Patient safety projects often interface with multiple influencing factors not necessarily anticipated in their original design. Clean and explicit results, as traditionally identified in controlled experimentation may be elusive due to institutional dynamics in leadership, facilities, and priorities. This presentation will describe two years' experience with patient safety projects at UCLA, during an interval when several significant changes occurred. The effects on data-driven projects will be described. Several ideas will be shared for group discussion regarding how to manage and interpret patient safety projects in a real-time world of constant change.