This paper tries to unravel the following question: why do we sometimes obtain results that are worse than expected despite having used technologies that are provenly efficacious or effective and having eliminated major groups of causes leading to poor performance? Inductive analysis and synthesis based on nine areas of health service research show that to effectively adopt some health technologies, it is not enough to simply choose an efficacious or effective change strategy. It sometimes becomes necessary to change the behavior of the health workers that will use it, and to modify certain environmental elements. Technology's effectiveness also depends on intervening simultaneously at various levels. Using a mix of evidence-based change (multifaceted) strategies is often mandatory. Improving health service calls for permanent, not sporadic, efforts; for ensuring universal access to information, and for adopting regulations to prevent poor or potentially harmful service delivery. A portion of the service improvements that have been attained have resulted from the use not of isolated change measures, but of combinations of the most effective measures as part of a single integrated intervention. To further reduce the gap between observed and expected effectiveness we should pilot behavior change strategies before adopting some health technologies, and to permanently install in our health system the multifactorial, integrated technology adoption mechanisms that we still lack. Failure to do this will mean being inefficient and pursuing short-term results at the expense of feasible and legitimate medium-term objectives.