The Ethical Conduct of Scientific Research
This essay was written by Rod King and was first published in the 1998 Mill Hill Essays.
Scientists are human beings, and are subject to the same foibles as any member of other sections of society. They are just as capable of cheating, lying, sharp practice, self deception and the full spectrum of crimes and misdemeanours as anyone else. So why should we believe anything that they tell us? What checks and balances exist within the scientific endeavour to ensure, as much as is humanly possible, that what is reported as scientific fact or as medical breakthrough is not just the result of wishful thinking on the part of the authors? And what about those jealousies that exist in every walk of life. How do they show themselves in science, and how do they affect lives and careers? Why should laboratory animals suffer for our benefit, and what precautions are in place to ensure that when they are used they suffer as little as possible? If a pharmaceutical company is paying for an investigator to run an expensive trial of a new drug which shows that the treatment produces little or no improvement on one that already exists, does the investigator have a duty to inform the medical profession if the company goes ahead and markets the new drug at an inflated price?
Ethical questions of this sort have to be faced by every scientist at some time in their professional lives. The code of conduct by which scientists exist, the ethics of the profession, is not tabulated so that it can be consulted, like the United States Constitution, whenever a problem arises. It is an evolving entity, subject to constant interpretation and refinement, which is under increasing scrutiny as those who provide the funding for research demand an accounting for the money which they invest. A classic response from scientists in training, when asked to think about these concepts for the first time, is that it is all a matter of individual choice. This is clearly wrong, since although there may be more than one correct, or to be more accurate, acceptable solution to a particular ethical problem, it is equally likely that there will be at least one solution which is clearly unacceptable to the individual or society at large. For this reason there is now pressure for training in ethics to be part of the curriculum for young scientists in the same way that it already is for young doctors.
How do these ethical considerations manifest themselves in the day-to-day world of research? The output of scientific research is knowledge. This knowledge is communicated to the rest of the scientific community, and to the world at large, mainly through the medium of the scientific literature. Thousands of scientific articles are published every week in journals and books of various kinds. Because career and future financial support for research depends largely on a scientist’s published output it could be seen as a great temptation to attempt to publish speculative or even downright untrue results in an effort to boost publication lists or reputation. Over the years the editors of journals have developed a quality control mechanism, a system of refereeing whereby experts in the field are asked to judge whether an article which has been submitted is suitable for publication in the journal concerned. Part of this judgment is whether the experiments performed could actually lead to the results and conclusions reached. The system is clearly not foolproof, but it is buttressed by the collegiality of scientific endeavour, a sort of honour code where deliberately misleading your colleagues, by committing the sins I describe below, will get you blackballed from the club faster than anything else.
The three cardinal sins of scientific research are fabrication, falsification and plagiarism. Fabrication, making up results, is not only dishonest, it is also stupid because it will almost inevitably be discovered. If the results you publish appear to be interesting and important, and what would be the point of going to all that trouble for something nobody was going to read, then equally inevitably someone will want to do the same experiments, or something very closely related to them. If others are unable to reproduce results already published then, in the very nature of scientific enquiry, they will want to know why. Once the cloud of suspicion has descended it can take a very long time to lift. Of course the fabricator can get lucky, the educated guess can be right, but like the gambler who has won once, the temptation to try their luck again will probably lead to disaster.
Falsification of results is a rather more subtle crime, and usually results from a kind of scientific wishful thinking. By its nature this crime can be very difficult to detect. “If only that point on the graph were a bit closer to the line then it would describe the result perfectly using the current ideas within an acceptable margin of error. This paper must be published before our competitors publish their version so there is no time to repeat the experiment. If the point is left out, or just moved a bit closer, it will look so much more convincing”. Indeed there are some famous examples of classical discoveries, Mendel’s laws of genetic inheritance and Millikan’s measurement of the charge on the electron, where data trimming of this sort, making the results look better than was actually possible at the time, is strongly suspected by modern historians of science. However, these temptations have to be resisted. Experimental data, carefully and properly collected, are the life blood of science and must be respected. If a result whose validity there is no reason to doubt falls outside the currently accepted model then it is time to examine the accuracy of the model, not the data.
