The need for consensus, consistency, and core outcome sets in perioperative research

Abstract

Aims: Bariatric surgery requires evaluation with consistent outcome reporting to improve study comparisons, data synthesis and clinical decision-making. We examine adverse event reporting after bariatric surgery to inform the development of a core outcome set (COS). Methods: A systematic review and an evaluation of the National Bariatric Surgery Register (NBSR) identified verbatim outcomes and definitions. Verbatim outcomes were classified into outcome domains (broad aspects of health measured to assess an intervention) by independent reviewers and outcome definitions considered by study design. Results: 66 studies (18 randomised, 48 non-randomised) identified 1073 verbatim outcomes; a further 21 were in the NBSR. 912 outcomes were reported only once. Outcomes were categorized into 9 domains: surgical complications (n=357), obesity-related disease (n=267), anthropometry (n=117),haematological/biochemical markers (n=80), treatment-pathway outcomes (n=75), observer assessed (n=65), peri-operative technical outcomes (n=45) and mortality (n=30).Of the 357 surgical complications reported the commonest were late complications, re-operation and band slippage reported in 17 (26%), 11 (17%) and 6 (9%) of studies respectively. Definitions were reported for fewer than 50% of outcomes, were varied and contradictory. For example, 'late complications' was defined 9 times (reported 17 times), only 3 were adequate. Outcomes were more frequently defined in randomised versus non-randomised studies (34% vs 24%, p=0.0001). Conclusions: Outcome reporting for bariatric surgery focuses on adverse events. Outcomes are inconsistent, ill-defined and lack consensus limiting the value of published data. Future work will survey key-stakeholders to priortise adverse events and outcomes and agree a COS to report as a minimum in all studies.