Hypertension is a major risk factor for cardiovascular disease [1] and the most common single problem managed in Australian general practice. [2] For more than a decade expensive new antihypertensive drugs have been prescribed more frequently than the older and more cost effective thiazide diuretics. [3–6] Newer antihypertensive drugs are among the highest volume and cost items for the Australian Pharmaceutical Benefits Scheme (PBS). [7] It is estimated that adherence to guidelines favouring older, less expensive agents in 1998 in Australia may have saved $45 to $108 million. [3] These guidelines promote antihypertensive drugs to augment lifestyle change, but not as a substitute for such change. [1, 8, 9]

The content of print advertising for antihypertensive drugs mirrors the trends away from cost-effective prescribing. [10] There is observational evidence of associations between doctors' prescribing and doctors' exposure to advertisements. [11–16] Thus, heavy promotion may be at least partly responsible for the more frequent prescribing of newer drugs. Indeed, recognition that marketing influences prescribing has prompted government-funded social marketing campaigns to encourage more cost-effective choices. [17] Doctors may not be aware of the degree to which their prescribing is influenced by advertising. [18]

The Medicines Australia Code of Conduct states that all promotional information "must be current, accurate, balanced and must not mislead either directly, by implication, or by omission". [19] The Australian National Medicines Policy (NMP) states that each partner (including the medicines industry) "accepts that all must be engaged in a cooperative endeavour to bring about better health outcomes for all Australians, focusing especially on people's access to, and wise use of, medicines." The NMP definition of "quality use of medicines" includes taking into account "the potential risks and benefits of treatment, dosage, length of treatment, and cost."[20] We believe evidence-based guidelines are a useful standard for evaluating how well advertisements support, rather than lead away from, wise use of medicine. Thus, the aim of our study was to determine whether print advertisements for antihypertensive medications in Australia promote prescribing for hypertension that is concordant with evidence-based guidelines. To answer this, we reviewed evidence-based guidelines for important prescribing messages, and then looked for these messages in a sample of advertisements from Australian medical publications.

At the time of performing our study, we were unaware of any other similar studies regarding messages in advertisements for antihypertensive medicines. Since then a study of antihypertensive drug advertising in a Dutch journal has been published. This study found 35% of the advertisements contained claims unsupported by evidence. [21] Our research builds on previous studies of advertisements and promotional brochures which have found overemphasis on relative statistical measures[22, 23], and claims lacking clarity, accuracy, balance and substantiation. [22–25]

Using these key messages from hypertension management guidelines, we developed a quality checklist to extract data from the advertisements (see additional file 1: checklist.pdf). The checklist was used to measure promotion of lifestyle modification, promotion of thiazides as first line agents, promotion of cardiovascular risk assessment, promotion of use of the drug for specific subgroups of patients, and the mention of alternative antihypertensive classes as treatment options (these issues are all drawn from the quality prescribing messages in table 1). We regarded advertisements as not fully promoting guideline-concordant care if they did not mention all of the items on the check list. We also counted the frequency of mention of drug harms and drug prices, because those factors are included in the NMP definition of quality use of medicines. Finally, we assessed statistical claims, because use of relative rather than absolute measures may have a stronger influence on doctors' prescribing,[36, 37] and previous research has demonstrated a lack of reporting of absolute measures in pharmaceutical advertisements and brochures. [22, 23]

We sampled advertisements for antihypertensive drugs from four advertisement-rich, general practice-oriented Australian medical publications. We drew these from issues published between February and July of 2004, inclusive. We examined every issue of the monthly publications (Australian Family Physician and Medicine Today) and one randomly selected issue per month of each of the weekly publications (Australian Doctor and Medical Observer). The random selection was based on randomly-generated integers from a web-based random number generator. [38] Our analysis included the largest advertisement (by page area) for each antihypertensive drug in each issue. Where two or more equally sized advertisements for the same drug existed, both or all were included. We excluded "short" advertisements (as defined by the Medicines Australia Code of Conduct). [19]

We evaluated text and graphs in the advertisements, but not imagery (thus, pictures of sports shoes, people swimming, and so on were not counted as promotion of physical activity). We examined each entire advertisement including contiguous "fine print", but we did not assess further information on other pages of the journals. (Only six advertisements in our sample of 113 referred readers to such non-contiguous fine print.) For some items, we stratified our findings according to whether information was present in the "main body" of the advertisement; for these purposes we excluded the "fine print" and "PBS Information" sections of the advertisement (see additional file 1: checklist.pdf).