Originality is a much prized commodity in science. Demonstration of originality by a candidate presenting for examination for the degree of Doctor of Philosophy is written into the Regulations of every University. The award of this particular degree is the sign that a new scientist has come through a period of training, in much the same way as the apprentice in times past learned their trade at the master’s knee, and is ready to branch out on their own. Citations for scientific prizes invariably stress the originality of the winners, and anyone who has been asked to give a reference for a colleague will know that a critical appraisal of their potential for originality is usually requested. So plagiarism, stealing another persons ideas, is the worst kind of professional theft. The opportunities to steal others ideas are legion. Members of panels reviewing grant applications have prior knowledge of what the applicants are intending to do and can steal ideas at that stage. Referees of papers submitted to learned journals can hold up publication of a rival’s work while they publish their own findings in a different journal. While apocryphal stories claiming this downside of the refereeing process to be true are told over drinks at the conference bar, there is, as far as I know, little real evidence that it happens with any great frequency.
Being first to publish a discovery, particularly if it can be exploited for commercial gain, has long been important in Industry. There the ownership of intellectual property, in the form of patents is the prime objective of the research laboratory. This has not always been the case in academic laboratories. While it was still important, from the point of view of reputation, to publish first and to be seen as the discoverer, the intellectual property generated thereby was available to anyone who wished to use it. This free sharing of information has sustained research and promoted collaboration for almost the whole of this century. With the increased emphasis on exploiting results obtained using funds provided by sponsors and on obtaining funding from commercial organisation, a whole new industry, whereby the technological discoveries of the University or Research Institute are transferred to the market place has been generated. This demands secrecy and embargoes placed on potentially profitable discoveries until patents can be obtained. This can have seriously damaging effects on the career of a young scientist, the publication of whose work can be held up for a period of up to two years. In the competitive world of academic research, unlike that of Industry, publication in academic journals, not patent record is important for advancement and this career disappointment is not usually offset by generous financial compensation. This is a relatively new problem for academic researchers and needs to be given serious ethical consideration by those who wish the funding organisation to profit from the efforts of the toilers at the wheel.
As the competition for research funds and academic employment has become fiercer the pressure to publish as much as possible has landed some people in a great deal of trouble and led to some almost comical practices. There have been examples of senior scientists accepting guest authorship on publications to which they have made little or no contribution, only to discover later that the work is fabricated. Other practices have involved such things as putting your colleague’s name on papers coming out of your laboratory as long as they reciprocate by putting your name on theirs. Fighting to have your name as first author on a paper, because citations of papers invariably involve the first author and not always everyone else who contributed, can cause enormous difficulties. In some cases it is not always clear whose name should go on to a publication as an author. This may seem odd to anyone who has not worked in a research laboratory. Surely the authors are the people who wrote the manuscript? Not necessarily. What about the senior scientist who obtained the grant, did no experimental work, but contributed to the discussions and brainstorming sessions which thrashed out the course of the investigation. Likewise the talented post-doctoral worker who can handle the delicate manipulations essential to the success of the endeavour, but finds writing up the results in a way that is acceptable to the editors of journals a complete trial. Each needs to have their contribution acknowledged, and individual’s assessment of the relative worth of their own contribution will not always agree. Nevertheless, scientists who work in teams have to sort out these problems among themselves, and in a well-run laboratory the practice is to write the front page of the manuscript, which contains the list of authors names, first. What is not acceptable under any circumstances is to include the results of original work done by members of the team without any recognition.
Another extensive area of ethical importance, about which hundreds of books and articles, and millions of words have been written is the use of animals in scientific and medical research. No matter how long, detailed and protracted the debate, eventually an individual choice has to be made between agreement or disagreement with the use of one species, a laboratory animal, for the benefit of another, man. There is effectively no middle ground, and those who disagree with experimentation on animals will not be swayed by arguments citing prevention of possible dangers to human beings or even benefits to other animals, and certainly not by cost-benefit analysis. Their argument, simply put, is that we humans have no right to use or to sacrifice the lives of animals in laboratory experiments for our own benefit. The laboratory scientist who has made the opposite decision would respond that without the use of animals in research, society would not be able to enjoy the fruits of the knowledge which has been obtained; would not for example have vaccine protection against infections, and would not have relief of pain or repair by surgery.
