Amphibian population declines and range constrictions are widespread but poorly understood. Effective conservation planning relies on accurate distribution data to develop a fundamental understanding of causal changes in species distributions. However, conventional detection methods for cryptic and elusive amphibians suffer from imperfect detection. Environmental DNA (eDNA) has emerged as an innovative and powerful conservation tool for detecting aquatic species presence; however, comparative studies aimed at fully understanding eDNA detection probabilities are lacking. In this study, I used environmental DNA (eDNA) sampling methods and detailed historical records to identify drivers of extirpation and sampling method efficiency for an imperiled, long-lived giant salamander, the Eastern Hellbender (Cryptobranchus alleganiensis alleganiensis) in West Virginia, USA. I used a site occupancy and detection modeling framework (SODM) to test the influence of current land use, historical mining, hydrogeomorphic, and water quality variables on model-based predictions of occupancy and detection. Hellbenders are extirpated from 51% of the 49 historical sampling sites, and the top-ranked model indicated watershed-scale road density was the strongest predictor of Eastern Hellbender extirpation, and water turbidity and electrical conductivity were the best predictors of detection. Detection probability estimates for eDNA (84%) and conventional sampling methods (28%) suggest that eDNA provides a substantial performance advantage over conventional detection methods. Integrating eDNA data within a SODM framework allowed me to accurately and thoroughly assess causal changes in Eastern Hellbender distribution throughout their historical range in West Virginia, which will aid conservation planning. This study emphasizes the impacts of anthropogenic land alterations on freshwater ecosystems and the sensitivity of long-lived amphibian species to rapid environmental change.