The psychometric properties of the HAPI-A were examined at intake and 90-day follow-up in consumers with mental illness (MI) or chronic addiction (CA) being served at one of 11 treatment facilities (n = 1168). A 4-factor subscale structure was confirmed and factor invariance tests indicated a single model for the CA and MI samples. Internal consistency and inter-rater reliability were good (ICCs = 0.72-0.88). Factor scores correlated with GAF ratings and predicted membership in severity/cost groupings proposed for setting reimbursement rates. Sensitivity to change was confirmed for all four factors for the CA sample and for three of four factors for the MI sample.