Early warnings are allowing HTM to teams prevent unexpected downtime and plan repairs before equipment fails.
By Alyx Arnett
For about five years, MultiCare Health System has used a monitoring platform to identify potential problems with its imaging equipment before they lead to unexpected downtime. The platform, Glassbeam, allows the clinical engineering department to spot warning signs of CT tube failures, cooling system issues, and other equipment problems, giving technicians time to prepare for repairs.
“The idea is we want to know that there’s a problem before the clinician tells us there’s a problem,” says Don Davenport, manager of clinical engineering at MultiCare.
Those advance warnings have allowed MultiCare to obtain replacement parts before equipment fails, reducing the time equipment may be out of service.
MultiCare is among health systems incorporating proactive monitoring and predictive maintenance into their equipment service strategies. While not every equipment failure can be anticipated, says Lisann Goodin, operations vice president of clinical engineering at Trimedx, the technology is giving healthcare technology management (HTM) teams more opportunities to identify developing problems and plan repairs. Adoption still varies among health systems “because the transition takes time, system-level acceptance, and integration,” says Goodin.
Imaging Leads Current Adoption
Predictive maintenance has become increasingly practical as improvements in device connectivity, remote monitoring, and data analysis make it possible to identify developing equipment problems earlier, Goodin says.
Advanced imaging equipment, particularly CT, MR, and interventional radiology systems, is among the strongest candidates for predictive maintenance, Goodin says. These systems generate substantial operating data, and unexpected downtime can disrupt patient care and be costly for health systems.
“Predictive maintenance works best when degradation happens gradually and detectable changes—cooling issues, tube degradation, power supply abnormalities, or declining image quality—precede a failure,” Goodin says.
As monitoring capabilities have advanced, manufacturers have also become better at identifying which equipment changes warrant attention. Cameron Sieber, head of strategy in the customer services business at Siemens Healthineers North America, says systems can increasingly distinguish routine fluctuations from patterns that indicate a genuine equipment problem. A brief rise in humidity, for example, may put a room outside equipment specifications without posing a meaningful risk to the system.
“We’ve gone from a few years ago being able to detect that we’re out of spec … to detecting the pattern that we’re seeing that’s actually meaningful,” she says.
One application of that capability is Siemens Healthineers’ TubeGuard, which analyzes CT X-ray tube performance to identify signs of an impending failure. Sieber says the service can sometimes provide approximately a week’s warning, giving teams time to prepare for a replacement.
That advance notice can reduce the disruption caused by a failure. According to data Siemens Healthineers shared from customer-facing CT materials, TubeGuard can reduce annual unplanned downtime associated with CT X-ray tube failures by 72%, from 23.2 hours to 6.5 hours per scanner.
“The ultimate goal for us in service is really not just about maximizing that warning time,” Sieber says. “It’s about being reliable and getting enough warning to take meaningful action.”
Expanding the Benefits of Equipment Monitoring
At MultiCare, several years of equipment monitoring have shown how early warnings can help technicians address developing problems before they become more serious. Davenport recalls one instance in which Glassbeam detected incremental errors indicating that a CT table was slipping, even though clinical staff hadn’t noticed a problem. The warning allowed MultiCare to address the issue before it developed into a larger failure.
Monitoring environmental conditions has also been useful, Davenport says. Changes in cooling water temperature, for example, can indicate a chiller problem, while power monitoring can help identify conditions that might contribute to equipment outages.
Sieber also emphasizes the value of monitoring environmental conditions when facilities are closed overnight or on weekends.
“We will often have customers who close down at night, and so their system, in particular an MRI system, might just be sitting there overnight with no one coming in,” she says. “One of the things that we do for magnet monitoring is we are consistently looking at the condition, the pressures, the temperature of the room, also of the water going into the system, and also looking at helium boil off.”
Sieber recalls several instances in which the company detected power outages at facilities and alerted customers in time for someone to intervene before the magnet quenched, potentially preventing significant downtime and helium loss.
MultiCare is looking to extend those kinds of monitoring benefits to more equipment. “I want it everywhere, and it takes time,” he says.
Davenport says Glassbeam now supports mammography systems, and the health system is beginning to explore monitoring additional X-ray systems, including portable devices. Tracking battery health, he says, is one potential application that could help identify problems before equipment becomes unavailable.
MultiCare is also working to build an in-house imaging service team. Davenport expects the monitoring tools to help the new team respond more proactively to developing equipment problems and reduce unplanned downtime.
Putting an Alert Into the Workflow
When adopting predictive maintenance tools, HTM departments need to establish clear workflows for responding to alerts, Sieber says. That includes determining who receives a warning, who evaluates it, who orders replacement parts, and who performs the repair.
“There has to be this really clear collaboration that ties into that operational workflow between an HTM team, facilities at the hospital or imaging center, the IT team, as well as with the OEM partner,” Sieber says.
For OEM-provided monitoring and predictive maintenance services, Sieber says the alert process depends on the hospital’s service arrangement. A warning may go to the in-house HTM team, the manufacturer’s service specialists, or both. Technical experts may review the information before recommending action, while issues involving cooling, power, or environmental conditions may also require coordination with facilities or IT.
