Principles of Educational Outreach ('Academic Detailing') to Improve Clinical Decision Making

The Harvard community has made this article openly available. Please share how this access benefits you. Your story matters

Citation Soumerai, Stephen B. 1990. “Principles of Educational Outreach (’Academic Detailing’) to Improve Clinical Decision Making.” JAMA: The Journal of the American Medical Association 263 (4) (January 26): 549. doi:10.1001/jama.1990.03440040088034.

Published Version doi:10.1001/jama.1990.03440040088034

Citable link http://nrs.harvard.edu/urn-3:HUL.InstRepos:32692212

Terms of Use This article was downloaded from Harvard University’s DASH repository, and is made available under the terms and conditions applicable to Other Posted Material, as set forth at http:// nrs.harvard.edu/urn-3:HUL.InstRepos:dash.current.terms-of- use#LAA Principles of Educational Outreach ('Academic Detailing') to Improve Clinical Decision Making Stephen B. Soumerai, ScD, Jerry Avorn, MD

With the efficacy and costs of medications rising rapidly, it is increasingly reach (sometimes known as "public in¬ important to ensure that drugs be prescribed as rationally as possible. Yet, terest detailing" or "academic detail¬ ' choices of drugs frequently fall short of the ideal of precise and cost\x=req-\ ing") is a newer approach that has been effective decision Evidence indicates that such decisions can be im- evaluated as part of a four-state ran¬ making. trial.4"7 proved in a of A number of theories and of communica- domized, controlled Prescribing variety ways. principles that could be such tion and behavior can be found that underlie the success of problems targeted by change use of with in approaches include drugs pharmaceutical manufacturers influencing prescribing practices. Based on low benefit/risk ratios when safer alter¬ this behavioral science and several field it is to define the trials, possible theory natives are available, use of ineffective to and practice of methods improve physicians' clinical decision making to or marginal therapies for treatable con¬ enhance the quality and cost-effectiveness of care. Some of the most important ditions, use of excessive numbers of techniques of such "academic detailing" include (1) conducting interviews to medications in vulnerable populations investigate baseline knowledge and motivations for current prescribing patterns, (eg, children and the elderly), use of (2) focusing programs on specific categories of physicians as well as on their high-cost agents when less expensive opinion leaders, (3) defining clear educational and behavioral objectives, (4) medications would work as well, and a underuse of effective agents for major establishing credibility through respected organizational identity, referencing illnesses in the elder¬ authoritative and unbiased sources of information, and both sides of (eg, hypertension presenting In one of educational out¬ controversial active in educational ly).5 study issues, (5) stimulating participation reach, two brief visits to physicians by interactions, (6) using concise graphic educational materials, (7) highlighting and clinical trained in effective essential and repeating the messages, (8) providing positive reinforcement of techniques of communication and per¬ improved practices in follow-up visits. Used by the nonprofit sector, the above suasion as well as pharmacotherapy techniques have been shown to reduce inappropriate prescribing as well as were found to reduce inappropriate pre¬ unnecessary expenditures. scribing of cerebral vasodilators, cepha- (JAMA. 1990;263:549-556) lexin, and propoxyphene by 14% in com¬ parison with controls (P=.0001).4 A formal economic analysis of these find¬ ings estimated benefits (reductions in unnecessary expenditures) and costs of THE HIGH cost of medical care and sions. Some efforts have emphasized re¬ the intervention and considered stabil¬ increasing concern about appropriate strictions and economic disincentives, ity of effects over time, use of substitute use of medical technology have resulted but evidence is emerging that in some medications, targeting of high prescrib- in a new wave of interest in how physi¬ applications such strategies may threat¬ ers, and quality effects.6 The data indi¬ cians can be encouraged to make more en the quality of care at least as much as cated that an operational-scale program accurate and cost-effective clinical deci- they reduce expenditures.1,2 Continuing would produce savings two to three education has received much less atten¬ times greater than the costs of mount¬ From the Program for the Analysis of Clinical Strate- tion as a means of behavior change, per¬ ing such a program, without Harvard Medical counting gies, Department of Social , because it often is ineffective in in the of care. A School and Beth Israel Hospital, Boston, Mass haps improvements quality Reprint requests to Harvard Medical School, 643 this regard.3 follow-up analysis7 indicated that while Huntington Ave, Boston, MA 02115 (Dr Soumerai). University-based educational out- physician characteristics such as age,

