D71. Meta-analyses in Plastic Surgery Demonstrate Need for Adequate Assessment of Publication Bias

Abstract

PURPOSE: Publication bias describes a phenomenon in which significant positive results have higher likelihood to be published compared to negative or non-significant results. Publication bias can confound the estimated therapeutic effect in meta-analyses and needs to be adequately assessed in Plastic Surgery literature. METHODS: A review of meta-analyses published in five Plastic Surgery journals from 2002-2022 was conducted. Inclusion criteria for meta-analyses were factors that demonstrated obligation to assess publication bias, such as interventions with comparable treatment groups and enough power for statistical analysis. Acknowledgment of publication bias risk, quality of bias assessment, methods used in assessment, and individual article factors were analyzed. RESULTS: 318 unique meta-analyses were identified in literature search, and after full-text reviews, 143 met inclusion criteria for obligation to assess publication bias. 64% of eligible meta-analyses acknowledged the confounding potential of publication bias, and only 46% conducted a formal assessment. Of those who conducted an assessment, 49% used subjective inspection of funnel plots alone, while 47% used any statistical testing in analysis. Overall, only 9/143 (6.3%) assessed publication bias and attempted to correct for its effect. Journals with a higher average impact factor were associated with mention and assessment of publication bias, but more recent publication year and higher number of primary articles analyzed were not. CONCLUSION: This review identified low rates of proper publication bias assessment in meta-analyses published in five major Plastic Surgery journals. Assessment of publication bias using objective statistical testing is necessary to ensure integrity of literature within Plastic Surgery.