The emergence of Cutaneous Leishmaniasis (CL) has been associated with the presence and destruction of forests, leading to the view that forest environments typically increase infection risk. Here we examine county data for CL between 1995-2000 in Costa Rica, a country that has had proportionally the largest rate of landscape transformation in the new world.  Statistical analyses show that landscape cover by itself does not explain the spatial distribution of the disease; a key factor instead is social exclusion when the multidimensionality of this human disease is properly taken into account.  In particular, a higher proportion of people living close to the forest edge diminishes the risk of infection.  This fraction, rainfall and social marginalization exhibit
break-points, or threshold values, where the qualitative shape and magnitude of the
association changes strongly.  In addition, forest cover appears to modulate the
temporal effect of ENSO at the county level.  Our results contradict the simple view of
forest cover as a major risk factor for this infectious disease and reveal a more
complex interplay of environmental and social factors.