Downloaded From: http://jamanetwork.com/pdfaccess.ashx?url=/data/journals/jama/9228/ by a Harvard University User on 02/21/2017 board certification, and specialty did the broadly defined components of sales might be manipulative or serve only to not affect response to the educational presentations, as derived from survey increase sales of a given product. intervention, a second reinforcement information. visit substantially increased compliance The following analysis of common Defining Specific Problems with prescribing recommendations. techniques and principles employed by and Objectives Other studies also have docu¬ such are on controlled change-agents based in- A key first step in any program to mented the effectiveness and receptiv¬ depth interviews with and videotapes influence clinical decision making is to ity of physicians to face-to-face counsel¬ of former pharmaceutical representa¬ define the area(s) to be addressed and ing.5,8"10 This article discusses the tives, other written accounts of the de¬ the specific behaviors to be encouraged such and our own principles and techniques of ap¬ tailing process, experience or discouraged. Such agenda setting is proaches to physician behavior change. in mounting several trials of "public in¬ more straightforward in the pharma¬ The prescribing of drugs has served terest detailing" in five states to reduce ceutical industry, where the fiscal "bot¬ as a useful model for studying ways of excessive or inappropriate prescrib¬ tom line" is, by definition, the outcome improving the accuracy of physician de¬ ing.4"710 As discussed below, these tech¬ of concern and increasing sales of specif¬ cision making. Problems of misutiliza- niques are also based on research in the ic branded products is the means toward tion of drugs stem from many sources, fields of adult learning, diffusion ofinno¬ that end. Sponsors of other progams only some ofwhich relate to the pharma¬ vations, and persuasive communica¬ face more complicated choices. For a ceutical industry.5 Nonetheless, the in¬ tion. nonprofit health maintenance organiza¬ dustry's marketing strategies may BEHAVIOR CHANGE tion, employer, or insurer responsible serve as an instructive model (both good STRATEGIES for all aspects of medical care, the out¬ and bad) for those interested in improv¬ come of concern may be the highest pos¬ ing the rationality of clinical decision The effectiveness of the commercial sible quality of care at the lowest possi¬ making. Detailing, the promotional ac¬ detail visit is the result ofmore than 100 ble cost—goals that may conflict and tivities of pharmaceutical company years of sales experience. Generally that often cut across multiple categories sales representatives, accounted for lasting less than 10 minutes, the presen¬ of service (eg, drugs, hospitalization, more than half of the estimated $3 bil¬ tation is likely to contain several effec¬ and physician visits may all be affected lion spent annually on promotion by US tive strategies for behavior change that by use of histamine antagonists). It pharmaceutical companies in the early serve to reinforce and complement one therefore is necessary to begin by iden¬ 1980s (Wall Street Journal. May 25, another.1819 Some similarities exist be¬ tifying the specific cases of practice that 1984:31, 49).11 Pharmaceutical repre¬ tween pharmaceutical marketing goals are of highest concern in terms of quali¬ sentatives have been operating for and several objectives of nonprofit or¬ ty and/or cost and to design a program more than 100 years and number about ganizations. For example, clinical de¬ that will target these decisions initially, 28 000 in the United States alone (Wall partment chairpersons, quality assur¬ specifying clear alternatives to be en¬ Street Journal. May 25, 1984:31, 49),12 ance staff, and health care administra¬ couraged. For example, Ray et al22 used or approximately one representative tors and payers would like to encourage Medicaid claims data to identify inap¬ for every 10 office-based practitioners.13 prescribing behavior that would contain propriate prescribing of tetracycline to At least $5000 is spent per physician costs without sacrificing quality of care children by general practitioners in each year in support of detailing activi¬ and, ideally, would improve it. In previ¬ Tennessee. Subsequent educational ef¬ ties, excluding journal and direct mail ous studies the use of inexpensive but forts focused on the objectives of reduc¬ advertising expenditures (Wall Street effective antibiotics has been promoted ing this practice as much as possible Journal. May 25, 1984:31, 49).n"13 The to replace costly cephalosporins in among targeted groups due to the risk detailer/physician ratio is even higher in the treatment of uncomplicated infec¬ of permanent discoloration of develop¬ some other western countries, approxi¬ tions4,6,7,9; the use of potentially danger¬ ing teeth and substituting safer antibi¬ mating 1:3 in Brazil and 1:4 in Mexico.14 ous antibiotics to treat minor infections otics for this population. A secondary One form of evidence for the effec¬ in ambulatory practice has been cur¬ (but less measurable) objective was to tiveness of these traveling personnel is tailed (eg, chloramphenicol for viral up¬ reduce prescribing of any antibiotics for that pharmaceutical companies contin¬ per respiratory tract infections)9; and simple upper respiratory tract infec¬ ue to invest such large sums of money in safe and effective over-the-counter an¬ tions. this approach, based on their own cost- algesics, such as aspirin, have been sub¬ benefit calculations. Several studies in¬ stituted for less effective and potential¬ Market Research dicate that for many physicians, the de- ly problematic prescription drugs such One of the most important lessons tailer is generally the first source of as propoxyphene.6 Other interventions that can be derived from pharmaceuti¬ information about new therapies.15 In¬ have focused on the way a drug is pre¬ cal marketing strategy is the need to deed, even when commercial messages scribed; for example, certain effective understand the motivations that under¬ on product efficacy differ markedly drugs are prescribed too frequently or lie physician use of a particular product. from objective, scientific sources of in¬ in excessive dosages, causing unneces¬ One failure of traditional continuing formation, the beliefs concerning sever¬ sary risks of adverse drug reactions and medical education has been the assump¬ al drugs of a large, randomly selected increased costs.20,21 While important dif¬ tion that the transmittal of rational in¬ sample of primary care physicians cor¬ ferences exist between pharmaceutical formation alone, independent of how it responded more closely to the commer¬ marketing and medical education or is provided, will predictably result in cial information.16 Despite the fact that quality assurance, there is reason to be¬ improved clinical decisions.3 While the detailers shape so much prescribing be¬ lieve that many of the techniques de¬ provision of new knowledge and infor¬ havior, there is little in the scientific scribed below could be adapted to such mation may be sufficient in some cases, literature about the techniques or ap¬ purposes. Our goal is to identify those considerable research has made appar¬ proach of these representatives and aspects of detailing that can be put in ent the need to consider other "nonsci- how they might be adapted to reduce the service of supporting the physician entific" factors such as attitudes toward inappropriate prescribing. The few in making better therapeutic decisions, particular diseases or patients, habits, studies that do exist14,17 simply describe while leaving aside any techniques that peer influence, patient demand, time

