Honey bees contribute substantially to the world economy through pollination services and honey production. In the U.S. alone, honey bee pollination is estimated to contribute at least $11 billion annually, primarily through the pollination of specialty crops. However, beekeepers lose about half of their hives every season due to disease, insecticides, and other environmental factors. Here, we explore and validate a spatiotemporal statistical model of Varroa destructor mite burden (in mites/300 bees) in managed honey bee colonies, exploring the impact of both environmental factors and beekeeper behaviors. We examine risk factors for Varroa infestation using apiary inspection data collected across the state of Illinois over 2018-19, and we test the models using inspection data from 2020-21. After accounting for spatial and temporal trends, we find that environmental factors (e.g., floral quality, insecticide load) are not predictive of Varroa intensity, while several beekeeper behaviors (e.g., smaller colony density, supplemental feeding, and mite monitoring/treatment) are protective against Varroa. Interestingly, while monitoring and treating for Varroa is protective, treating without monitoring is no more effective than not treating at all. This is an important result supporting Integrated Pest Management (IPM) approaches.