What interventions have been tried to counter promotional activities, and with what results?

This review reports on research on interventions to control or counter promotion, and the effects of such interventions. It is not a comprehensive review of interventions, because there are many descriptive reports on these in the database, which are not included here. These reviews only describe research, not descriptive reports.

4.1 Guidelines, codes and regulations for printed and broadcast material

This is by far the most commonly researched intervention.

Few studies describe regulations and guidelines about printed and broadcast promotional material and how they are monitored in different countries. Lexchin describes the situation in the UK, Australia, and Canada. He suggests that in most industrialised countries the day to day control of promotion rests with voluntary national industry association (930). He examines the enforcement of the four codes (there are two in Canada) according to five critical aspects. These are mechanisms for recognising violations, the composition of monitoring committees, sanctions for violation, the quality and quantity of information in reports issued about complaints and violations, and the circulation of these reports. He argues that codes that rely on complaints are inadequate because too many violations are missed. Instead he recommends proactive monitoring of a random sample of promotional activities. He also suggests that the majority of members of committees should be from outside the industry, that sanctions should be raised so that they deter companies from misleading promotion, that offenders should be required to correct false information, and that information about the enforcement of the code should be publicly available (930). His analysis is particularly clear and useful.

Herxheimer and Collier described the situation in the United Kingdom (3020), where the Association of the British Pharmaceutical Industry (ABPI) introduced a Code of Practice for the promotion of prescription medicines in 1958. This is administered by a code of practice committee consisting of one independent barrister, 12 representatives from member companies, and two independent doctors (and now also a representative of a patients’ organisation). The ABPI secretariat examines advertisements in a random selection of journals to see if they comply with the code, and since 1985 a medical consultant also makes an independent scrutiny of a random selection of advertisements. Possible breaches of the code are pursued by informal correspondence with the company which placed the advertisements. The code of practice committee deals with cases not resolved by this informal correspondence, plus all complaints originating outside the ABPI. This process was largely secret until 1982. Herxheimer and Collier argue that the evidence suggests that the code has failed to deter promotional excesses. The ABPI's wish to secure compliance with the code seems weaker than its wish to pre-empt outside criticism and action: its self regulation seems to be a service to itself rather than to the public. They suggested that the code of practice committee should become publicly accountable, that the majority of its members should represent the health professions and the public, and that effective sanctions were needed.

Some journals also have policies about pharmaceutical advertisements. Wilkes and Kravitz (21900) surveyed 221 editors of medical journals in North America. Twelve percent acknowledged occasional conflicts of interest between editorial decisions and the wishes of advertisers. Most journals published pharmaceutical advertisements (67%) and 41% of the editors of these journals reported having a great deal of control over advertisements. However only eight journals had required advertisers to make a change to an advertisement in the last five years. Forty percent of the editors thought that journal advertisements should be subject to the same peer review process as scientific articles.

CONCLUSION: Neither the self regulatory systems that have been studied nor review by journal editors provide effective controls on drug advertising.

Many studies show that printed advertisements do not meet regulations and guidelines in force in various countries.

In the United Kingdom, Morgan et al. (1590) suggested in 1976 that the ABPI guidelines were “not being observed strictly”. Herxheimer and Collier (3020) reviewed 302 cases considered by the ABPI code of practice committee between January 1983 and December 1988, and found that the ABPI code was commonly broken. Breaches were identified in 192 of the complaints from competing companies, doctors and pharmacists. In total there were 379 breaches, of which 270 were also possible breaches of the Medicines Act. Misleading claims or comparisons, and misleading or unsubstantiated information were the most common breaches. Herxheimer and Collier argued that the code had no obvious deterrent effect. The code of practice committee had no power to require retraction or correction of misleading statements and gave no adverse publicity to those responsible for breaches. Although they required an undertaking from offenders that the breach would not be repeated, this was not always honoured. The authors noted that there was no consumer input into the code, and complainants had no opportunity to respond to defendants’ arguments.

