Improving the safety, quality, and efficiency of care with the help of clinical decision support tools is one of the core objectives in the meaningful use of Electronic Health Records. Successful adoption of support systems depends on the quality of delivered information, its relevance to the clinical task and individual patient, integration of the system with the entire clinical workplace, and ease of use of the system. This paper presents continuous development and evaluation, as well as lessons learned in development and maintenance of an evidence-based system that supports development of individualized patient care plans. Since its deployment in August 2009, the Evidence-Based Practice InfoBot (EBP InfoBot) system is in daily use at the NIH Clinical Center and responds to 21 requests a day, on average.