Whether it is tests on drug safety and efficacy by drug companies or environmental studies by oil companies, there is a (perhaps understandable) tendency to mistrust the results of science funded by major corporations with a vested interest in the results. Should we automatically assume that studies funded by industry are biased, or is there good science being done?
There are two major reasons why a major corporation will fund a study:
- They have to. For example, a pharmaceutical company can not put a new drug on the market without first doing extensive testing to show that it has a beneficial effect and is “safe” (or at least that any harmful side effects are outweighed by the beneficial effects). This testing costs on the order of 100 million dollars or more, so obviously no one else is going to pay to have it done. Nor should anyone else be responsible – if a company wants to market a new product it makes sense that they should be the ones paying for the testing, and not taxpayers.
- The study is expected to benefit the corporation. This could be for something like regulations which are believed to be unnecessarily restrictive; for example, many environmental guidelines err on the side of conservatism when there are gaps in the knowledge about pollutants or environmental processes; funding science to address these gaps might cost the corporation less than complying with more stringent regulations.
In either case the corporation would obviously like to see favourable results from the science. How can we tell whether the science has been done properly, or has been influenced by the desires of the corporation?
In some cases, such as with drug and chemical testing, there are safeguards in place to help ensure that studies are done properly. OECD (the Organization for Economic Cooperation and Development) and other agencies have published “Good Laboratory Practice” (GLP) methods and guidelines which have to be followed for many of these studies. These guidelines require the testing laboratory to follow rigorous procedures, and include requirements related to the characterization of test materials, the number of animals required (for animal tests), endpoints to be measured, test conditions, and how to evaluate and report the results. The raw laboratory data are available for review.
Studies conducted using GLP use “tried and true” methodologies, are rigorous, transparent, and give regulatory agencies a high level of confidence in the results. Since the raw data are available, these studies can generally be relied on. It is, of course, possible for an unethical laboratory to fake results; however, the consequences of this are quite severe for everyone involved. Perhaps of more concern are inherent limitations from the rigid GLP methods; there has been some concern raised about the effectiveness of GLPs for some endocrine disruptors, such as BPA for example (the issue of conflicting results for BPA and different regulatory interpretations is something I might tackle in another post in the future).
More “academic” studies do not have the same standardization as drug and chemical testing. In these cases new methodologies often have to be developed. The limitations of science journals may also result in the raw data not being available, though separate industry reports may include the data, and journals are getting better about including online supplemental material. This does not invalidate the studies though – it just means that more attention may need to be given to the methods and whether they are appropriate. Personally I think the larger concern might be studies with unfavourable results never seeing the light of day, but that problem extends beyond just industry-funded science.
Ultimately, every study has to be judged on its own merits. I have seen very good studies which were funded by industry, and very poor studies that were not. Except in the (hopefully rare) case of unethical and illegal faking of data, the data itself can be relied on; for any study, regardless of who funds it, any conclusions drawn from the data by the authors should be treated with caution unless it is evident that they follow from the data. It is also important to note that, while a corporation may fund a study, it is normally done by independent researchers or laboratories – and good corporate citizens will use reputable researchers.
I think there are a few things that can be done to improve the reliability of industry-funded science:
- Ensuring thorough reporting of funding and any potential conflicts of interest of the researchers up front. In general this has been improving, particularly as it relates to corporate ties or funding (certain other groups, such as alternative medicine proponents and various advocacy organizations, tend not to be as good about reporting conflicts of interest).
- Working to improve GLP guidelines to reflect modern toxicological knowledge, as well as acknowledging the need to go beyond GLP methods where existing knowledge of a chemical suggests they are not sufficient.
- Ensuring that those conducting science do not have a financial stake in the results (i.e. the researchers must be paid the same regardless of whether the results are favourable or unfavourable). In an ideal world I’d even suggest taking this a step further and removing any interaction between the funding corporation and researchers, possibly by having an independent organization act as the go-between, but this would require international cooperation and probably isn’t going to happen anytime soon.
- Training scientists in ethics. Engineers have to take courses in ethics and, at least in Canada, write an ethics exam before they can practice as a Professional Engineer. Some science professional associations are starting to look at ethics exams too, but generally for scientists there is no requirement for ethics training or testing in order to practice.
- Increasing the availability of raw data and detailed documentation of studies.