In developing countries, frequent breaches of codes have also been found. In French-speaking West Africa, Chirac et al. (5740) found that 13.5% of the advertisements they studied did not meet even the most liberal interpretation of the IFPMA code. Dikshit and Dikshit (810) found that the Indian Issue of the Monthly Index of Medical Specialities (MIMS) did not seem to follow the code of the IFPMA, which it is covered by. Vo-Kyung and Ok Kim (5810) report a high level of violations of South Korean regulations on the advertising of medicines. Lal et al. (4620) found that only 2% of the 585 advertisements in their study cited adequate references, and therefore conformed with the WHO Criteria for Ethical Medicinal Drug Promotion. These advertisements were supplied by sales representatives to clinical departments of an Indian hospital. Both the IFPMA and the WHO Ethical Criteria for Medicinal Drug Promotion include the concept of ‘reminder’ ads, which do not need to contain full prescribing information. Tomson and Weerasuriya (5820) analysed advertisements in the Ceylon Medical Journal and found that only a quarter contained information on adverse effects and a quarter on contraindications. However they suggest that 68% would pass both the IFPMA and the WHO codes if they were classified as reminder advertisements.

In the USA, Roth (3150) found that 65% of the newspaper and magazine DTCA advertisements reviewed by two pharmacists presented a ‘fair balance’ of information, as required by the FDA. Stryer and Bero concluded that 42% of the 482 items received by an internal medicine residency programme, an HMO and a private internist’s office, did not meet the FDA’s requirements. Others failed to adequately meet the requirements for ‘fair balance’, or instructions for use, or included unapproved uses. However the article does not clearly outline how they operationalised the FDA requirement for ‘fair balance’. In a similar study Rothermich et al. (2100) suggested that 42% of the advertisements they reviewed from medical journals in 1984, 1988 and 1992 did not appear to comply with FDA requirements. In Wilkes et al.’s very thorough study (18590) published as (3690) advertisements were reviewed by experts in relevant fields of medicine and by clinical pharmacists, and 40% of advertisements were considered by at least two reviewers to be not balanced.

In Canada, Lexchin and Holbrook (920) found problems with the methodological quality of studies referenced in drug advertisements, despite the existence of guidelines.

In Australia, Moulds et al., in a letter to the Medical Journal of Australia reported that a review by clinical pharmacologists of advertisements in Australian journals found only 53% of the advertisements to be acceptable. Technical breaches of the Australian Pharmaceutical Manufacturers’ Association Voluntary Code of Conduct were found in 22% of the advertisements (320). Carandang and Moulds (3700) assessed 127 advertisements found in four issues of Australian medical journals against the Association’s code. Forty percent were judged to be technical breaches, and these had increased since previous surveys. Seven percent of the advertisements made unacceptable claims, fewer than in previous surveys, but the difference could have arisen from use of different judges rather than being due to a real difference over time.

In New Zealand, Lexchin (6410), in a letter to the editor, reports that only around half of the non-reminder advertisements in four 1987 issues of the New Zealand Medical Journal contained each of the kinds of information required by the PMA Code of Practice. In a small study Peacock (15330) found that many advertisements did not comply with the voluntary code of the Researched Medicines Industry (formerly the PMA) or the Medicines Regulations.

These studies all show high levels of non-compliance with codes of practice and guidelines about promotional material, indicating that these are poor ways of controlling promotion. However they did not study the impact of changing codes or regulations, nor do they provide evidence (from different countries or different medicines) of the advantages or disadvantages of different forms of regulation. Three studies that do this are Sencan and Ustel (5840), Najman et al. (1580) and the Canadian Drugs Directorate (18530).

Sencan and Ustel (5840) examined the introduction of Turkish regulations that required advertisers to include basic information about medicines in advertisements. They found a considerable increase in the percentage of advertisements in a Turkish medical journal that contained basic information, after the introduction of the regulations. In 1990, the year before the regulations were introduced, only 41% of advertisements included adverse effects, while in 1991, 60% did. The figures for contraindications were 37% and 57%, for contraindications 26% and 51%, and for price 1% and 46%. The proportion of advertisements from foreign countries increased significantly (58% to 89%).

A particularly useful study by Najman et al. (1580) investigated the impact of legislative or voluntary codes of practice on advertisements in three countries. In the US, legislative requirements have been in place since 1961, while in the UK and Australia pharmaceutical manufacturers have voluntary codes of practice. Comparing advertisements in the major medical journals of those three countries for 1961, 1967, 1973 and 1977 Najman et al. found it hard to see evidence of the impact of the voluntary codes. Advertisements in the USA outlined the dangers of medicines much more than those in the UK and Australia. Fisherow (18250) notes that the FDA has considerable power in its interactions with companies over advertising, because it is also the regulatory body which decides whether to allow drugs onto the American market.

