To determine if hospital-based clinical pharmacy services and pharmacy staffing continue to be associated with mortality rates. A database was constructed from 1998 MedPAR, American Hospital Association's Annual Survey of Hospitals, and National Clinical Pharmacy Services databases, consisting of data from 2,836,991 patients in 885 hospitals. Data from hospitals that had 14 clinical pharmacy services were compared with data from hospitals that did not have these services; levels of hospital pharmacist staffing were also compared. A multiple regression analysis, controlling for severity of illness, was used. Seven clinical pharmacy services were associated with reduced mortality rates: pharmacist-provided drug use evaluation (4491 reduced deaths, p=0.016), pharmacist-provided in-service education (10,660 reduced deaths, p=0.037), pharmacist-provided adverse drug reaction management (14,518 reduced deaths, p=0.012), pharmacist-provided drug protocol management (18,401 reduced deaths, p=0.017), pharmacist participation on the cardiopulmonary resuscitation team (12,880 reduced deaths, p=0.009), pharmacist participation on medical rounds (11,093 reduced deaths, p=0.021), and pharmacist-provided admission drug histories (3988 reduced deaths, p=0.001). Two staffing variables, number of pharmacy administrators/100 occupied beds (p=0.037) and number of clinical pharmacists/100 occupied beds (p=0.023), were also associated with reduced mortality rates. The number of clinical pharmacy services and staffing variables associated with reduced mortality rates increased from two in 1989 to nine in 1998. The impact of clinical pharmacy on mortality rates mandates consideration of a core set of clinical pharmacy services to be offered in United States hospitals. These results have important implications for health care in general, as well as for our profession and discipline.