than explicit (Stivers 2002b). Nevertheless, “patient demand” is usually cited by physicians as the most important factor contributing to overuse, and there are anecdotal reports of patients’ demanding specific, broad-spectrum antibiotics.

The use of “patient satisfaction” tools may reinforce the incentive to please patients by prescribing antibiotics. Some health plans use patient satisfaction measures in contracting decisions or to determine bonus payments. Other plans release satisfaction scores directly to patients, who may use the information when selecting physicians. Little, Rumsby et al. (2005) found that when patients with lower respiratory tract infections were randomized to receive an immediate antibiotic, a delayed antibiotic, or no antibiotic, the proportion of patients who were very satisfied was 14 percent higher in the immediate-antibiotic group than in the no-antibiotic group. A study by Christakis, Wright et al. (2005) reported that parents were more satisfied when their children received antibiotics for cough and cold symptoms, and a survey of parents found that 40 percent would be somewhat or extremely dissatisfied if watchful waiting was recommended (Finkelstein, Stille et al. 2005). Other work suggests that the time spent with the physician (Ranji, Steinman et al. 2006) or physician-patient communication (Mangione-Smith, McGlynn et al. 1999), rather than receipt of an antibiotic, is a more important determinant of patient satisfaction. Gonzales, Steiner et al. (2001) found that a combined patient and physician education effort lowered antibiotic consumption but did not adversely affect patients’ satisfaction with care. Fear of malpractice lawsuits is sometimes cited as a factor contributing to overprescribing. A child who suffers from meningitis that could have been averted with an antibiotic prescription would make a sympathetic victim. However, there are no documented cases in the medical literature of a physician being sued for failing to prescribe an antibiotic, and a board member of a malpractice insurer reported that he was unaware of such cases (Kenneth E. Thorpe, personal communication, May 3, 2006). In surveys, physicians do not rate liability concerns highly compared with other nonclinical factors driving inappropriate antibiotic use APUA 1999; Bauchner, Pelton et al. 1999). The prospect of lawsuits is, however, an emerging issue in inpatient settings (see Chapter
4), where lax hand washing and poor hygiene contribute to the spread of both resistant and sensitive bacteria.

Pay for performance

Over the past several years, many health plans have started to use bonuses tied to quality measures, or “pay for performance” programs, to influence physician behavior. Quality measures are typically based on patient satisfaction surveys, measures of physicians’ adherence to medical practice guidelines, or even patients’ outcomes. Antibiotic prescribing is not a focus of these efforts, which instead emphasize cancer screening and quality and process measures for chronic conditions like diabetes. Although there is a great deal of enthusiasm among payers, employers, and policymakers for pay-for-performance programs, it can be surprisingly difficult to design effective incentive schemes. Rosenthal, Frank et al. (2005) studied a pay-for-performance program in a California health plan sician groups based on the level of cervical cancer screening, breast cancer screening, and hemoglobin A1c testing. They found that the program basically rewarded physician groups whose levels were high to begin with, and improvements in quality were apparent for only one of three measures. Pay for performance can also lead physicians to avoid high-risk or noncompliant patients or overuse certain technologies. For example, programs that measure quality based on cancer screening rates may lead physicians to recommend screening for older, sicker persons who are unlikely to benefit from early detection (Walter, Davidowitz et al. 2004).

Quality bonuses could potentially be used to influence antibiotic prescribing rates. It would be difficult, however, to differentiate between appropriate and inappropriate prescribing in such a way as to discourage miscoding and misclassification. Diagnoses for acute respiratory conditions are somewhat interchangeable, so it is simple for physicians to substitute an “appropriate” ICD-9 code for an “inappropriate” code when bonus funds are at stake. In some cases the appropriateness of prescribing an antibiotic depends on symptom duration, which is not recorded in insurers’
administrative claims files.

Even if patients’ complaints were documented accurately in administrative data, it is unclear how to link bonuses to physician behavior. Should bonuses be tied to absolute prescribing levels or changes in prescribing patterns? Should bonuses be tied to the number of prescriptions or prescribing rates (i.e., the number of prescriptions divided by the number of patients seeking treatment for colds, ear infections, etc.)? Each option presents problems. For example, tying payments to prescribing rates may lead physicians to encourage patients with minor symptoms to seek care, thereby inflating the denominator of the prescribing rate ratio.

