Evaluating spin in the abstracts of systematic reviews and meta-analyses on cannabis use disorder

Identifying spin bias of nonsignificant findings in biomedical studies

OBJECTIVE

The purpose of this research note is to share a technique for the identification of spin bias that we developed as part of a living systematic review on the cardiovascular testing of e-cigarette substitution for cigarette smoking. While some researchers have remarked on the subjective nature of ascertaining spin bias, our technique objectively documents forms of spin bias arising from the misrepresentation of nonsignificant findings and from the omission of data.

RESULTS

We offer a two-step process for the identification of spin bias consisting of tracking data and findings and recording of data discrepancies by describing how the spin bias was produced in the text. In this research note, we give an example of the documentation of spin bias from our systematic review. Our experience was that nonsignificant results were presented as causal or even as significant in the Discussion of studies. Spin bias distorts scientific research and misleads readers; therefore it behooves peer reviewers and journal editors to make the effort to detect and correct it.

Evaluating the relationship between industry sponsorship and conflicts of interest among systematic review authors on treatments for cannabis use disorder

Background: Cannabis is the most commonly used illicit drug worldwide. In addition to potential adverse effects, an estimated 9% consistent cannabis users are likely to become dependent and may develop a cannabis use disorder (CUD). Methods: This cross-sectional study developed a search strategy using Ovid, MEDLINE, and Ovid Embase for systematic reviews and meta-analyses focused on CUD treatment in June 2020. These reviews were evaluated for conflicts of interest (COIs) per previously developed classification scheme. Our primary objectives were to (1) evaluate the presence of disclosed or undisclosed COI of systematic review authors, regarding treatment of CUD; and (2) determine whether overall summary effect estimates, narrative results and conclusions were influenced by the presence of disclosed or undisclosed COIs among systematic review authors. Results: Our systematic search returned 560 articles which 9 systematic reviews were eligible for data extraction. We found 77.8% (7/9) contained at least one author with a COI. From the 51 authors included, 29.4% (15/51) were found to have a COI. Forty-four percent (4/9) were funded, 22.2% (2/9) were not funded, and 33.3% (3/9) had no funding statements. Out of the 7 systematic reviews with one or more authors containing COI, 14.2% (1/7) included results favoring the treatment group and 28.6% (2/7) included conclusions favoring the treatment group. Our results showed no significance between funding source and results (p = 0.429) or conclusions. Additionally, we found no significance between the presence of COIs with the favorability of results (p = 0.56) or conclusions. Conclusion: Multiple studies favored the treatment of cannabis-containing products, even though COIs were found in the majority of the systematic reviews. COIs have the ability to sway results of a study, which can affect clinical decision-making. Stricter guidelines should be enforced among authors displaying COIs in systematic reviews studying CUD treatment.