Grant: Catalyst Grant: Ethics
McGill University
$95,250.00
Transparent Science: using computational tools to assess conflicts of interest and their consequences in biomedical publications
FY 2014-15
Summary
Conflict of interest (COI) in medical publishing - the situation in which an author's private interests or affiliations compete with his or her responsibilities to the scientific community, readers and society - is a growing concern. Researchers with conflicts of interest may be more likely to publish positive results, suppress negative findings, advocate positions unsupported by evidence, and produced biased clinical practice guidelines and policy recommendations. While many journals have instituted COI disclosure policies, their application in practice is uneven, and there is evidence that many authors fail to adequately disclose COI.
While interest in COI is increasing, little is known about three key problems: (1) the extent of COI across the medical literature; (2) the sufficiency of current COI disclosure policies to identify actual COI; and (3) the impact of COI disclosure policies on readers' interpretation of medical literature. One of the barriers to investigating these issues is the labor-intensiveness of identifying and extracting COI information from published literature.
We will use computational and crowdsourcing techniques to identify, extract, and assess COI disclosure statements from medical literature. This approach will enable us to analyze patterns of COI disclosure at the scale of tens of thousands of articles over periods of decades, offering the unique opportunity to observe how COI practices have changed over time, differ across subfield of biomedical research, journals, and researchers.
We will also use this platform to conduct a series of experiments to assess individuals' ability to identify and interpret COI information. Again, the scale of this study affords new ways of assessing the relative and absolute effectiveness of different COI reporting strategies.
Department
Program
Catalyst Grant: Ethics
Recipient
Location
Quebec, Canada
Recipient Type
Academia
Type
Grant
Source
CIHR Funding Decisions Database
As published
Id
314784
Name
Ruths Derek
Region
Québec
Country
Canada
Orgtype
Institution Paid; Research Institution
Instname2
Health Services and Policy Research
Theme2
Clinical
Prcname2
Catalyst Grant : Ethics (2016-2017)
Projecttitle
Transparent Science: using computational tools to assess conflicts of interest and their consequences in biomedical publications
Programtype2
Operating Grants
Orgnameinp2
McGill University
Cihramount2
$95,250
Cihrequipment2
$4,750
Programname2
Catalyst Grant: Ethics
Keyworddelim
Conflict Of Interest; Crowd Sourcing; Data Mining; Financial Disclosure; Machine Learning; Textual Analysis
Abstract
Conflict of interest (COI) in medical publishing - the situation in which an author's private interests or affiliations compete with his or her responsibilities to the scientific community, readers and society - is a growing concern. Researchers with conflicts of interest may be more likely to publish positive results, suppress negative findings, advocate positions unsupported by evidence, and produced biased clinical practice guidelines and policy recommendations. While many journals have instituted COI disclosure policies, their application in practice is uneven, and there is evidence that many authors fail to adequately disclose COI.
While interest in COI is increasing, little is known about three key problems: (1) the extent of COI across the medical literature; (2) the sufficiency of current COI disclosure policies to identify actual COI; and (3) the impact of COI disclosure policies on readers' interpretation of medical literature. One of the barriers to investigating these issues is the labor-intensiveness of identifying and extracting COI information from published literature.
We will use computational and crowdsourcing techniques to identify, extract, and assess COI disclosure statements from medical literature. This approach will enable us to analyze patterns of COI disclosure at the scale of tens of thousands of articles over periods of decades, offering the unique opportunity to observe how COI practices have changed over time, differ across subfield of biomedical research, journals, and researchers.
We will also use this platform to conduct a series of experiments to assess individuals' ability to identify and interpret COI information. Again, the scale of this study affords new ways of assessing the relative and absolute effectiveness of different COI reporting strategies.
Competitiondate
201409
Approvedterm2
2 yrs 0 mth
Pinamesdelim
Ruths, Derek; King, Nicholas B
Approvedterm
2 ans 0 mois
Cihrcontribution2
~ $95,250~ $4,750~