What is publication bias in a systematic review?

When published research results differ systematically from unpublished investigations’ outcomes, this is referred to as publication bias. The term “dissemination bias” has also been proposed to capture all types of biases in the research dissemination process, such as outcome-reporting bias, time-lag bias, grey-literature bias, full-publication prejudice, language bias, citation bias, and media-attention bias. We can assess publication bias by comparing the findings of published and unpublished studies on the same topic. Following up on cohorts of research from conception and comparing publishing levels in trials with statistically significant or “positive” outcomes revealed that those with such results had a higher likelihood of formal publication than those without. Within reviews, funnel plots, and related statistical approaches can be used to detect the existence or absence of publication bias, though these methods are not without limitations. 

Methods for minimizing publication bias are explored and appraised, including detecting and incorporating unpublished results and research. Searching without regard to outcome, searching prospective trial registers, searching informal sources such as meeting abstracts and PhD theses, searching regulatory body websites, contacting authors of included studies, and getting pharmaceutical or medical device companies for additional studies are examples of these. Including unpublished research frequently changes effect sizes, although it does not necessarily eliminate publication bias.  

The mandatory registration of all clinical trials at the outset is a significant step forward, but it might be difficult for reviewers to get data from unpublished studies found in this manner. Instead, journals can eliminate publication bias by publishing high-quality research regardless of novelty or unexciting outcomes, as well as methods or full-study data sets. In addition, researchers, patients, journal editors, peer reviewers, study funders, research ethics committees, and regulatory and legislative authorities must all work together to eliminate the multiple activities involved in publication bias fully. 

References 

Song, Fujian, Lee Hooper, and Yoon Loke. “Publication bias: what is it? How do we measure it? How do we avoid it?.” Open Access Journal of Clinical Trials 2013.5 (2013): 71-81.