A Canadian report (18530) looked at how often magazine advertisements for non-prescription drugs complied with regulatory requirements. This is interesting because these advertisements are not subject to mandatory pre-publication clearance, and are monitored by a complaints-only system. The study found an extremely high level of non-compliance with the requirements. Of the 51 advertisements only 37% complied fully, and 39% contained major violations.

CONCLUSION: Government regulation of advertising can be more effective than voluntary regulation.

4.2. The ‘Fair balance’ requirement

In the USA Kefauver-Harris Amendment of 1962 required that prescription drug advertisements contained a fair balance of both negative and positive information. This resulted in the inclusion of a ‘brief summary’ of side effects, warnings and contraindications. For a good history of the regulation of DTCA in the US, see Wilkes et al. (21540). There is a surprisingly large amount of research on the possible effect of the ‘fair balance’ requirement, of which the high quality work was done by Louis Morris from the FDA. Some of the research on this topic looks at advertisements for consumers, some at advertisements for doctors.

In three articles Morris and different co-authors look at ways of conveying risk information to consumers in DTCA. This kind of work is important for working out how best to regulate DTCA, what kind of risk information should be presented and how. Both papers describe aspects of a study that experimented with different forms of presenting risk information in mock DTCA advertisements for two fictitious prescription drugs. Morris et al. (18240) focused on TV advertisements in which the amount of information, the content (either general or specific), the format (video, audio) and the placement of information about risks were varied. Six hundred and seventy six people, (of whom 50% had one or more of the conditions which the drugs treated), viewed the advertisements. They found that more risk messages were recalled if more were included in the advertisement, if these were more specific, if they were both written and spoken, and if they were spread out throughout the advertisement. When people remembered high levels of risk information they tended to have less knowledge and awareness of the benefits of the medicines. Morris et al. (4150) outlined results from both the TV advertisements and magazine advertisements. Presenting the entire patient package insert in magazine advertisements led to lower knowledge of risks than advertisements that included the information in other forms. Morris et al. (21870) also described the results related to both TV and magazine advertisements, and looked at differences among the sample. For example, older people were more positive about both the drug and the advertisement.

Similarly, but in a much smaller and less rigorous study, Larson et al. (3480) tested a small number of doctors’ assessments of the believability of mock advertisements which included or did not include brief summaries. They also looked at whether the country in which the advertisement was found (US or Mexico) affected believability, and found that it did not. Although the inclusion of brief summaries led to statistically significant increases in believability, this was relatively unimportant in explaining the variance in believability. They use attribution theory to explain their results. Larson and Smith (2950) appear to describe the same study. Hurd (3370) described a small study were pharmacy students were asked to rate the credibility and honesty of a sales representative, and whether they would purchase the product, in a series of hypothetical situations. Including negative information about the product did not affect their preference for the product, but they rated the sales representative’s honesty more highly.

Cady and Larson (3310) attempted to find out whether the inclusion of a brief summary made ads more believable. They used advertisements for two mock products and asked students to rate their believability with and without the brief summary. The major claim in one advertisement was more believable in the version with the brief summary, but the results were somewhat equivocal overall. Schommer et al. (3340) assessed what patients had learnt after viewing an DTCA advertisement which included both positive and negative information about an antihistamine. The study used patients waiting for appointments at a clinic, and only 34% of those approached agreed to participate. The participants were confused by the claim that the antihistamine gave ‘non-drowsy’ relief, and the statement that about one in a hundred people experience drowsiness. The authors suggest that it may be confusing to have both risk and benefit information in the same advertisement. (However these claims do seem to be contradictory and the viewers’ confusion seems entirely appropriate). Tucker and Smith (3360) asked 192 people in a shopping mall to view four different versions of an advertisement for a fictional influenza vaccine. Those which included risk information were rated as more ‘appealing’ than those without, but those with no warning or a general warning (all medicines have side effects, please consult your doctor), were rated as providing more ‘security’.

It is hard to interpret the practical significance of these studies. While Morris et al. usefully explore different formats for conveying risk information many of the other studies simply compare advertisements with and without risk information.

CONCLUSION: Communicating risk information to achieve “fair balance” requires great care and testing.

4.3 Guidelines for sales representatives

There is surprisingly little discussion in the literature about attempts to regulate the behaviour of sales representatives. This suggests that, compared to print and broadcast advertisements, sales representatives’ activities are more difficult to document and study. That also makes them more difficult to regulate.

