Publication bias in gastroenterological research

To assess whether publication bias is present in gastroenterological research, determinants of publication were studied in abstracts submitted to the American Gastroenterological Association. A random sample of 326 clinical trials (CCT), 455 abstracts on other clinical research (OCR) and 219 basic science reports (BSS) was evaluated. Publication rates were examined by a literature search and by a mail survey of abstract authors. Survey information was available on 499 abstracts. The overall estimated publication rate was 57%. A higher number of previous publications and multi-center status were associated with higher publication rates. Significant differences in publication of negative study results were identified only after correction for under-ascertainment and response bias. Statistically significant studies were published faster and in higher impact journals. The most frequent reason cited for non-publication was lack of time.