Abstract In teaching centers, primary failure of thoracic epidural analgesia can be due to multiple etiologies. In addition to the difficult anatomy of the thoracic spine, the conventional end point—loss-of-resistance—lacks specificity. Furthermore, insufficient training compounds the problem: learning curves are nonexistent, pedagogical requirements are often inadequate, supervisors may be inexperienced, and exposure during residency is decreasing. Any viable solution needs to be multifaceted. Learning curves should be explored to determine the minimal number of blocks required for proficiency. The problem of decreasing caseload can be tackled with epidural simulators to supplement in vivo learning. From a technical standpoint, fluoroscopy and ultrasonography could be used to navigate the complex anatomy of the thoracic spine. Finally, correct identification of the thoracic epidural space should be confirmed with objective, real-time modalities such as neurostimulation and waveform analysis.