In Australia the Australian Pharmaceutical Manufacturers Association has a code of conduct covering sales representatives. Roughead et al. (460) look at whether sales representatives in Australia conform to this. Although the code does not state what kind of information sales representatives must provide, it does insist that their presentations be current, accurate and balanced. Roughead et al. recorded and analysed meetings between sixteen sales representatives and seven GPs. These included 33 presentations of prescription medicines. They found that omission of risk information was common, and that adverse reactions and interactions were mentioned only in statements that minimised the risk of the product being detailed. Thirteen of the 16 presentations included at least one inaccuracy, and four mentioned unapproved indications. This is a really useful study and a simplified version of the method could form the basis of a system for routine monitoring of the quality of representatives’ presentation. A fuller account of the study is available in (15910).

In France a network of volunteer general practitioners and specialists monitor the activities of sales representatives. After s/he is visited by a sales representative, each doctor completes a questionnaire on whether the indications and dose regimen given by the sales representative matched the Summary of Product Characteristics (as they are required to); whether contraindications, precautions for use, interactions and adverse effects were mentioned by sales representatives; and the arguments and incentives used. The completed questionnaires are analysed and a summary published in La revue Prescrire. Prescrire International (16600, 16630) has discussed these findings in English. This is discussed further below (in ‘Monitoring/countering promotion).

At the practice level, Becker et al.’s ethnographic study (21620) found that practices with policies and guidelines about when sales representatives could visit appeared to find interactions with them more useful and less intrusive.

CONCLUSION: Studies of promotion by drug company representatives suggest that the guidelines and regulations that should control them are not effective.

4.4 Guidelines for post-marketing surveillance

In the UK in 1988 the Committee on Safety of Medicines, the Association of the British Pharmaceutical Industry, the British Medical Association, and the Royal College of General Practitioners agreed voluntary guidelines on post-marketing surveillance. These relate to observational cohort studies, in which doctors prescribe drugs under normal conditions. Under the guidelines, information about these studies, and the results, must be submitted to the Medicines Control Agency. Waller et al. (4310) reviewed the studies that had been carried out under these guidelines. They found that studies were characterised by weak designs and had considerable difficulties in recruiting participants, and suggest that they could make only a limited contribution to assessing the safety of new medicines. Others have suggested that in reality these studies are a promotional tool (3220, 3400).

CONCLUSION: The only regulatory system for post-marketing surveillance that has been studied has not been successful.

4.5 Guidelines on conflict of interest in research

The previous review described a number of studies that showed that published studies which are funded by manufacturers are likely to support the manufacturers’ drugs. Published work does not always disclose relationships between manufacturers and researchers. Two studies have looked at research institutions’ policies about conflict of interest.

Lo et al. (22150) looked at ten medical schools that received the largest amounts of National Institutes of Health funding in the US. They found wide variation and significant limitations in policies about conflict of interest in clinical trials. Five universities required disclosure of all possibly conflicting financial interests, independent of value. Universities varied in whether their policies covered non-faculty research staff. The authors recommend that university-based investigators and research staff be prohibited from holding stock, stock options, or decision-making positions in a company that may reasonably appear to be affected by the results of their clinical research.

McCrary et al. (22140) surveyed medical schools and other research institutions that received over $5 million annually from the National Institutes of Health or the National Science Foundation, 48 journals and 17 federal agencies. They found five medical schools and ten other research institutions had no policy on conflicts of interest. There was marked variation amongst the rest. Less than half of the journals (43%) had policies requiring disclosure of conflict of interest. The management of conflicts and penalties for nondisclosure were almost always totally discretionary. Only three institutions required financial interests to be disclosed to research subjects.

Many authors have called for researchers to explain clearly sponsoring company involvement in clinical trials (eg, 21770). There have also been calls for a prospective register of all trials to be set up, before the results are known. This would reduce the problems of bias in published findings. Stern and Simes point out that this should not be difficult in countries where ethical approval is required for trials, because the register could be added on to this process (21680).

CONCLUSION: many organisations, including many medical schools, research institutions and medical journals lack adequate policies for dealing with conflicts of interest. There is a strong case for all trials to be listed on a public register at the time they are set up.

4.6 Guidelines for package inserts and compendia

It seems that in the US, FDA control over the content of Physicians’ Desk Reference (PDR, the commonly used compendium of prescribing information) leads to greater inclusion of important information. Alloza and Lasagna found that compared to compendia in Spain, Brazil and Mexico, the PDR included three times as many precautions as the mean for the other three countries, and four times the number of adverse effects. Their study focused on anti-inflammatory drugs marketed in the four countries (5750).

