OBJECTIVES: High-risk, non-muscle-invasive bladder cancer (NMIBC) is heterogeneous in its presentation, resulting in patient subpopulations with diverse treatment options. A comprehensive model estimating costs and health outcomes with various NMIBC treatment strategies is needed for diverse patient subpopulations. METHODS: A Markov model simulating patient outcomes was developed based on published treatment guidelines. Health states encompass high-risk NMIBC, tumor-free, muscle-invasive progression (MIP), and metastasis. Four patient populations were considered: (1) high-risk T1 or Ta tumors; (2) Carcinoma in situ (Cis) only; (3) high-risk T1/Ta tumors with concomitant Cis; (4) general NMIBC population with high-risk T1/Ta tumors and/or Cis. Treatment options include trans-urethral resection (TUR) and adjuvant intravesical bacillus Calmette-Guérin (BCG), mitomycin C (MMC) or valrubicin for populations (1), (3) and (4), or intravesical treatment alone for population (2). The model assesses treatments as first- or second-line, or as alternatives for patients intolerant of or refractory to other treatment. Radical cystectomy is performed after MIP or repeated treatment failure. Response (in patients with Cis) and recurrence rates and percentages of BCG-intolerant/refractory patients were estimated from the literature. Costs were obtained from publicly available sources. The model was validated against published epidemiology and cost data. RESULTS: The lifetime cost per person of treating high-risk T1/Ta patients with BCG was within 20% of costs estimated in published economic models. The lifetime cost per person of treating BCG-intolerant/refractory Cis patients with valrubicin was estimated to be 45% higher than for cystectomy. Adverse event costs for the general high-risk NMIBC population treated with BCG account for 17% of the total treatment cost compared with 5% for MMC. Approximately 40% and 15% of patients treated with BCG eventually undergo cystectomy and MIP, respectively.  CONCLUSIONS: This validated model can be used to estimate costs and health outcomes associated with existing treatment strategies as well as the cost-effectiveness of novel intravesical therapies.