OBJECTIVES: The total cost of care (TC) for children with cerebral palsy (CP) who use botulinum toxin type A (BTX-A) was compared to the TC for CP children not using BTX-A. METHODS: A nested case control design compared BTX-A users and non-users in the South Carolina Medicaid program from 1995 through 2001. Patients with at least one CP diagnosis between 1996 and 1999 who were age 18 or younger were included. They were followed for one year prior to study enrollment, and 24 months after enrollment. A 1:6 match of BTX-A users to CP patients (non-BTX-A users) was performed using propensity scores from a logistic regression model predicting BTX-A use with demographics, CP severity, analgesia use, inpatient hospitizations, Chronic Disease Score, CP comorbidities, and pre-period TC. Cases and controls were compared using logistic regression on post-period data to estimate the same model predicting BTX-A use. RESULTS: BTX-A users (n=58) were identified and matched to 348 randomly selected non-users. After matching, the cases and controls were not statistically different on any of the model variables used in the matching procedure. Estimation of the final model revealed that only CP severity was significantly related to BTX-A use (diplegia p = 0.0107, hemiplegia p <0.0001, and quadraplegia p <0.001). The Hazard Ratio for these variables revealed the likelihood of BTX-A use relative to the lowest CP severity level (diplegia HR = 5.91, hemiplegia HR = 23.88, and quadraplegia HR = 10.22). While BTX-A is most frequently used in more severe patients, a statistically significant difference in total cost of care was not found (p = 0.528). CONCLUSIONS: While a greater percentage of CP children treated with BTX-A had more severe diagnoses, their TC was not different from CP children not using BTX-A after controlling for prior-period conditions using propensity scores.