Sieber also emphasizes the importance of secure, reliable connectivity so equipment data can flow between monitoring systems and service teams. And once a predictive maintenance program is in place, she says health systems should track whether it is reducing downtime and improving equipment performance.
Another consideration is determining which alerts warrant action. At MultiCare, Davenport says Glassbeam alerts have generally been accurate, but not every notification requires an immediate service response. He says the health system has worked with Glassbeam to refine its alerts and determine which problems need attention.
“It’s been a learning curve of filtering through,” Davenport says.
At Trimedx, predictive alerts are incorporated into the maintenance workflow through work orders that include recommended repair actions, documentation, parts information, and service guidance. Technicians evaluate the warnings before deciding how to respond.
“Human expertise, judgment, and an understanding of clinical priorities will remain essential,” Goodin says. “AI serves as a tool to help teams make faster, better-informed decisions.”
The Data Behind the Prediction
Predictive maintenance relies on having enough reliable equipment data to recognize patterns that may signal an impending failure, Goodin says.
“Data quality matters most of all because AI is only as good as the data we feed it,” she says.
Trimedx draws on more than 7.4 million device records and over 27 years of history, according to Goodin, allowing the company to compare patterns across equipment types and healthcare environments. A 2025 Trimedx case study describes how the company’s Predictive Work System forecast a tube failure in the only emergency department CT scanner at a 147-bed hospital. The team ordered a replacement before the component broke down, and the new tube was on site when the failure occurred. Trimedx estimated that the preparation avoided 24 to 36 hours of unexpected downtime.
Individual hospitals may not have access to that volume of data, but they can still use their own maintenance records to identify potential equipment problems, says Danielle McGeary, vice president of HTM at AAMI.
McGeary says some CMMS platforms are incorporating AI-supported features that can analyze service histories and identify patterns in equipment failures. For hospitals with older equipment that lacks built-in monitoring capabilities, these tools offer another way to make use of the information they’re already collecting.
“AI takes data, learns from it, and uses it to predict what’s coming next,” she says. “That’s very different from sorting a spreadsheet and pulling out numbers.”
But for CMMS platforms to provide useful predictive insights, technicians need to document equipment problems in sufficient detail. McGeary says that if a work order simply records that an infusion pump was broken and repaired, the system has little information to work with.
“AI is only as good as the data you give it. If a technician closes every work order with ‘I fixed it,’ then the next time that device fails, the system has nothing to learn from. It’s just going to tell you to fix it,” she says.
Documenting which component failed and how it was repaired gives the system information it can use to recognize similar problems in the future, she says.
Putting Predictive Maintenance to Work
For HTM departments considering predictive maintenance, Davenport recommends starting with their existing equipment inventory and identifying where additional monitoring could provide the greatest value. CT and MR systems are natural candidates, he says, but the right starting point depends on the equipment a hospital operates and the service arrangements already in place. He advises departments to review their OEM service agreements before investing in a separate platform.
“See what you’re paying for your vendors if they have [predictive tools],” he says. “Try to make sure that you’re not totally duplicating the service, but make sure you’re seeing value out of those.”
Implementation requirements are another consideration. At MultiCare, deploying Glassbeam involved setting up a virtual server to collect equipment data, obtaining security approvals, configuring imaging systems, and installing environmental sensors. Davenport says the process was relatively straightforward because of the health system’s working relationship with its information services department.
HTM departments should also consider how predictive maintenance fits alongside their existing preventive maintenance (PM) programs. At MultiCare, Davenport says the health system has not used monitoring data to adjust scheduled PM intervals.
“I think it’s so important to keep up on your PMs per the manufacturer on this type of equipment,” he says. “It’s more of, if we see something failing, we can be ahead of it.”
Sieber agrees that predictive capabilities are not yet a substitute for scheduled maintenance. She says moving toward condition-based maintenance will require extensive validation and evidence that equipment can continue operating safely and reliably.
“We’re not at that turning point to eliminate scheduled preventative maintenance,” she says.
Davenport encourages HTM departments to explore what is available and determine whether the technology fits their equipment and resources.
“Just open the door for the conversation and see where it can fit,” he says.
The Next Phase
As predictive tools become more sophisticated, service teams could gain greater visibility into developing equipment problems and more opportunities to plan maintenance before a failure disrupts patient care, Sieber says.
She sees potential in technologies such as digital twins, which use virtual representations of physical equipment to simulate how systems might perform under different operating conditions.
“I think artificial intelligence and these digital twins are really going to help us not only recognize patterns but start to predict more what we can expect to come out of the system and when we should really take care of something,” she says.
Goodin expects these advances to reshape how HTM departments approach equipment service.
“Predictive maintenance will continue to help HTM teams spend less time reacting to unexpected breakdowns and more time on planned interventions, risk prioritization, and strategic asset management,” she says.
McGeary also encourages HTM professionals to become familiar with the technology as its capabilities expand.
“The HTM professionals who embrace it will be the most successful,” she says. “If we can make our processes more efficient and enable ourselves to provide better maintenance on our equipment, that ultimately provides better care for the patients.”
ID 435214133 © Sorapop Udomsri | Dreamstime.com
Alyx Arnett is chief editor of 24×7 Magazine. Questions or comments? Email [email protected]
