Wednesday, August 1, 2018

Misgudied and controversial recommendations on "ethics dumping" from a Horizon 2020 funded project



This book is a result of the European Commission funded research project, TRUST, under the Horizon 2020 program. It contains a collection of case studies that are supposed to illustrate what the authors call “Ethics Dumping”. It is not clear exactly what the authors mean by this. According to a footnote in the book

The term was introduced by the Science with and for Society Unit of the European Commission: ‘Due to the progressive globalisation of research activities, the risk is higher that research with sensitive ethical issues is conducted by European organisations outside the EU in a way that would not be accepted in Europe from an ethical point of view. This exportation of these non-compliant research practices is called ethics dumping’ (European Commission nd).

The central phrase here is “export of non-compliant or unethical research practices”, although it is not exactly clear what is meant by “export” in this context. It seems that the consortium would want to include all research they identify as “unethical” in a low- or middle income country where researchers, sponsors or funders from high-income countries are involved. Typically, though, “dumping” is associated with activities that confer some advantage on the agent that carries out the “dumping”. In the research context this would be research that is for the primary benefit of people in rich countries, but is done in low- and middle-income countries because it is easier to do there. The problem is that many of the examples in the book are about research that is done for the benefit of people in low-income countries. Another major problem is that no attempt is made to acknowledge that the ethics judgments the authors make are controversial and highly contested. This would not be a problem if it was simply a report by the project group. After all, participants in a research project are free to reach any conclusion they feel is justified by their research. The project has, however, also developed a Global Code of Conduct to counter ethics dumping that is going to be used in the evaluation of research projects funded by the Commission, according to a press release:

The European Commission’s Ethics and Research Integrity Sector in the Directorate-General for Research and Innovation intends to propose the code as a reference document for future research projects seeking funding under the EU’s Framework Programmes for research and innovation.

The cases in the book are supposed to illustrate what practices the Code of Conduct should prevent. If the Code and the accompanying cases will be used to make decisions about what projects the European Commission will fund, we’d better be sure that there is general agreement about the recommendations, and that they continue to promote research that will improve health conditions of people in low-and middle income countries. The problem is that this is not the case. Let me illustrate this with one example from the book.

One chapter deals with three randomized controlled clinical trials to establish which screening program for cervical cancer was most suitable for India. The trials compared traditional cytology (“Pap-smears”), visual inspection after application of acetic acid (VIA), and testing for the presence of HPV. The trials also included a non-intervention arm. The trials were carried out in the 2000s and were funded by the US National Institutes of Health and the Bill and Melinda Gates Foundation. The criticism of the trials by the authors of the chapter is based on two premises.
           
First, is the claim that the “international standard of screening is cytology” (p.35). Second, is the claim that researchers have an “ethical obligation to provide <the international> standard care to participants in the control arm, as they are under their direct care during the course of research” (p. 44). It is the second premise that is contentious. Let me explain why.
   
There is no doubt that the “international standard” (Pap-smears) is effective at preventing cervical cancer and deaths. However, the test is not perfect. It has low sensitivity and specificity, requires highly trained personnel and qualified laboratories. The only way that this test can be a reasonable predictor of cancer is for it to be done regularly every few years. Even if the test misses a cancer once, repeating it will increase the likelihood that it will catch the cancer before it is too late. Given the low specificity the test also will give a high number of false positives requiring follow up procedures. For these reasons it has been generally agreed that a nationwide implementation of a screening program in India using this test was not feasible when the trials started, and is still not feasible. The infrastructure is not in place for an effective population screening program. However, the test has been available for people in India who want it and can pay for it themselves, or through their insurance. This means that it is effectively available only for a small proportion of Indian women.
        
VIA was proposed as an alternative to Pap-smears. In the early 2000s studies comparing these two methods had shown that they had similar sensitivity, but VIA had lower specificity. This has led Ruth Macklin to conclude that the effectiveness of VIA as a screening method had already been established in 2002, but this is simply a mistake. The same source that professor Macklin quotes to support her claim that VIA is effective, goes on to say that

Further research is required to improve its specificity without compromising sensitivity. Information from ongoing studies regarding its longitudinally-derived sensitivity, efficacy in reducing incidence/mortality from cervical cancer, its cost-effectiveness and safety will be useful in formulating public health policies to guide the organization of VIA-based mass population-based screening programmes in developing countries (p. 36)

In this paragraph there is a call for research that will establish its “longitudinally-derived sensitivity, efficacy in reducing incidence/mortality from cervical cancer, cost-effectiveness and safety”. This is quite important, because the critics, including Ruth Macklin, claim that all of this has already been established in 2002. For them, the only open research question is whether VIA as a screening method is implementable in a low income country such as India. But this clearly was not the prevailing expert opinion. The experts believed that it was still an open question whether VIA is effective as a screening program in low-income countries. This, of course, was precisely what the three trials that started in India was trying to address, thus meeting the call for research by the WHO consultation.

In order to justify these trials one still needs to establish that it was necessary to include a non-treatment arm. Obviously, a large number of experts believed that it was. I believe that a non-treatment arm was necessary, but will not argue for that here (will follow later).
            To sum up the argument so far: There was general agreement among experts that further research into the effectiveness of alternative methods for cervical cancer screening was necessary, and at least a substantial number of experts believed that a non-treatment arm was necessary. To complete the argument that these trials were justified, a final point needs to be established: Is it justified to withhold a known, effective intervention in the control group even though it is necessary for scientific reasons? The TRUST project participants believe that it is never justified to do so when the research deals with a major health endpoint such as cancer, or survival. Again, I shall not address the substance of the argument here, as this is also not addressed in the TRUST book.  Basically they point to agreement in international guidance documents as their main argument for their position on this issue. They claim that

Ethical guidelines governing the use of placebo in research severely restrict the use of placebo or “no intervention” if an effective treatment or test already exists for the disease being studied. This is to ensure that research participants in the control arm do not receive a lower standard of care than is already known to be effective, and are not therefore disadvantaged by their participation in the study.

The severe restriction they refer to is that no-treatment is allowed only if this would case non-serious harm. The problem with this view is that it is not a fair description of the content of international guidance documents. One should at least acknowledge that major guidelines such as both the current and last version of the CIOMS guidelines (see commentary to Guideline 5) recognize that there may be circumstances where research ethics committees may allow for exceptions to this rule. Basically, these exceptions deal with circumstances where the aim of the research is to produce knowledge that may be relevant in the local context, precisely the context of these three trials of cervical cancer screening programs in India.

And here is the problem. If the European Commission is going to use the Code of Conduct developed in the TRUST research program, with the accompanying case studies that illustrate how the Code is to be applied, as a basis for funding decisions of their research program for poverty related diseases, the Commission will use a misleading and controversial moral analysis as a basis for their decisions. There are, of course, some people who think that is good and agree. But there are also reasonable people who will disagree, and point out that the effect will be that some beneficial research to alleviate poverty-related diseases will not be funded. The TRUST project has reached their conclusion with regard to the moral issues involved, and the Commission is of course free to adopt their recommendations as a basis for their funding decisions. But at least one hopes that they recognize the controversial nature of the TRUST recommendations.

No comments:

Post a Comment