Background: Disease management programmes (DMPs) are structured treatment programmes for chronically ill patients. The German Institute for Quality and Efficiency in Health Care (IQWiG) prepares synopses of evidence-based guidelines to identify relevant recommendations for DMPs; in this context, strong recommendations are particularly important. Guideline groups sometimes issue strong recommendations based on weak evidence ('discordant recommendations') if they regard their implementation to be important and wish to promote implementation. Analyses of guidelines of the Endocrine Society and the World Health Organization showed that about 35% of recommendations were discordant.

Objectives: To examine the frequency of discordant recommendations in evidence-based guidelines using the example of coronary heart disease (CHD) and to evaluate whether, in recent years, a trend in this frequency can be identified.

Methods: The analysis was based on 3 systematic searches for evidence-based guidelines conducted for 3 chronological guideline synopses on CHD. The guidelines were published between 1/2002-6/2007, 6/2007-9/2010, and 12/2011-12/2016. All recommendations and the corresponding Grade of Recommendation (GoR) and Level of Evidence (LoE) were extracted and checked for discordant recommendations. A quantitative descriptive analysis was performed.

Results: 21, 14 and 41 guidelines were included. The first synopsis (2007) included 816 recommendations, of which 61 (16%) were allocated to a high GoR in combination with a low LoE – 19 (31.1%) of these 61 were not supported by literature. The second synopsis (2010) included 403 recommendations, of which 48 (18%) were discordant – 16 (33.3%) of these 48 were not supported by literature. The discordant recommendations referred especially to diagnostics, coronary angiography, interventional and surgical coronary revascularisation and pharmacological therapy, but so far can only be identified by evaluating GoR and LoE. The analyses for the third synopsis (2016) are ongoing.

Conclusions: Discordant recommendations also exist in guideline synopses on CHD. To assess them they should be easy to identify and explained.