Hospitals That Don't Stop: Seismic Isolation in Critical Care

The Morning After Is When Hospitals Matter Most

At 4:31 a.m. on January 17, 1994, the Northridge earthquake hit Los Angeles. By 5:00 a.m., emergency rooms across the region were filling with crush injuries, heart attacks, and people who had been thrown out of bed. By noon, the question was no longer how strong each hospital had been built. The question was which ones could still take patients.

The Los Angeles County / USC Medical Center, a conventional fixed-base hospital, sustained roughly $389 million in damage and had to evacuate. The USC University Hospital, less than a kilometer away and supported on 149 lead rubber bearings supplied by DIS, kept operating. Surgical procedures already underway when the earthquake hit were completed without interruption. Many patients reportedly only learned that an earthquake had occurred when staff turned on televisions later that morning.

That contrast, two hospitals, the same earthquake, the same neighborhood, with two completely different outcomes, is the single most cited piece of evidence in the case for hospital seismic isolation. It also reframed the entire conversation about what hospital seismic design should aim for.

Operability, Not Just Survivability

For most buildings, the goal of seismic design is what engineers call "life safety": the structure may be damaged, but it should not collapse, and occupants should be able to evacuate. That is a reasonable standard for an office building, where a few months of repairs are an inconvenience but not a public health crisis.

For a hospital, life safety is not enough. A hospital that survives an earthquake but cannot operate is, for the patients who needed care that morning, functionally the same as a hospital that collapsed. They are going somewhere else, in an ambulance, on roads that may also be damaged.

That is why the seismic design goal for hospitals has shifted from survivability to operability or immediate occupancy. In California, OSHPD (now HCAI) hospital seismic standards have moved progressively in this direction for three decades, and similar standards are now common in Japan, Taiwan, Turkey, and elsewhere. The functional question is no longer "will it stand up?" It is "will the elevators still run, the operating rooms still be sterile, the imaging equipment still calibrated, the data center still online, the gas and water lines still intact?"

A conventional code-compliant hospital is designed for roughly 10 percent probability of collapse in 50 years. An isolated hospital is designed to keep operating through the same earthquake that would force the conventional one to evacuate.

Why Isolation Is the Right Tool for This Job

Hospitals are uniquely well suited to base isolation, for reasons that are partly structural and partly about contents.

Structurally, hospitals tend to be 4 to 10 stories, on firm or moderately firm soil, with relatively heavy floor systems because of mechanical loads, lab equipment, and shielding. That is exactly the building geometry where period shift works best. Pushing a hospital's effective period from 0.8 seconds to 3.0 seconds drops the inertial force demand by a factor of three or more.

Contents-wise, modern hospitals are full of expensive, fragile, calibration-sensitive equipment. An MRI scanner that survives the earthquake but needs to be recalibrated is an MRI that is offline for weeks. Surgical robots, linear accelerators for cancer treatment, blood bank refrigeration, pharmacy automation, and the IT infrastructure that runs all of it are damaged by acceleration, not displacement. Base isolation cuts floor accelerations roughly in half compared to a fixed-base structure of the same size, and that is the variable that protects equipment.

The wider point is that for a hospital, the structure itself is rarely the largest component of cost. The contents are. The replacement value of a hospital's medical equipment can easily exceed the construction cost of the building. Isolation is the cheapest way to protect that investment.

Four Hospitals That Make the Case

Over 40 hospitals worldwide use DIS lead rubber bearings as their primary seismic isolation system. Four of them illustrate why the technology has become the default.

Tan Tzu Medical Center, Taichung, Taiwan

The Tan Tzu Medical Center, also called the Taichung Tzu Chi General Hospital, is the largest seismically isolated building in the world. It covers 1.7 million square feet of hospital and support space, includes about 400 beds, and sits on 386 DIS lead rubber bearings. Taiwan is on the boundary of the Eurasian and Philippine Sea plates, with major earthquakes every few decades. The hospital was designed for operability through a maximum considered earthquake. The bearings allow up to roughly 60 cm of horizontal displacement under that event.

Arrowhead Regional Medical Center, Colton, California

Arrowhead is a 920,000 square foot Level II trauma center built on a site that is roughly two miles from the San Jacinto fault and nine miles from the San Andreas. Five of its six buildings are independently base-isolated, using a hybrid system that combines elastomeric bearings with viscous damping devices. The facility is designed to remain self-sufficient for at least three days following a magnitude 8-plus event on the San Andreas, with its own power, water, and medical gas capacity.

