Speech Recognition from the Cloud for Everyone

The Catholic Hospital Association Hellweg has transitioned from digital dictation to cloud-based speech recognition, resulting in faster processes for physicians and patients, as well as significantly lower IT operating costs.

Innovation and modernity are core principles at the Catholic Hospital Association (KHV) Hellweg, demonstrated through state-of-the-art infrastructure in both medical and administrative areas. “Our IT penetration level would be at EMRAM level 6,” says IT manager Matthias Kloos. To further enhance their capabilities, the facilities are currently implementing electronic medication management including closed-loop and unit dose systems using Hospital Future Act (KHZG) funding.

All four hospitals in the network use ORBIS as their Hospital Information System (HIS), which has partially replaced other systems. “We pursue a holistic system approach. ORBIS is our central system, and other applications should also come from Dedalus when possible, if they fit functionally,” explains Kloos.

Since May 30, 2023, KHV Hellweg has implemented integrated digital speech recognition with ORBIS Speech REC hosted. “Our priority is to improve the diagnostic processes and physician documentation,” says Kloos. The network previously used digital dictation but wanted to replace it with a cloud-based speech recognition solution.

There were two serious alternatives under consideration. “Since we’re using KHZG funding for this project, the new solution needed voice-controlled navigation capabilities within documents. This is one of the unique selling points of the selected solution. From a functional perspective, it was also our first choice,” said Yves Krell, physician at the Centre for Orthopaedics and Trauma Surgery at Marienkrankenhaus Soest. After thoroughly examining critical data protection questions, the hospital group proceeded confidently with the implementation.

Strong and Reliable Partnership

For the network-wide rollout, the IT department worked with Dedalus experts to set up the system, establish cloud connectivity, configure ORBIS integration, and deploy it to all physician workstations within just one week.

Before the rollout, Dedalus conducted user training—very effectively, as Kloos emphasizes: “All users were trained quickly and efficiently. Dedalus was present at all five locations, offering hourly training sessions from 8:00 to 17:00 without requiring additional support from our team. We were able to go live with minimal effort on our part, and doctors began dictating immediately. Three weeks after the initial training, additional sessions were offered for users who had been on holiday or unable to attend previously.” This approach enabled successful training of all 419 physicians across the hospital network in a very short timeframe.

Dedalus proved to be a strong and reliable partner beyond just training. “With ORBIS Speech REC hosted, the experience has been consistent with all previous ORBIS modules: smooth and professional implementation. Agreements and schedules were honoured, and the collaboration was extremely pleasant,” Kloos praises.

Before purchasing, KHV Hellweg tested the speech recognition solution as an on-premises installation. The disadvantage of this approach was that the network would need to provide and manage the complete server infrastructure. With numerous servers already in use and a busy IT department, they opted for the cloud-based solution. “We operate a central data center at Christian Hospital Unna with a Citrix environment that delivers ORBIS as an application. We provide access to all users at all workstations, with cloud connectivity through our central internet gateway,” Kloos explains. The facilities interconnect via 10-Gigabit fiber optic lines, with a microwave link serving as backup.

Significant Time Savings

“As a user, digital speech recognition doesn’t fundamentally change my workflow,” Krell notes. “The procedure remains largely the same—we previously used digital documents for dictation. The main difference is that I now format the text myself instead of relying on the transcription department. Despite this, we’re working faster today.”

Physicians primarily use speech recognition for creating discharge summaries, surgical reports, and outpatient documentation. Krell also sees quality improvements in clinical documentation. “Previously, we might write ‘No fractures or degenerative changes,’ but today we provide more detailed diagnoses and physical examination documentation. We can do this because speech recognition makes us faster,” the doctor explains. In the past, he and his colleagues relied on text templates, but these didn’t always accurately reflect what they intended to document.

“ORBIS Speech REC hosted offers significant process acceleration. I speak faster than I can type, and the documentation is immediately available to all authorized individuals in the clinical information system,” explains Krell. During ward rounds with mobile workstations, speech recognition isn’t used, which Kloos agrees with: “Live documentation at the bedside via speech recognition raises data protection concerns, as unauthorized persons could potentially overhear.”

Physicians who document for all their patients after completing rounds do so at workstations in doctors’ offices—preferably using speech recognition. This demonstrates the high acceptance of ORBIS Speech REC hosted. “It certainly helped that the medical staff had been requesting speech recognition,” Kloos notes with a smile.

The advantages are particularly evident in outpatient clinics. With speech recognition, hardly any patient leaves without a report. This also significantly reduces workload in the emergency department. “Evaluating X-ray images takes considerably less time. We dictate faster than we write—not to mention the improvement over traditional digital dictation. This saves both patients and doctors unnecessary waiting times,” says Yves Krell, highlighting another benefit of the new solution.

“Voice-controlled navigation in the document is one of the unique selling points of ORBIS Speech REC hosted,” adds Yves Krell, Catholic Hospital Association Hellweg.

Convincing Advantages

Krell acknowledges the transitional phase following implementation. “We’re now responsible not only for the clinical content but also for correcting and formatting the text. Previously, medical transcriptionists handled these tasks, paying close attention to phrasing and grammar. Speech recognition can’t fully replace that expertise. This presents new challenges, especially for non-native speaking colleagues, in maintaining documentation quality. However, this additional step also takes time,” the physician explains.

The hospital group has implemented both general medical vocabulary and specialized radiology terminology. Users have reported some difficulties with procedural terminology and proper nouns. “Overall, though, the standard medical vocabulary is well-developed and reliable,” says Kloos. “We don’t receive negative feedback from users.” The software training process is simple and intuitive—doctors mark incorrectly transcribed terms either manually or by voice command, enter the correct terminology, and save it.

Initially, leadership needed to persuade some physicians to embrace speech recognition as the future solution. Even now, some colleagues occasionally prefer typing to digital dictation. “But it’s just a matter of time with ORBIS Speech REC hosted,” Kloos believes, citing ambulatory care and emergency departments as examples. “When we deploy tablets for our physicians across the board, digital voice input will become the standard. Who would want to type then?” the IT manager confidently asserts. Nevertheless, he acknowledges that both text creation methods will likely continue to have their place, as some employees may type just as quickly as they dictate.

After three years, when KHZG funding expires and the system must demonstrate its value, KHV Hellweg will critically evaluate their use of digital speech recognition. “But we won’t be questioning the system itself, only individual licenses that we might be able to optimize,” emphasizes Matthias Kloos. His department is already analyzing overall usage patterns and monitoring the costs of internal and external transcription services. In departments showing limited adoption, they’re identifying reasons and implementing appropriate measures to increase acceptance.

Central Government Programme activities

The techUK Central Government Programme provides a forum for government to engage with tech suppliers. We advocate for the govtech sector, evangelise tech as a solution to public sector challenges, facilitate market engagement, and help make the public sector an easier market to operate in. Visit the programme page here.