Does your work need IRB approval? Better check, says author of retracted paper

screen-shot-2016-09-25-at-9-51-48-pm

Does an article that discusses anonymized student projects about how to catalog data count as research on human subjects?

One of the students included in the paper thought so, and complained to the journal after learning that it had published the case study of the program without the approval required for studying people. The authors admitted they didn’t get consent from participants, because they didn’t realize the work required it. The mix-up has prompted both them and the journal to reconsider their policies regarding ethics approval of studies.

In the meantime, “A Project-Based Case Study of Data Science Education” has been retracted, with this notice:

It has come to the attention of the Data Science Journal Editors that the authors of the paper “A Project-Based Case Study of Data Science Education” (Turek, Suen and Clark, 2016) did not seek the necessary approval for research involving human subjects prior to conducting their study. In addition, they failed to obtain consent from research participants before publication. The article has therefore been retracted. In the interests of protecting the identity of the research participants, we have also withdrawn the contents of the article from the published record.

Data Science Journal requires that all research involving human subjects is conducted in accordance with the Declaration of Helsinki (or equivalent framework) and, if appropriate, has been approved by the local institutional research ethics committee.

The article included a discussion of data science projects completed by teams of students, Caroline Wilkinson, an editor at publisher Ubiquity Press told us. The article didn’t include the names of participants, but the journal still felt they were “potentially identifiable.” Further, the researchers told us, they used some data from one of the teams that was not publicly available.

For these reasons, the journal decided to retract and remove the article, rather that issue a correction, which “would not have adequately protected the identity of the participants,” Wilkinson said.

She told us how the journal learned about the issue:

We received a complaint from one of the participants involved in the study that the authors had not sought their consent prior to publication. The specific concern of the complainant was that the study was unfairly critical of their team’s work and that this could affect their reputation and future job prospects. Following further investigation, we confirmed that the authors had not followed appropriate procedures for research involving human subjects prior to conducting their study.

It seems that the authors made an honest mistake, Wilkinson told us, and that they “were very keen to rectify the error once it came to light.”

She said that there were no red flags during the submission process — which has now led the journal to take more extensive measures to ensure that articles are up to snuff, ethically:

Upon submission to Data Science Journal, authors are asked to confirm that they have obtained the necessary ethical approval for their research. We also ask peer reviewers to comment on any potential ethical issues when evaluating a paper. In this case, no concerns were raised. This case has, however, prompted us to re-evaluate our procedures for identifying potential ethical issues with submitted articles. We will shortly be introducing a requirement for all authors to include a statement outlining the steps taken to ensure the research was conducted ethically.

Last author Dav Clark, currently at the Kennedy Krieger Institute in Maryland, told us that he had conducted research involving human subjects before, however:

I heuristically believed that [what we were doing this time] was not human subjects research.

In the aftermath of the retraction, he sought to develop a “bright line test” to indicate when researchers, who are doing anything involving people, need to check in with their university’s Institutional Review Board:

The bright line ended up being, “always.” That’s the only bright line there is.

Clark advises that other researchers contact an administrator before beginning a research project to ensure that they are following ethical protocol, even if it seems like overkill.

Like Retraction Watch? Consider making a tax-deductible contribution to support our growth. You can also follow us on Twitter, like us on Facebook, add us to your RSS reader, sign up on our homepage for an emailevery time there’s a new post, or subscribe to our daily digest. Click here to review our Comments Policy. For a sneak peek at what we’re working on, click here.

7 thoughts on “Does your work need IRB approval? Better check, says author of retracted paper”

  1. I always say the following: If you are doing research involving persons, and are not sure if IRB approval is needed, it is best to ask the IRB if their approval is needed. That does not mean filling out the full paperwork to get that question answered. Usually, the IRB director or administrative staff can address a general question. Also, most if not all IRBs will NOT grant RETROSPECTIVE approval. If you collect the data and ask the IRB later if it was OK, you are likely to get an answer that you do not want to hear. Ask first.

  2. This is encouraging to read. I have an on-going research project on the subject of proof of ethics board review for published research. My data shows that skipping ethics board review is quite common and publish or perish pressure is driving it. I’ve identified published research that did not receive ethics board approval and nothing was done by the author’s universities or the journals. Journals need to require proof of ethics board review. In my field, journals typically ask authors to check a box in the submission form to confirm that the research has gone through appropriate ethics board review. This is inadequate. At a minimum, journals should require authors to submit a scan of the ethics board approval for that research.

    1. In my research, I found North American university researchers who deliberately chose to conduct research in Germany as a method to avoid IRB oversight.

  3. This is discouraging to read.

    The Declaration of Helsinki applies to medical research, only. That’s what it says, anyhow. “Anonymized student projects about how to catalog data” doesn’t sound like medical research, to me.

    Additionally, if “Data Science Journal requires that all research involving human subjects… if appropriate, has been approved by the local institutional research ethics committee,” I wonder what they think is “appropriate” for researchers who are not affiliated with any institution which has an “institutional research ethics committee?” That presumably wouldn’t be a problem for Dr. Turek, who has a berkley.edu email address, but it could be a problem for many other researchers.

  4. Copy/pasted from http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0163251 (‘Personality traits are associated with research misbehavior in Dutch scientists: a cross-sectional study’; published 29 September 2016):

    “The medical Ethics Review Committee of VU University reviewed the study protocol and decided that the Medical Research Involving Human Subjects Act (WMO) does not apply to our study and that an official approval of our study by this committee is therefore not required.”

  5. Zachary Schrag’s book “Ethical Imperialism” ought to be required reading for everyone involved in human subjects research, including members of IRBs. It is simply untrue that all research involving human subjects is subject to IRB approval — the long-standing exceptions include journalism, and more recent exceptions include oral history — and the dilemma of soliciting an exception from one’s IRB is the mission-creep and overreach that raise the classic problem of who is policing the police?

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.