3 ms·
I've always thought it would be an interesting -- but not technically difficult -- project to do a data scrape of PubMed.gov and simply count up and classify al
by danso 11y ago
I've always thought it would be an interesting -- but not technically difficult -- project to do a data scrape of PubMed.gov and simply count up and classify all of the studies that contain a conflict of interest disclosure...but while the scraping would be easy, I'm not sure how accurate or complete the disclosures are.
For example, from the OP:
> Last week, the Pennington Biomedical Research Center in Louisiana announced the findings of a large new study on exercise in children that determined that lack of physical activity “is the biggest predictor of childhood obesity around the world.”
> The news release contained a disclosure: “This research was funded by The Coca-Cola Company.”
This is the link to the press release, which contains the Coca Cola disclosure in the last line:
https://www.pbrc.edu/news/?ArticleID=284 https://www.pbrc.edu/news/?ArticleID=284
The press release links to the PubMed page for the study: http://www.ncbi.nlm.nih.gov/pubmed/26173093 http://www.ncbi.nlm.nih.gov/pubmed/26173093
...Maybe I'm missing it but I don't see the Coca-Cola disclosure in the PubMed abstract. However, the Coca-Cola mention is in the full text of the study: http://onlinelibrary.wiley.com/enhanced/doi/10.1002/oby.21152/ http://onlinelibrary.wiley.com/enhanced/doi/10.1002/oby.2115...
(note: this above link is visible for me because I'm on a university network. It may not work for you otherwise)
On the other hand, the full text of the study also mentions the ClinicalTrials.gov ID number: NCT01722500.
Which means it can be found here: https://clinicaltrials.gov/ct2/show/NCT01722500 https://clinicaltrials.gov/ct2/show/NCT01722500
So maybe the data collection approach should begin with ClincalTrials.gov, which does have endpoints intended for convenient bulk data analysis: https://clinicaltrials.gov/ct2/resources/download https://clinicaltrials.gov/ct2/resources/download
- streptomycin 11y agoDisclosures basically never appear in abstracts on PubMed.
- jonlucc 11y agoAlso, it might not be so easy as just scraping the disclosures. I work at a major pharmaceutical company, and my name is on a paper, along with my affiliation. There are also non-pharma scientists on that paper. The paper contains a line stating that those two non-pharma scientists declare no conflict of interest. A scraper might dismiss that as non-industry, but more than half of the authors work in industry.
- rgejman 11y agoUnfortunately, < 40% of trials on ClinicalTrials.gov report results after completion :(. http://www.nejm.org/doi/full/10.1056/NEJMsa1409364 http://www.nejm.org/doi/full/10.1056/NEJMsa1409364
- danso 11y agoThat's true, and is its own problem for science...but for this situation, it'd only be enough that the clinical trial is registered and accurately reports its sponsors, regardless of its results.