Objective: Little is known about medical expenditures in the multiple sclerosis population after the introduction of expensive disease modifying agents in the 1990s. This study examines new data from 2000-2005 population-based survey of MS to provide estimates of health services utilization by disease severity and controlling for other risk factors. Methods: We used a subsample (n=919) of patients with relapsing remitting and secondary progressive MS from the Sonya Slifka Longitudinal Multiple Sclerosis Study that follows a nationally representative cohort of MS patients. We examined utilization of hospital and outpatient care, emergency room (ER), therapy, mental health services, alternative medicine, home health and personal care. For most utilization categories, we used log-linear negative binomial regression models to estimate mean utilization, accounted for possible correlation of observations for the same person by using robust standard errors, and applied baseline sampling weights to obtain estimates representative for the entire US population with MS. For provider visits and amount of personal care (conditional on using it), we estimated similar GLS Gaussian models with logarithmic transformation of the number of services as dependent variable. Results: Use of most medical services increased significantly with the worsening of MS. However, patients with severe MS (EDSS >7.0) used the same or fewer services than patients with EDSS 6.5-7 except for home health and personal care. Having relapses significantly increased use of most categories of health services. Longer time since diagnosis, after adjusting for MS stage, was associated with a lower level of utilization of ER, hospital outpatient care, therapy, and alternative medicine. Patients who were not married generally used more services than married patients. Conclusion: While MS places a substantial financial burden on individuals as well as private and public payers, use of specific categories of health services varies significantly with MS severity, presence of relapses, and availability of informal help.