Electronic medical records

Continued use of an antibiotic or use of antibiotics from a given class increases the risk that patients will develop primary resistance. Prescribing guidelines, such as the Infectious Diseases Society of America’s (IOSA) guideline for pneumonia (Bartlett, Breiman et al. 1998), recommend that physicians select an antibiotic from a different class for patients who recently received antibiotics. For example, if a patient received a macrolide previously, the physician should prescribe a penicillin or cephalosporin for the next prescription. Analysis of prescribing patterns indicates that the recommendation to rotate or switch antibiotics is often not followed in practice (Wu, Howard et al. 2006). One of the difficulties in adhering to the guideline is that physicians may be unaware of the types of antibiotics that patients have been prescribed previously. Electronic medical records, which follow patients across settings, may help improve prescribing patterns by allowing physicians to access past medications. Electronic records could be programmed with electronic flags to alert physicians when they do not prescribe guidelinerecommended therapy. Also, patients who use a large number of antibiotics over a certain period could be targeted for special educational interventions or even subjected to higher cost-sharing amounts.

From a societal standpoint, electronic records would be useful for surveillance and for implementing mixed prescribing strategies (Laxminarayan and Weitzman 2002), where many antibiotics are used simultaneously to avoid excessive selection pressure on any particular antibiotic or antibiotic class. Although it is not clear how a mixed strategy could be enforced across settings or payers, large clinics or health plans could use electronic record systems to encourage and monitor adherence with mixed prescribing for their patients and enrollees. For example, a clinic could add a prompt to electronic medical records that randomly assigns penicillinbased antibiotics for some patients and macrolides for others with a given condition, taking into account patients’ allergies and recent antibiotic usage.

Managing drug company–physician interactions

Pharmaceutical company marketing, which encompasses print advertising and physician detailing, has both brandexpanding and business-stealing effects. For antibiotics, which are a mature product, we can assume that marketing has been primarily a business-stealing effect. There is little information on the magnitude of companies’ marketing efforts on behalf of antibiotics. A casual inspection of pediatric and family medicine journals indicates that antibiotics are advertised and that many of the advertisements are for broad-spectrum drugs. Many advertisements make reference to the drug’s activity against resistant strains. Companies also promote antibiotics with physician detailing (i.e., interactions between companies’ sales forces and physicians) and by distributing samples.

To the extent that marketing leads to overuse of antibiotics and broad-spectrum drugs in particular, policies to counteract the impact of advertising and detailing may slow the development of resistant strains. “Academic detailing” (or alternatively, “counter detailing”) is generally considered the most effective educational method of influencing prescribing patterns (Gross and Pujat 2001; Bloom 2005). In academic detailing, a representative from a neutral organization, such as a professional society, hospital, or public health agency, meets with physicians one-on-one to discuss prescribing practices (Soumerai and Avorn 1990). Typically, the detailer is a peer or someone knowledgeable in the clinical area being discussed and is trained in effective communication. Academic detailing has been used successfully to reduce overuse of several types of drug classes (Soumerai and Avorn 1990; Bloom 2005), including antibiotics (Gross and Pujat 2001; Solomon, Van Houten et al. 2001).

Interventions to improve diagnostic accuracy

Though most physicians recognize that antibiotics are overprescribed in aggregate, it is not easy to determine the appropriateness of antibiotics for individual patients. Few 0physicians can rule out bacterial infection with certainty based on the often nonspecific complaints of patients suffering from respiratory conditions. There is always a chance that antibiotics will help, and in rare cases failure to prescribe antibiotics may lead to serious complications, like meningitis.

Development of decision rules and diagnostic tests that enable physicians to determine the etiology of infections and symptoms can reduce use of antibiotics and broad-spectrum drugs. Decision rules employ easily observable patient characteristics to predict the response to antibiotic therapy. Though most decision rules have fairly low sensitivity and specificity, they are inexpensive and can be easily incorporated into clinical practice. Diagnostic tests are more accurate but are also more expensive and invasive. Several promising tests are under development (Christ-Crain, Jaccard-Stolz et al. 2004; Esposito, Tremolati et al. 2005), but most currently available tests are too slow or entail too much discomfort to patients to be of much practical use in office settings.

