You write that “many medical journals already require anonymized data sharing as a condition of publication”. Most medical journals do not require this and there has been pushback against the idea. To quote from an article I reported for Nature in 2023:
“In 2016, the International Committee of Medical Journal Editors (ICMJE), an influential body that sets policy for many major medical titles, had proposed requiring mandatory data-sharing from RCTs. But it got pushback — including over perceived risks to the privacy of trial participants who might not have consented to their data being shared, and the availability of resources for archiving the data. As a result, in the latest update to its guidance, in 2017, it settled for merely encouraging data sharing and requiring statements about whether and where data would be shared.
The ICMJE secretary, Christina Wee, says that “there are major feasibility challenges” to be resolved to mandate IPD sharing, although the committee might revisit its practices in future. Many publishers of medical journals told Nature’s news team that, following ICMJE advice, they didn’t require IPD [individual patient data] from authors of trials.“
Ah fair enough, but in my anec-datal experience the number of journals that require data-sharing from clinical trials is *extremely* low, and this is really something worth highlighting; it shouldn't really be portrayed as 'many journals do'. (I say anec-datal: in 2023 I contacted five or six publishers, who in turn publish dozens of medical journals, to ask about this, and none of them said they required IPD data-sharing for clinical trials; though perhaps individual journals have a different policy).
In that JAMA Network Open study you linked to about highly-cited clinical trials (correctly noting that data-sharing from these trials is low), not one of the listed journals in table 1, showing where these trials tended to be published, has a policy that requires data-sharing! They all follow the ICMJE.
No wonder that those conducting the clinical trials tend not to share their individual patient data, if journal publishers don't require it.
You write that “many medical journals already require anonymized data sharing as a condition of publication”. Most medical journals do not require this and there has been pushback against the idea. To quote from an article I reported for Nature in 2023:
“In 2016, the International Committee of Medical Journal Editors (ICMJE), an influential body that sets policy for many major medical titles, had proposed requiring mandatory data-sharing from RCTs. But it got pushback — including over perceived risks to the privacy of trial participants who might not have consented to their data being shared, and the availability of resources for archiving the data. As a result, in the latest update to its guidance, in 2017, it settled for merely encouraging data sharing and requiring statements about whether and where data would be shared.
The ICMJE secretary, Christina Wee, says that “there are major feasibility challenges” to be resolved to mandate IPD sharing, although the committee might revisit its practices in future. Many publishers of medical journals told Nature’s news team that, following ICMJE advice, they didn’t require IPD [individual patient data] from authors of trials.“
Source: https://www.nature.com/articles/d41586-023-02299-w
Thanks! I didn't mean to imply most journals require this, since I don't know what the proportion is across all journals.
I linked to this paper later on (https://jamanetwork.com/journals/jamanetworkopen/fullarticle/2811814) for the same idea - highly cited clinical trials tend to be published in journals that have data sharing agreements as anonymized IPD.
Thanks for the write up and link, I hadn't seen that!
Ah fair enough, but in my anec-datal experience the number of journals that require data-sharing from clinical trials is *extremely* low, and this is really something worth highlighting; it shouldn't really be portrayed as 'many journals do'. (I say anec-datal: in 2023 I contacted five or six publishers, who in turn publish dozens of medical journals, to ask about this, and none of them said they required IPD data-sharing for clinical trials; though perhaps individual journals have a different policy).
In that JAMA Network Open study you linked to about highly-cited clinical trials (correctly noting that data-sharing from these trials is low), not one of the listed journals in table 1, showing where these trials tended to be published, has a policy that requires data-sharing! They all follow the ICMJE.
No wonder that those conducting the clinical trials tend not to share their individual patient data, if journal publishers don't require it.