Introduction Clinical practice guidelines aim to improve patient care. They are based on best available evidence and are frequently viewed as “gold-standard” care for the disease or intervention that they address. The aim of this study was to determine the overall quality of the evidence supporting current British Society of Gastroenterology (BSG) guidelines.

Methods Guidelines were retrieved from the BSG website on 6th January 2012. Those posted after 2006 were considered current. The quality of supporting evidence was graded in accordance with the systems initially used to assess the primary literature. Adherence to the BSG's advice on guideline writing issued in 2010 was assessed in guidelines published thereafter.

Results 18 BSG guidelines currently exist addressing topics in endoscopy (n=7), luminal gastroenterology (n=8), and hepatology (n=3). Four guidelines published in the study period were updates of previous guidance. These were published a median of 7.5 years after the initial guidance. Of a total of 434 evidence-based recommendations the quality of evidence was low in 42.8% (range 7.1%–85.7%), that is, from case studies or consensus opinions. High quality evidence-based recommendations (consistent data from randomised controlled trials) accounted for only 14.3% of all recommendations (range 0–45.5%). Overall, there was significant heterogeneity between guidelines. These were developed using four different evidence-grading systems. In those published since 2010 only one out of eight guidelines adhered to the evidence grading system advised by the BSG Clinical Services and Standards Committee.

Conclusion

1. Evidence-based recommendations in current guidelines are most frequently based on low quality evidence, reflecting a lack of available high quality evidence.

2. There is significant heterogeneity in the current guidelines despite BSG committee advice on guideline writing.

These findings risk reducing the utility of these guidelines and highlight important areas of unmet research need that should be addressed as a matter of priority. Studies co-ordinated through loco-regional networks could substantially improve the evidence base and ultimately patient outcomes.

Competing interests None declared.