17 Citations (Scopus)
259 Downloads (Pure)

Abstract

Introduction Randomised controlled trials (RCTs) in surgery are frequently criticised because surgeon expertise and standards of surgery are not considered or accounted for during study design. This is particularly true in pragmatic trials (which typically involve multiple centres and surgeons and are based in ‘real world’ settings), compared with explanatory trials (which are smaller and more tightly controlled).

Objective This protocol describes a process to develop and test quality assurance (QA) measures for use within a predominantly pragmatic surgical RCT comparing minimally invasive and open techniques for oesophageal cancer (the NIHR ROMIO study). It builds on methods initiated in the ROMIO pilot RCT.

Methods and analysis
We have identified three distinct types of QA measure: (i) entry criteria for surgeons, through assessment of operative videos, (ii) standardisation of operative techniques (by establishing minimum key procedural phases) and (iii) monitoring of surgeons during the trial, using intraoperative photography to document key procedural phases and standardising the pathological assessment of specimens. The QA measures will be adapted from the pilot study and tested iteratively, and the video and photo assessment tools will be tested for reliability and validity.

Ethics and dissemination
Ethics approval was obtained (NRES Committee South West—Frenchay, 25 April 2016, ref: 16/SW/0098). Results of the QA development study will be submitted for publication in a peer-reviewed journal.
Original languageEnglish
Article numbere026209
Number of pages7
JournalBMJ Open
Volume9
Issue number3
Early online date28 Feb 2019
DOIs
Publication statusPublished - Mar 2019

Research Groups and Themes

  • BTC (Bristol Trials Centre)
  • BRTC
  • Centre for Surgical Research

Fingerprint

Dive into the research topics of 'Protocol for developing quality assurance measures to use in surgical trials: an example from the ROMIO study'. Together they form a unique fingerprint.

Cite this