Very little research has looked specifically at the effect of promotion on attitudes, much more has examined the effect of promotion on knowledge. The studies here are part of a field of research into the determinants of prescribing – how doctors learn about drugs, and how they come to prescribe new products.
Most of the studies discussed in this review are really about how much doctors report using promotion as a source of information (either for all drugs, or particularly for new drugs) rather than about effects of promotion on attitudes or knowledge. They are included because they provide information relevant to the question of whether promotion affects prescribers’ knowledge.
Some studies look directly at the impact of promotion on attitudes and knowledge, by using an experimental approach (3920), by interviewing people about previous exposure (680), or by following up participants in a promotional event (1690). Others approach the question in a more sophisticated or indirect way. Ziegler et al. (3530) look at whether doctors notice and remember errors in promotion. Sangiry et al. (21260) look at whether consumers are aware of information missing from advertisements. Others (3520), (3670) look at doctors’ attitudes or knowledge in areas where there is disagreement between commercial and scientific information and infer the impact of promotion from this. Ferry et al. (4510) directly assessed knowledge of prescribing for the elderly and looked at it in relation to self-assessed reliance on promotion.
The methods that have been used in this area are not capable of producing certainty about causal relationships. Firstly both exposure to promotion, and knowledge and attitudes about drugs, are often assessed using self-report data. Secondly the relationship between them is often also assessed using self-report. That is, doctors are asked how much their prescribing is influenced by promotion. Self-report can be misleading when doctors’ beliefs are inaccurate (eg, they may believe that they are exposed less often than they are), or when their answers to questions are biased towards being more socially acceptable than what they really believe.
2.1 Reported use of promotion as a source of drug information
In a 1974 FDA survey in the US, 64% of all doctors, and 80% of GPs and paediatricians reported using materials from sales representatives as a source of drug information. Fifty percent of doctors reported using journal advertisements (4040). Christensen and Wertheimer (3580) found that sales representatives were reported to be the first source of information for one of the two drugs they studied. Advertisements in journals were the third source for the other drugs. This small study (29 doctors) is now very old (1975-6).
Two studies found doctors in developing countries relied very heavily on industry-based sources of information. Ahmad and Bhutta (5690) found 95% of the doctors they interviewed in Karachi relied upon industry promotional material as their main source of information about drugs. They also found extremely high levels of irrational prescribing and dispensing for children. Similarly, Tomson and Angunawela (5730) describe heavy reliance on industry sources of information, and much polypharmacy in a peripheral clinic in Sri Lanka. In contrast, in Osiobe’s two Nigerian studies (5610) (5590) health professionals and health professional faculty members reported low use of commercial information.
Some differences have been described between different kinds of drug information, and over time. In Hatton et al.’s study sales representatives were used more as a source of general information about drugs rather than pregnancy related information (2010). Williams and Hensel (3030) claim from their review of studies on sources of information about drugs, that commercial sources of information had declined in importance over time. They do not describe the search methods that they used to locate the articles included in their review. The studies included in this review were all surveys and a social acceptability bias as the cause for the results that they found cannot be excluded, i.e. over time it may have become less acceptable to claim reliance on commercial sources of information.
Some studies have explored differences between doctors in how far they say they rely on commercial sources of drug information. In McCue et al.’s study (3540) doctors who had been practising more than 15 years used drug sales representatives as a source of information about new drugs more frequently than other doctors did. In Abate et al.’s study (3820) academic medicine physicians used drug industry sources for their drug information questions less than private practice physicians did. Drug sales representatives were rated the most important source of information about advances in anti-rheumatic drugs by doctors who qualified in the 1950s, the second most important by those who qualified in or after 1960 in Murray-Lyon’s study of GPs in Scotland (710). Gaither et al. (650) found that, among the 108 Michigan HMO doctors they surveyed, those who were not Board certified were more likely to intend to use sales representatives and literature from the drug industry than others, for information about a fictitious new drug. Also those with more than five colleagues at their work site were less likely to use industry literature.