Hospital de Zona, Xindian (New Taipei), Taiwan

The Xindian hospital is another major Taiwanese facility in the DIS-isolated portfolio. Like Tan Tzu, it sits in a high-seismicity zone with a code requirement for post-earthquake operability. Lead rubber bearings carry the heavy core hospital tower and decouple it from the worst of the high-frequency content of subduction-zone earthquakes.

Erzurum Regional Hospital, Turkey

Erzurum sits in eastern Anatolia, on the North Anatolian fault system, in a region with a long and brutal earthquake history. The Erzurum Regional Hospital uses DIS isolators in a base isolation system designed to keep the facility online following a major event. The February 2023 Kahramanmaras earthquakes, Mw 7.7 and Mw 7.6, sit roughly 700 km southwest of Erzurum, but they served as a powerful reminder of why Turkish hospital codes have moved toward isolation for critical facilities.

What "Operational After the Quake" Actually Requires

The bearings alone do not make a hospital operational. They are the most visible piece of a coordinated design strategy that includes:

• Flexible utility connections at the isolation interface. Every pipe, conduit, duct, and cable that crosses the moat has to accommodate the design displacement, often 30 to 50 cm, without rupture. DIS works with mechanical and electrical designers to detail these crossings.
• Restraint of nonstructural components. Even with halved floor accelerations, ceiling tiles, ductwork, light fixtures, and equipment racks must be braced. Hospitals that have failed in earthquakes more often than not failed at the nonstructural level.
• Backup power and water sized for days, not hours. An isolated hospital with no fuel is no better off than a conventional hospital.
• Seismic moat clearance. The displacement gap around the building must remain unobstructed, with detectable curbs and no landscape encroachment.

This is where the DIS engineering team's involvement extends beyond bearing supply. The isolation system has to be designed in coordination with the mechanical, electrical, and architectural teams from concept stage. Retrofitting an isolation interface into a finished hospital design is expensive and inefficient. Designing it in from the start is, in most cases, cost-neutral or cost-negative relative to the equivalent fixed-base hospital.

The Cost Conversation

The first question hospital administrators ask about isolation is almost always whether they can afford it. The honest answer is that the right time to ask is at concept design, not at construction documents.

When DIS is brought in during concept design, total project cost is typically reduced by up to 30 percent relative to a comparable fixed-base hospital that meets the same operability target. That reduction comes from smaller foundations, lighter structural framing, and reduced bracing on nonstructural components, all of which become possible once you cut floor accelerations roughly in half.

If isolation is added late, after the structural system is set, the cost premium can be 2 to 5 percent. That is what the USC University Hospital paid in the early 1990s. After Northridge, the calculation was easy: the isolation premium on USC was a small fraction of the $389 million damage bill at LA County a kilometer away. The math has only gotten more favorable since.

A Note on Seismic Moats and Daily Operations

One question that comes up early in any isolated hospital project is whether the moat creates operational problems day to day. In practice, the moat is the most invisible part of the design once construction is finished. It is typically a 30 to 60 cm gap around the perimeter of the building, covered at grade by sliding plates or hinged grating that pedestrians walk over without noticing. Vehicle entrances use heavier sliding plates rated for ambulance and delivery truck loads.

Utility connections crossing the moat are the more interesting detail. Water, sewer, medical gas, fuel oil, fire suppression, electrical conduit, fiber, and HVAC ductwork all have to allow the design displacement, often 30 to 50 cm, without rupturing or losing service. DIS coordinates with the project's MEP engineers to specify flexible connections, expansion loops, and slack cable runs at the isolation interface. None of this is exotic technology, but all of it requires deliberate detailing. A hospital that has its bearings designed correctly but its utility crossings designed sloppily will lose service in an earthquake even if the structure is fine.

Why This Is Now Standard, Not Experimental

When the first isolated hospital opened in the early 1990s, isolation was viewed as an emerging technology. Three decades and 40-plus DIS-isolated hospitals later, it is the default option for new critical care construction in high-seismicity regions. The 1994 Northridge earthquake, the 2011 Tohoku earthquake, the 2016 Kumamoto earthquakes, and the 2023 Turkey earthquakes have each added to the data set of isolated hospitals that stayed open while their conventional neighbors did not.

What has changed in the last decade is that the conversation has moved from "should we use isolation?" to "how do we use it best on this specific facility?" The technology no longer needs to prove itself. The remaining question is whether the hospital you are planning today is being designed for the morning after the next earthquake, or only for the moment of the shaking itself.