Preprint, peer-reviewed pairs of studies generally concordant

Characteristics, results, and final interpretations are generally consistent for clinical studies posted as preprints on medRxiv and subsequently published in peer-reviewed journals, according to a study published online Dec. 9 in JAMA Network Open.

Guneet Janda, from the Yale School of Medicine in New Haven, Connecticut, and colleagues examined the concordance in preprints of clinical studies posted to medRxiv that are subsequently published in peer-reviewed journals (preprint-journal article pairs) in a cross-sectional study.

The researchers found that 77.0 percent of the 1,399 preprints first posted on medRxiv in September 2020 had been published as of Sept. 15, 2022, at a median of six months after preprint posting. Overall, 53.6 percent of the 547 preprint-journal article pairs describing clinical trials, observational studies, or meta-analyses were related to COVID-19. Most (86.4 percent) of the 535 pairs reporting sample sizes in both sources were concordant; larger samples were seen in the journal publication for 58.9 percent of the 73 pairs with discordant sample sizes.

With respect to primary end points, 97.6 percent of pairs were concordant and 2.4 percent were discordant. Overall, 535 pairs had numerical results for the primary end points: 81.1 percent had concordant primary end points, while 65.3 percent of the 101 discordant pairs had effect estimates in the same direction and were statistically consistent. Concordant study interpretations were seen for 96.2 percent of the pairs, including 82 of 101 pairs with discordant primary end point results.

"For preprint-journal article pairs with discordant results, most changes were minor numerical changes, often owing to sample size differences," the authors write.

Several authors disclosed ties to industry, including law firms; several authors disclosed ties to medRxiv.

Copyright © 2022 HealthDay. All rights reserved.