From 1976 the Nigerian government required drug manufacturers to produce leaflets and package information to be included with medicines. These were to be inspected before drugs were registered. Osifo (5640) compared the content of package inserts found in 28 prescription drugs obtained from pharmacies in Benin City with the American PDR. Those for the four products from American firms were the same or similar to the PDR entries. Those from foreign subsidiaries and affiliates of US firms included more indications and fewer precautions. Osifo suggests that the Nigerian health authorities have failed to adequately enforce the drug labelling controls.

A major study by the Office of Technology Assessment of the US Congress (15620) found that the label and package inserts for at least half of a sample of 241 products sold by US-based companies in four countries--Brazil, Kenya, Panama and Thailand--failed to provide sufficient information for doctors to use the drugs safely and effectively.

CONCLUSION: Product inserts tend to be more informative in the USA than in other countries. In Africa product inserts produced by US-based companies tend to be more informative than those produced by companies based in other countries.

4.7 Guidelines about gifts

In the USA the American Medical Association has guidelines about gifts from the pharmaceutical industry incorporated in its Code of Ethics. These suggest that gifts to doctors should primarily benefit patients and should not be of substantial value. The American College of Physicians also suggests that a useful criterion for determining acceptability is whether doctors would “be willing to have these arrangements generally known”.

In Gibbons et al.’s study of attitudes to gifts from the drug industry to doctors, only 62% of doctors were aware of any guideline about accepting gifts (3790). Awareness of a guideline was the only predictor of doctors reporting that gifts were not appropriate. However these are self-report data, so those who knew about guidelines may have felt more social pressure to say that gifts were unacceptable.

Drug samples, although not intended as gifts to doctors, may in fact be used in this way. Westfall, McCabe and Nicholas (3850) found that in their family practice residency almost all staff, including medical practitioners, office staff etc had used samples provided by drug sales representatives for their personal or family use. The total retail cost of these was over US$10,000. As a result of their findings they instituted new controls over access to the drug samples.

CONCLUSION: Not enough is known about the impact of guidelines for gifts to reach any conclusions.

4.8 Guidelines for trainee doctors and for hospitals

As noted before, much North American research has been on trainee doctors. Many studies have looked at codes of practice that regulate the contact between trainee doctors and the industry.

Mahood et al. (1400) surveyed all sixteen Canadian family medicine training programmes about their policies and practices about contact with the pharmaceutical industry, and what they taught in their curriculum about relationships with the industry. Four programmes had formal policies about interaction with the industry, and one had a policy on research involvement only. Thirteen of the programmes taught critical appraisal of industry products and claims. One programme did not allow sales representatives in training units, and one only allowed contact with sales representatives if faculty members or pharmacists were present. All programmes had some degree of industry sponsorship. Keim et al. (1060) surveyed 80 directors of US emergency medicine training programmes. They found 61% of programmes had specific guidelines to limit interaction between trainees and students. In 34% such interaction was not allowed in the clinical department. Only 17% of programmes allowed unrestricted group presentations by sales representatives to students. Bucci and Frey (4430) surveyed directors of family practice residency programmes in the US in 1989. They found 20.5% of their respondents included methods to evaluate material provided by sales representatives in their curricula, 79% limited contact between sales representatives and faculty and residents, and about 30% had printed guidelines for sales representatives. In 1991 US family medicine residency programmes were again surveyed, by Brotzman and Mark (4480). They examined a range of promotional activities and found that for each, most programmes allowed it within informal guidelines. Fifty-eight percent of programmes had a written policy about at least one aspect of promotion. They concluded that programmes were more concerned with regulating access to and information from sales representatives, rather than gifts.

In two articles (21570, 3430) Thomas describes a survey of pharmacy directors from 446 US hospitals about their hospitals’ policies towards drugs sales representatives. Sixty-seven percent of hospitals had written policies about drug representatives. Policies included the need for representatives to make appointments, to sign in at each visit, limitations on the areas in which they could detail their products, and controls on samples. Some hospitals used fines or loss of exhibit privileges as penalties for those who broke guidelines. Between late 1983 and 1986 about a fifth of hospitals increased their restrictions and many anticipated more restrictions. Most appeared to be motivated by a desire to contain costs. Small hospitals were less likely to have increased their restrictions, but changes were planned across all sizes and types of hospitals. These were most often restrictions on the products which sales representatives could detail and sample, and the people who they could have business contact with.

