Understanding the drivers of health and economic cost for the treatment of mental health conditions is critical to meet the accelerating demands for care. We conducted an economic evaluation of real-world healthcare-systems data from 27,540 patients receiving care for a mood or an anxiety disorder within the UK National Healthcare Service. Using Markov models built on discrete health states to compare the cost-effectiveness of different interventions, we show that the principal drivers of healthcare cost relate to waiting times and treatment effectiveness. We find that internet-delivered cognitive behavioral therapy has a ‘dominant’ incremental cost-effectiveness ratio relative to standard care, offering similar clinical effectiveness but with shorter treatment times. In most healthcare systems, the clinical effectiveness of mental healthcare remains unquantified, and long treatment times are common. The potential for these findings to inform mental healthcare policy is substantial, particularly around immediacy of access and the importance of outcomes-focused quality management.