Information from drug companies and opinion leadersBMJ 2003; 326 doi: https://doi.org/10.1136/bmj.326.7400.1156 (Published 29 May 2003) Cite this as: BMJ 2003;326:1156
- Alessandro Liberati, associate professor (, )
- Nicola Magrini, director
- University of Modena and Reggio Emilia, Via Campi 287, 44100 Modena, Italy
- Centre for Evaluation of the Effectiveness of Health Care (CeVEAS), Via Le Huratori, 44100 Modena, Italy
Medicines can offer enormous health benefits if choices for treatment are made appropriately, and availability of valid information is a necessary condition. The asymmetry in the information available to health professionals and consumers is a fundamental barrier to rational and informed choice. Good quality information, however, is a rationed commodity for health professionals also, and the use of different standards in its dissemination represents a major determinant of the failure of the therapeutic chain.1 Healthcare systems make limited investments to provide independent information, and pharmaceutical companies—who fund most clinical research—therefore become major players in the dissemination of information to health professionals and the public. Do pharmaceutical companies and the researchers acting as opinion leaders for them behave fairly and consistently or do they adopt double standards when they write in peer reviewed journals and talk to practitioners? We know that this form of information asymmetry exists.2 Two recent examples—a document from the European Federation of Pharmaceutical Companies3 and the debate generated by the ALLHAT study, a landmark trial in the treatment of hypertension—illustrate this danger.4–8
The document from the European Federation of Pharmaceutical Companies identifies 20 diseases and conditions, such as dementia, asthma, hepatitis C, rheumatoid arthritis, some cancers, and osteoporosis, for which “potentially achievable benefits are not achieved.”4 According to the document, this happens because patients are denied access to important therapeutic interventions due to poor diagnosis, limited awareness of patients of effective drugs, and strict cost containment by healthcare systems.
The document is worth reading for the way it is written—it overlooks basic principles of synthesis of scientific information. In 98 pages and 184 citations readers are warned against an alleged underuse of effective drugs. None of the 20 conditions is discussed with reference to systematic reviews. In Clinical Evidence and the Cochrane Library one can find 5–15 systematic reviews for each of the conditions discussed.9 10 For each disease the document quotes only some of the positive studies—negative ones are ignored. This selectivity is not limited to studies of effectiveness and extends to those on appropriateness. Not a single study on overuse—the dimension of quality of care predominantly addressed by the literature—is quoted, and only research looking at underuse is mentioned. In the section on Alzheimer's disease, for example, the document from the European Federation of Pharmaceutical Companies says that with second generation acetylcholinesterase agents patients can enjoy an increase in their quality of life, with massive economic benefits for the society. Only one reference (in German) supports the statement. Available systematic reviews, however, do not support this view, given the inadequate follow up, questionable end points, and lack of data about quality of life in the studies, and conclude that the practical importance of changes to patients and carers is unclear.10
The ALLHAT study
The second example is the different approach chosen by pharmaceutical companies and some opinion leaders in the scientific discussion in peer reviewed journals5 6 and in the rejection of the ALLHAT study's results published in national journals targeted at prescribers and cardiologists.7 8 This double standard represents a predefined strategy aimed at challenging unwelcome results.11 ALLHAT had immediate responses in the Journal of Hypertension, whose editors asked specialists to discuss its methodological strengths and weaknesses.5 The invited commentaries substantially praised the strengths, identifying only minor weaknesses.11–14
Full scale attacks directed at the design, analysis, and interpretation of ALLHAT appeared instead in the journal of the Italian Society of Hypertension (Ipertensione & Prevenzione Cardiovascolare)7 and in a weekly medical magazine.8 According to the Italian editorial, the drawbacks of ALLHAT were a lack of previous experience in research trials in most centres, the short duration of the study, the lack of a washout period before randomisation, and the unusual combination of drugs associated with the experimental ones. The editorial also implied that the ALLHAT investigators had their own competing interests since a priority for the public institutions that funded ALLHAT (National Institutes of Health in the United States) is keeping the costs of care low.7
In the other article the major flaws in ALLHAT were listed as the ineffectiveness of monotherapy (actually 55% of patients were well controlled with monotherapy; the need to add a second drug, which could have favoured chlortalidone; the fact that 90% of the population was already treated with antihypertensive drugs (doesn't this happen in almost all hypertension trials?); that heart failure was not a prespecified end point (it was, and a totally validated one too); and that follow up was too short to show differences that could become important with long term treatments (ALLHAT has one of the longest follow ups among hypertension trials). In its conclusion the article says that ALLHAT's results should not be applied to Italian or European clinical practice, given the countries' low percentage of black patients (who benefit most from a thiazide diuretic treatment).8 Interestingly, in ALLHAT's main report, a subgroup analysis indicated that results did not change according to race.4 However, none of these arguments was present in the letters to the editors published by JAMA a few months after the ALLHAT report.6
These two cases underscore the need for a common set of rules if pharmaceutical companies and opinion leaders want to save their credibility. Firstly, pharmaceutical companies need to agree to collaborate on an explicitly prioritised research agenda where questions related to care are more relevant than drug specific issues. Secondly, a transparent code of conduct against publication bias or selective suppression of information should be developed. Since only a minority of randomised controlled trials gets published, a registry of ongoing studies following the example of the US National Cancer Institute's physician data query database (www.nci.nih.gov/cancer_information/pdq/) is warranted, together with disclosure of investigators' competing interests and their roles in the design, conduct, analysis, and interpretation of the study. Finally, both pharmaceutical companies and opinion leaders should recognise their ethical duty to avoid speaking two different languages—the scientific one in peer reviewed journals and the language stuffed with personal opinions when they speak directly to practitioners or policy makers.
Competing interests AL and NM work in a national health service centre whose main activity is to provide information on the effects of healthcare interventions to health professionals. Their daily struggle for truly independent information and its effective delivery be the cause of their scepticism.