By David Tuller, DrPH
Yesterday I reviewed an account of a publishing dilemma that had been submitted to the forum of the Committee on Publication Ethics. The COPE forum offers advice on thorny situations submitted anonymously by members. In this case, the submission appeared to be from BMJ Open and it appeared to be discussing Professor Esther Crawley’s school absence study. That study was exempted from ethical review on the specious grounds that it qualified as “service evaluation.” BMJ Open has defended its decision to publish the paper without ethical review.
I was pleasantly surprised by the COPE forum response, quoted in full below. It was reasonable, given the misinformation conveyed in the BMJ Open account. The COPE forum response also clearly justifies the concerns raised about the study’s lack of ethical review. My remarks, posted after the response, will focus on the first paragraph.
FORUM ANSWER: The Forum suggested that perhaps the issue is not whether or not the service evaluation is research, but was the evaluation carried out in human subjects (which would require a sound ethics approach) or were the data contained in registries where the patient data were anonymised. It would appear that the latter is the case and that this is a secondary data analysis, but the editor could ask for clarification from the author on the methodology as it needs to be adequately described. Was this a dataset developed out of a research project that had ethics approval for human subjects? If so, the secondary analysis might not need new ethics approval if additional analyses were covered in the initial approval. The methodology is confusing the issue of whether ethics approval was required. The Forum suggested these points need to be clarified before a decision on whether to add a correction on the article or to respond to the blogger.
This issue often arises with audit articles, which is often a term used for service evaluations. There is a contradiction in that journals publish research articles and yet audits or service evaluations are not thought of as research requiring ethics approval. However, it is up to the ethics committees and their procedures to decide what is research for the purposes of ethics approval. Separately, journals need to decide what they can publish so it is the editor’s decision on what to publish in their journal, irrespective of the decision of the ethics committee.
The Forum suggested that the journal may need to provide more information or specific guidelines for authors on what they mean when they say they accept waiving of ethics approval for service evaluations. What is meant by service evaluations?
The Forum agreed that posting a correction may be excessive and perhaps a short editor’s note would be more appropriate. The Forum advised against responding to the blogger and getting into a spiral of communication that could become problematic. A suggestion was to write an editorial on the concepts more broadly and how the journal’s policy is going to evolve in the future regarding secondary research being conducted as service evaluations/audits/quality improvement reporting and what the ethics requirements will be in the future. What are the expectations of the journal for future submissions of service evaluations?
The Forum suggested the editor may wish to consult the SQUIRE 2.0 (Standards for QUality Improvement Reporting Excellence) guidelines on how to publish quality improvement studies, which can be found on the Equator Network website (https://www.equator-network.org/reporting-guidelines/squire/).
MY RESPONSE: The first paragraph indicates that the COPE forum members remained confused about the data collection procedures involved in the school absence study. This is not surprising, since the account presented by BMJ Open was opaque on the matter, as I noted in my comments yesterday. In responding, the COPE forum members apparently assumed that the school absence study involved “secondary data analysis” from anonymized registries, which could legitimately have qualified it as service evaluation.
This of course was not the case, but it is telling that BMJ Open wrote its account as if it could be interpreted that way. The COPE forum position is clear: If this study did not involve secondary data analysis but instead was carried out on human subjects, it “would require a sound ethics approach.” Given that the pilot intervention investigated in the school absence study was carried out on human subjects who were identifiable to the main researcher, it required ethical approval and should not have been considered service evaluation. It is still not clear why BMJ Open was ever confused about this issue.
As the COPE forum members conclude in the first paragraph: “The methodology is confusing the issue of whether ethics approval was required. The Forum suggested these points need to be clarified before a decision on whether to add a correction on the article or to respond to the blogger.”
The COPE forum has asked the journal editors to clarify the methodology used for the paper. Have they done that yet? All it would require is a review of the BMJ Open paper itself, which describes clearly how Professor Crawley and her team found those identified through the pilot intervention. Perhaps if BMJ Open editor Adrian Aldcroft or editor-in-chief Trish Groves re-read what was published in their own journal, they would finally acknowledge that their continuing defense of the decision to publish it without ethical review is not viable.
The rest of the COPE forum response is about a side issue, as least as it relates to this case: Whether and how the journal should clarify what it means by the terms “research” and “service evaluation.” But these suggestions presume that the problem in this case was one of confusion—on the part of the authors and the publisher—about the definitions of these terms. In fact, I doubt anyone was confused at all.
To sum up: The activities in the school absence study do not qualify as “service evaluation” by any standard. As this response from the COPE forum makes clear, the investigators should not have exempted the study from ethical review. Readers of the paper can see that for themselves. The editors at BMJ Open know it as well. Their efforts to pretend otherwise are unconvincing and will continue to damage the journal’s reputation.