Abstract
Purpose: :
To determine recent rates of subsequent full publication of reports of clinical trials presented at ARVO annual meetings; to analyze characteristics which may determine the likelihood of such publication; to evaluate resources with which to identify these publications.
Methods: :
The US Cochrane Center’s CEVG@US project indexes ARVO abstracts which report human clinical trials as defined by Cochrane standards: the trial involves an intervention and control tested in living human subjects, and allocation to the arms of the trial must be randomized. These abstracts are available in the Central Register of Controlled Trials in the Cochrane Library, and are retrievable by a search of the ‘source’ field for "IOVS." 294 trials were identified from the 2004 abstracts, and 330 from the 2005. Subsequent publications were identified by searches of Medline (PubMed), Embase, Web of Science, Scopus and Google Scholar; Visionet, VisionCite, PsycINFO, BIOSIS, INSPEC and regional medical indices such as LILACS and AIM were also searched when subject or authors indicated. Published articles were judged to be related to the ARVO abstract when at least one abstract author was acknowledged on the published paper, and additional criteria including study populations, interventions and outcome measures were congruent. In some cases, authors acknowledged the earlier ARVO presentation.
Results: :
: Full publication of slightly fewer than half of the abstracts for these two years were identifiable as of December, 2007. PubMed indexes the majority of these publications, but other indices and search engines contribute to comprehensive searches.
Conclusions: :
Our findings do not differ substantially from earlier studies, which have examined subsequent publication of a sample of abstracts from each section of the 1985 ARVO meeting (Juzych, IOVS, 1993), from the 1985 ARVO and 1984 American Academy of Ophthalmology abstracts (Juzych, Ophthalmology, 1991) and of RCT reports presented at ARVO and American Academy of Ophthalmology conferences in 1988 and 1989 (Scherer, JAMA 1994). A recent systematic review on full publication across disciplines (Scherer, Cochrane Database of Systematic Reviews, 2007) reaches similar findings. Full publication of clinical trial research, including reports of studies showing no difference or negative outcomes, should be the standard for human clinical reseach. Clinical trial registries provide a useful adjunct to publication.
Keywords: clinical research methodology • clinical (human) or epidemiologic studies: biostatistics/epidemiology methodology