Evaluation of plans and policies is a key element in their administration and must be performed under real conditions. Such evaluation is complex, as plans and policies include a diverse set of components that operate simultaneously. Moreover, external factors frequently influence those same issues that programs attempt to change. Unless plans and policies are evaluated under real conditions, a policy that effectively reduces the effects of a problem may be deemed ineffective (if the problem increases due to the influence of factors the program does not attempt to affect), or a policy that is unable to influence the problem it attempts to solve may be judged useful (if the magnitude of the problem is being reduced through the influence of factors other than the policy). The present article discusses evaluation of health policies, plans or complex programs, with emphasis on effectiveness assessment, using data from real examples. Among other issues, the need to identify the distinct components of policies and plans is reviewed. This article also describes how to evaluate the outcome or results of a program with indicators from other sources. Aspects related to the timing of evaluation and assessment indicators are analyzed. We discuss situations in which the launch of a new policy or intervention is followed by an increase in the reported magnitude of the problem it attempts to solve. These situations are illustrated by cases in which this increase is attributable to improved detection and by others in which the increase is related to factors external to the intervention. The frequent confusion of the effects of the intervention with other events is covered, with data from some examples. Finally, evaluation of plans that include a wide range of objectives is also addressed.