Objective: To identify sociodemographic, clinical, and physician/practice factors associated with deep brain stimulation (DBS). DBS is a proven surgical therapy for Parkinson disease (PD), but is recommended only for patients with excellent health, results in significant out-of-pocket costs, and requires substantial physician involvement. Methods: Retrospective cohort study of more than 657,000 Medicare beneficiaries with PD. Multivariable logistic regression models examined the association between demographic, clinical, socioeconomic status (SES), and physician/practice factors, and DBS therapy. Results: There were significant disparities in the use of DBS therapy among Medicare beneficiaries with PD. The greatest disparities were associated with race: black (adjusted odds ratio [AOR] 0.20, 95% confidence interval [CI] 0.16-0.25) and Asian (AOR 0.55, 95% CI 0.44-0.70) beneficiaries were considerably less likely to receive DBS than white beneficiaries. Women (AOR 0.79, 95% CI 0.75-0.83) also had lower odds of receiving DBS compared with men. Eighteen percent of procedures were performed on patients with PD who had cognitive impairment/dementia, a reported contraindication to DBS. Beneficiaries treated in minority-serving PD practices were less likely to receive DBS, regardless of individual race (AOR 0.76, 95% CI 0.66-0.87). Even after adjustment for demographic and clinical covariates, high neighborhood SES was associated with 1.4-fold higher odds of receiving DBS (AOR 1.42, 95% CI 1.33-1.53). Conclusions: Among elderly Medicare beneficiaries with PD, race, sex, and neighborhood SES are strong independent predictors of DBS receipt. Racial disparities are amplified when adjusting for physician/clinic characteristics. Future investigations of the demographic differences in clinical need/usefulness of DBS, ease of DBS attainment, and actual/opportunity DBS costs are needed to inform policies to reduce DBS disparities and improve PD quality of care.