CONCLUSION: Doctors’ own reports suggest that promotion is often used as a source of drug information, less so by doctors who qualified more recently or who practise in an academic rather than a private setting.
2.2 Reported use of promotion as a source of information in adopting new medicines
Hibberd and Meadows (4000) found 85% of the UK doctors they interviewed said they used MIMS (a commercial source) to learn about new drugs, but most used non-commercial sources to find out about efficacy. Similarly, the British doctors in Eaton and Parish’s study (2260) reported using sales representatives as a source of information about new drugs, but relying on them less to establish whether a drug was useful and should be prescribed.
In a study of British GPs, by Strickland-Hodge and Jepson (1980), three commercial sources were rated in the top five sources used to alert respondents of new drugs, but five professional sources were the most popular for providing information to evaluate drugs. GPs who worked alone cited sales representatives as a source of information for evaluating drugs more often than GPs who worked in group practices.
In Peay and Peay’s 1994 paper (4450) about specialists and high-risk drugs, they found that commercial sources of information played little or no role in the adoption of drugs in the doctor’s primary area of expertise, but suggest that these sources may provide information about new drugs outside this area of expertise.
Manning and Denson (3570) looked at how US internal medicine specialists first learnt about cimetidine. Their study was performed soon after cimetidine was launched in the US in 1977. Drug sales representatives were rated as the sixth or seventh most commonly mentioned source in each stage of learning about cimetidine. However advertisements declined in importance. They were the eighth most common source for first knowledge of the drug, tenth for learning principles of using it, and thirteenth for providing update information.
Parboosingh et al. (1230) interviewed specialists attending annual scientific meetings and asked them to identify two or three changes they had made in their clinical practice in the last two years, and the factors involved in these changes. Eighty-one of the 192 changes made were changes in prescribing. Sales representatives were noted as initial sources of information for less than 20% of the changes, and very infrequently noted as precipitating the changes. Like other studies, this suggests that commercial information may be more important in alerting doctors to drugs, and less important in later stages of decisions to adopt new drugs.
Williamson (3420) draws on literature on risk assessment to examine general practitioner prescribing of new drugs. He concludes that the level of risk which a doctor perceives determines how much external validation he or she requires in order to prescribe the drug. From a small survey, Kleinman claims to show that doctors’ preferred information sources vary with the perceived riskiness of medicines. He argues that sales representatives are the most important source for low risk drugs, but are less important for higher risk drugs. Both studies are too limited to provide conclusive evidence but their theory deserves more testing and their approach of drawing on other social science literature is one that other researchers should follow.
A US survey of 680 doctors (4380) found that 9% of doctors rarely or never met with sales representatives. The study suggested that documented evidence of a product’s efficacy and applications was the major factor in doctors’ decisions to switch or increase the frequency of a particular medication. The report of the survey in Pharmaceutical Executive provides no detail about the methods used so the study is hard to evaluate.
In one of the few qualitative studies on promotion, Jones et al. (21430) interviewed 38 consultants in Birmingham hospitals and 56 GPs who regularly referred patients to the teaching hospital. They also monitored the prescribing of specific drugs by the GPs, and in the hospital. They reported that sales representatives were an important source of information for both GPs and specialists. Jones et al. suggest that prescribers were not consistent in their definition of ‘prescribing a new drug’. They were unsure whether this meant adding this drug to their regular prescribing repertoire, or whether it could involve prescribing it only a few times. In addition, GPs were unsure whether to include new drugs that they were prescribing because the hospital had started a patient on them. This suggests definitions need to be very clear in quantitative studies in this area, so that results are consistent.
CONCLUSION: Self reports indicate that promotion is often used as a source of information about new drugs, especially for indications for which the doctor has less expertise.
