Balgrist Campus: Integrative research into low-back pain with the support of Google Cloud
About Balgrist Campus
Balgrist Campus, Switzerland, is an internationally renowned research institute for musculoskeletal issues.
Tell us your challenge. We're here to help.
Contact usAbout partners
Balgrist University Hospital
Balgrist University Hospital is a center of excellence for the diagnosis, treatment, and aftercare of musculoskeletal problems.
Ergon Informatik
Ergon Informatik develops customized applications and advanced software solutions.
Balgrist Campus turned to Google Cloud to help it create an agile hybrid infrastructure for expanding its research.
Google Cloud results
- Boosts security and eliminates the need for a VPN
- Enables team growth through scalable infrastructure
- Provides additional computing power for research
- Establishes architecture for future projects
- Allows colleagues to collaborate more effectively
Enables a hybrid on-premises and cloud architecture
Lower-back pain is an experience most of us can relate to. It’s the most common musculoskeletal condition in adults, affecting up to 84% of us at some point in our lives. Around 577 million people experience lower-back pain each year, costing an estimated 0.1%–2% of gross domestic product annually in European countries, in terms of healthcare costs and lost productivity. Yet, with multiple physical and psychological factors potentially at play, the exact cause of lower-back pain, particularly chronic lower-back pain, isn’t always clear, and determining the precise source, or sources, can be problematic.
“The cloud environment we were using wasn’t well integrated with our network, resulting in a number of challenges, particularly when it came to security, scalability, and managing our infrastructure. Google Cloud was much more suitable and offered us a greater degree of flexibility.”
—Michael Meier, Principal Investigator, Balgrist University HospitalThe Integrative Spinal Research (ISR) group located at the Balgrist Campus of the Balgrist University Hospital is working to address this situation, gathering data from lower-back pain patients across a broad range of symptom durations. Rather than drawing on big data sources, the researchers directly recruit groups of between 50 and 100 patients for their studies, which allows detailed patient characterization. For around 90% of the people who attend the clinic, the cause for their pain remains unclear. Experts at the facility had been using an on-premises cloud setup linked to a network of supercomputers to process and analyze the information they gathered, including anatomical and functional MRI (magnetic resonance imaging) scans, to explore brain mechanisms related to chronic lower-back pain. But this setup had limitations. On the initiative of the ISR Head of Research Petra Schweinhardt and the support of IT Specialist Torsten Bergander, the team looked to Google Cloud to overcome them.
“The cloud environment we were using wasn’t well integrated with our network,” explains Michael Meier, Principal Investigator, Balgrist University Hospital, “resulting in a number of challenges, particularly when it came to security, scalability, and managing our infrastructure. Google Cloud was much more suitable and offered us a greater degree of flexibility.”
Powering team growth
Many of the challenges that the ISR research team was hoping to address stemmed from the success of the research projects it had already run. The ISR Group was founded in 2016 and is the research arm of the Department of Chiropractic Medicine at Balgrist University Hospital headed by Schweinhardt. The ISR Group integrates different approaches to investigate the mechanisms underlying acute and chronic back pain: from pain behavior, pain modulation, and pain assessments to biomechanical aspects and cortical representation of the spine, including its alteration in recurrent pain. A further research focus deals with the clinical course of patients treated in the chiropractic outpatient clinic and possible mechanisms of chiropractic treatment methods. In an interdisciplinary approach, the ISR Group combines different disciplines, such as neuroscience, psychology, epidemiology, and movement science. The ISR Group has built a strong reputation for research, attracting talent from across Switzerland and further afield, but this growth presented problems.
“Our team is continuously growing,” says Meier, “this means we have limited space for workstations. And the amount of data we gather is also increasing, which means we need more storage space, the ability to share data more effectively, and more computing power to analyze it. Google Cloud enabled us to achieve this.”
“We’re making the same tools that previously demanded expensive hardware available through Compute Engine. Everyone has their own data pipelines and does things slightly differently, but we’re able to ensure that people can collaborate, whether they’re using a VM or on-prem hardware.”
—Simon Carisch, Research Assistant, Balgrist University HospitalEnabling a hybrid approach
Increasing the availability of workstations was the obvious solution to the first of these issues. Given the expense of purchasing its previous infrastructure, however, the ISR Group was reluctant to entirely replace its existing supercomputers or to spend money on more equipment if a more cost-effective solution was available. The flexibility that Google Cloud afforded in being able to add virtual machines (VMs) via Compute Engine, which can be accessed via inexpensive PCs, laptops, and other hardware, alongside its existing supercomputers, was just one of the attractions for the team.
