The direct relationships and associations among clinical pharmacy services, pharmacist staffing, and medication errors in United States hospitals were evaluated. A database was constructed from the 1992 National Clinical Pharmacy Services database. Both simple and multiple regression analyses were employed to determine relationships and associations. A total of 429,827 medication errors were evaluated from 1081 hospitals (study population). Medication errors occurred in 5.22% of patients admitted to these hospitals each year. Hospitals experienced a medication error every 22.04 hours (every 19.13 admissions). These findings suggest that at minimum, 90,895 patients annually were harmed by medication errors in our nation's general medical-surgical hospitals. Factors associated with increased medication errors/occupied bed/year were drug-use evaluation (slope = 0.0023476, p=0.006), increased staffing of hospital pharmacy administrators/occupied bed (slope = 29.1972932, p<0.001), and increased staffing of dispensing pharmacists/occupied bed (slope = 19.3784148, p<0.001). Factors associated with decreased medication errors/occupied bed/year were presence of a drug information service (slope = -0.1279301, p<0.001), pharmacist-provided adverse drug reaction management (slope = -0.3409332, p<0.001), pharmacist-provided drug protocol management (slope = -0.3981472, p=0.013), pharmacist participation on medical rounds (slope = -0.6974303, p<0.001), pharmacist-provided admission histories (slope = -1.6021493, p<0.001), and increased staffing of clinical pharmacists/occupied bed (slope = -9.5483813, p<0.001). As staffing increased for clinical pharmacists/occupied bed from the 10th percentile to the 90th percentile, medication errors decreased from 700.98 +/- 601.42 to 245.09 +/- 197.38/hospital/year, a decrease of 286%. Specific increases or decreases in yearly medication errors associated with these clinical pharmacy services in the 1081 study hospitals were drug-use evaluation (21,372 more medication errors), drug information services (26,738 fewer medication errors), adverse drug reaction management (44,803 fewer medication errors), drug protocol management (90,019 fewer medication errors), medical round participation (42,859 fewer medication errors), and medication admission histories (17,638 fewer medication errors). Overall, clinical pharmacy services and hospital pharmacy staffing variables were associated with medication error rates. The results of this study should help hospitals reduce the number of medication errors that occur each year.