Downloaded From: http://jamanetwork.com/pdfaccess.ashx?url=/data/journals/jama/9228/ by a Harvard University User on 02/21/2017 advantages and disadvantages of rec¬ sages.27,28 Physician awareness of the ommended vs competitive products. profit motive and its potential impact on These factors eventually became major the objectivity of promotional messages messages to communicate in face-to- presents a formidable obstacle for com¬ face details and printed advertisements mercial representatives. (eg, the relative efficacy of one product Sponsorship of educational outreach over another, patient convenience, side programs by neutral professional effects profile, and so on). groups represents a significant advan¬ If the facilitator is sufficiently skilled, tage for "nonprofit" detailers. Two tri¬ he or she may discover some key non- als of face-to-face prescribing education clinical motivations that might other¬ utilized the imprimatur of respected wise remain hidden owing to their nega¬ medical organizations, a state medical tive social or ethical tint (eg, prescribing society and a medical school drug infor¬ tranquilizers to quiet demented elderly mation program.4,9 Such unbiased orga¬ patients at the request of nursing home nizational identities were an important staff). Practice-driven motivations may factor in achieving high rates of physi¬ cause many physicians, although fully cian acceptance of the personal visits aware of the pharmacology involved, to (92% of physicians randomized to this use marginal therapies because of pa¬ group agreed to meet with the medical tient demand factors or intentional use school-based "detailer").4 Another im¬ of active placebos.23 For this sizable portant aspect of both programs was group, an intervention aimed only at their utilization of academically based presenting efficacy data would yield educators bringing the latest clinical minimal results. In our prior work, we findings ofimportance to the physician's instead developed a set of "patient man¬ doorstep. In another setting, the teach¬ agement strategies" and education di¬ ing hospital, credibility also can be dem¬ rected at the patient (Fig 1) to help onstrated through identification with these physicians to cope better with in¬ expert consultants (eg, antibiotic use appropriate requests. guidelines approved by the infectious The focus group approach may fail if disease service).20 the group of physician-volunteers (or paid consultants) sampled is not gener- Targeting 'High-Potential' alizable to the larger group of practitio¬ Physicians ners. To avoid this problem, several Pharmaceutical sales representa¬ such groups are convened and larger, tives often prioritize their sales calls in Fig 1.—Patient education material designed to re¬ structured or unstructured surveys of relation to the magnitude of potential duce patient demand for vasodilators for claudica¬ more representative samples of target sales to individual These tion. for an "exercise is physicians.29 Space prescription" provided are inside. physicians conducted before devel¬ determinations are made through com¬ opment of program content. Thus, fit¬ pany profiles of physicians' previous ting the relevant information to the val¬ prescribing patterns and other data on ues, attitudes, and beliefs of practicing physician specialties and interests. constraints in busy office practices, and physicians is an ongoing task both in Thus, a representative can maximize economic incentives as prerequisites to defining the general content of commu¬ his potential sales per hour of effort by the design of educational messages.523"26 nications in the planning stage and in spending the most time with physicians For example, providing more informa¬ the process of one-on-one education. whose total prescribing volumes are tion on the therapeutic efficacy of aspi¬ high enough to yield potentially high rin may be irrelevant to the physician Establishing Credibility sales of the detailer's products or with who prescribes prescription analgesics Establishing and maintaining a credi¬ physicians who are opinion leaders. primarily in response to patient de¬ ble identity is an essential ingredient in Similarly, from a noncommercial stand¬ mands. effective communication. In the phar¬ point, it makes no economic sense to Several approaches can be used to maceutical industry, credibility is blanket all physicians with face-to-face determine physician motivations for sought by reference to prestigious clini¬ interventions when many of these phy¬ their therapeutic decisions, including cal research centers that evaluated the sicians do not prescribe either inappro¬ focus group interviews, surveys of pre- drug and journals that reported the re¬ priately or in high enough volumes to scribers, and ongoing communication search. According to representatives yield substantial savings or improve¬ between detailers/educators and physi¬ we interviewed, the "aura" surrounding ments in care. cians. In focus group interviews,26 the identification with clinical trials at such In previous controlled studies of uni¬ market researcher assembles a small research centers may contribute as versity-based pharmaceutical educa¬ group of consumers (in this case, physi¬ much to adoption of the drug as the tion for community physicians,4,6,7 9 mod¬ cians) and attempts to stimulate a frank scientific aspects of the studies. Phar¬ erate- to high-volume prescribers of and open discussion of the important maceutical companies also sponsor sci¬ target drugs were identified using data motivations for choosing particular entific symposia for practicing physi¬ from the Medicaid Management Infor¬ products over others. On one level, an cians to convey product-specific in¬ mation Systems in five states. While attempt is made to identify the major formation and to enhance the academic physicians reduced excessive prescrib¬ product characteristics (eg, efficacy or image of the company and its represen¬ ing to the same degree regardless of safety) that physicians value in their tatives, although data suggest that the baseline levels of use, the benefit/cost choice of similar products as well as phy¬ content of such symposia occasionally ratio was the highest (approximately sician perceptions of the comparative may be colored by product sales mes- 3:1) in the group with the highest initial

