False findings in JAMA Psychiatry

Here’s an email I sent JAMA Psychiatry editor Dost Öngür regarding false findings that he recently allowed to be published in his journal. I sent it on March 25, and his March 28 reply is below (ed. 03-29-23). 

***

Hi Dr. Öngür,

This evening, JAMA Psychiatry turbo-rejected my comment (submitted at 7:45 p.m. rejected at 8:32).

Is this how you do dialogue about ethics when someone informs you that your journal has published false findings?

Thanks and best regards,
Vera

***

“Research Misconduct Generated False Findings Requiring Retraction”

Inaccurate data in scientific papers constitutes research misconduct when “data are altered, omitted, manufactured, or misrepresented in a way that fits a desired outcome” (1). This applies to Zandberg et al’s recent JAMA Psychiatry article “Association Between State-Level Access to Reproductive Care and Suicide Rates Among Women of Reproductive Age in the United States” (December 28, 2022; (2)).

The authors reported enforcement of abortion restriction laws “was associated with higher suicide rates among reproductive-aged women (β = 0.17; 95% CI, 0.03 to 0.32; P = .02).” But eTable 5 shows that, in their analysis, two of three reproductive-age category bins returned insignificant results. The results presented as primary appear to drop the oldest bin and combine the younger two: p-hacking. The authors misrepresented this misleading result as pertaining to all reproductive-age women, and misrepresented their findings as having “remained significant when using… different age categorizations.”

In addition to p-hacking and misrepresentation of results, the authors did not accurately represent their reported findings’ practical significance. Suicide is rare. The expected number of suicides in their N = 1022 subsample of interest was zero. So the real-world implication of their claimed 5.81% suicide rate increase is zero. This illustrates why we probably want to look at counts and not rates when it comes to rare events.

The authors also inflated that claimed rate increase. The correct 95% compatability interval estimate is 1.03-1.0581 (3). The reported finding corresponds to the upper bound.

Abortion is reliably associated with about a 2x statistically and clinically significant suicide risk increase (4). We don’t know if the link is causal or not, but we know it’s substantial. Women deserve to know that. Fabricating results that tell the opposite story, as Zandberg et al did, doesn’t serve them.

Science is not self-correcting (5). But scientists reward authors who report their own honest mistakes (6). So I asked Zandberg to retract based on these concerns, and received no reply (7).


Citations

(1) “The Prevalence of Inappropriate Image Duplication in Biomedical Research Publications,” Elisabeth M. Bik, Arturo Casadevall, and Ferric C. Fang, ASM Journals: mBio, Vol. 7, No. 3, https://journals.asm.org/doi/full/10.1128/mBio.00809-16.

(2) “Association Between State-Level Access to Reproductive Care and Suicide Rates Among Women of Reproductive Age in the United States” Jonathan Zandberg, Rebecca Waller, Elina Visoki, and Ran Barzilay, JAMA Psychiatry, 2023;80(2):127–134. doi:10.1001/jamapsychiatry.2022.4394.

(3) “Science Fiction: Bad abortion-suicide research turns risk upside-down, but will the authors retract?” Vera Wilde, SubStack, March 15, 2023, https://wildetruth.substack.com/p/science-fiction.

(4) “Abortion Myths, Part 1: Dubious science downplays substantial possible risks of abortion,” Vera Wilde, SubStack, Feb. 23, 2023, https://wildetruth.substack.com/p/abortion-myths-part-1

(5) “The natural selection of bad science,” Paul E. Smaldino and Richard McElreath, Royal Society of Open Science, Vol. 3. No. 9, September 2016, https://royalsocietypublishing.org/doi/10.1098/rsos.160384.

(6) “The consequences of retraction: Do scientists forgive and forget?” Alison McCook, June 16, 2015, Retraction Watch, https://retractionwatch.com/2015/06/16/life-after-retraction-in-many-cases-its-forgive-and-forget/.

(7) Author correspondence, March 20, 2023, https://twitter.com/vk_wilde/status/1637718347113541633.

 

***

Hello

 

Thank you for your message.  We monitor posts on our website to make sure they are collegial and constructive and we concluded that your comments did not meet that criterion.

 

I see that you have tweeted and also have a blog post about this specific concern. As stated in the journal’s commenting policy, we do not post comments that “duplicate what a commenter has already said in one or more previously published comments.” If you wish to provide an original comment that you have not posted elsewhere and that is without unfair allegations of research misconduct, we may consider it.

 

Sincerely,

Dost Ongur

Share