Physician education and quality improvement

Most ongoing efforts to reduce antibiotic prescribing use various types of educational interventions and “quality improvement” strategies targeted at physicians and patients. Educational materials typically describe situations where antibiotic prescribing is and is not appropriate and attempt to raise physicians’ awareness of the problem of resistance. Strategies to disseminate information include educational mailings to physicians, educational pamphlets for patients, and academic detailing (discussed above).

Since 1995 the Centers for Disease Control and Prevention has funded a national educational campaign to reduce overuse (see http://www.cdc.gov/drugresistance/community/). Originally named the National Campaign for Appropriate Antibiotic Use in the Community and renamed Get Smart, this program focuses on changing prescribing patterns for patients with upper respiratory infections and targets  providers, patients, and parents. National trends in prescribing rates and antibiotic use suggest that it and similar efforts to raise awareness about the problem of resistance have been effective. From 1980 to 1992, antibiotic use rates in the population increased (though the increases were not statistically significant at the 5 percent level) (McCaig and Hughes 1995). In more recent years, however, studies have documented substantial declines in antibiotic use (McCaig, Besser et al. 2002; Finkelstein, Stille et al. 2003; Steinman, Gonzales et al. 2003; Stille, Andrade et al. 2004; Roumie, Halasa et al. 2005; Miller and Hudson 2006). For example, McCaig, Besser et al. (2002) report that the number of antibiotics prescribed annually in physicians’ offices to children under age 15 declined from 45.5 million in 1989–1990 to 30.3 million in 1999–2000. Prescribing rates declined from 838 per 1,000 children to 503 per 1,000 children. Likewise, Steinman, Gonzales et al. (2003) report that the proportion of outpatient visits where the patient received an antibiotic decreased from 13 percent to 10 percent between 1991–1992 and 1989–1999 for adults and from 33 percent to 22 percent for children. The total number of antibiotic prescriptions decreased from 230 million to 190 million. While antibiotic use has declined, however, the market share of broad-spectrum drugs has been increasing (Steinman, Gonzales et al. 2003; Stille, Andrade et al. 2004). Howard (2005) argues that concerns about resistance are partly responsible for the shift toward broad-spectrum drugs. Thus, efforts to publicize the impact of overuse on patient outcomes will decrease antibiotic use overall but may have the unintended consequence of increasing consumption of broad-spectrum antibiotics.

Programs to reduce antibiotic use may be effective in reducing rates of resistance among the most common respiratory pathogen, Streptococcus pneumoniae. Recent surveillance data indicate that although resistance to fluoroquinolones is becoming more common in S. pneumoniae, resistance to other drug classes is decreasing (Doern, Richter et al. 2005). This pattern is consistent with trends in prescribing patterns. A French study reported that region-level interventions to reduce antibiotic use were associated with reductions in the rate of colonization with penicillin-resistant S. pneumoniae (Guillemot, Varon et al. 2005).

Controlled studies of educational interventions provide additional evidence on the effectiveness of education and quality improvement. Unlike studies of antibiotic use trends, they control for other factors that may affect antibiotic use over time (e.g., publicity surrounding the “hygiene hypothesis,” which posits that lack of early exposure to pathogens leads to immune-related disorders later in life). A recent summary of these studies concluded that educational interventions, which are often accompanied by feedback, reduced prescribing rates by about 9 percent and increased adherence to guidelines on antibiotic choice by more than 10 percent (Ranji, Steinman et al. 2006). To put these results in perspective, the review noted, “For a 100,000-member health maintenance organization (HMO), these studies suggest that a QI [quality improvement] strategy targeting all ARIs [antibiotic-resistant infections] for patients of all age groups would result in a savings of approximately 3000 to 8000 antibiotic prescriptions per year.”

