OBJECTIVES: In the HAART era, timely identification and treatment of non-AIDS morbidities can enhance survival.  Awareness of disparities in rates and causes of death can inform public health strategies. METHODS: Patients studied had ≥2 visits to any of 10 HIV clinics during 1996-2009.  Death rates and causes were stratified by age, race, sex, healthcare payer and calendar period; trends were compared using Poisson regression. RESULTS: Among 8519 patients there were 993 deaths, 430 from AIDS, 469 from non-AIDS and 94 with unknown causes.  Overall death rates per 100 person-years of follow-up declined during 1996-2000, 2001-2004, and 2005-2009 from 3.85 to 2.12 to 1.33, respectively (p<0.01 for trend) as did AIDS death rates (1.96, 0.91, 0.33; p<0.01 for trend) and non-AIDS death rates (1.55, 1.09, 0.78; p<0.01 for trend).  Approximately one-third (31%) of mortality declines during 2005-2009 were from non-AIDS death reductions and 65% from AIDS death reductions; non-AIDS death rates remained higher than AIDS deaths among patients <45 (0.5 vs. 0.3, p<0.01) but not for patients ≥45 years. Overall 2005-2009 death rates were higher among blacks than non-blacks (2.03 vs. 1.01, p<0.01) and among persons with public healthcare payers (PUB) than non-public payers ([nPUB], 2.66 vs. 0.60, p<0.01). During 2005-2009, non-AIDS deaths accounted for 64.7% of deaths among blacks and 61.1% among PUB.  More than 88% of patients received HAART during 2005-2009 in all subgroups; rates differed by race (85.6% among blacks, 90.0% among non-blacks, p<0.01) but not by payer (88.2% among PUB, 88.8% among nPUB; p = 0.57) CONCLUSIONS: In the HOPS, continued mortality declines resulted from decreases in non-AIDS and AIDS-related deaths.  Excess mortality persisted among blacks and the publicly insured, with the majority of deaths from non-AIDS causes, suggesting that improved detection and management of chronic non-AIDS diseases are needed to achieve further mortality reductions.