Reporting of outcomes in gastric cancer surgery trials: a systematic review

Background The development of clinical guidelines for the surgical management of gastric cancer should be based on robust evidence from well-designed trials. Being able to reliably compare and combine the outcomes of these trials is a key factor in this process. Objectives To examine variation in outcome reporting by surgical trials for gastric cancer and to identify outcomes for prioritisation in an international consensus study to develop a core outcome set in this field. Data sources Systematic literature searches (Evidence Based Medicine, MEDLINE, EMBASE, CINAHL, ClinicalTrials. gov and WHO ICTRP) and a review of study protocols of randomised controlled trials, published between 1996 and 2016. Intervention Therapeutic surgical interventions for gastric cancer. Outcomes were listed verbatim, categorised into groups (outcome themes) and examined for definitions and measurement instruments. Results Of 1919 abstracts screened, 32 trials (9073 participants) were identified. A total of 749 outcomes were reported of which 96 (13%) were accompanied by an attempted definition. No single outcome was reported by all trials. 'Adverse events' was the most frequently reported 'outcome theme' in which 240 unique terms were described. 12 trials (38%) classified complications according to severity, with 5 (16%) using a formal classification system (Clavien-Dindo or Accordion scale). Of 27 trials which described 'short-term' mortality, 15 (47%) used one of five different definitions. 6 out of the 32 trials (19%) described 'patient-reported outcomes'. Conclusion Reporting of outcomes in gastric cancer surgery trials is inconsistent. A consensus approach to develop a minimum set of well-defined, standardised outcomes to be used by all future trials examining therapeutic surgical interventions for gastric cancer is needed. This should consider the views of all key stakeholders, including patients.