Studies that combined profiling and education did not report better results than studies of education alone. Educational interventions that employed active learning strategies, like academic detailing, led to larger reductions than interventions that relied on passive learning strategies, like mailings—a finding echoed in earlier reviews (Gross and Pujat 2001; Sbarbaro 2001). As mentioned above, studies with a patient education component did not report better results than studies that examined only physician education.

The median duration of studies included in the Ranji, Steinman et al. review (2006) was only six months, raising concerns about the durability of the interventions. At the population level, the reductions in antibiotic use noted earlier have been sustained, suggesting that even if the effect of a single, small-scale intervention is reversible, the cumulative effect of multiple interventions and nationwide education efforts is not. Based on the published literature to date, it is difficult to determine whether interventions to reduce antibiotic use are cost-effective.

Going forward, it is unclear whether additional educational programs can further reduce antibiotic prescribing rates. Gonzales, Corbett et al. (2005) studied the impact of a recent educational initiative in Colorado. They concluded that although the program reduced antibiotic prescribing rates for adult patients, “there appears to be little room for improvement in antibiotic prescription rates for children with pharyngitis.” A study of the same intervention, limited to elderly enrollees in a Medicare managed care plan, found that it did not affect prescribing rates in this population. (Gonzales, Sauaia et al. 2004)

Conclusion and issues for future research

Policies to reduce antibiotic use in outpatient settings can appeal to extrinsic motivations (i.e., self-interest), intrinsic motivations (i.e., wanting to do good for its own sake), or a combination of both. Policies that appeal to extrinsic motivation include cost sharing and pay for performance. Policies that appeal to intrinsic motivations include education and feedback. Historically, policies to reduce antibiotic use and improve health care quality more generally have relied on intrinsic motivation based on the theory that physicians want to do the right thing—they just lack information and the right tools. Along the same lines, poor quality was viewed as a manifestation of poorly designed systems, rather than the fault of specific individuals. In the past few years, attitudes toward quality improvement have shifted somewhat, and appeals to clinicians’ and patients’ extrinsic motivations are becoming more common. Policies that rely on extrinsic motivation are often powerful, but there is always the risk that appeals to pecuniary motives may crowd out intrinsic motivations. Also, policies that appeal to extrinsic motivations can lead to unintended consequences. For example, as discussed earlier, a policy to tie compensation to prescribing rates could lead physicians to encourage visits by patients with minor symptoms to make the prescribing rate appear to be lower.

Evidence from clinical trials and national trends in prescribing patterns suggests that policies designed to appeal to intrinsic motivations, such as intensive physician and patient education, have led to substantial reductions in antibiotic use. It is unclear whether additional interventions or policy changes are necessary. The decline in antibiotic use may not be sustainable, or it may be sustainable but further reductions in antibiotic use may be difficult to achieve through educational programs alone.

Price-based interventions, including changes in demand-side cost sharing by patients and the use of bonuses for physicians tied to practice patterns, are an increasingly popular method of lowering costs in primary care. Antibiotic use has not been a focus of these efforts, and aside from the RAND Health Insurance Experiment, their impact on antibiotic use is untested. Research from other clinical areas suggests that price-based interventions have an impact on physician and patient behavior, but incentives are a blunt instrument, and it can be difficult to structure incentives to bring about desired results. With antibiotics, incentive-based programs may fail to differentiate between “appropriate” and “inappropriate” use and can be subject to gaming by physicians. Interventions to make substitutes for antibiotics more readily available, perhaps by giving physicians “cold relief packs” to distribute in their offices, may lower prescribing rates without adversely affecting patient outcomes. The impact of making substitutes to antibiotics more readily available and other nonprice interventions to change the relative cost of antibiotics to patients has not been well studied.

Use of policies to reduce transmission of bacteria and viruses in the community has received little attention in the medical literature, though there is an extensive literature on infection control in inpatient settings. Perhaps clinicians perceive that opportunities for reducing transmission rates are limited, or that it is impossible to marshal public support for meaningful changes unless the disease in question is life threatening. Currently, many day-care centers require parents to keep sick children home. The net impact on antibiotic demand is unknown, since parents may seek antibiotics to speed their children’s return (and their own return to work). Vaccination presents an attractive mechanism for reducing antibiotic demand, since in most cases private and public incentives for obtaining vaccination are aligned (though not necessarily of the same magnitude). Future work should examine the potential of pneumococcal vaccination for reducing resistance rates directly, via the vaccine’s differential activity against resistant serotypes, and indirectly via the impact of the vaccine on antibiotic demand.

