OBJECTIVES: To assess the validity of Medicaid Analytic Extract (MAX)-2009, Outpatient (OT) and Provider Characteristics (PC) data for use in mental health research. METHODS: Medicaid issue briefs, Medicaid managed-care enrollment report, MAX-PC evaluation reports and MAX-validation tables for the year 2009 were reviewed to assess the validity of MAX-OT and MAX-PC data. Validity of the data was defined in terms of availability [OT: substantial number of encounter claims; PC: substantial number of OT billing/servicing provider IDs], completeness [OT: number of claims per PME, % enrollees with encounter record, number of claims per service user; PC: % provider IDs with primary taxonomy code] and usability [OT: primary diagnoses and/or procedure code reported per record]. Each state was ranked on these criteria as 1: Good, 2: Fair and 3: Poor for both OT and PC data and a total combined-score was calculated. The states with minimum total combined-score, which met all the criteria, were considered appropriate for research. RESULTS: Thirty-six states delivered some/all mental health services through managed care organizations (MCOs) and the rest relied on fee-for-service (FFS). Eleven states contracting with comprehensive MCOs, 4 states contracting with behavioral health organizations (BHOs), and 6 FFS states have usable OT-encounter data. PC data was available for all 51 states, of which 15 states have usable data with >90% claims with OT billing/servicing provider IDs. Twelve of these states have data with ≥75% provider IDs with primary taxonomy codes. Three comprehensive MCO states (KY, NJ, NY), 2 BHO states (AZ, FL) and 2 FFS states (MS, LA) met all the validity criteria for OT and PC files. CONCLUSIONS: Although, most states provide mental health services through comprehensive MCOs, OT-encounter data for these services is usable for limited states only. Researchers should exercise caution while combining OT and PC files for mental health research.