“We’re making the same tools that previously demanded expensive hardware available through Compute Engine,” says Simon Carisch, Research Assistant at the ISR Group. “Everyone has their own data pipelines and does things slightly differently, but we’re able to ensure that people can collaborate, whether they’re using a VM or on-prem hardware.”
Balgrist Campus approached Google Cloud in early 2020, and the ISR Group is now one of several research groups seamlessly blending on-premises and cloud-based systems. This is partly achieved through the use of HTCondor, an open source platform that supports high-throughput computing through coarse-grained distributed parallelization of computationally intensive tasks.
This hybrid approach brings with it many benefits, particularly in terms of maximizing the amount of computing power available at any time. It also brings a consistent experience wherever staff work by enabling easier hot-desking within the office, as well as more seamless remote working, particularly helpful during the early days of the COVID-19 pandemic. This hybrid approach also delivers security benefits, bypassing the need to use a VPN to dial into the previous cloud setup, which also caused performance issues, and providing for a future where barely any data is stored locally.
Moving to a cloud-first infrastructure
Eventually, the researchers hope to move to an entirely cloud-based setup. Wilhelm Kleiminger at Ergon Informatik helped in the early stages of Google Cloud integration on campus. He organized workshops on campus to clarify the research team’s needs and to share industry best practices on how to make use of Google Cloud. The first phase in this process, currently underway, is ensuring that when new starters begin working with the team, or an existing machine is no longer fit for purpose, that a VM is provided as a replacement. One benefit of this approach is that new starters receive a consistent experience when it comes to training for and accessing their new equipment. The team also expects it will lead to cost savings long term, as expensive supercomputers are replaced with VMs. It believes it will begin to see the financial results of these infrastructure modernization efforts by the end of 2021.
Increasing computing power
With an increasing team size, as well as advances in the technology used to conduct research, the ISR Group also needed more computing resources. Once more it turned to Google Cloud. The Group now uses an Ubuntu instance on Compute Engine equipped with 16 CPUs, 64 gigabytes of RAM, and 10 terabytes of storage space on a persistent disk, plus two SSDs powering its VMs. It is also using Cloud Storage buckets to store data from completed projects. And all of these elements can be further scaled up at any time, in line with increasing demand. Furthermore, new VMs can be created and seamlessly integrated into the Group’s HTCondor cluster on demand.
“The apps we use to analyze MR images and data are advancing rapidly,” says Meier. “We really needed that extra computing power. It means we can run far more sophisticated analytical processes, such as machine learning. This can mean we need to repeat a predictive model with different parameters ten or 100,000 times to get a null distribution. That’s something we can now easily achieve with our Google Cloud environment.”
“Google Cloud is really the tool we need to ensure the future of our research, especially when it comes to predictive models regarding the development and persistence of lower-back pain. It provides the architecture necessary to enable the kind of data gathering and analysis that we hope to perform.”
—Michael Meier, Principal Investigator, Balgrist University HospitalLooking to the future
The ISR team is at the start of its research phase and its partnership with Google Cloud. The team hopes to share what it has learned through its adoption of Google Cloud so that other research groups at the Balgrist Campus can enjoy the same benefits. Meier also has ideas on how the benefits the team sees with Google Cloud could be applied more broadly. That includes the integration of new supervised machine learning algorithms and unsupervised methods, such as clustering or deep learning approaches with neuroimaging and behavioral data.
“Google Cloud is really the tool we need to ensure the future of our research, especially when it comes to predictive models regarding the development and persistence of lower-back pain,” says Meiers. “It provides the architecture necessary to enable the kind of data gathering and analysis that we hope to perform. We’ve already seen huge benefits from our move to Google Cloud, and we hope to share them with colleagues. We'd love to be able to conduct multicenter studies, using Google Cloud as the infrastructure to share data between different clinics and to improve our predictive models.”
Tell us your challenge. We're here to help.
Contact us
About Balgrist Campus
Balgrist Campus, Switzerland, is an internationally renowned research institute for musculoskeletal issues.
About partners
Balgrist University Hospital
Balgrist University Hospital is a center of excellence for the diagnosis, treatment, and aftercare of musculoskeletal problems.
Ergon Informatik
Ergon Informatik develops customized applications and advanced software solutions.