An important goal for policy, besides reducing the level of antibiotic use, is improving patterns of use. The medical literature emphasizes the need to hold broad-spectrum antibiotics in reserve, using them only when older drugs have failed, but the emerging economics literature on resistance finds that mixed prescribing strategies are optimal. Designing policies to improve prescribing patterns and implement mixed prescribing strategies is an important goal for future research. Electronic medical records, which allow third parties to monitor prescribing patterns across settings and physicians offices, may be helpful in this regard.

 

 

References

Adam, P., M. Stiffman, et al. (1998). “A Clinical Trial of Hypertonic Saline Nasal Spray in Subjects with the Common Cold or Rhinosinusitis.” Archives of Family Medicine 7(1): 39-43.

APUA (1999). Massachusetts Physician Survey. Pilot Survey of Primary Care Physicians in Massachusetts, 1998. http://www.tufts.edu/med/apua/Research/physicianSurvey1-01/physicianSurv... (accessed March 9, 2006). Alliance for the Prudent Use of Antibiotics.

Arnold, S. R., U. D. Allen, et al. (1999). ”Antibiotic Prescribing by Pediatricians for Respiratory Tract Infection in Children.” Clinical Infectious Diseases 29(2): 312-17.

Arroll, B. (2005). “Antibiotics for Upper Respiratory Tract Infections: An Overview of Cochrane Reviews.” Respiratory Medicine 99(3): 255-61.

Barden, L. S., S. F. Dowell, et al. (1998). “Current Attitudes Regarding Use of Antimicrobial Agents: Results from Physician’s and Parents’ Focus Group Discussions.” Clinical Pediatrics 37(11): 665-71.

Bartlett, J. G., R. F. Breiman, et al. (1998). “Community-Acquired Pneumonia in Adults: Guidelines for Management.” Clinical Infectious Diseases 26(4): 811-38.

Bauchner, H., S. I. Pelton, et al. (1999). “Parents, Physicians, and
Antibiotic Use.” Pediatrics 103(2): 395-401. Belshe, R. B., P. M. Mendelman, et al. (1998). “The Efficacy of Live Attenuated, Cold-Adapted, Trivalent, Intranasal Influenzavirus Vaccine in Children.” The New England Journal of Medicine 338(20): 1405-12.

Bloom, B. S. (2005). “Effects of Continuing Medical Education on Improving Physician Clinical Care and Patient Health: A Review of Systematic Reviews.” International Journal of Technology Assessment in Health Care 21(3): 380-5.

Christ-Crain, M., D. Jaccard-Stolz, et al. (2004). “Effect of Procalcitonin-Guided Treatment on Antibiotic Use and Outcome in Lower Respiratory Tract Infections: Cluster-Randomised, Single-Blinded Intervention Trial.” Lancet 363(9409): 600-7.

Christakis, D. A., J. A. Wright, et al. (2005). “Association Between Parental Satisfaction and Antibiotic Prescription for Children with Cough and Cold Symptoms.” The Pediatric Infectious Disease Journal 24(9): 774-7.

Doern, G. V., S. S. Richter, et al. (2005). “Antimicrobial Resistance Among Streptococcus pneumoniae in the United States: Have We Begun to Turn the Corner on Resistance to Certain Antimicrobial Classes?” Clinical Infectious Diseases 41(2):139-48.

Esposito, S., E. Tremolati, et al. (2005). “Evaluation of a Rapid Bedside Test for the Quantitative Determination of C-reactive Protein.” Clinical Chemistry and Laboratory Medicine 43(4): 438-40.

Finkelstein, J. A., C. Stille, et al. (2003). “Reduction in Antibiotic Use Among US Children, 1996-2000.” Pediatrics 112(3): 620-627.

