Background: The Canadian Agency for Drugs and Technologies in Health’s (CADTH’s) Canadian Optimal Medication Prescribing and Utilization Service (COMPUS) is a collaborative, pan-Canadian service funded by Health Canada. In partnership with federal, provincial, and territorial health ministries, COMPUS identifies and promotes the use of optimal therapy. COMPUS produces optimal therapy recommendations through the COMPUS Expert Review Committee (CERC). In 2007, COMPUS applied the Grades of Recommendation, Assessment, Development and Evaluation (GRADE) method to develop recommendations for use of insulin analogues in the management of diabetes mellitus (DM). GRADE is a relatively new method for the evaluation and presentation of evidence to assist those responsible for the development of clinical guidelines and recommendations. However, there is little published information describing the use of GRADE, and this approach has rarely been used in Canada. Objectives: This presentation demonstrates how COMPUS implemented the GRADE approach to develop recommendations for insulin analogue use in Canada. Methods: Outcomes relevant to prescribing decisions for the use of insulin analogues in the management of DM were identified and ranked by an expert panel. Unique GRADE Evidence Profiles were developed for each treatment comparison and specific patient population. Each Evidence Profile presented meta-analytic estimates of effect size and, where data permitted, pharmacoeconomic results. Quality and internal and external validity of the evidence was assessed. Recommendations were developed, and the strength of each was determined. Results: Thirty-two outcomes were identified as relevant to prescribing decisions. Estimates of effect size derived from 94 meta-analyses. Thirty-three Evidence Profiles were produced from which recommendations were generated. Modifications were made to the application of the GRADE approach to add to its effectiveness in the development of the recommendations. Conclusions: Incorporating evidence from single studies and integrating resource utilization data presented unique challenges when using GRADE. However, GRADE provides a systematic and transparent process for identifying, analyzing and presenting a large body of evidence for the development of evidence-based optimal therapy recommendations.