Defining Clinical Process Value In Radiation Oncology: A Pilot Study Using A Real Time Location System And Discrete Events Simulation Technology

Operations strategy decisions like staffing and machine allocation impact access to care, patient wait times, and resource utilization in radiation oncology. Care delivery processes in radiation oncology are uniquely complex: patient visits merge radiation treatments and clinic time into the same visit, and resources like clinic spaces are shared among a changing set of staff. This makes it especially difficult to predict the impact of major operational decisions. We hypothesized that a Real Time Location System (RTLS) that tracked staff times with patients could feed sufficient data to enable robust modeling of the impact of operational changes for a typical “OTV day,” allowing for assessment of efficiency, costs, and the overall value of specific operations decisions. We mapped the process for on-treatment visits (OTVs) for a Head and Neck clinic, spanning machine treatment and clinic. Staff wore badges that emitted infrared signals captured by detectors on ceilings. Locations of staff and time durations were fed into a central computer by staff type. Patient queuing timestamps from Mosaiq were used to account for patient flow. We mapped all times to discrete events in the process. A discrete-event simulation (DES) was conducted to model the impact of adding /subtracting a room, nurse, attending, and treatment machine. Operational endpoints analyzed included patient throughput (treated patients per hour), patient wait times, and overtime. Per-minute costs were extrapolated from publicly available median salary data. Industry standard values were used for amortized treatment machine costs and room space. Process Value (PV) was defined as throughput divided by total cost. Results were scaled such that PV is 100 for the base case. The base case scenario mirrored the true operating environment of 4 clinic rooms, 3 head and neck nurses, 4 treatment machines, and 2 attendings. 10,000 clinic days are simulated for each scenario. Base case metrics were: average throughput of 3.88 patients/hour, total cash and delay costs of $10.593.40, and utilization levels of 69.08% for rooms, 75.03% for nurses, 52.75% for attendings, and 80.19% for machines. Adding/(losing) a unit of a resource changed PV by -4.15% (-0.42%) for a room; +3.91% (-40.20%) for a nurse, +12.19% (-42.46%) for a machine and -7.21% (-11.26%) for an attending. Integrated use of RTLS and DES allowed for value assessment of operational decisions. Our model specifically demonstrated that adding a nurse and a treatment machine would be cost-effective. The method can be used at the institutional level to guide staffing and resource allocation, or at the policy level to guide permits for new facilities to meet population health needs. The approach will incorporate care quality metrics in its next iteration, and should be studied prospectively against real costs for validation.