"The intervention is safe": recognizing and rectifying spin for harms in systematic reviews

Article type
Authors
Bero L1, Boutron I2, Chou R3, Cornelius V4, Doshi P5, Golder S6, Li T1, Mayo-Wilson E7, Meursinge Reynders R8, Naaman K9, Page M10, Quan N1, Qureshi R1, Vassar M11
1University of Colorado Anschutz Medical Campus, Denver, CO, United States of America
2Universite Paris Cite, Paris, France
3Oregon Health and Science University, Portland, OR, United States of America
4Imperial College London, London, United Kingdom
5University of Maryland, Baltimore, MD, United States of America
6University of York, York, United Kingdom
7University of North Carolina, Chapel Hill, NC, United States of America
8Amsterdam University Medical Center, Amsterdam, The Netherlands
9Indiana University Bloomington, Indianapolis, IN, United States of America
10Monash University, Melbourne, New South Wales, Australia
11Oklahoma State University, Stillwater, OK, United States of America
Abstract
Background: “Spin” refers to misleading reporting, interpretation, and extrapolation of findings in research. The study of spin primarily focuses on beneficial outcomes and primary studies. Spin for “harms” – the opposite of benefits – in systematic reviews (SRs) is particularly concerning for public health and clinical practice because SRs are used to inform coverage and reimbursement decisions, shape guidelines, and influence clinical care.

Objectives: Our objectives were to develop a framework for identifying spin associated with harms in SRs of interventions, apply our framework to a set of reviews to identify examples of how spin might manifest in published SRs, and revise the spin examples to provide guidance on how spin can be rectified.

Methods: We searched PubMed (August 2021) to identify studies of spin in SRs. We extracted existing frameworks for spin in SRs and developed a specific framework for harms through an iterative process that engaged an international working group of researchers specializing in spin and reporting bias. We assessed a random sample of intervention SRs for spin using our framework and identified an example of each type of spin in our framework. We revised the examples to remove spin, taking into consideration the context (e.g., medical discipline and source population), findings for harms, and methodological limitations of the original SRs.

Results: From 32 included studies, we identified 10 frameworks with 26 potential examples of spin related to harms in SRs. We refined this list to 12 specific types (Table). Of the 100 randomly sampled intervention SRs, 59 reviews assessed harm and 41 did not, of which 27 (46%) and 6 (15%), respectively, had at least one type of spin for harms. The box presents an example each type of misleading reporting, interpretation, and extrapolation of harms from published SRs as well as the justification for why it is spin and a suggested revision (with tracked changes) to eliminate the spin.

Conclusions: This framework includes guidance for authors, peer reviewers, and editors in recognizing and rectifying or (preferably) avoiding spin, ultimately enhancing the clarity and accuracy of harms reporting in SR publications.

This project did not involve consumer input.