Karim,
The results cannot be interpreted in isolation but you will need thorough background information about the health and nutrition interventions in the sampling frame, information on access and programme coverage as well as the kind of sampling that was applied for the survey.
Due to issues of ‘mandate’, technical capacity of the ministry of health and partner organizations, resources/supply constraints, it is possible to have interventions largely targeting children with moderate malnutrition and not adequately those with severe acute malnutrition. In the short term, this may translate to a manageable proportion of children with moderate acute malnutrition and a more than proportionate number of children with severe acute malnutrition. Hence the scenario you are describing, GAM 10 % with > 2% SAM.
Alternatively, poor quality interventions targeting children with SAM (e.g. lack of in-patient care, low coverage, poor or no defaulter follow up, high degree of non-response). All these will contribute to a high prevalence of SAM resulting to the disproportionate balance you are describing.
This however does not rule out the possibility of poor quality data, often contributed to by biased sampling.