Heterogeneous distributions are a fundamental principle of ecology, manifesting as natural variability within ecological levels of organization from individuals to ecosystems. In disease ecology, variability in biotic and abiotic factors can result in heterogeneous patterns of transmission and virulence—broadly defined here as the negative consequences of infection. Still, our classic theoretical understanding of disease dynamics comes from models that assume homogeneous transmission and virulence. Here, we test this assumption by assessing the contribution of various sources of individual and spatial heterogeneity to patterns of transmission and sublethal measurements of virulence in two lizard–malaria systems: a three‐parasite assemblage (Plasmodium floridense, Plasmodium leukocytica, and Plasmodium azurophilum) infecting the lizard Anolis gundlachi in the rainforest of Puerto Rico and a single‐parasite system (P. floridense–Anolis sagrei) in Florida. Using a Bayesian model selection framework, we evaluated whether individual host differences (i.e., body size and sex) or spatial variability (i.e., habitat type and local‐scale host spatial structure) drive heterogeneity in the probability of infection or its associated health costs (i.e., body condition, blood chemistry). We found that the probability of infection increases with increasing lizard body size in both systems. However, in Florida, we found the relationship to be subdued in deforested habitats compared to the adjacent urban hydric forests. Furthermore, infection was spatially clustered within sampling sites, with “hot” and “cold” spots across the landscape. Nevertheless, we found no clear evidence of costs of infection on lizard health in any of the measures assessed and hence no grounds for inference regarding heterogeneous virulence. Ultimately, the consistency of our results across systems suggests prominent roles of individual and spatial heterogeneities as driving factors of transmission of vector‐borne diseases.