For She had Eyes and Chose Preprints - Social Media Impact, Citation and Publication rates of Ophthalmology preprints

BackgroundPreprinting, is the sharing of non-peer reviewed, unpublished scholarly manuscripts. Across many fields of medicine, an exponential rise of manuscripts being posted to preprint servers has been observed. This has exploded during the COVID19 pandemic where early dissemination of information was critical, or where COVID19 priority disrupted the publication dynamics and priorities in other areas. We examined the characteristics of ophthalmology related preprints in this study. MethodsWe searched the bioRxiv and medRxiv servers for preprints relating to field of ophthalmology. Preprints were screened by title and abstract to ensure they were related to the field of ophthalmology. Outcomes included number of tweets, upper bound followers of tweeters, number of citations, news outlets reports and dates posted/published. If a preprint was published the same outcomes were collected for the published version to allow for comparisons, as well as journal publisher and cite score. ResultsAfter screening, a total of 720 preprints met our inclusion criteria. 420 of the preprints went on to be published. The publication rate of the preprints was 58.3%. The median number of tweets received on preprints was 3.5, IQR 5.5. 98.75% of preprints were tweeted about. Citation rate was 40.14%. The average number of days from date posted on server to date published was 180.5{+/-}124 days. Regression analysis revealed that tweets as a preprint predicts tweets as a published article, P=0.0135. citations as a preprint predicts citations as a published article, P<0.001. The average cite score was 8.5{+/-}6.2 among those preprints that we published in journals. 16.66% of papers were published by Elsevier (n = 70). 3.19% of preprints were COVID19 related, with significant differences found between them and non-COVID19 preprints in terms of number of tweets and number of citations. ConclusionsOphthalmology preprints are increasing across preprint servers. This maybe to bypass publication times and allow early dissemination of work, as well as increase visibility and citations. We identified that preprint citations and tweets predict published version citation and tweets.