Correcting the record to enhance patient care
Retractions play a vital role in maintaining the integrity of academic publishing and ensuring research credibility. While they are often viewed negatively due to concerns about reputational damage and career impact, BMJ Group sees them as crucial to scientific progress—reinforcing trust, refining evidence, and ultimately improving patient care.
A recent example is The BMJ‘s retraction and replacement of an article on unexpected weight loss as a potential cancer warning sign. After identifying a methodological error, Dr Brian Nicholson and his team at the University of Oxford worked closely with the journal to retract and publish the new version of the study, Prioritising primary care patients with unexpected weight loss for cancer investigation: diagnostic accuracy study. This collaboration ensured that the research findings were accurately represented, reinforcing trust in the evidence used to guide clinical decision making.
The updated analysis revealed a striking shift in findings. Expanding the study population from 64,000 to over 326,000 patients, the 2023 update found that 4.8% of those with unexpected weight loss were diagnosed with cancer within six months—compared to just 1.4% in the original 2020 results. These revised findings led to critical changes in clinical recommendations, now advising urgent cancer referrals for all men aged 50+, women aged 60+, and younger patients with specific symptoms. They include an increase in recognised clinical features associated with cancer in men (from 10 to 20) and women (from 11 to 17), improving diagnostic accuracy. Recognising the significance of their updated results, Nicholson and his colleagues proactively approached The BMJ.
“We found the journal very responsive to our approach for advice about what we should do next.”
Brian D Nicholson
Academic Clinical Lecturer, Associate Professor, and General Practitioner, speaking to Retraction Watch, BMJ [Group], doing the right thing
In March 2023, The BMJ published an expression of concern, followed by a retraction and new publication of the article in The BMJ in October. The journal informed the UK’s National Institute for Health and Care Excellence (NICE) about the retraction, as the study had the potential to shape recommendations on referring individuals with unexplained weight loss in primary care to specialist services.
This study has gained significant attention, with an Altmetric Score of 178, ranking it in the top 5% of research outputs. It has been covered by 15 national news outlets, referenced in two blogs, and discussed by 89 X users, helping expand its reach to health professionals and the public. Academically, it has been cited 19 times and saved by 99 reference manager Mendeley readers.
Beyond its academic influence, the study’s findings showed that certain patient groups—such as men over 50 and women over 60—had a cancer risk above the 3% threshold recommended by NICE for urgent referral. Additionally, the study expanded the list of clinical features associated with cancer in these patients, increasing the number of relevant features from 10 to 20 in men and 11 to 17 in women.
Additionally, by demonstrating the real-world implications of NICE’s recommendations on cancer investigation, Nicholson’s paper provides valuable evidence to help assess and refine guideline implementation.
Helen Beynon
Research Integrity Manager, and advisor to the Committee on Publication Ethics (COPE)
This case exemplifies The BMJ’s leadership in fostering trust, accountability, and the highest standards of research, setting a benchmark for evidence-based improvements in patient care. Through effective collaboration with authors, this approach has led to a shared success that ultimately serves the best interests of both patients and the public.
Retractions at BMJ Group
At BMJ Group, retractions are considered by journal editors in cases of evidence of unreliable data or findings, plagiarism, duplicate publication, and unethical research. We may consider an expression of concern notice if an article is under investigation. All retraction notices explain why the article was retracted. The retraction procedure depends on the publication stage of the article:
Online First publication
A new article version containing just the metadata will be posted, with a retraction note replacing the original text. A retraction notice will also be published in the next available print issue. The original text will remain accessible.
Publication in an issue or a continuous publication journal
A replacement version of the article containing just the metadata will be posted, with a retraction note replacing the original text. The PDF will be replaced with a version watermarked with “Retracted”, but the original text will remain accessible. A retraction notice will also be published in the next available print issue.
In rare cases, we may have to remove the original content for legal reasons. In such cases, we will leave the metadata (title and authors) and replace the text with a note saying the article has been removed for legal reasons. A retraction notice will also be published online and/or in print.
Retraction notices are indexed and linked to the original records in Medline and Web of Science.
