OBJECTIVES: Drug administration via intravenous infusion is an important incident-to service for numerous U.S. physician practices. Effective staffing and patient scheduling can be better achieved by recognizing the individual components of the infusion activity. This project built a detailed infusion model and applied it to multi-site infusion observations. METHODS: A model of drug administration activity via intravenous infusion as performed in the physician's office was created that defined a series of individual tasks divided into sequential stages. The model was then applied to observations of infusions in 75 physician offices located in 26 states. Observation instruments were pretested. The sampling frame was a broad-based convenience sample. Measures were taken to ensure a representative sample as to practice type and size. Observation data were recorded in a database. Unit of analysis was the individual physician office. Statistical analysis yielded descriptive statistics. RESULTS: The activity analysis model consisted of sequential service types (pre-; intra-; and post-service) comprised of 48 tasks. Intraservice tasks, representing the overall infusion activity, were divided into four stages. On-site observers applied the model in time-and-motion observations. Mean direct labor time totaled 129.7 minutes, of which 111.9 minutes per patient (86%) was intraservice. Intraservice therapy administration was divided into two sub-stages. Sub-stage 1 consisted of tasks to commence the infusion, while sub-stage 2 consisted of patient monitoring after the drug began to flow. The mean direct labor time for infusion sub-stage 1 totaled 20.5 minutes (16%), while infusion sub-stage 2 totaled 79.9 minutes (62%). Only minor variations in infusion protocol were observed. CONCLUSION: Physician office decision-makers need actionable methods in order to better manage drug administration services and to increase patient satisfaction. This study provides important detailed activity analysis information and process mapping that can be readily applied to practice-specific relevant scheduling, staffing and management decisions.