What a Tertiary Hospital in China Learned Staffing 19 Porters vs 10 Robots

Medicine and specimens moving slower than the wards needed

A tertiary hospital in China with 22 open wards depended on human logistics staff to carry medicines, supplies and laboratory specimens between the pharmacy, the wards and the laboratory. Every trip was a walk, a wait and a hand-off, and every hand-off was a chance for something to go missing or arrive late.

Before the trial, the hospital ran a team of 19 manual porters. Nurses felt the gaps first, when a sample sat in a corridor or a drug order arrived after the round had moved on. Managers needed a way to weigh adding robots against adding staff, backed by numbers that would hold up to scrutiny.

Speed was only part of the problem. Omissions and verification errors carried clinical weight, so any new flow had to match or beat the people it replaced on accuracy, not just on pace.

• Delivery time from pharmacy to ward and from ward to laboratory
• Omission and verification accuracy on every hand-off
• Nurse and patient satisfaction with service
• Fleet reliability across the working day

A six month parallel trial with a 10 robot fleet

The hospital ran a single center, parallel controlled study from June 1 to November 30, 2025. Ten autonomous mobile robots served the 22 open wards while manual logistics personnel continued the existing service, so both groups worked the same hospital over the same six months.

Each robot navigated with SLAM mapping, using a global path planner and a local planner that handled obstacles in the moment. A queuing model coordinated the fleet so no single unit became a bottleneck. Each unit carried a 200 kg payload and ran about seven hours on a charge, which shaped the charging rhythm around shift changes.

Measurement was built to treat both groups the same way. Robot task times came from dispatch platform timestamps at second level precision, while manual times were logged on handheld devices and checked through regular audits.

• Map the hospital and set routes between pharmacy, wards and laboratory
• Run robots and manual staff side by side for six months
• Dispatch tasks through one central platform with automatic timestamps
• Track delivery trips, error rate and uptime every month

Faster routes, more trips and fewer missed items

Delivery time fell by 32% to 36% against manual delivery, a difference the trial reports as statistically significant (p below 0.001). Pharmacy to ward trips averaged 21 minutes with robots against 31 minutes manually, and ward to laboratory trips averaged 23 minutes against 38.

Volume rose sharply. Over six months the ten robots completed 7.3 times as many delivery trips as the 19 manual workers. Monthly trips climbed from 10,500 in June 2025 to 14,751 in November, while average delivery time settled at 21.0 minutes from July onward, down from 23.9 minutes in June.

Accuracy held up as volume grew. The robot group reached 100% verification accuracy against 97% for manual staff, and its omission rate fell to 0.004% from 0.056% (p = 0.011). The robot system's error rate dropped from 0.012% in June to 0.003% in November, and uptime rose from 94.2% to 99.0% over the same months.

Nurse and patient satisfaction rated significantly higher in the robot group (p below 0.05). The trial's ten year lifecycle model estimates savings of 6.928 million RMB, but that is a modeled figure rather than a measured cash result, and it should be read that way.

What a fleet decision looks like for a hospital operator

The hospital's result is a reminder that a fleet is never just a machine decision. Route design, charging windows, dispatch software and the people who keep the robots running all decide whether delivery gains hold after the first month. Service Robot Co. treats that as an integration problem: we select the right autonomous mobile robot across manufacturers, then deploy, integrate, train and service every unit through a nationwide US engineer network.

For a US hospital weighing a similar pilot, we begin with a site assessment and mapping, then quote one fixed monthly payment on a three year lease. Our robots are leased, not sold, so capital is not tied up before the fleet proves its worth. Maintenance and spare unit coverage sit inside the same partnership, so a hospital does not chase separate vendors when a unit goes down.

The trial also shows why uptime planning matters as much as top speed. Some hospital robot makers now lease on fixed monthly fees rather than selling machines outright, a shift described in a recent press report on hospital robot capacity.