The fact is that animals are used for research purposes in all kinds of research establishments. Their use in the United Kingdom is regulated by the Home Office under the “Animals (Scientific Procedures)” Act of 1986, which set up a pyramidal administrative structure to be observed by those who wish to use animals, and an Inspectorate to monitor the working of the Act in practice. Any establishment where animal experimentation is taking place must have appointed a person who is ultimately responsible for implementation of the conditions of the Act within the establishment. Experiments involving animals can only be carried out under the terms of a licence, which has to be approved by the Home Office Inspectorate, and which after the beginning of April 1999 will also require approval through a local ethical process, in much the same way as experimentation on human patients or material has to have ethical approval. Some establishments have already done this voluntarily and have working ethical committees containing scientists and non-scientist members, to which licence applications must be submitted. Those who are granted licences are responsible for the research carried out under it and the observance of the ethical limitations placed on the work. In addition, those whose task is to carry out the experiments must demonstrate their competence by obtaining a personal licence from the Home Office.
The Act also requires scientists to observe three guiding principles; Reduction, Refinement and Replacement. Reduction of the number of animals used, refinement of the techniques being used to gain a particular item of information, and replacement of animals by other forms of experimental system if at all possible. Scientists do not usually need encouragement to do their utmost to subscribe to these principles. Animal experimentation is expensive; animals have to be fed, housed and have their cages cleaned. The standards of housing and environmental control for experimental animals are often more demanding than they are for the scientist in the laboratory. When the equipment which maintains these conditions in the animal house breaks down it must be repaired immediately, and if this happens on Christmas day someone has to be on call to put it right. On top of this, until very recently, except for naturally identical multiple births, almost every animal was a genetic individual. The influence of this “genetic background” of the animal is yet another variable that has to be accounted for in the interpretation of the experimental results.
How much easier it would be to work with a culture of identical cells in a dish. If the information the scientist is seeking is accessible through experiments on cultured cells then there is no reason and no excuse for using animals at all. When such a step forward is made through refinement of techniques then old methods should be superseded by new after a minimum period of validation in which the identity of the results using the two methods can be verified. But some experiments can not yet be performed in this way, and in our never ending search for knowledge about how our bodies work, and how they can be put right when they go wrong, we owe a huge debt of gratitude to the laboratory animal.
As I mentioned above, scientists can find themselves in situations where they face a conflict of interest. This can afflict all members of the profession. Imagine for example the very senior scientist, the world expert on a particular disease who has an undisclosed financial interest in a small therapeutics company manufacturing drugs to alleviate the symptoms of the disease. This scientist is asked by the Government to advise on what treatment of the disease the Ministry of Health should sanction for general use in the Health Service. Clearly, from the Government’s point of view, the scientist is the right person to ask, but will his or her judgment be swayed by the financial interest? Resolution of conflicts of interest is an area of human interaction where lawyers tell us that rapid and complete disclosure of interest by all of the parties involved solves problems in the shortest time. It allows people to make informed decisions while taking account of all the likely influences. Honesty really is the best policy.
No essay on the ethics of scientific research would be complete without mention of the fate of those who accuse others of misconduct. Some recent high profile cases have shown that those who blow the whistle on unethical conduct of others, particularly if the accused are very senior in the profession, can suffer severely even when their worst fears turn out to be completely accurate. In the past the immediate reaction of the scientific community, whether governed by a genuine disbelief that the person accused could be guilty of misconduct, or that the accuser has some grudge against the accused, or even possibly a collective guilty conscience, has seemed to be one of closing ranks, excluding the accuser and denigrating the investigating team. The over-zealous nature of some past investigators of accusations of misconduct did nothing to improve this situation. More recently, Institutions all over the world have understood this problem and have written codes of conduct for investigating accusations without revealing the identity of the accuser, at least at the beginning. If the accusation, made in good faith, can be shown at an early stage to be mistaken then a minimum amount of harm is done to both parties. If the accusation is in fact properly grounded then it may require a large degree of courage on the part of the whistle-blower to continue right through to the end of the process.
Numerous learned books, articles, pamphlets, policy statements, editorials and codes of conduct have been written about the application of ethical standards to the practice of scientific research. However, as I said at the beginning of this essay, scientists are human and individual. Relatively few have actually read the well-meant guidance, and many opportunities exist for them to stray from the narrow path. Despite my co-editor’s advice that absence of evidence is not evidence of absence, I like to think that it is greatly to the credit of the profession in general and scientists and their teachers in particular that so little wrong doing actually comes to light.