Downloaded From: http://jamanetwork.com/pdfaccess.ashx?url=/data/journals/jama/9228/ by a Harvard University User on 02/21/2017 volume of target drug use.6 As hospi¬ peutic approach, their credibility may Logical, meaningful questioning is an excel¬ tals, health maintenance organizations, be damaged and physicians may even lent way to give you credibility and to get the and other health care organizations con¬ feel offended that their beliefs and expe¬ physician into a pattern of dialogue and a mindset towards to a tinue to systems riences were treated as invalid. In¬ responding positively develop computerized will the to track resource utilization on a the detailer is to closing challenge. Probing get physi¬ physi¬ stead, likely briefly cian to assimilate the information and require it will become the reasons for cian-specific basis, acknowledge physician's him to actively follow your thought direction. to the often elicit¬ increasingly practical identify physi¬ using competitive drug, As he answers your questions, his thought cians with particular utilization prob¬ ed by asking him. However, he will then will begin to crystallize. Remember, a person lems and present them with focused deemphasize them in comparison with does his best listening when he is talking. educational interventions. in the of his However, advantages company's prod¬ a of research the above studies these were Similarly, long history data not uct. For the same reason, the detailer and in the field of adult learn¬ used on an individual basis feed back practice to may briefly mention the disadvantages ing has suggested that stimulating information indicating that particular of the drug he is promoting—a message learner involvement interac¬ were outside of es¬ communicated the through physicians operating probably already by tion achieves a higher likelihood of at¬ tablished norms. Surveillance and feed¬ competition in any case. The same prin¬ educational and behavioral ob¬ back of this sort is another to educational ef¬ taining approach ciple applies nonprofit as well as satisfaction.34 that work best in certain forts. For in our to jectives greater may settings example, attempts Our educational detailers were also with established lines of an urban teach¬ authority (eg, persuade physicians at trained to ask physicians, during both teaching hospitals and staff model ing hospital to substitute gentamicin initial and how and health maintenance when for we follow-up visits, they organizations). appropriate tobramycin, their patients reacted to prescribing learned that physicians believed the un¬ recommendations. These often Involvement of 'Opinion Leaders' claim that to¬ queries substantiated marketing led to identification of correctable ob¬ bramycin was less nephrotoxic than stacles. For that A key pharmaceutical marketing Rather than this example, discovering gentamicin. ignoring some used as strategy is to identify and involve local issue (or for physicians propoxyphene leaders: those individ¬ important waiting physi¬ an analgesic because of a lack of aware¬ physician opinion cians to it we ac¬ be bring up), actively ness of its risk/benefit attributes made uals who tend to early adopters of the and innovations and knowledged controversy pro¬ it possible to tailor educational materi¬ pharmaceutical impor¬ vided literature references that sup¬ tant and sources of influence als to these deficits. respected the conclusion of no target knowledge in their ported alternative for other physicians communi¬ difference in ties. Previous research by Coleman et meaningful nephrotoxi- Repetition and Reinforcement city. This process was accomplished in a al15 documented that a relatively small tactful manner that did not excessively of a few is of leaders are the Repetition major points proportion physician criticize these another often in con¬ most into the medical infor¬ misperceptions. technique ignored integrated ventional education that aids mation network and are also the most continuing Promoting Active Learner in retention and be¬ likely to be extremely influential per¬ Involvement memory promotes suaders of other, less-integrated physi¬ havior change.35,36 In a commercial de¬ cians. Through these informal social Successful behavior change often is tailing visit, a few principal advantages on the of the educator of the are to be networks, a "contagion" response is dependent ability product likely repeated to relate the (or to the several times the seen after adoption by the opinion lead¬ message product) during detailing visit; and interests of the name itself is further im¬ aided the repre¬ beliefs, needs, values, product er, by pharmaceutical the individual the at¬ in the mind sentatives who may refer to these re¬ physician. Thus, printed physician's through to engage the in two- from the manufacturer and adver¬ spected peers in their own detailing of tempt physician gifts other The involvement of way communication is an important way tisements with product identification. practitioners. a such can be carried such opinion leaders in the design and of building physician-specific presen¬ While approaches to tation32 at the same time as it increases trivial extremes, there is some value to implementation of quality of care or cost-containment educational interven¬ the involvement of the physician in the such mnemonic devices. For example, interaction. For the one we considered (but nev¬ tions can also increase their effective¬ example, asking "giveaway" he or she uses a er for our was a ness. Stross and Bole30 observed that physician why given adopted) early study and with the label "The Best of could drug makes it possible to identify free sample kit surveys community physicians then relevant motivations and Antibiotic for Viral the most target Therapy Upper easily identify educationally as well as Infections" emblazened influential physicians. These individ¬ clinical circumstances to Respiratory the of their beliefs.5 In over a whose unit-dose uals were subsequently enlisted in a gauge strength blister-pack the commercial this also makes cells were all empty. Rationale and evi¬ program to their colleagues' setting, upgrade it for manufacturers to use de¬ dence for these claims would on arthritis treatment decisions during in¬ possible appear as of informa¬ the back of the This formal "teachable moments," when tailers sources marketing package. principle in reac¬ of on a small number of their advice was on related tion, particularly determining concentrating being sought soon release is an un¬ clinical problems. tions to products after their important messages based on to the market.25 However, detailers are derstanding that educational inter¬ Two-Sided Communication careful not to overquestion physicians, changes successfully transmit but a few as this may lead some physicians to feel points, no matter how complex the mes¬ Research has also shown that "inocu¬ that their time is being wasted or, worse sages being sent. This is equally true lation against counterarguments," or yet, that the detailer is trying to test or with physicians, who have a hard presenting both sides of an issue, aids manipulate them.19 A sales-maximizing enough time keeping up with the clutter learning, particularly in situations version of some of these principles is ofinformation deluge in a rapidly chang¬ where the receiver is knowledgeable embodied in an article entitled "Credi¬ ing medical practice. Attempting to about counterarguments.31 If detailers bility: The Key to Long-term Success," communicate too many or complex mes¬ completely discount the positive attri¬ which appeared in a publication for sages often will fail to achieve retention butes of an alternative product or thera- pharmaceutical sales personnel33: of the most important points. Graphic

