Reading a 10-Robot Hospital Study Before You Copy It

What did the Chinese hospital study find?

A six-month study at a Chinese tertiary hospital, published in Scientific Reports on April 24, 2026, compared 10 autonomous mobile robots with 19 manual workers moving drugs and specimens. According to the paper, robot delivery reduced delivery time by 32% to 36% (p < 0.001) compared with manual delivery.

The headline is real, but it is narrow. It describes one hospital, one building, one fleet, and one side-by-side comparison of manual and robot routes. It does not predict what a hospital in Ohio or Arizona will see.

The figure also reached readers through a secondary market-report summary, and that summary changes at least one number. Read the paper first, then the summary.

How was the comparison built?

The study used a single-center, parallel-controlled design. The robot group and the manual group ran at the same time, which is a stronger test than comparing this year with last year. Pharmacy-to-ward times were 31 minutes plus or minus 6.2 for manual delivery and 21 minutes plus or minus 4.2 for robots, with a 95% confidence interval of 8.5 to 11.5 minutes.

Navigation relied on established methods: Dijkstra's algorithm for global path planning, A* combined with the Dynamic Window Approach for local navigation, and an M/M/c queuing model to coordinate the fleet. The queuing model deserves the most questions from a buyer, because it decides which robot takes which request as call volume rises.

Why do the trip count and quality numbers matter?

Speed per trip is only part of the case. The paper reports that the 10-robot fleet completed 7.3 times as many delivery trips as the 19 manual workers over six months. That is a capacity finding, because a robot fleet can keep running through shifts in a way a staffed route cannot.

Quality moved as well. The robot group reached 100% verification accuracy and item integrity, and its omission rate was significantly lower (p = 0.011). Nurse and patient satisfaction were also significantly higher in the robot group (p < 0.05).

Uptime tells a rollout story. The paper's monthly table shows system uptime rising from 94.2% in June 2025 to 99.0% in November 2025. A pilot judged in its first weeks will look worse than one judged after the ramp.

Where does the market-report summary go wrong?

The summary describes a six-month tertiary hospital study in China with 10 robots against 19 manual workers and delivery time down 32% to 36%, which matches the paper. It then says the robots completed 7.3 more delivery trips per week. The paper reports 7.3 times the trips over six months. Those are different claims, and the weekly framing overstates what the paper measured.

The same summary cites a 2025 pilot with 615 medications and 415.8 staff hours saved, but no primary source for that pilot appears in the record. Treat it as unverified. Its forecast of USD 9.75 billion by 2034 at a 15.7% compound annual growth rate is a vendor-reported projection, not a measured result.

Other market pages carry similar gaps. One cites a 27% fall in average robot prices from 2020 to 2024 and a 36% logistics labor cost cut at a hospital that is not named. Neither claim has a verifiable source behind it.

What does a Chinese hospital setup tell a US facility?

Local conditions shape the result as much as the robot does. The study ran in a Chinese tertiary hospital with its own elevator access, corridor layout and staffing model. Its costs were reported in renminbi, with a total investment of RMB 3.5 million covering 10 robots, 10 charging stations and IoT infrastructure. A US building will have different elevators, dock positions and labor costs, so do not convert any of those figures into a forecast for your site.

Elevators are a known weak point. A 2025 feasibility study at a tertiary hospital, covering 122 non-urgent medication missions, quantified how elevator operating rate affected delivery success and delay. A US pilot that ignores elevator waits will misread its own numbers.

Other settings point the same way. A systematic review of robots in critical care found that a TUG robot reduced mean pharmacy cycle time by 29.6%, and it noted that every robot in the review had technological limitations and hidden costs. A clinical lab study compared specimen turnaround on robot deployment days in May 2025 with the same period in 2024, when transport was manual.

What should a local pilot measure?

Measure the baseline first, on the same routes and hours you plan to automate. Averages hide the slow tail, so record spread and worst cases alongside the mean, the way the paper reported confidence intervals.

Track these measures for both manual and robot routes, using the same definitions:

• Request-to-delivery time by route and hour, including elevator wait time
• Trips completed per robot per shift, and idle time between requests
• Omissions and item verification errors, logged the same way for both methods
• Fleet uptime by week, with the cause of every stop
• Nurse and patient satisfaction, surveyed before and after the pilot
• Full cost, including charging stations, network work and staff training

How does a site assessment shape the test?

Before a pilot starts, the physical facts need pinning down: floor plans, elevator access points, dock positions, charger locations and any Wi-Fi dead zones. A site assessment captures those details and sets the boundaries of the test, so the pilot measures the building you actually have.

Service Robot Co. is an OEM-neutral integrator. We compare robots across manufacturers for the route, then handle deployment, integration, training and service through a nationwide US engineer network. Customers receive a robot on a three-year lease with one fixed payment, quoted after the site assessment, so the cost comparison in the pilot stays clear of the brand question.

What can a US hospital conclude from this study?

The study supports a narrower claim than the headline. In one tertiary hospital, a well-run fleet delivered faster, handled more trips and reported fewer omissions than a manual baseline over six months. It does not establish a percentage for your building.

Use the paper to justify a measured pilot, not a purchase or lease decision. If your own data show the same direction, you have a method you can defend to finance and nursing leadership. If it does not, the pilot has still shown you which variables to fix.