Each advertisement was rated independently by two authors, with disagreements resolved by consensus. By designing a checklist which was largely "yes or no" in available responses, we tried to minimise subjective interpretative difficulties. Inter-rater reliability was not formally calculated. The few inter-rater differences we found were usually due to oversights, and consensus was easily achieved. Descriptive results in the form of counts and frequencies were calculated using Microsoft Excel.

In our paper, we use the term "thiazide class drug" to refer to traditional thiazides (bendrofluazide and hydrochlorothiazide) as well as the other "thiazide-like" low-ceiling diuretics chlorthalidone and indapamide. Guidelines tend to recommend these interchangeably, and meta-analysis evidence suggests they have equivalent effects. [39]

113 advertisements met our inclusion criteria. These were composed of 27 unique advertisement designs which appeared between 1 and 14 times each.

To our question of whether Australian antihypertensive advertisements promote quality prescribing, our results offer a mixed answer, demonstrating some important missed opportunities for the promotion of quality use of antihypertensives.

We found that messages to prescribers about lifestyle interventions were almost absent from the advertisements. This bias towards promotion of drugs rather than lifestyle change is problematic because lifestyle interventions should be first line treatment for many hypertensive patients, and the foundation of ongoing treatment even when medications are used.

Much information that is important for quality care was underemphasised or not provided by the advertisements. PBS costs of medications were frequently reported in small print but cost comparisons within and between drug classes were missing. In most advertisements, information about harms was confined to the fine print. There was a lack of emphasis on assessing and modifying patients' overall cardiovascular risk.

The promotion of medications for specific target groups may help the individualisation of therapy for the patient in an evidence-based manner. However, advertising claims based on data from higher-risk patients may artificially heighten a prescriber's sense of a drug's efficacy for patients at lower risk. Similarly, the over-emphasis on relative rather than absolute statistics may create over-optimistic impressions of efficacy and thus encourage over-prescribing. [36, 37] Also, if doctors and patients perceive drugs as very effective, then they may overlook the need for lifestyle change.

The absence of advertisements for beta-blockers may be helpful, given that recent meta-analyses show less favourable outcomes for beta-blockers (particularly atenolol) than other antihypertensive classes. [41, 42] (While these meta-analyses were not published at the time of our sampling of advertisements, most of the primary trials on which they were based were published.)

In our sample, thiazides were the most advertised class, contrasting strongly with a previous study which found that thiazide advertising in the New England Journal of Medicine declined from 4.2% of advertising pages in 1985 to 0% in 1996. [10] In our study, the most frequently advertised thiazide was indapamide. This drug is priced almost twice as high as some other equally effective thiazides. The only other thiazide advertised was hydrochlorothiazide, which was only advertised in combination medications. Clearly, the advertisements do not encourage the prescribing of the most cost-effective drugs as a first-line monotherapy.

Our study has several limitations. The validity of our results may be limited by the number of advertisements surveyed. The generalisability of our results is uncertain because we focused only on advertisements in Australia during 2004. We did not examine the role of imagery and emotional appeals in the advertisements, because it is difficult to assess these objectively. We did not attempt to assess the accuracy of therapeutic claims made in the advertisements. We also did not attempt, for those advertisements promoting antihypertensive agents for particular patient subgroups, to determine whether a particular class of antihypertensive would have been the most appropriate choice for that particular subgroup (in fact, such issues are often contested in the literature).

We used evidence-based guidelines as a gold standard for evaluating advertisements. The industry may advocate acceptance of lower standards but we believe higher standards for pharmaceutical promotion are required by the Australian NMP.

Guidelines may not be perfect. Sometimes recommendations vary between guidelines. The guidelines selected for our study varied in their recommendations about first-line drugs so our decision to favour thiazides is open to discussion. Guidelines may be biased by vested interests including drug companies. [43]

Our study describes the content of the advertisements, but is not designed to show association or causation between exposure to promotion and prescribing behaviour. Our study should be read, though, in the context of existing observational evidence of such associations. [11–15, 44–46]

Despite these limitations, we believe our results indicate that many Australian print advertisements for antihypertensive medications lack elements important for cost-effective care consistent with evidence-based guidelines. Our findings lend support to calls for increased availability of independent prescribing information, reform of the incentives for, and regulation of, pharmaceutical advertising,[47] and avoidance of promotional information by prescribers. [48] These results also support calls for medical publications to be helped to develop alternative business models rather than relying on drug promotion. [49]