Downloaded From: http://jamanetwork.com/pdfaccess.ashx?url=/data/journals/jama/9228/ by a Harvard University User on 02/21/2017 support with headlines, diagrams, and photographs is an important compo¬ nent, exposing the physician to visual as well as verbal presentation of the main points. Feedback of improved behavior with reinforcement is another central compo¬ nent of effective detailing techniques that can be used by noncommercial drug information programs. Reinforcement can occur when physicians discuss their experiences with implementing recom¬ mended changes in prescribing behav¬ ior during follow-up visits. Detailers can verbally encourage and applaud their successes as well as discuss prob¬ lem cases and their resolution (eg, pa¬ tient demand). This principle is corrobo¬ rated by data that suggest that a reinforcement visit was associated with a twofold increase in the change in pre¬ scribing behavior in an educational experiment.7 The commercial detailer may encour¬ age the physician to "try out" the prod¬ uct on his patients and to report back his or her experiences at a subsequent visit, often facilitating this with free samples. If these experiences are positive, fur¬ ther discussion and positive reinforce¬ ment of the product ensue. However, it is easy for such a "case study" approach to cross over into pseudoscience. Some pharmaceutical representatives will pursue this strategy further by inviting the physician to conduct a "study" ofthe product in his or her own practice by giving several patients the promoted drug to compare with others who are taking alternative treatments. In a de¬ scription ofa day spent with a successful drug representative reported in the Wall Street Journal (November 8, 1982:1, 25), a physician was encouraged Fig 2. —Educational material used in changing prescribing habits of a cephalosporin antibiotic. Pharmacoki- to conduct his own small-scale "clinical netic considerations were emphasized over cost issues. trial" comparing one company's new nonnarcotic analgesic with other pain medications in patients: "Try it, com¬ receptionists are given gifts to gain ac¬ and protocols, drug bulletins, commer¬ pare it for yourself, you'll like it better." cess to the physician; physicians are giv¬ cially prepared brochures, and two-sid¬ This technique can set up an opportuni¬ en both gifts and free samples of the ed "unadvertisements," have produced ty to reinforce "positive clinical experi¬ drug; patients receive both samples and little or no evidence supporting their ences" (possibly due to placebo effects) educational pamphlets, also donated by effectiveness in changing practice pat¬ in subsequent visits. Despite scientific the detailer. One particularly offensive terns when used alone.5,10 Nevertheless, invalidity of such nonblinded, small- use of incentives was the recent cam¬ well-illustrated materials that empha¬ scale trials in nonrandomized samples of paign by Ayerst, New York, NY, to size the main clinical recommendations patients, this approach purports to ele¬ promote use of its brand of ß-blocker by in a straightforward way are important vate the office physician's status to the offering physicians "frequent prescrib- adjuncts to face-to-face education that level of clinical researcher, with all of its er bonuses" for each patient prescribed can be mailed, posted, or inserted in positive connotations. Through the the product for whom a small reporting patient charts as well as used as a visual drug to be marketed, the physician's form was completed. These could then device during detailing visits. Figure 2 observations are equated with "sci¬ be traded in for free air travel in the gives an example of such an educational ence," further validating both the phy¬ United States and to the Caribbean. pamphlet used at Beth Israel Hospital, sician and the drug. Brief Graphic Print Materials Boston, Mass, to reduce incorrect use of In addition to such suspect approach¬ the antibiotic cefazolin.20 The front cov¬ es, commercial detailers often employ Recent comprehensive reviews of er of the monograph supports the pre¬ incentive strategies to reinforce desired studies that evaluate a wide variety of scribing change message in both head¬ behavior.37 These incentives are direct¬ mailed continuing educational materi¬ lines and graphics; in addition, the ed to all the actors involved: nurses or als, including prescribing guidelines credibility provided by support ofhospi-

