A blog from the Centre for Research Ethics & Bioethics (CRB)

Month: September 2015

Articles may be retracted if ethics is neglected

Pär SegerdahlWhen a scientific article is retracted, it means that the article should never have been published and that data and conclusions from the study should not be used to underpin future research.

Articles are often retracted when it is found that the authors acted fraudulently. They may have been careless, or cheated, or have plagiarized someone else’s (or their own!) previous work. Retracted articles may still be available for reading, but with a notice that they are retracted, and with explanations of the reasons behind the decision.

A rarer and less known reason to retract scientific articles is that the study reported does not satisfy ethical requirements for the protection of research participants.

Human research participation should be voluntary and research on humans must first be approved by an ethical review board. Editors of medical journals are bound by the same requirements. They increasingly require that authors state that the research they want to publish has an ethics approval.

How common is it that published articles are retracted because ethical requirements were neglected? How do editors motivate their decision? And what happens afterwards – are the articles cited and used despite the retraction?

Ethical retractions are uninvestigated, but in an article in the journal Accountability in Research Yusuke Inoue (former guest researcher at CRB) and Kaori Muto, present a study of articles retracted for ethical reasons:

One difficulty they mention is that unethical research may still produce scientifically valid data, results and conclusions – although neglect of ethics is a strong warning sign that other demands may have been neglected. Editors must therefore strike a balance between the requirement to retrospectively protect research participants and the scientific value of the article and its results. And if one decides to retract the article for ethical reasons, the research study may have to be repeated with new participants, which is also ethically problematic.

Yusuke Inoue and Kaori Muto studied retracted medical papers in English in the period 1981-2011. They found that the first ethical retractions did not occur until 2000 (2 articles). The number was then relatively constant (14 articles 2001-2010), but increased dramatically in 2011 (83 articles) – most of them related to a research scandal around anesthesiology researcher Joachim Boldt.

Most retraction notices stated as reason for the decision, simply “lack of ethical review.” However, editors rarely explained the decision more closely, for example, if they judged that the whole study was fraudulent, or judged that the study was well done but lacked ethical review. It then becomes unclear how to assess the contents of the retracted article.

Inoue and Muto also found that the majority of articles that were retracted for ethical reasons continued to be quoted. In some cases, it could be established that citations were deliberately misleading (as when authors cite their own retracted articles without mentioning that they are retracted). In other cases, however, retracted articles were cited perfectly legitimately, to specify that data from them had been excluded.

Inoue and Muto’s conclusion is that editors need to explain more clearly the reason behind their ethical retractions, so that future researchers can better assess the content of the articles. Moreover, discussion is needed on how data from articles that were retracted for ethical reasons may be used.

While we’re discussing scientific misconduct, I take the opportunity to link to an American dissertation that shows that often when misconduct is revealed by the Office of Research Integrity, it does not lead to the retraction of articles:

The number of retracted articles thus gives a poor measure of the extent of scientific misconduct. There are many “fraudulent articles” in circulation!

Pär Segerdahl

We recommend readings - the Ethics Blog

Dissertation on palliative care of children with cancer

Pär SegerdahlApproximately every fifth child who gets cancer in Sweden dies from their disease. In her dissertation work at CRB, Li Jalmsell studied the care of these children at the end of their life from both the child’s and the parents’ and siblings’ perspectives.

One of her findings is that one doesn’t generally recognize that the child’s cancer is beyond cure until very close to death, giving little time to plan palliative care based on personal preferences.

Jalmsell also did surveys with parents and siblings who lost a child/sibling, and interviewed children with cancer. The children themselves emphasize in the interviews that they want honest information, even when it is bad. But they also want the conversations to be hopeful and contain a plan ahead; and they want to be informed simultaneously with the parents (not after the parents).

The psychological suffering of parents and siblings who lost a child/sibling seems to be influenced by different factors. Parents’ suffering after the child’s death is much dependent on how they experienced the child’s suffering near the end of life. The parents’ suffering also tended to increase if the child underwent bone marrow surgery before death, perhaps because of the hope of a cure that such an intense treatment awakens.

Siblings generally felt ill-informed and unprepared for the child’s death. Siblings who didn’t get opportunity to talk about what they could expect tended to feel anxiety long after the child’s death.

Jalmsell also stresses the importance of parents talking about death with their child. Other studies have shown that parents who don’t talk often regret this afterwards; while parents who talk with the child about death don’t regret it. In Jalmsell’s own study the parents say that the initiative to talk about death often came from the child, often through stories. The child understands its situation.

If you want to read Li Jalmsell’s dissertation, you can find it here:

It emphasizes the importance of open communication with the whole family.

The public examination is on Friday, September 25, at 09:00, at the Uppsala Biomedical Centre (BMC), room A1:111a. The examination will be conducted in English. Welcome to listen and ask questions!

Pär Segerdahl


Interesting Big Data-symposium on video

Pär SegerdahlMany posts on the Ethics Blog are about how new possibilities to collect and process large amounts of data change the horizon for medical research.

But “Big Data” makes its entry also in the humanities and social sciences. How does the horizon change there? How is the understanding of humans and of society affected when processing large amounts of data opens up a new field of vision for humanists and social scientists?

A symposium in Gothenburg last summer took up the issues, I saw at Christian Munthe’s blog (“Philosophical Comment”). He links to a video recording from the symposium and I link to Christian’s blog post; that way you’ll find both the blog and the video:

When you have time, take a look – the presentations are exciting!

Pär Segerdahl

We like challenging questions - the ethics blog

Bioethicists suggest broad consent for biobank research

Pär SegerdahlIt is still unclear what kind of consent should be used when collecting biological samples for future research. Different forms of consent are practiced, which creates another uncertainty: which research is actually permitted with the collected samples?

This haphazard situation leads to unintended constraints on research. But it also leads to research sometimes being carried out without consent.

Against this background, the US National Institutes of Health (NIH) organized a workshop to discuss whether it is ethically reasonable to manage these uncertainties by using broad consent for future research when collecting biological samples.

The group of bioethicists who attended the workshop, including Mats G. Hansson, recently published their thoughts and conclusions in the American Journal of Bioethics:

The group’s proposal is that broad consent is ethically reasonable and often the best option, if it has three components:

  1. Consent is conducted initially, in connection with sample collection.
  2. There is a system for oversight and approval of future research.
  3. As far as possible, there should be ongoing communication with, and information to, donors.

Biological samples are collected in a variety of contexts. It is here that the haphazard situation arises, if different forms of consent are used, or perhaps no consent at all. By initially informing potential donors of the wide range of research that can be carried out, they can take a position on risks and benefits of donation (given the oversight and the general conditions of the future research that they are informed about).

The group emphasizes that broad consent gives donors control over the use of samples, while minimizing costs and burdens for both donors and researchers.

They also point out that empirical studies show that most people want to decide if their samples may be used for research. Most respondents also say that the decision is not influenced by the specific details of the future research (e.g. what diseases are studied, what techniques are used, or which parts of the sample are studied).

Of course there are examples of research that can be perceived as controversial, such as human cloning. But broad consent can be combined with specific restrictions. Oversight moreover considers whether research proposals can be said to comply with the donors’ values.

If donors still hesitate, they are free to choose not to donate the sample.

Pär Segerdahl

Approaching future issues - the Ethics Blog