In the last decade, the epidemiology of malaria has changed quite significantly in the Americas. Between 2000 and 2009, the number of cases reported for the region decreased by 50%, from 1.18 million to 526,000, while the variation within the countries included a 90% decrease in Suriname and almost a 200% increase in Haiti. The ratio between the number of cases of Plasmodium vivax malaria and Plasmodium falciparum malaria did not vary greatly, holding steady at approximately 3:1 .
The Amazon subregion continues to report about 90% of the total malaria cases for the Americas, as well as the majority of P. falciparum cases, although the absolute number of cases in 2000-2008 decreased from 992,061 to 515,440, or 50% , similar to the regional trend.
Evaluating the impact of the different control strategies on the reduction of malaria described above is an important task; however, it is an extremely complex task with high costs. In addition, the positive changes described probably do not result solely from the control strategies, but are also the result of a combination of factors that include changes in the environmental and epidemiological conditions determining the transmission of malaria.
Because public health interventions are complex, it is important to be clear about the specific aspects, programmes, and interventions to be evaluated. Habicht et al.  propose a logical framework to evaluate the performance and impact of public health interventions. This framework has been used in various international studies [4–7]. The design comprises three levels, ranging from simple to complex, depending on the evaluation designs and type of inference one wishes to make: (a) adequacy, (b) plausibility and, (c) probability. The adequacy level is the most basic and refers to evaluation of public health interventions relative to criteria generally corresponding to technical aspects of the interventions, quality of implementation and coverage of services. The inference that is made in an adequacy evaluation is whether the interventions are being implemented as planned and achieving the desired objectives (i.e., coverage, quality or others). The plausibility level requires a control group, and the inference indicates the possibility that the effects observed in the intervention group result from the intervention and not external factors. The probability level requires that the control group be assigned randomly, and the inference is presented in terms of effects in the intervention group being results of the intervention with known certainty and statistical reliability, and not results of mixed variables, bias or chance.
The Habicht et al.  framework is also useful in identifying the various levels of evidence required to make decisions about a public health intervention. For example, if one has no knowledge or evidence that an intervention is being implemented adequately, according to technical standards and with the necessary resources, then a probabilistic evaluation cannot be considered because any effects and impact the evaluation finds cannot be attributed to the intervention. In addition, if the intervention has not been implemented in an adequate manner, one can anticipate that it will have no impact. In these cases, one must first evaluate the adequacy of the interventions. Application of the logic described above also contributes to avoiding unnecessary expenses in evaluation studies when simpler evaluations can be carried out .
Based on the framework proposed by Habitch et al. , the adequacy level was chosen for this study which involved evaluating whether the malaria control strategies are implemented in a technically correct manner, with the necessary resources and required quality.
The following malaria control strategies were evaluated: (a) indoor residual spraying, (b) insecticide-treated bed nets, (c) timely diagnosis and, (d) artemisinin-based combination therapy (ACT). The five countries studied were Bolivia, Colombia, Ecuador, Guyana and Peru.