Finkelstein, J. A., C. J. Stille, et al. (2005). “Watchful Waiting for Acute Otitis Media: Are Parents and Physicians Ready?” Pediatrics 115(6): 1466-73.

Flynn, C. A., G. H. Griffin, et al. (2004). “Decongestants and Antihistamines for Acute Otitis Media in Children.” Cochrane Database of Systematic Reviews (3): CD001727.

Foxman, B., R. B. Valdez, et al. (1987). “The Effect of Cost Sharing on the Use of Antibiotics in Ambulatory Care: Results from a Population-Based Randomized Controlled Trial.” Journal of Chronic Diseases 40(5): 429-37.

Gonzales, R., K. K. Corbett, et al. (2005). “The ‘Minimizing Antibiotic Resistance in Colorado’ Project: Impact of Patient Education in Improving Antibiotic Use in Private Office Practices.” Health Services Research 40(1): 101-116.

Gonzales, R., A. Sauaia, et al. (2004). “Antibiotic Treatment of Acute Respiratory Tract Infections in the Elderly: Effect of a Multidimensional Educational Intervention.” Journal of the American Geriatrics Society 52(1): 39-45.

Gonzales, R., J. F. Steiner, et al. (1997). “Antibiotic Prescribing for Adults with Colds, Upper Respiratory Tract Infections, and Bronchitis by Ambulatory Care Physicians.” JAMA 278(11): 901-4.

Gonzales, R., J. F. Steiner, et al. (1999). “Decreasing Antibiotic Use in Ambulatory Practice: Impact of a Multidimensional Intervention on the Treatment of Uncomplicated Acute Bronchitis in Adults.” JAMA 281(16): 1512-1519.

Gonzales, R., J. F. Steiner, et al. (2001). “Impact of Reducing Antibiotic Prescribing for Acute Bronchitis on Patient Satisfaction.” Effective Clinical Practice 4(3): 105-11.

Gross, P. A. and D. Pujat. (2001). “Implementing Practice Guidelines for Appropriate Antimicrobial Usage: A Systematic Review.” Medical Care 39(8 Suppl 2): II55-69.

Guillemot, D., E. Varon, et al. (2005). “Reduction of Antibiotic Use in the Community Reduces the Rate of Colonization with Penicillin G-Nonsusceptible Streptococcus pneumoniae.” Clinical Infectious Diseases 41(7): 930-8.

Hillman, A. L., M. V. Pauly, et al. (1999). “Financial Incentives and Drug Spending in Managed Care.” Health Affairs 18(2): 189-200.

Hoberman, A., D. P. Greenberg, et al. (2003). “Effectiveness of Inactivated Influenza Vaccine in Preventing Acute Otitis Media in Young Children: A Randomized Controlled Trial.” JAMA 290(12): 1608-16.

Howard, D. H. (2005). “Life Expectancy and the Value of Early Detection.” Journal of Health Economics 24(5): 891-906.

Huskamp, H. A., A. M. Epstein, et al. (2003). “The Impact of a National Prescription Drug Formulary on Prices, Market Share, and Spending: Lessons for Medicare?” Health Affairs 22(3): 149-158.

Huskamp, H. A., R. G. Frank, et al. (2005). “The Impact of a Three-Tier Formulary on Demand Response for Prescription Drugs.” Journal of Economics & Management Strategy 14(3): 729-753.

Hutchinson, J. and R. Foley. (1997). Influence of Nonmedical Factors on Antibiotic Prescription Rates. Paper presented at the American Society of Microbiology (ASM ) 37th Interscience Conference on Antimicrobial Agents and Chemotherapy. September 1997, Toronto, CA.

Jacobs, M. R. (2002). “Prevention of Otitis Media: Role of Pneumococcal Conjugate Vaccines in Reducing Incidence and Antibiotic Resistance.” The Journal of Pediatrics 141(2): 287-293.

Johannes, L. (2005). “Choosing a Pill for That Cold.” The Wall Street Journal, Dec. 27, D4.

Klugman, K. P. (2004). “Vaccination: A Novel Approach to Reduce Antibiotic Resistance.” Clinical Infectious Diseases 39: 649-51.

