Background The administration of chemotherapy in a hospital setting is intrinsically complex, involving a highly regulated chain of processes, including : i) medical prescription, ii) pharmaceutical preparation and iii) administration. Although partially digitalized, these processes are often managed by heterogeneous and non-integrated systems which generate bottlenecks, redundant information, and time misalignments between different phases, compromising overall efficiency and the quality of the service provided. Lean Thinking and the Value Stream Map (VSM) can allow managers to focus on the activities that truly generate value for the patient and to identify any waste.

Aim To properly analyze patient flow and evaluate performance indicators, a quantitative and analytical model was developed to study the current behavior of the system and the effects of any necessary changes.

Methods According to VSM principles, critical information was systematically recorded regarding patient arrival times, blood draws, the average time required to perform a blood draw and subsequently obtain the results, therapies administered, averages of the different durations of therapies performed,and the average time between the end of one patient and the arrival of the next. We derived key performance indicators, including Flowtime and ValueTime for each patient, and the respective ValueTime for each chair. The No-Value Time (NVT) percentage of patients was defined as the portion of time spent by each patient waiting compared to their Flowtime. The No-Value Time percentage (NVTp) of the chairs was defined as the ratio between the sum of thetimes during which the chair remains unused during the daily work horizon. The overall inefficiency of the department was measured by considering the ratio between the average No-Value Time and the average Flowtime, for both patients and available chairs.

Results Three different scenarios were studied, designed to evaluate the effects of different resource allocation and operational planning strategies on department performance. In scenario 1, the department is filled to saturation, meaning it maximizes available resources to provide patient care and therapy, leading to a 95-minute NVT and 27% NVTp. If, in the same scenario, patients are invited to arrive at the center at different times (e.g., 40 patients every 2 hours throughout the day), NVT can be reduced to 69 minutes with an increase in NVTp to 52%. In scenario 2 treatments were offered in 4 different rooms (A-B-C-D) based on the median length of infusion and level of patient complexity, and not the underlying disease. Thus, room A can be dedicated to longer infusion times (> 2 hours, like CHOP, ABVD, obinotuzumab etc), B room for intermediate therapies (30 min-2 hours, like carfilzomib, decitabine), room D for short therapies (less than 10 minutes, like bispecific antibodies, daratumumab, bortezomib), room C for device maintenance and blood drawing, obtaining a further NVT reduction to 57 minutes but a NVTp worsening to 48%. In scenario 3, room A operates independently from the other three rooms, while in rooms B, C, and D, all different activities can be carried out without any distinction, achieving 96 minutes NVT but a significant NVTp improvement of 23%. Modifying the scenario with a scheduling of arrival times achieves maximum efficiency, reducing the time patients spend waiting to receive the necessary care, achieving 50 minutes NVT and 38% NVTp.

Based on these scenarios, we developed an integrated software solution with AI functionality based on the automation of infusion station assignment, eliminating the distinction by pathology and length of treatment, reducing the average patient waiting time to 13 minutes, the in-patient stay to 23 minutes for short-term therapies, such as daratumumab and bispecifics, and 45 minutes for intermediate therapies such as isatuximab infusion.

Conclusions Our approach proposes leveraging existing infrastructure resources, particularly infusion chairs, which are often underutilized or suboptimally managed manually. This approach takes into account dynamic allocation, which takes into account variables such as the patient's clinical priority, expected duration of therapy, and drug type.Reducing waiting times between administrations, combined with the ability to dynamically accommodate urgent or unplanned patients, allows the system to be more flexible and responsive.

This content is only available as a PDF.
2025
Sign in via your Institution