2.3 Impact of promotion on self-reported attitudes and knowledge
Engle (3920) carried out a large study on the effect of a single-advertiser publication on doctors’ attitudes and expected prescribing behaviour. Four first-edition hardcover books, about 100 pages long were mailed, one at a time, to 19,200 doctors. These were on topics related to medicine but were mostly non-technical and enjoyable to read. The books each included 18 pages of advertising for a broad-spectrum antibiotic. Questionnaires were sent to random selections of 1200 of the doctors 45 days before the first book was sent, and about one month after each of the other books was sent. Engle compared the attitudes and expected prescribing of readers (those who had read at least one book, those who had read at least two books) with non-readers (baseline and those who had not received the books). Readers were significantly more positive toward the company that sent the books than non-readers, but not more positive toward other companies. Readers were more likely to expect their frequency of prescribing the company’s product to increase. This was statistically significant and seemed to affect only the sponsor’s product. Engle suggests that the campaign may have been successful because the books were probably read cover to cover, unlike technical journals. This large and ambitious study provides before and after study evidence that a promotional campaign can significantly affect prescribers’ attitudes. The study design could have been improved by including a randomly selected control group.
Sandberg et al. (680) present evidence that students given textbooks by pharmaceutical sales representatives are unlikely to remember the name of the company or its products. They interviewed 205 fourth year medical students, of whom 90% had received one or more textbooks from companies. Most could remember the title of the book, but only 25% could remember the company or a product associated with the gift. Most of the students were interviewed during their personal interview for admission to a residency. This seems a far from ideal interview situation. Students may have been very nervous and this may have affected their recall, and they are likely to have had a strong desire to give the answer that they thought the interviewer wanted. More importantly, this study did not explore one of the key points about giving gifts to students. It is likely that the effects of gifts on students include establishing habits, e.g. a willingness to receive gifts and the development of positive attitudes towards drug companies. Because prescribing cannot be influenced immediately, the memory of the link between a gift and a specific product or company is less important.
Spingarn et al. (1690) found that house staff who had attended a grand round on Lyme disease, presented by a drug company, were more likely than non-attenders to prefer the use of the expensive parenteral drug made by the speaker’s company. They reported that they would choose this drug for Lyme disease, even in mild cases where it would not be the best choice. This was in spite of the fact that the speaker did not recommend this style of treatment. The authors speculate that the effect may have come about because the speaker devoted extra time to the late complications of Lyme disease. These cases are infrequent but require the more expensive treatment. They note that this may also occur in non-commercial presentations. In addition, most of those who were present claimed that they did not know at the time that the speaker was from a drug company, even though he was introduced in this way.
CONCLUSION: Promotion influences attitudes despite some evidence that the details are not always remembered.
2.4 Research designs that aim to avoid the limitations of self-report data
Research by Ziegler, Lew and Singer (3530) suggests that doctors may not be very critical of verbal promotional information. They found 12 inaccurate statements (i.e. statements which contradicted the Physicians’ Desk Reference or literature quoted by or handed out by the sales representative) in brief presentations given by sales representatives at industry-sponsored lunches. When they later surveyed 27 doctors who attended these presentations only 7 recalled hearing the representative make a claim that they knew to be false.
Similarly, Sansgiry et al. (21260) found that consumers may not be very critical of advertisements. They compared consumer’s (students in non-health related subjects) assessments of 14 advertisements for over the counter medicines with those of experts (clinical pharmacists). Each advertisement was viewed by nine participants. The consumers rated the advertisements as more factual, good and complete than the experts, even though information on contraindications and side effects was missing. Consumers were not able to identify misleading and inaccurate information. Only 20% of consumers identified side effects correctly and 14% contraindications.
