POS1448 EVIDENCE BASED PRACTICE: WHAT IS THE EVIDENCE THAT BRITISH SOCIETY FOR RHEUMATOLOGY GUIDELINES ARE EVIDENCE BASED?

Background:Clinical practice guidelines are designed to ensure that patients are treated according to best evidence, with the goal of optimizing clinical outcomes and reducing unwarranted variation in care. They compile, rate and translate the data available into recommendations that form the basis of evidence-based practice for most clinicians. Despite their importance, the evidence base informing different guidelines varies in quality. A recent study of American College of Rheumatology (ACR) Practice Guidelines demonstrated only 17 of 35 class I (strong benefit to harm ratio) recommendations were supported by level A evidence (high quality randomized controlled trails or meta-analyses)1.Objectives:To review the evidence supporting the British Society for Rheumatology (BSR) guidelines.Methods:Thirteen sets of guidelines that were available on the BSR website as of October 16th 2019 were reviewed (https://www.rheumatology.org.uk/practice-quality/guidelines). A range of methodologies (including Grading of Recommendations Assessment, Development and Evaluation (GRADE), Scottish Intercollegiate Guidelines Network (SIGN), EULAR and Royal College of Physicians (RCP) recommendations) were used to assess the quality of evidence and strength of recommendation. For comparability between guidelines the level of evidence was converted to a score between I (highest quality) and IV (lowest quality) and the strength of recommendation was converted to a rating between A and D. The polymyalgia rheumatica guideline was not assessed due to unclear methodology and lack of level of evidence for all recommendations.Results:Of the 12 BSR guidelines assessed, there were 554 recommendations in total. The number of recommendations per guideline ranged between 13 and 80. Across all assessed guidelines, 94 recommendations (17.0%) were classified as level I, 161 (29.1%) as level 2 and 299 (54.0%) as level 3 or 4. These figures are similar to those reported in the ACR guidelines (23%, 19% and 58% respectively)1. The proportion of level I evidence varied from 46.2% (Axial Spondyloarthropathy guideline) to 0% (Hot Swollen Joint guideline).Conclusion:Over half of all BSR guideline recommendations have level of supporting evidence of III/IV. A wide range of methodologies are used to generate BSR guidelines (GRADE, SIGN, RCP / EULAR). This makes it challenging for readers unfamiliar with these approaches to interpret evidence and hinders comparisons between guidelines. A standardized methodology for future guideline development would overcome these barriers.References:[1]Duarte-Garcia A, Zamore R & Wong JB. The Evidence Basis for the American College of Rheumatology Practice Guidelines. JAMA Intern Med, 2018 Jan 1;178(1):146-148.Disclosure of Interests:None declared