Relationships between the environment and health outcomes are complex and likely nonlinear in nature. However, until recently, most studies used ordinary linear regression to model these relationships. The overall goal of this research was to investigate nonlinear relationships between the environment and health. To accomplish this goal, we used several large, national datasets across varying populations and local environments. Destination accessibility is an important measure of the built environment that is associated with active transport and body mass index (BMI). In the first study, we sought to determine the relationship between the density of nonresidential destinations (a proxy for walkability) and BMI, allowing for the possibility of a nonlinear relationship. We merged information from 17.2 million driver's license records with the locations of 3.8 million nonresidential destinations and census tract socioeconomic data from six states. BMI peaked in the middle density, with significantly lower values in both the low and high-density extremes -- a markedly nonlinear relationship. Next, we confirmed our previous nonlinear findings in an independent sample of 2,405 primary care patients with multiple chronic conditions from 13 states, and extended our analysis to include mental and physical health outcomes, in addition to BMI. Several statistical methods were used to confirm the nonlinear relationship between nonresidential destinations and BMI. We also established novel nonlinear relationships between nonresidential destinations and mental health. All three health measures were significantly worse in middle density areas with better values on either extreme. Then, we extended the previous analyses to the natural environment. We used data on 3,409 adults from 119 US counties and the natural amenities scale, a county-level measure of the natural environment, to assess the relationship between the natural environment and health at the intersection of various demographic and social factors, allowing for the possibility of a non-linear relationship. Health was generally worse in areas with poor natural environments; however, this relationship was not linear. In areas with low natural amenities, greater amenities were associated with better physical and mental health, but only for advantaged populations. Meanwhile greater amenities in high amenity areas was associated with a decrease in mental and physical health for disadvantaged populations. Finally, in the review paper, we described the current state of the literature on the nonlinear relationships between walkability and health. We argue that using linear regression techniques to model nonlinear relationships could introduce bias and be partially responsible for the conflicting findings in the literature. We conclude that there are nonlinear relationships between the environment and health. Complex relationships require complex modelling. Ignoring the possibility of a nonlinear relationship could obscure the true relationship and lead researchers and public health officials to draw incorrect conclusions. Future research should confirm these findings and investigate the mechanisms driving these relationships.