Kumar, S., P. Little, et al. (2003). “Why Do General Practitioners Prescribe Antibiotics for Sore Throat? Grounded Theory Interview Study.” BMJ 326(7381): 138.

Kyaw, M.H., R. Lynfield, et al. (2006). “Effect of Introduction of the Pneumococcal Conjugate Vaccine on Drug-Resistant Streptococcus pneumoniae.” The New England Journal of Medicine 354(14): 1455-63.

Laxminarayan, R. (2003). Economic Responses to the Problem of Drug Resistance. The Resistance Phenomenon in Microbes and Infectious Disease Vectors: Implications for Human Health and Strategies for Containment—Workshop Summary. Knobler, S. L., S. M. Lemon, et al. (eds). Washington, DC: National Academies Press, 121-129.

Laxminarayan, R. and M. L. Weitzman. (2002). “On the Implications of Endogenous Resistance to Medications.” Journal of Health Economics 21(4): 709-718.

Leibowitz, A., W. G. Manning, et al. (1985). “The Demand for Prescription Drugs as a Function of Cost-Sharing.” Social Science & Medicine 21(10): 251-277.

Little, P., K. Rumsby, et al. (2005). “Information Leaflet and Antibiotic Prescribing Strategies for Acute Lower Respiratory Tract Infection: A Randomized Controlled Trial.” JAMA 293(24): 3029-35.

Macfarlane, J., W. Holmes, et al. (1997). “Influence of Patients’ Expectations on Antibiotic Management of Acute Lower Respiratory Tract Illness in General Practice: Questionnaire Study.” BMJ 315(7117): 1211-4.

Macfarlane, J., W. Holmes, et al. (2002). “Reducing Antibiotic Use for Acute Bronchitis in Primary Care: Blinded, Randomised Controlled Trial of Patient Information Leaflet.” BMJ 324(7329): 91-94.

Mangione-Smith, R., E. A. McGlynn, et al. (1999). “The Relationship Between Perceived Parental Expectations and Pediatrician Antimicrobial Prescribing Behavior.” Pediatrics 103(4 Pt 1): 711-8.

McCaig, L. F., R. E. Besser, et al. (2002). “Trends in Antimicrobial Prescribing Rates for Children and Adolescents.” JAMA 287(23): 3096-102.

McCaig, L. F. and J. M. Hughes. (1995). “Trends in Antimicrobial Drug Prescribing Among Office-Based Physicians in the United States.” JAMA 273(3): 214-9.

Metlay, J. P., J. A. Shea, et al. (2002). “Tensions in Antibiotic Prescribing: Pitting Social Concerns Against the Interests of Individual Patients.” Journal of General Internal Medicine 17(2): 87-94.

Miller, G. E. and J. Hudson. (2006). “Children and Antibiotics: Analysis of Reduced Use, 1996-2001.” Medical Care 44(5 Suppl): I36-I44.

Park, S., S. B. Soumerai, et al. (2005). “Antibiotic Use Following a Korean National Policy to Prohibit Medication Dispensing by Physicians.” Health Policy and Planning 20(5): 302-9.

Paul, I. M., K. E. Yoder, et al. (2004). “Effect of Dextromethorphan, Diphenhydramine, and Placebo on Nocturnal Cough and Sleep Quality for Coughing Children and Their Parents.” Pediatrics 114(1): e85-90.

Poehling, K. A., T. R. Talbot, et al. (2006). “Invasive Pneumococcal Disease Among Infants Before and After Introduction of Pneumococcal Conjugate Vaccine.” JAMA 295(14): 1668-74.

Rabago, D., B. Barrett, et al. (2005). “Nasal Irrigation to Treat Acute Bacterial Rhinosinusitis.” American Family Physician 72(9).

Rabago, D., A. Zgierska, et al. (2002). “Efficacy of Daily Hypertonic Saline Nasal Irrigation Among Patients with Sinusitis: A Randomized Controlled Trial.” The Journal of Family Practice 51(12): 1049-55.

Ranji, S., M. Steinman, et al. (2006). Antibiotic Prescribing Behavior. Rockville, MD: Agency for Health care Research and Quality.

