We evaluated direct relationships and associations among clinical pharmacy services, pharmacist staffing, and drug costs in United States hospitals. A database was constructed from the 1992 American Hospital Association's Abridged Guide to the Health Care Field and the 1992 National Clinical Pharmacy Services database. Multiple regression analysis, controlling for severity of illness, was employed to determine the associations. The study population consisted of 934 hospitals. Four clinical pharmacy services were associated with lower drug costs: in-service education, $77,879.19+/-$56,203.42 (a total of $48,518,735.37 for the 623 hospitals offering this service, p=0.016); drug information, $430,579.84+/-$299,232.76 ($90,852,346.24 for the 211 hospitals offering this service, p=0.015); drug protocol management, $137,333.67+/-$98,617.83 ($45,045,443.76 for the 328 hospitals offering this service, p=0.049); and admission drug histories, $213,388.21+/-$201,537.85 ($5,548,093.46 for the 26 hospitals offering this service, p=0.011). As staffing increased for hospital pharmacy administrators (p<0.0001), dispensing pharmacists (p<0.0001), and pharmacy technicians (p<0.0001), drug costs increased. As staffing increased for clinical pharmacists, drug costs decreased (p=0.018). The results of this study show that increased staff levels of clinical pharmacists and some clinical pharmacy services are associated with reduced hospital drug costs.