Hospital Staff-Hour Claims: A Checklist Before You Commit

Does a staff-hour claim tell you anything on its own?

Not much, until the count is defined. A staff-hour figure is the output of a counting rule, and that rule often sits in a footnote, a slide appendix or a conversation nobody wrote down. Before a facility commits, it needs three answers: what counted as a trip, what the hours were compared against, and who checked the result.

Rochester Regional Health makes a useful test case, provided it is treated as a worked example. Finger Lakes 1 reported on August 21, 2026 that the health system estimates more than 50,000 staff hours returned to its hospital staff through its Moxi fleet. The estimate is the system's own, and I found no independent audit of it. That makes Rochester a set of questions to ask, not evidence that the same result will appear at your site.

What counts as a trip?

The trip is the unit everything else multiplies, so its definition drives the total more than any other choice. A pickup and a drop-off is the easy case. The harder cases are a run with several stops, a robot that waits minutes for an elevator or a door, a delivery the recipient refused, and an empty return. Each can be logged as one trip or several, and the total moves with that choice.

Rochester's reported figures show scale without showing the rule. According to the same Finger Lakes 1 report, the robots completed more than 100,000 trips across Rochester General Hospital and Unity Hospital since 2023, including more than 75,000 at Rochester General and more than 25,000 at Unity. CBS News reports that the hospital began using the robot in 2023 and operated eight devices. Neither source explains whether a robot stalled at a door counted as a trip, and that is the first question to put to any vendor.

Which baseline are the hours measured against?

Hours saved mean nothing without a before picture. The baseline might be a time study of runners before the robots arrived, a sample of staff logs, recollection from supervisors, or a model that assumes a walking speed and a route length. Each produces a distinct figure, and a baseline built from memory is the softest of the four.

• Which units, shifts and weeks were sampled, and were any of them chosen after the fact?
• Did the comparison period include a seasonal peak, a staffing gap or a unit closure?
• Were robot hours taken from the robot's own log or estimated by staff?
• Does the baseline cover the whole job the robot takes over, or only the easiest slice of it?

Who checks the result, and how?

The 50,000-hour figure is an estimate published by the health system that runs the robots, which is the usual path these numbers take into public view. That is not an accusation. It is the reason an independent check belongs in the decision before a budget leans on the number.

A check can be modest. Ask for the raw trip log, then compare a sample of it with the schedules that nurses or runners actually worked during the same weeks. A hospital's quality or operations office, an internal audit group, or an outside time-study firm can each review the method, as long as none of them is paid by the vendor.

Rochester's 55,000 miles and 120 million steps: how to read them

Rochester also reports about 55,000 miles of walking avoided and roughly 120 million steps. The reviewed report does not say how those figures were derived from the trip count, so they deserve the same caution as the hours. When a vendor offers a walking or distance number, ask how it was computed and whether anyone measured the walk directly.

How do you test a claim on your own floor?

A short pilot tells you more than a brochure. Choose one unit, one route and one shift. Log every trip with its start, stop and any wait, then time the same route by hand over a comparable period. Put the two sets of numbers side by side before anyone signs.

Plan the pilot to include a normal week and a busier stretch, and write the sampling dates into the record before the first trip is logged. This is the kind of test we run with clients. Service Robot Co. is a full-service commercial robot integrator that is OEM-neutral, so we select the robot for the task rather than from one manufacturer's catalog. We then finance, deploy, integrate, train and service each unit through a nationwide US engineer network. Robots come to a facility on a 3-year lease, with one fixed lease payment quoted after a site assessment.

What does a discontinued program signal?

A nurse.org report says MultiCare, a Washington state health system, discontinued its Moxi robots. A program that ends is not proof that the claims were wrong, and the reviewed material does not give the reasons. It is still a reason to ask a reference site what it measured, what it stopped measuring and what it changed.

A checklist to run before a facility commits

Put these questions to any vendor, including the one that gave you the staff-hour figure.

If the vendor cannot answer the first three items in writing, treat the headline figure as a starting point for a pilot, not as a finding. The fourth and fifth items show whether anyone with no stake in agreeing has tested the figure.

• A written definition of a trip that covers waits, failed deliveries and returns.
• The named baseline, its sampling period, and who collected the data.
• The raw trip log as an export, not only a summarized hours total.
• The name of the party that checked the method, and whether the vendor pays that party.
• Two reference sites willing to discuss their log method, including any site that stopped.