Rosenthal, M. B., R. G. Frank, et al. (2005). “Early Experience with Pay-For-Performance: From Concept to Practice.” JAMA 294(14): 1788-93.

Roumie, C. L., N. B. Halasa, et al. (2005). “Trends in Antibiotic Prescribing for Adults in the United States—1995 to 2002.” Journal of General Internal Medicine 20(8): 697-702.

Sbarbaro, J. A. (2001). “Can We Influence Prescribing Patterns?” Clinical Infectious Diseases 33 (Suppl 3): S240-4.

Schwartz, R. K., S. B. Soumerai, et al. (1989). “Physician Motivations for Nonscientific Drug Prescribing.” Social Science & Medicine 28(6): 577-82.

Solomon, D. H., L. Van Houten, et al. (2001). “Academic Detailing to Improve Use of Broad-Spectrum Antibiotics at an Academic Medical Center.” Archives of Internal Medicine 161(15): 1897-902.

Soumerai, S. B. and J. Avorn. (1990). “Principles of Educational
Outreach (‘Academic Detailing’) to Improve Clinical Decision Making.” JAMA 263(4): 549-56.

Steinke, D. T., T. M. MacDonald, et al. (1999). “The Doctor-Patient Relationship and Prescribing Patterns: A View from Primary Care.” Pharmacoeconomics 16(6): 599-603.

Steinman, M. A., R. Gonzales, et al. (2003). “Changing Use of Antibiotics in Community-Based Outpatient Practice, 1991-1999.” Annals of Internal Medicine 138(7): 525-33.

Stille, C. J., S. E. Andrade, et al. (2004). ”Increased Use of Second-Generation Macrolide Antibiotics for Children in Nine Health Plans in the United States.” Pediatrics 114(5): 1206-11.

Stivers, T. (2002a). “Participating in Decisions about Treatment: Overt Parent Pressure for Antibiotic Medication in Pediatric Encounters.” Social Science & Medicine 54(7): 1111-30.

Stivers, T. (2002b). “Presenting the Problem in Pediatric Encounters: ‘Symptoms Only’ versus ‘Candidate Diagnosis’ Presentations.” Health Communication 14(3): 299-338.

Sutter, A. I., M. Lemiengre, et al. (2003). “Antihistamines for the Common Cold.” Cochrane Database of Systematic Reviews (3): CD001267.

Talbot, T. R., K. A. Poehling, et al. (2004). “Reduction in High Rates of Antibiotic-Nonsusceptible Invasive Pneumococcal Disease in Tennessee After Introduction of the Pneumococcal Conjugate Vaccine.” Clinical Infectious Diseases 39(5): 641-8.

Tamblyn, R., R. Laprise, et al. (2001). “Adverse Events Associated With Prescription Drug Cost-Sharing Among Poor and Elderly Persons.” JAMA 285(4): 421-429.

Taylor, J. A., T. S. C. Kwan-Gett, et al. (2003). ”Effectiveness of an Educational Intervention in Modifying Parental Attitudes About Antibiotic Usage in Children.” Pediatrics 111(5): e548-554.

Tomooka, L., C. Murphy, et al. (2000). “Clinical Study and Literature Review of Nasal Irrigation.” Laryngoscope 110(7): 1189-93.

Vanden Eng, J., R. Marcus, et al. (2003). “Consumer Attitudes and Use of Antibiotics.” Emerging Infectious Diseases 9(9): 1128-35.

Walter, L. C., N. P. Davidowitz, et al. (2004). “Pitfalls of Converting Practice Guidelines into Quality Measures: Lessons Learned From a VA Performance Measure.” JAMA 291(20): 2466-70.

Whitney, C. G., M. M. Farley, et al. (2003). “Decline in Invasive
Pneumococcal Disease after the Introduction of Protein-Polysaccharide Conjugate Vaccine.” The New England Journal of Medicine 348(18): 1737-46.

Wu, J., D. Howard, et al. (2006). “Adherence to Infectious Disease Society of America Guidelines on Empiric Therapy for Patients with Community Acquired Pneumonia in a Commercially Insured Cohort.” Clinical Therapeutics 28(9): 1451-61.