Jewesson and Herer (940) describe the Vancouver Hospital and Health Science Centre’s (VHHSC) policy to restrict and monitor the activities of sales representatives at the hospital. Representatives are required to register on arrival and departure, wear an identification badge throughout their visit; and record the purpose of their visit, who they will contact at the hospital, which company they represent, and copies of any literature they distribute.

CONCLUSION: Guidelines for regulating contacts between drug companies and medical trainees vary greatly between institutions.

4.9 Knowledge of these guidelines and their effect on attitudes

Only 23% of family medicine registrars surveyed by Sergeant et al. (980) had read the Canadian Medical Association’s (CMA) guidelines on appropriate interactions between doctors and representatives. Although six of the seven training centres had policies or referred to guidelines, only 26% of residents thought their institution did. Although accepting a private dinner provided by a company contravenes the CMA guidelines, residents who knew about the guidelines or were aware that their training centre had a policy about this, were no less likely to say they would accept such a dinner. However at one centre where company sponsored lunches were not allowed, and where there was a teaching session on interacting with industry representatives, residents were less likely to see the literature from sales representatives as useful, were less likely to say they would accept the private dinner or see representatives when they were practising.

Brotzman and Mark (3640) also looked at whether residency programmes’ policies on drug sales representatives affected the attitudes of residents in the programmes. They randomly selected fourteen US family medicine residency programmes, of which seven had written policies and at least one restriction on the activities of sales representatives. Residents in programmes with no policy were four times as likely to see detailing as a helpful source of information, and twice as likely to see journal advertisements as helpful, as residents in programmes with policies (the latter difference was not statistically significant). Those in programmes without policies felt gifts were more acceptable, and had more interactions with sales representatives. Brotzman and Mark note several possible reasons for this association between policies and residents’ attitudes. Residents may interpret the absence of a policy as approval by the programme directors of promotional activities; programmes may differ in their culture regarding promotion; residents in programmes without policies may be exposed to more promotion, be more used to it and therefore more accepting. In addition programmes with policies may also include other interventions which shape residents’ attitudes. However, when Brotzman and Mark removed from the analysis the three programmes with an explicit curriculum in this area, the results remained the same.

Ferguson et al. (2300) asked practising internists (internal medicine specialists) in the US whether they had trained in an institution with an enforced policy about sales representatives, and whether they now saw sales representatives and accepted samples. They found no differences in whether doctors now saw sales representatives and accepted samples between those who had trained in an institution with an enforced policy and others. This suggests that any impact of policies during training may not persist over time.

McCormick et al. (21420) reached different conclusions. They compared doctors who had trained at a Canadian medical school which did not have a policy restricting sales representatives’ access to residents, with those who trained at another medical school which did have such a policy. They also compared doctors who had trained at the second school before the policy had been introduced. Those who trained while the policy was in force were less likely to find information from sales representatives beneficial in guiding their practice than other doctors several years after they graduated. The authors speculate that this could be due to the educational environment, to strong faculty opinions, or the doctors never having learnt to interact in a constructive way with sales representatives.

CONCLUSION: There is conflicting evidence about whether guidelines affect the attitudes of trainee doctors and if so whether any effects persist over time. Guidelines alone seem to have no strong influence on the attitudes of trainee doctors, but can be effective together with active faculty support in an academic setting.

4.10 Education about promotion

Several authors describe programmes to teach health professionals how to interact with sales representatives and interpret promotional information. Palmisano and Edelstein (3470) briefly describe a seminar programme used for a range of health professionals. This included a simulation of a sales representative’s sales pitch. Anastasio and Little (670) describe a programme for family practice residents which aimed to improve their ability to obtain useful information from sales representatives. Over three hours, students were taught skills for controlling interactions with sales representatives, and for critically analysing promotion, and they discussed ethical issues. Students had two appointments with sales representatives to practice their skills, and these were watched and evaluated by others. Students rated themselves as more confident in their skills after the course.

Kelcher et al. (1020) describe a programme to provide family medicine residents with a structured approach to dealing with sales representatives. This consisted of a one-hour educational seminar, and five visits from sales representatives. The sales representatives were asked to give their presentations according to a structured plan and during their presentations the residents completed evaluation forms and discussed the advertising techniques used. These were followed by discussions with preceptors where the residents discussed what they had learnt about the drugs. According to questionnaires most of the small number of residents (15) who had completed the programme thought it should continue. They authors argue that this approach is better than restricting residents’ access to sales representatives. The project was funded by two drug companies.