Downloaded From: http://jamanetwork.com/pdfaccess.ashx?url=/data/journals/jama/9228/ by a Harvard University User on 02/21/2017 tal experts and references to clinical studies published in respected journals are also presented prominently. It is essential to pilot-test materials with physicians who are similar to those in the planned audience to be certain that the content is actually received as intended. Figure 3 presents the front cover of one of six unadvertisements used in a controlled trial of a face-to-face educa¬ tional program to reduce excessive psy- choactive drug use in nursing homes. The headline and photograph promote a gentle touch and "tucking in" of elderly residents at bedtime as an alternative to sedative use. On the reverse side, other practical strategies for promoting sleep are emphasized, such as avoiding day¬ time caffeine and napping, and increas¬ ing exercise and mobility. It also en¬ courages more careful selection and dosing of sedatives if they are used. Thorough references to the literature are included. When such graphically il¬ lustrated data are referred to in face-to- face sessions they can provide an inter¬ esting and salient respite from verbal presentations and also involve the prescribing physician in two-way communication. Offering Practical Alternatives Behavioral science research in nu¬ merous domains documents that it is far easier to get a subject to relinquish a behavior if an alternative behavior is proposed as a replacement, rather than simple inaction. Thus, it is important to offer the prescriber an alternative to the practice being discouraged—either a more rational medication choice or a nonpharmacologic alternative. Phar¬ maceutical representatives ensure that local stock the alternative Fig 3.—Cover side of "unadvertisement" encouraging nursing home staff to substitute gentle touching and pharmacies reassurance of elderly residents at bedtime for sedative use. drug recommended to replace specific products prescribed by physicians. Sim¬ ilarly, it is important that academic de¬ could be tailored to the patient's physi¬ background, and often none exists. tailing programs consider the practical¬ cal condition (Fig 1). For this approach Training is a crucial element in success¬ ity and availability of other treatments to be most effective, there should be no ful pharmaceutical marketing; it is also to replace targeted drug therapies. financial disincentives to the patient the most carefully guarded. Hemminki Nondrug alternative behaviors, in par¬ that act to discourage alternative treat¬ and Pesonen,14 in a survey of pharma¬ ticular, may be more difficult to achieve ments. For example, overuse of expen¬ ceutical representatives in Europe, re¬ if they require more time to implement. sive prescription analgesics may be dif¬ ported that the mean length of training Unfortunately, it may be easier and ficult to reduce in a Medicaid program time before commencing detail work more convenient for the busy physician that does not reimburse for recom¬ was 14.3 weeks. This was divided be¬ to prescribe a benzodiazepine than to mended over-the-counter alternatives tween sales techniques and a segment engage in a counseling session concern¬ such as aspirin or acetaminophen. on basic medical sciences and pharma¬ ing the patient's family problems. On Selection and Training of cology. Closed-circuit television and the other hand, encouraging patients Academic Detailers role playing are common elements of with peripheral vascular insufficiency training used to perfect presentation to take short walks was a reasonably According to the pharmaceutical styles. In addition, fledgling represen¬ actionable alternative to use of "periph¬ marketing literature, commercial detail tatives may spend some time trying out eral vasodilators" prescribed for this people are selected on the basis of both their presentations on groups of physi¬ condition.4 To help facilitate this behav¬ science background and potential for cians. This approach is particularly use¬ ior, a patient education pamphlet was sales,'* although in practice the latter ful in increasing the representative's prepared as a physician "prescription" factor often predominates. There is no ability to answer questions convincingly of a specific exercise program, which standard requirement for any science and to tailor the presentation to the par-

Downloaded From: http://jamanetwork.com/pdfaccess.ashx?url=/data/journals/jama/9228/ by a Harvard University User on 02/21/2017 ticular style of the physician. Fair trade physician-researcher (J.A.) who se¬ al aspect of the commercial setting, but and advertising laws are emphasized as quentially acted the role of a resistant not generally used in the not-for-profit they apply to the presentations. and hostile physician, an indifferent setting, is the provision of financial in¬ Management control and feedback to physician, and a receptive physician. centives that reward pharmaceutical detail people are carried out principally Following this, approximately a half representatives who achieve major pre¬ by district sales managers on the basis day was spent practicing on one anoth¬ scribing changes that result in high of ongoing communications from detail¬ er, with videotaped feedback. When the sales of their company's products (For¬ ers, including information contained in -educators and investiga¬ tune. October 1972:99-102). From the required call reports as well as reports tors were satisfied that all trainees were perspective of many physicians, such of sales from pharmacies and in hospi¬ performing adequately, they pretested financial incentives probably reduce the tals.38 Since representatives are gener¬ their presentation on randomly selected detailer's credibility and the perceived ally responsible for product sales in par¬ community physicians. These encoun¬ objectivity of his claims concerning ticular geographic regions, companies ters were discussed by telephone with product efficacy. In fact, these incen¬ can track the amount of a given medica¬ the project manager before actual de¬ tives have, on occasion, led to misinfor¬ tion sold in those areas to monitor repre¬ tailing of study physicians began. In mation by overzealous detailers.27,4042 sentatives' sales performance. The general, these initial experiences indi¬ This may be one reason that physicians sales manager is also the key person in cated that physicians were more recep¬ thus are reluctant to admit that their providing emotional support to detail tive than was expected based on the prescribing decisions are influenced by people. Although most representatives experiences of commercial represent¬ commercial representatives.16 are salaried, the majority also receive atives. The principles embodied in this ap¬ incentives for greater sales through After every physician visit, detailed proach also conform with research con¬ commissions and bonuses. encounter forms were completed to as¬ ducted in the more general field of adult Based on reports in the medical mar¬ sess physician receptivity, problems en¬ education. For example, one synthesis keting literature,13 most pharmaceutical countered, methods of overcoming ob¬ of this literature identified five condi¬ sales representatives probably make stacles, call records, exposure times, tions of adult learning: motivation to between three and eight physician con¬ etc. These were reviewed periodically change, active involvement of the learn¬ tacts per day. The contact time per visit by the project director (S.B.S.) to at¬ er, relevance to past experience, feed¬ varies enormously, but an average of 5 tempt to correct particularly difficult back, and an informal atmosphere.43 to 8 minutes is a reasonable estimate of problems experienced by some detail¬ Similarly, the extensive literature on average physician exposure time.13,19 ers. We found that the clinical pharma¬ patient compliance also documents the Because of this limited time, most rep¬ cists needed only a minimal amount of effectiveness of face-to-face educational resentatives rarely attempt to promote training and supervision to perform ad¬ interactions and the general failure of more than two or, at most, three prod¬ equately in this function. This success printed materials when used alone.44 ucts during one visit. One product usu¬ was evidenced by the prescribing data This is particularly sobering in view of ally represents the major topic to which outcomes: physicians reduced target the recently legislated "practice guide¬ most of the representative's time is drug use by 14% compared with controls lines" initiative that will be sponsored committed during the visit, with a sec¬ after only two brief visits (P = .001).4 by the US Department of Health and covered "on the ondary drug way out COMMENT Human Services and is intended to im¬ the door." prove the use and outcomes of common In noncommercial community pro¬ The long-standing effectiveness and medical practices nationwide. If such grams, studies have documented the ef¬ growth of pharmaceutical detailing in guidelines are not well integrated into fectiveness of clinical pharmacists both recent decades illustrate that consider¬ formal and informal person-to-person with4 and without39 doctoral-level train¬ able change in physician prescribing educational programs they may not ing as well as consultant physicians.9 practice is possible without regulatory have any practical value. In coming Both pharmacists and specially trained interventions. The behavioral science years, it is likely that computer-based physicians (eg, infectious disease spe¬ literature and an increasing number of decision support systems will become cialists) have been used in hospitals5; field studies indicate that a number of sophisticated and widespread enough more study is needed, however, on the important principles of adult learning that some (but not all) of the interac¬ comparative cost-effectiveness of vari¬ and persuasion may be adaptable to tions described above will be done elec¬ ous personnel for particular problem ar¬ noncommercial programs designed to tronically. However, such capacity is eas. Doctoral-level clinical pharmacists improve clinical decision making.5,10 not likely to be available in most prac¬ who were primarily junior faculty in This is particularly important at a time tice settings in the immediate future. schools ofpharmacy have been shown to when powerful new agents are available Some may argue that restrictions and be effective as part-time, academically that have real toxic potential as well as regulations, rather than education, are based detailers.4"7 Because of their pre¬ unique therapeutic effectiveness (eg, simpler and more effective strategies vious training in pharmacology and isotretinoin). Regulation can ensure for improving prescribing practice. In drug information, only about 2 days of that drugs with unacceptable risk/bene¬ fact, restrictive hospital formularies are formal training was necessary for a pro¬ fit relationships are not permitted to be an effective means of reducing improper gram targeting three drugs. This fol¬ used, but no current regulatory appara¬ use of some medications. However, for lowed self-study of a training manual tus can ensure that marketed drugs are many medications the problem is mis¬ covering important clinical issues, im¬ prescribed intelligently. It should be use of an otherwise effective agent (eg, portant principles of communication possible to take the best of these educa¬ antibiotics). Removing access to the and persuasion, and major recommen¬ tional approaches and put them at the drug through a restrictive formulary or dations concerning the target drugs and service of optimal care, while leaving reimbursement restrictions cannot ad¬ their alternatives. In the training, vid¬ behind any techniques that are not in dress this issue. Further, drug restric¬ eotapes were shown of a former phar¬ keeping with the spirit of professionally tions are not practical without effective maceutical representative detailing the based continuing medical education. sanctions and enforcement. In addition, target drugs and their alternatives to a For example, one important structur- no extensive literature exists on the ef-

