Resolving conflicts of interest in medicine

Chris Winchester

At a workshop on managing conflicts of interest in medicine, Chris Winchester, CEO of Oxford PharmaGenesis and co-founder of Open Pharma, argued for openness.  

EBMLive this week was a chance for academics, clinicians, patients and journal editors with an interest in evidence-based medicine to gather in Oxford and review the status of the field. One prominent theme was conflicts of interest (COIs): Margaret McCartney (GP, Glasgow) and Carl Heneghan (Professor of Evidence-Based Medicine at the University of Oxford) led a workshop on developing a COI statement in research. After lively breakout groups, we had the chance to submit our conclusions. This blog post summarizes mine.

What is the most important issue involving COIs in medicine?

The first step in managing COIs in medicine is to define what we mean by a COI. Currently, all the attention is on the conflicting interest, but a conflict needs two sides. We need to define the primary interest because how we define this determines how we identify potential COIs. For example, if we define the primary interest as patient care, then anyone responsible for balancing budgets will have a conflict. If we define the primary interest as maximizing patient care in a socially responsible way, what does this mean for private health insurers and other institutions seeking profit from providing healthcare in the USA? Even in the UK, NHS employees may be considered to have an interest in delaying the availability of high-priced drugs (to help to balance their budgets), or in promoting the effectiveness of flu vaccination and other preventative measures (for which they receive financial incentives). Yet such COIs are rarely discussed.

This may be because a COI is in the eye of the beholder. Indeed, some definitions state that COIs include anything that an observer may think has influenced the judgement of those involved. But who is the observer? There is a danger that definitions of COIs are used to separate an ‘in group’ (‘one of us’, whose views are sound) from an ‘out group’ (‘one of them’ whose motives are suspect and whose views can be ignored). I believe that this has been part of the reasoning for excluding patients from decision-making in the past. Even a seemingly objective definition of a COI, “a situation in which someone cannot make a fair decision because they will be personally affected by the result” (Cambridge Dictionary), can be used to disenfranchise individuals such as patients who are directly affected by outcomes. There is a real danger that, with unclear definitions of COIs and people with COIs not knowing what to disclose, people with legitimate interests could be denied a voice.

What potential solutions can reduce the impact of COIs in medicine?

The obvious way to reduce the impact of COIs in medicine is to exclude people with potential COIs from participating in activities where those COIs are relevant. An alternative approach is to be transparent by making any potential COIs known to those who may need to take them into account. In practice, it makes sense for the two approaches to co-exist. But which COIs are acceptable in which circumstances? I believe that, as well as being defined, potential COIs need to be classified or graded based on type and magnitude (e.g. of potential gain or loss in value) to explain prospectively what is likely to be acceptable in different situations. For example, it seems clear that accepting travel expenses to advise a pharma company on clinical trial design should not preclude someone from being involved in guideline development, whereas owning shares in a biotech company could do. A clear classification system would help to clarify which COIs are acceptable when contributing to a debate, and which are acceptable when making a decision.

What should a COI statement include?

I propose that the COI statement includes clear recommendations based on the responsibilities of the different stakeholders, for example:

  • researchers – disclose COIs prospectively on their institutional website or a platform such as ORCID, and update them at least annually
  • institutions – develop and report on an institutional key performance indicator for COIs that summarizes the performance of that institution’s researchers individually and by department
  • funders – take researchers’ and institutions’ transparency about COIs into account when making funding decisions.

I would like to see a COI statement recognize that engagement between different stakeholders with different interests brings benefits as well as risks. Working together, academics and industry have delivered cures for hepatitis C and Helicobacter pylori infection, and developed vaccines to prevent infection with human papillomavirus (HPV), reducing the suffering and death caused by these Group 1 carcinogens. The treatments now available for patients with rare diseases are another excellent example of progress through collaboration. Given that academic engagement with industry leads to the development of better drugs for patients, we need to destigmatize such engagement by using a less judgemental term (e.g. potential COIs, disclosures or relevant interests). In other words, we need to maximize the benefit–risk profile of engaging with industry, rather than simply minimizing the risk.


Chris Winchester is an employee of Oxford PharmaGenesis Ltd, a Director of Oxford PharmaGenesis Ltd, Oxford PharmaGenesis AG, Oxford PharmaGenesis Inc., Oxford PharmaGenesis Pty Ltd and Oxford PharmaGenesis Holdings Ltd, and a shareholder in Oxford PharmaGenesis Holdings Ltd. He is a member and past Chair of the International Society for Medical Publication Professionals, a member of the European Medical Writers Association and a co-founder of Open Pharma.