T he challenge of how to speed and sustain improvements in a complex health system continues to be a vexing one. It's a challenge that has become even more urgent because of the current proposals for national health care reform. The success of that reform will depend in large part on whether we can learn how to transform our health care system into one that is more effective, reliable and efficient. The national meeting of the Quality Enhancement Research Initiative (QUERI) of the Department of Veterans Affairs, Veterans Health Administration (VA) in late 2008 1 offered a chance to reflect on what we have learned over a decade of studying and facilitating improvement for priority conditions within the VA. Although there are unique features to the VA due to the patients we serve and our integrated structure, many of the lessons are broadly applicable.The most obvious lesson is a familiar but painful one: change is hard and often slow. Many of the most impressive improvements that the VA and other health systems have made over the past 10 years-for example, improving cancer screening, immunizations and cardiac care with combinations of reminder systems, performance measurement, education and practice changes-are now behind us. 2 To maintain improvement going forward, we will have to address new challenges, and the gains may be more modest. In baseball parlance, if QUERI was born with visions of becoming the slugger who belts towering home runs, it has grown up into a solid utility player who scatters singles, infield hits and sacrifice flies. But as any fan knows, you can score more runs by a steady succession of hits than by always swinging for the fences. Meaningful progress is still possible, but it requires that we remain focused on specific goals, work to align all the different parts of the health care system and stay pointed in a consistent direction.The second lesson relates to the still young discipline of implementation science and translation research that forms the scientific core of QUERI. Implementation science has made important strides and is achieving growing attention, from new journals 3 and conferences 4 to the succession of Clinical and Translational Science Awards granted by the National Institutes of Health. At the same time, there still seems to be much more attention on translating basic science into clinical applicationsovercoming the "First Translational Block" (T1)-than on the equally important need to implement proven clinical applications more widely and reliably (T2). 5,6 If we want to attract more attention, resources and expertise to this effort, however, we need to do a better job of turning what we have learned into tangible tools and guidance for clinical managers and policy makers. It is time to move beyond attempts to develop the perfect theoretical model to explain the "black box" of the implementation process. We have numerous models to help organize and structure our observations and frame more specific hypotheses 7 . Common to all of them is an emphasis on the importance o...