A 1982 study by Avorn, Chen and Hartley is very commonly quoted as evidence of the negative impact of promotion (3520). They surveyed doctors about two drugs about which there was significant disagreement between scientific and commercial sources of information. There was no scientific evidence of benefit from, cerebral vasodilators and evidence of minimal efficacy for propoxyphene. However promotional material presented them as efficacious and reliable. Avorn et al. argued that by looking at which of these beliefs doctors held they could see which type of information source doctors were really influenced by. Most of the 85 Boston doctors they surveyed said that they relied mainly on academic sources of information, and that advertising, sales representatives and patient preference were minimal influences on their prescribing. However their beliefs about cerebral vasodilators and propoxyphene tended to be more consistent with the commercial literature than with the scientific consensus. Nearly half (48%) of the doctors who supported the use of vasodilators stated that they were more influenced by scientific rather than commercial sources of information. Avorn et al. say that this discrepancy between where the doctors’ beliefs seemed to come from, and their statements about what influenced them could be because doctors are unaware of how commercial sources influence them, or it could be because doctors are unwilling to admit this influence.
The Avorn et al. study is particularly important because it is very widely quoted. Therefore it is important to analyse it critically, and to suggest how further research might explore its findings. The study was simple, presumably inexpensive, yet cleverly designed. However it could be criticised in several ways. Avorn et al. focus on two sources of influence: ‘scientific’ and ‘commercial’. However more doctors in the study rated their own ‘training and clinical experience’ as a very important influence on their prescribing than rated either scientific or commercial sources in this way. At the time of the study there was a clear disagreement between the scientific and commercial views of the drugs studied, but the authors suggest incidentally that this had not previously been the case (eg the link between cerebral blood flow and senile dementia is “a concept now abandoned”). Doctors holding the ‘commercial’ belief could be holding on to a view taught in medical school or learnt from other doctors in the past and/or reinforced by the placebo effect in practice (their ‘clinical experience’). Avorn et al. do not convincingly demonstrate that the doctors’ ideas came from commercial sources. They may instead have resulted from their training and beliefs that ‘clinical experience’ is more valuable than scientific evidence-based medicine. It may be hard for doctors to exclude drugs from their prescribing repertoires if they learnt about them in medical school and they seem to work in practice.
Greenwood’s study (3670) included a sub-study that repeated Avorn et al’s method. He surveyed 332 GPs, in one area of England, about the use of four medicines where scientific and commercial views conflict. The commercial view was held by 77% of doctors for one drug, and by 55%, 28% and 13% for the others. This study was a PhD thesis, which is difficult to obtain. Summaries are published in HAI News, No 48, August 1989 (17300) and in Lexchin (1220).
Similarly, in Peru, Cardenas and Isenrich (2730) found that while doctors said they relied mostly on medical literature for their drug information, in fact their self-reported prescribing decisions were clearly not based on this.
Ferry et al. (4510) found that doctors who reported relying on advertising as a source of information achieved lower scores on a test of knowledge about prescribing for the elderly. This study had a low response rate.
CONCLUSION: Doctors’ attitudes are influenced by promotion much more than they think.
Directions for future research
There is no literature in the database on the impact of promotion on the attitudes and knowledge of other people such as consumers, pharmacists, nurses, or drug-store staff, all of whom may be important decision-makers about medicines. Future studies could include these groups. That the effects of promotion are likely to be great is suggested by long experience with promotion of breast-milk substitutes throughout the world. Publications in this area are of course outside the scope of this database.
Promising research designs such as that pioneered by Avorn et al. seem worth pursuing further. One possible approach would be to examine a treatment for which there is substantial scientific support, but little advertising, such as Oral Rehydration Solution (ORS). If such a study also found that doctors claimed to be influenced more by scientific rather than commercial information, but tended not to prescribe ORS (because there is little or no commercial information about its benefits), Avorn et al.’s conclusions would be much strengthened. Such a study would also avoid the difficulty of excluding a drug from one’s prescribing repertoire: because in this case a treatment is being added rather than deleted. One of the advantages of this type of study is that it is relatively cheap: essentially it involved a telephone survey of 85 doctors.
No comments:
Post a Comment