Hopper et al. (3810) measured the effect of a 40 minute lecture and discussion session about pharmaceutical promotion, on primary care and internal medicine-paediatric residents. Surveys were completed three weeks before and four weeks after the teaching session. Very small numbers of people were involved. After the intervention residents were more likely to believe that sales representatives may use unethical marketing practices, that gifts with no patient benefit may be inappropriate, and that other doctors’ prescribing habits may be negatively influenced by accepting gifts. This study suffers from very small numbers of participants. It is also unclear whether these reported attitude changes are likely to persist over time.

Suryawati and Santoso (17550) described the effect of a teaching module for medical students in Indonesia. This consisted of a one-hour lecture on commercial and non-commercial sources of drug information, the WHO Ethical Criteria for Medicinal Drug Promotion, and examples of misleading promotion. This was followed by a two hour workshop in which participants critically assessed advertisements from local medical journals. Participants in the course, and two control groups, completed tests that involved assessing 10 different advertisements. The intervention group significantly improved after the course and this effect was still apparent after 12 months. One control group, who were part of the same class, improved somewhat after 12 months, probably due to cross-contamination, and the other control group performed poorly at the test 12 months after the course.

Shaughnessy et al. (1910) reported on an educational intervention for resident doctors in the US. This centred around a Pharmaceutical Representative Evaluation Form which residents used to evaluate presentations by sales representatives. The form included the completeness of the presented information, the techniques of persuasion used, and the use of rational and irrational appeals. Using a modified version of the questionnaire developed by McKinney et al. (2480) the authors found that a year after the programme residents were more likely to disagree that sales representatives and gifts have no impact on prescribing. Similarly, Vinson et al. (660) found a change in students’ attitudes after a lecture on concerns about pharmaceutical marketing practices to second year medical students. However this change was measured by questionnaire only seven weeks after the lecture.

Wilkes and Hoffman (22210) describe an educational programme in which university pharmacists portrayed pharmaceutical company representatives to model a promotional presentation, that they designed to generate critical thinking among third-year medical students about the influence of pharmaceutical representatives on physicians’ prescribing practices. The authors found that the programme increased the uncertainty many students felt about the accuracy and ethics of standard drug "detailing." Compared to questionnaire responses that students provided before the exercise, the attitudes they expressed in course assessments completed 12 weeks after the session revealed much more uncertainty about the ethics and value of interactions with drug reps, and the number who stated that they wanted to interact with drug reps during their residency fell from 86% to 61%. This is potentially a valuable approach for inoculating medical students against some of the worst potential consequences of biased drug detailing presentations.

In some of these studies it is difficult to determine whether any reported changes in attitudes after educational sessions are due to actual changes in attitudes, or to participants perceiving it to be less socially appropriate to express their original attitudes. However this is less of a problem in the Suryawati and Santoso study, because they tested participants’ skills at assessing advertisements, rather than their attitudes. This study provides a useful model and convincing evidence of its effects.

CONCLUSION: Education about promotion appears to change attitudes and can improve skills. The impact of education about promotion on prescribing has not yet been tested.

4.11 Monitoring/ countering promotion

The Medical Lobby for Appropriate Marketing (MaLAM), based in Australia, asked people around the world to send complaints about medicines advertisements. MaLAM chose particularly serious complaints, and prepared a letter to the advertiser, quoting the advertisement, summarising the literature and asking the company to provide their best evidence to support the claims made in the advertisement. A copy of the letter was then sent to MaLAM’s subscribers (over 700 in 1991) and they were invited to sign and send it to the company. Most, but by no means all MaLAM subscribers were doctors (20200). From 1998, MaLAM produced a series of newsletters called ‘Healthy Scepticism’ for New Zealand GPs (17100, 17120, 17110). Each issue critically appraised advertising for one therapeutic group, by looking at the possible interpretations of the advertisements and marking these as unjustified, borderline, or justified by the evidence. The newsletter was funded by PHARMAC (Pharmaceuticals Management Agency). PHARMAC is the New Zealand government’s drug purchasing agency. It negotiates subsidies for medicines, and is now also involved in attempting to improve medicines use. MaLAM has now changed its name to Healthy Scepticism, and it no longer sends letters critiquing specific advertisements.

