51˶

Retract Now: Negating Flawed Research Must Be Quicker

— Incentives and streamlined processes can prevent the spread of incorrect science

MedpageToday
A photo of a magnifying glass lying on an open book.
Schneider is an associate professor of information sciences.

It's take back time.

A concerning stem cell research and transplantation that misled researchers and the general public for 15 years is among the retracted in 2023. These papers covered topics ranging from to to .

The the 2008 article came long after many medical experts knew it was wrong. Many knew for years that the treatment it proposed -- transplanting airways using a patient's own stem cells -- was not a "" but "" The first patient "" and "."

The paper's author, former surgeon Paolo Macchiarini, is now serving for unethical surgeries that harmed patients. This case inspired a and a season of "" Other recent cases of false research leading to retraction include and a by artificial intelligence (AI)-generated anatomically incorrect .

The rise in fraudulent or inaccurate research papers is an . It's costing publishers with . Advances in AI have between and tools to detect it. As an academic or medical healthcare provider, it is vital to know when it is possible to rely on research, and to what extent. Cleaning up the literature should not require a court case -- as it did in the Macchiarini case -- or extend over decades.

Another problem is that policymakers globally may already have made funding decisions or influenced medical practices (as recommended by the research) by the time a paper is retracted, even when the timeframe is modest, . Policymakers in elected positions and at academic institutions, as well as funders, research publishers, and developers of need alerts when articles are retracted and research is debunked. This is to ensure that when it does take over a decade to retract, it is possible to quickly update knowledge in the field and shift guidelines and policy in tandem.

The scientific record is a record of the paths taken over time. And the mistakes and falsification along the way are a valid part of the version of record. That includes flaws in from brain scans, and fraud such as an with fake data. Mistakes are understandable and research is continually updating itself, yet the scientific record is not always updated.

It is urgent to be faster and more responsive in retracting publications. To do this, it is worthwhile to streamline the processes for retracting research as necessary. Publishers of journals must include the right to retract in publication agreements. National funding agencies must set aside a fraction of a percent of their budget to fund national centers for investigation. Institutions need to tweak the incentive structures of career scientists. This is because researchers secure and advance their careers by writing and publishing papers; so retraction (which results in "losing" a paper) can impact the people (authors) and their career prospects.

Additionally, , and are becoming more important, while funders diverse as part of establishing career progression or documenting the output of science funding. There is incentive on all sides to publish and publish quickly. There is little incentive to retract.

No one entity can do the work to retract. While authors may request retraction, often institutional or publisher investigations are needed. It is important to keep retraction out of courts of law. The world can't wait 15 years to pull dangerous science from the literature.

is associate professor of Information Sciences at the University of Illinois Urbana Champaign, and a Public Voices fellow of . She leads the Alfred P. Sloan Foundation project, ""

Disclosures

Schneider has non-financial affiliations with Retraction Watch.