Downloaded From: http://jamanetwork.com/pdfaccess.ashx?url=/data/journals/jama/9228/ by a Harvard University User on 02/21/2017 fectiveness of drug restrictions in com¬ research and practice has begun to tests, and other diagnostic resources munity settings, and the few extant emerge based on the concept of academ¬ and the decision to hospitalize. In this studies suggest that restrictions of one ic detailing that may present a rational way, improved precision in clinical deci¬ drug can lead to generally inappropriate response to these pressures. Such edu¬ sion making can present an attractive use of substitute medications.5,45 Educa¬ cational outreach activities by universi¬ alternative to the twin extremes of un¬ tional approaches thus are more flexible ties, medical centers, and professional acceptable costs and unacceptable and adaptable in influencing most real- associations are increasingly proposed restriction. world clinical decisions that do not fall or being implemented under the spon¬ into neatly defined categories of correct sorship of hospitals, public-sector This work was supported in part by the John A. and incorrect practice. programs, health maintenance organi¬ Hartford Foundation and the National Center for Health Services Research HS 03880 and HS The cost of new is increas¬ and other insurers. The (grants therapies zations, appli¬ 04933). at the same time that of this offers ing sharply pres¬ cability approach promise Figures 1 through 3 were reproduced with per¬ sure is escalating to control health care to other areas beyond prescribing, such mission from Beth Israel Hospital, Boston, Mass. costs. A small but important body of as the use of blood products, laboratory References

