BACKGROUND : The effectiveness and value of teledermatology and face-to-face workflows for diagnosing lesions are not adequately understood. OBJECTIVE : We compared the risks of biopsy and cancer diagnosis among 2 face-to-face workflows (direct referral and roving dermatologist) and 4 teledermatology workflows. METHODS : Retrospective study of 59,279 primary care patients presenting with a lesion from January through June 2017. RESULTS : One teledermatology workflow achieved high-resolution images with use of a dermatoscope-fitted digital camera, a picture archiving and communication system, and image retrieval to a large computer monitor (in contrast to a smartphone screen). Compared with direct referral, this workflow was associated with a 9% greater probability of cancer detection (95% confidence interval [CI], 2%-16%), a 4% lower probability of biopsy (relative risk, 0.96; 95% CI, 0.93-0.99), and 39% fewer face-to-face visits (relative risk, 0.61; 95% CI, 0.57-0.65). Other workflows were less effective. CONCLUSIONS : Differing proficiencies across teledermatology workflows and selection of patients for direct referral could have caused bias. CONCLUSIONS : Implementation is critical to the effectiveness of teledermatology.