The increasing skepticism about drug development, clinical trials, and publications goes in parallel with the popular view on pharmaceutical industry, which is far from unbiased. It has been reported that people see pharmaceutical companies as business corporations with low ethical standards such as the arms industry. The paradox is that never in history there were as many regulations, constraints and supervision of drug development, approval, and marketing strategies as nowadays. Hence, this climate comes up from mistakes made by several agents in this drama: the pharmaceutical industry, of course, but also opinion leaders, medical journals, regulatory bodies, politicians, and even clinicians. The raise of evidence-based medicine may also be partly responsible, because evidence is only available for questions receiving funding, and most of the funding comes from companies expecting refunds from their investments. Some of us believe that this would be fair as far as strict regulations and public funding are able to counteract against the risk of relying almost exclusively on company-sponsored evidence-based medicine. Otherwise, evidence-base may become evidence- bias.

Bias may come from two main sources: biases in trial design, and biases in results dissemination. Trial design biases are easier to counteract: for instance, regulatory bodies as the Food and Drug Administration (FDA) in the United States or the European Agency for the Evaluation of Medicinal Products (EMEA) in Europe have set their own trial design guidelines for marketing approval (1,2). This strategy has been successful to avoid marketing of potentially ineffective or unsafe drugs via sophisticated designs or statistical analyses. However, those trials have characteristically high internal validity but poor external validity (3), providing little information about the use and effectiveness of a given drug in clinical practice. Non-regulatory trials, on the other hand, may be more generalizable, but are commonly biased in favour of the novel (expensive) drug. Examples of trial design biases include false non-inferiority designs, enriched designs, underpowered comparator samples, unfair comparator doses, inclusion of patients who are non-responsive to the comparator, unfair rescue medication rules, and "creative" outcome measures favouring the drug of choice. Biases in results dissemination are more difficult to ascertain. The most well-known one is publication bias: positive trials are published while the negative ones remain forever as "data on file", or at most they are presented at a small meeting as a poster or shown at a website in a very concise format. Conversely, positive trials are re-analysed, subanalysed, and repeatedly published and presented at scientific meetings. Publishing negative trials and making that information available to society is not only the responsibility of pharmaceutical companies, but also of researchers and opinion leaders, particularly those who sit on companies' advisory boards and have access to privileged information. Confidentiality rules apply to matters that may have to do with competitive research, but if the companies do not put their negative data in the public domain within a reasonable time period, the rights of the patients who voluntarily participated in those trials are being broken. Other sources of bias not as evident as selective publication include unbalanced presentations, which are especially common at but not exclusive of standalone meetings or satellite symposia: exclusion of the comparator arm in placebo-controlled trials, emphasis on certain (favourable) secondary measures, minimization of adverse events, and many more; again, not only the industry but also academics and speakers should avoid this kind of pseudoscience, and ideally clinicians should be able to identify and criticize those who promote it. Unfortunately, in most countries the only source of continuing medical education (CME) is the pharmaceutical industry, so, even when there are no biases as those mentioned so far, the focus of the meetings is rarely free of indirect or direct financial interest. Here, national governments and scientific societies also have their responsibility. In Europe, CME credits are not well implemented and this carries higher risk of unbalanced meeting programmes and poor attendance to scientific sessions. Finally, the issue of treatment guidelines deserves further discussion: while recusal of experts with potential conflicts of interest would leave them practically orphan of any expertise, the fact is that the evidence- bias discussed above, which has more to do with the dearth of independently sponsored trials than with the potential interests of the advisors, makes them often more supportive to newly marketed drugs than to cheap, old compounds (4,5), such as lithium, which has been decreasingly prescribed in many countries despite the evidence that it is effective and may have unique antisuicidal properties (6,7). Boyd and Bero (8) have recently reviewed the management of conflicts of interest in guidelines development and they recommend a standard policy requiring all financial ties to be made public in advance.

Everything that we discussed so far suggests that something should be done to increase the integrity and credibility of pharmaceutical companies, journals, meetings, authors, and presenters. Fava is right when he says that disclosure is simply the minimal requirement for scientific credibility. Systematic feedback is another advisable practice: I think that every scientific meeting should provide feedback forms including a specific score for scientific balance for every presentation. Scientific societies and editorial boards should have a conflicts of interest advisory committee to discuss the feedback provided by meeting attendees and readers (9). I do not think, though, that we should make the mistake of considering potential conflicts of interest as something necessarily bad. Having a potential conflict of interest does not mean at all that whatever that person says or writes is biased; excluding highly respected academics from editorial boards or meetings just because they have a potential conflict of interest would lead to the total fall of clinical research, with enormous impact on the number and quality of new drugs becoming available for the treatment of most conditions, including mental disorders. Consistently, most research subjects and patients understand and accept conflicts of interest (10) and some of them actually encourage them as far as they may carry benefit for the people (11). The best experts in certain conditions have generally multiple potential conflicts of interest, but at the end of the day what makes them attractive for patients, clinicians, trainees, governments and pharmaceutical companies is their credibility, and credibility is hard to achieve and easy to lose. At the end of the day, conflicts of interest are not always a bad thing: they correlate with interaction between public and private health care providers, and some of us believe that the best health care system is neither purely public nor purely private, but every effort should be made to ensure that the interest of a few does not go over the interest of society, the integrity of clinical research, and the progress of medicine.

Some potential sources of conflicts of interest do not seem to be as popular or scandalous as the ones that come up from the marketing of pharmaceutical companies, even though they may be an important source of bias in education and clinical practice. First of all, not all conflicts of interest are of financial nature; in fact, a very oppositional attitude towards the pharmaceutical business may carry political or professional benefits; others have actually made a lot of money with books reporting "drug compa- nies malpractice". Finally, some financial conflicts of interest that are rarely disclosed are the ones belonging to the public health care sponsors: the government, the local authorities, and the hospital managers, whose interest is generally to avoid spending money on the most expensive drugs. In several countries, physicians get supplementary income if they are able to save money from the pharmaceutical expense; in certain hospitals, expensive drugs are not provided even though they have an approved indication; and in the editorial arena, some government- sponsored publications show a bias against new expensive drugs which is only comparable to the bias of company advertisements in favour of their drug.