Two papers report on the success of MaLAM’s letters to companies requesting evidence to support claims made in advertisements (4710, 5880). Wade et al. (4710) report on Jan 1988 to June 1989. In this time ten companies were contacted about seventeen products, and fifteen replies were received. In four cases the advertising was changed, and in one the marketing of the product in developing countries was discontinued. Mansfield (5880) reports on the July 1989 to June 1990 period. Eight companies were contacted regarding nine drugs. No response was received from two companies, two said they would withdraw the claims, one that they would withdraw the indication, one that they would reformulate the product, and two that they would withdraw the products. In (20990) MaLAM reports on Hoechst’s discontinuation of Baralgin, presumably in response to a 1994 MaLAM letter. MaLAM operates on a very small budget, but seems to be very effective in lobbying companies to change the worst examples of their advertising.

The Prescrire Network (described above) is also an attempt to monitor promotion. They have found that the legal requirements for sales representatives in France (that Summaries of Product Characteristics, and Transparency commission reports be given to doctors) are frequently not met (16630). The Prescrire network has identified many problems in presentations given by sales representatives to individual doctors, but there is no evaluation of the impact of the network itself.

Berings et al. (4600) report on an intervention to improve the prescribing of benzodiazepines by providing industry-independent information. They randomly allocated the 128 doctors in the study to three groups. One received no information, one received a pamphlet on rational prescribing of benzodiazepines which was similar in format to drug company brochures, and one received the pamphlet and oral information from a visiting GP. Throughout the study benzodiazepine prescriptions were recorded on specially designed prescription forms. During the four week follow-up period the groups which received both written and oral information prescribed 24% less benzodiazapines, those who received written information only prescribed 14% less, and the control group prescribed 3% less. There was no significant increase in the use of other psychotropics: in fact in the experimental group these also decreased. Forty of the 44 doctors who received the visit said they would like more of these in the future.

4.12 Research as an intervention

Several authors, notably Milton Silverman, Philip Lee, and Mia Lydecker have published descriptions and analyses of promotion and its effects, which appear to have been instrumental in improving the quality of promotional material.

In 1976 the Silverman team published The Drugging of the Americas, which compared promotion of 40 products by 12 companies in the US and Latin America. Looking at standard, widely used drug compendia, they found that promotional claims were exaggerated, and warnings were limited, minimised or entirely ignored. The findings are summarised in (10480), which also notes the enormous media coverage the book received around the world. In 1982, Prescriptions for Death (16240), reported similar research in a wider range of countries: the USA, the UK, Africa, Indonesia, Malaysia, Singapore, Philippines, Latin America. Similar results were found. The groups’ 1982 article (10510) reports no readily apparent differences between US and other multi-nationals, multi-nationals and other domestic firms, brand and generic companies, or companies from the capitalist or the socialist block. Fieldwork for the third study was carried out in 1984, published in book form in 1986 and summarised in (10540). This involved 63 drugs, 1069 different products, 303 companies and 30 countries. They found noticeable differences between the results of their earlier work and the situation in 1983. Companies showed more restraint in describing the value of medicines and more willingness to disclose potential hazards. However, problems still existed, particularly in Latin American countries. The 1992 book ‘Bad Medicine’ (16000) presents a more positive picture. Fieldwork for this was done in 1987/8, and included 40 drugs, 1500 products, 400 companies, in the US, the UK and 74 developing countries. The authors concluded that most multinationals had improved considerably by the late 1980s. Local and domestic firms were now mainly responsible for inaccurate promotion. Silverman et al.’s books show a clear improvement over time in promotion in developing countries, and it likely that the books themselves, by drawing international attention to the topic, have been at least partly responsible for this improvement.

CONCLUSION: Publication of descriptions of deceptive promotion can lead to improvements.

Directions for future research

Greater linkage between research, interventions and evaluation is needed. Those planning interventions need to draw on previous research for designing and targeting their programmes. For example, previous reviews have suggested that some doctors rely heavily on promotion, and that their prescribing is also sometimes irrational. Interventions targeted at these doctors are likely to have a greater impact than those targeted at doctors in general, or particularly interventions which include volunteers only (likely to be those who are already sympathetic to rational prescribing messages). Interventions also need to be evaluated, and these evaluations need to be published so that others can learn from them. Reasonable follow-up times are needed, to show whether the effects of interventions persist over time.

Studies are also urgently needed comparing the effect of different regulatory frameworks. Najman et al.’s study (1580) was the only one included here which did this. Governments and others introducing policies to regulate promotional activities need good evidence of the advantages and drawbacks of different systems.

 

See Index | Date Entered : Monday, June 02, 2003