Evidence based surgery

: Evaluating Outcome Measurement, Methodology and Reporting Quality

Abstract

The body of work in this thesis presents the research and publications under a general theme of outcome measurement, evidence synthesis and reporting quality in surgery. The work highlights the author’s own personal contributions and publications under this theme, and collaborations with colleagues from Oxford University, Harvard Medical School, Imperial College London and University College London. This thesis provides an in depth commentary on evidence-based surgery, with discussion on the challenges of conducting randomised controlled trials in surgery. Systematic review and meta-analysis methodology is discussed, exploring the nuances and assumptions of random/fixed effect models, quality assessment using GRADE and assessment of risk of bias (RoB) using Cochrane’s ROBINS-I tool. The author has evaluated reporting quality in surgery, identifying suboptimal compliance with the CONSORT-NPT checklist. This work formed basis for change of policy and the requirement for mandatory completion and uploading of a CONSORT statement by authors when submitting articles to the peer reviewed journal, International Journal of Surgery (IJS), with significant improvement in compliance.The commentary also reviews plastic & reconstructive breast surgery, and provides an in-depth discussion on quality of life assessment, COSMIN, minimal important differences (MID), how to choose a questionnaire and particular domains in a study, and reporting using CONSORT and SPIRIT PRO checklists. Health-utility measures for cost-utility analyses are also discussed. The authors’ systematic reviews and meta-analyses are presented on clinical outcomes and PROs of DIEP versus implant-based reconstruction; and on immediate versus delayed reconstruction, in context of radiotherapy (RT). The former review provides a weak recommendation that DIEP reconstruction maybe more cost-effective and yield higher PRO scores, with suitable warnings in light of poor quality and serious risk of bias. The RT review identified no statistically significant difference in outcomes between immediate and delayed breast reconstruction (BRR), challenging dogma where majority of UK BRR are delayed, with significant heterogeneity in outcome measurement, suboptimal reporting of core outcome set and no grading of complications. The reviews have demonstrated paucity of high quality evidence and the need for future high quality studies. The PhD award will facilitate the author in establishing a research group and to undertake postdoctoral research. This will include conducting a national stream funded large prospective cohort study in evaluating immediate versus delayed autologous BRR, in context of radiotherapy, with robust reporting of BRR core outcome set and incorporation and measurement of disease-specific PROs and cost-effectiveness, addressing the limitations highlighted by the reviews.

Date of Award	21 Sept 2020
Original language	English
Awarding Institution	University of Portsmouth
Supervisor	Graham Mills (Supervisor)