The illuminating discussion by Drs Healy and Thase (Reference Healy and Thase2003) focuses on the magnitude of the pharmaceutical industry's influence on academic medicine. However, this discussion needs to be taken a step further, and evaluated in relation to patient care. From my perspective, the central question is: ‘Does the influence of the pharmaceutical industry on academia result in biased knowledge?’ Professionals are charged with serving the best interests of patients/clients. In order to accomplish this, professionals need unbiased knowledge that can lead to an accurate risk–benefit assessment and serve to guide clinical decisions. If available knowledge is biased, decisions will be affected and clients will suffer accordingly. The frequently touted disclosure of potential conflicts of interest in academic publications is a small step in addressing the much more difficult question of whether existing knowledge is biased. Recognising potential bias is an initial step towards assessing and removing it from the collective knowledge used to make decisions in practice. For example, registering clinical trials is an approach to reducing publication bias (Reference Dickersin and RennieDickersin & Rennie, 2003). Meta-analysis is an approach to removing bias from expert reviews of the literature (Reference BeamanBeaman, 1991), although expert reviews still retain influence in the formulation of some practice guidelines (e.g. American Psychiatric Association, 1997). As the field moves more towards the implementation of evidence-based practice guidelines, the importance of removing bias remains central to providing optimal clinical care. If the extensive financial arrangements between industry and academia resulted in no bias to knowledge, I would probably agree with Dr Thase that no new policies are necessary to ‘safeguard our integrity’ (p. 390). However a recent systematic review and meta-analysis of evidence bearing on this question found ‘strong and consistent evidence... that industry-sponsored research tends to draw pro-industry conclusions’ (Reference Bekelman, Li and GrossBekelman et al, 2003: p. 463). The question now becomes, ‘What safeguards should be implemented to remove this bias from the knowledge that guides clinical practice (cf. Reference BodenheimerBodenheimer, 2000)?’ Commitment to our patients’ well-being requires that we act from this integrity.