Objective. The objective of this study is to achieve consensus among experts on issues, apart from the scientific evidence, that should be considered when formulating recommendations in clinical practice guidelines.

Methods. The study was conducted from July 2003 until February 2004. An initial list of items was generated from relevant literature. The items were grouped into nine domains: (1) clinical relevance, (2) safety, (3) patient perspective, (4) availability of resources, (5) cost, (6) organisation of care, (7) professional perspective, (8) legal consequences, (9) industrial conflicts of interest. Experts in the field of guidelines development validated this list in a two round Delphi procedure. All items were scored using a 4-point Likert scale. In the second round, the participants were also asked to determine whether an item should be included or considered in the formulation of recommendations in clinical guidelines using the answering options yes or no.

Results. A total of 28 Dutch experts participated in the first Delphi round and 21 of them participated in the second round. The domains that received high priority were clinical relevance , safety , availability of resources and cost . The domain patient perspective was considered important as well but not all the items in this domain received a high score. The opinions varied most in the domain cost . The items in the domain professional perspective had relatively low scores. Using the threshold of 60% consensus on the value of yes , the initial pool of 49 items was reduced to 36 items grouped into 8 domains. The domain industrial conflicts of interest was excluded due to low scores.

Conclusions. This is the first systematic approach to explore other considerations apart from the evidence relevant in formulating the recommendations in clinical guidelines. The resulting 36 items can easily be used as a checklist during the process of guideline development. The next step is to test its validity and use in practice.