Dissemination and reporting bias of clinical and observational studies conducted in the Local health unit of Verona

Introduction: Bias in dissemination and reporting of clinical research findings has an impact on the pooled summary utilised to produce clinical-therapeutic guidelines and recommendations. This analysis aims to evaluate the dissemination and reporting biases of interventional and observational studies conducted in the setting of the Local health authority of Verona (Aulss). The possible correlation between both biases and profit versus no-profit sponsors was also evaluated. Methods: Verona's Aulss studies completed in the period 01.01.2014-31.01.2021 were extracted from the Clinical study register of the Veneto Region and any form of results' dissemination was identified and compared with the original protocol. Identified studies were stratified by profit or no-profit sponsor and results compared using the Chi-Square test. Results: 67 studies (29 profit; 38 non-profit) were included in this analysis. 58.2% of the studies (n=39/67) reports at least one type of findings' dissemination, for 22.4% data-analysis or publication is in progress, while 19.4% has not been published. Regarding the evaluation on reporting bias, 36 of the 39 published studies were considered (n=19 profit; n=17 non-profit): 64% (23/36) showed inconsistencies between the results reported in the manuscript and the protocol. The number of non-compliant profit studies (n=15/19; 79%) was statistically higher than the compliant ones [n=8/17; 47%; (p=0.049; χ2=3.845)]. Discussion: This study highlights that findings' dissemination occurs for the majority of the studies evaluated and that the odds of selective reporting are higher for industry funded studies than for publicly funded studies, affecting the quality of the research.