1. Soumerai SB, Avorn J, Ross-Degnan D, Gort\x=req-\ 15. Coleman JS, Katz E, Menzel H. Medical Inno- periments on Mass Communication. Princeton, maker S. Payment restrictions for prescription vation: A Diffusion Study. Indianapolis, Ind: NJ: Princeton University Press; 1949. drugs under Medicaid: effects on therapy, cost, and Bobbs-Merrill Co Inc; 1966. 32. Watts WA. Relative persistence of opinion equity. N Engl J Med. 1987;317:550-556. 16. Avorn J, Chen M, Hartley R. Scientific vs. change induced by active compared to passive par- 2. Lurie N, Ward NB, Shapiro MF, Brook RH. commercial sources of influence on physician pre- ticipation. J Pers Soc Psychol. 1967;5:4-15. Termination from Med-Cal: does it affect health? N scribing behavior. Am J Med. 1982;73:4-8. 33. Ragland B. Credibility: the key to long-term Engl J Med. 1984;311:480-484. 17. Hemminki E. Content analysis of drug-detail- success. Pharm Rep. 1985;15:14-15. 3. Haynes RB, Davis DA, McKibbon A, Tugwell ing by pharmaceutical representatives. Med Educ. 34. Long HB. Adult Learning: Research and P. A critical appraisal of the efficacy of continuing 1977;11:210-215. Practice. New York, NY: Cambridge University medical education. JAMA. 1984;251:61-64. 18. Roseman E. Make sure your sales force deliv- Press; 1983. 4. Avorn J, Soumerai SB. Improving drug-thera- ers the right message. Med Market Media. Sep- 35. Cook T, Insko C. Persistence of attitude py decisions through educational outreach: a ran- tember 1985:90-95. change as a function of conclusion reexposure: a domized controlled trial of academically based 'de- 19. Roseman E. Checking the scorecard. Med laboratory-field experiment. J Pers Soc Psychol. tailing.' N Engl J Med. 1983;308:1457-1463. Market Media. October 1985:40-44. 1968;9:332-328. 5. Soumerai SB, Avorn J. Efficacy and cost-con- 20. Avorn J, Soumerai SB, Taylor W, et al. Reduc- 36. Roseman E. Open your eyes to using visual tainment in hospital pharmacotherapy: state of the tion of incorrect antibiotic dosing through a struc- aids in selling. Med Market Media. October 1985:6\x=req-\ art and future directions. Milbank Mem Fund Q. tured educational order form. Arch Intern Med. 10. 1984;62:447-474. 1988;148:1720-1724. 37. EisenbergJM. General principles: a briefover- 6. Soumerai SB, Avorn J. Economic and policy 21. Buchwald D, Soumerai SB, Vandevanter N, view. In: Doctor's Decisions and the Cost of Medi- analysis of university-based drug 'detailing.' Med Wessels MR, Avorn J. Effect of hospitalwide cal Care. Ann Arbor, Mich: Health Administration Care. 1986;24:313-331. change in clindamycin dosing schedule on clinical Press Perspectives; 1986:91-97. 7. Soumerai SB, Avorn J. Predictors of physician outcome. Rev Infect Dis. 1989;11:619-624. 38. Smith MC. Principles ofPharmaceutical Mar- prescribing change in an educational experiment to 22. Ray W, Federspiel CF, Schaffner W. Prescrib- keting. Philadelphia, Pa: Lea & Febiger; 1968:299\x=req-\ improve medication use. Med Care. 1987;25:210\x=req-\ ing oftetracycline to children less than 8 years old: a 316. 221. two-year epidemiologic study among ambulatory 39. Avorn J, Soumerai SB, Everitt DE, Sherman 8. Soumerai SB. Improving the quality and econo- Tennessee Medicaid recipients. JAMA. 1977;237: DS, Beers MH, Salem SR. Reduction of psychoac- my of in-hospital prescribing: getting more for less. 2069-2074. tive drug use in nursing homes through 'academic Med J Aust. 1988;149:574-576. 23. Schwartz RK, Soumerai SB, Avorn J. Physi- detailing.' Clin Res. 1988;36:331A. Abstract. 9. Schaffner W, Ray WA, Federspiel CF. Improv- cian motivations for nonscientific drug prescribing. 40. Eastman M. The day the detail man talked. Am ing antibiotic prescribing in office practice: a con- Soc Sci Med. 1989;28:577-582. Pharm. 1978;NS18:8. trolled trial of three educational methods. JAMA. 24. Smith MC. Appeals used in advertisements for 41. Silverman M, Lee PR. Pills, Profits and Poli- 1983;250:1728-1732. psychotropic drugs: an exploratory study. Am J tics. Berkeley: University of California Press; 1974. 10. Soumerai SB, McLaughlin T, Avorn J. Improv- Public Health. 1977;67:171-173. 42. Darvon MD education campaign deteriorating ing drug prescribing in primary care: a critical anal- 25. Gorzalski EJ. Use your sales force to feed back rapidly, FDA scolds Lilly. Weekly Pharm Rep. ysis of the experimental literature. Milbank Q. In vital product launch information. Med Market Me- 1980;29:4. press. dia. October 1986:1827. 43. Long HB. Adult Learning: Research and 11. Rucker TD. Drug information for prescribers 26. Zelnio RB, Zimmerman JG. Match your data Practice. New York, NY: Cambridge University and dispensers: toward a model system. Med Care. collection method with your research needs. Med Press; 1983:244. 1976;14:156. Market Media. January 1985:51-55. 44. Haynes RB, Wang E, Gomes MDM. A critical 12. Rx detail forces will not keep pace with MD 27. Opren scandal. Lancet. 1983;1:219-220. review of interventions to improve compliance with growth during next six years, SRI predicts; MS & 28. Bowman MA. The impact of drug company prescribed medications. Patient Educ Counsel. D and SK & F are increasing detail forces against funding on the content of continuing medical educa- 1987;10:155-166. trends. F-D-C Reports. 1982;44:3. tion. Mobius. 1986;6:66-69. 45. Soumerai SB, Ross-Degnan D, Gortmaker S, 13. Roseman E. How to gain your fair share of 29. Labson D. Conversations with a detailperson: Avorn J. Withdrawing payment for non-scientific sales time. Med Market Media. February 1985:24\x=req-\ part II. Med Market Media. February 1979:16. drug therapy: intended and unexpected effects of a 28. 30. StrossJK, Bole GG. Evaluation of a continuing large-scale natural experiment. JAMA. In 14. Hemminki E, Pesonen T. The function of drug education program in rheumatoid arthritis. Arthri- press.OC9457 company representatives. Scand J Soc Med. tis Rheum. 1980;23:846-849. 1977;5:105-114. 31. Hovland CI, Lumsdaine AA, Sheffield FD. Ex-

Downloaded From: http://jamanetwork.com/pdfaccess.ashx?url=/data/journals/jama/9228/ by a Harvard University User on 02/21/2017