US Healthcare Firm Converts Multilingual Speech into Structured Medical Reports

Inspired Spine, a US healthcare organization specialized in spine health and surgery, has developed SURI (Smart Universal Resource Identifier), an AI system that converts multilingual speech into structured medical reports.

The Inspired Spine researchers explained that traditional electronic health record (EHR) systems require manual input, contributing to physician burnout and increasing the risk of errors. Existing solutions, such as dictation tools and scribes, offer partial relief but remain costly and inconsistent — especially for providers who speak languages other than English or major European languages.

According to them, “SURI enters this landscape as a transformative solution.” It combines automatic speech recognition (ASR) and large language models (LLMs) to generate detailed clinical documentation from doctor-patient dialogues in 21 languages.

The system records real-world clinical interactions, converts speech into text, and refines the output using a fine-tuned version of LLaMA. Trained on real, anonymized patient interactions, it recognizes specialized medical terminology and ensures linguistic accuracy across multiple languages.

Additionally, SURI continuously learns from user feedback. It tracks physician edits, detects recurring corrections, and integrates these refinements into future outputs. Over time, this iterative learning process minimizes manual revisions, saving physicians valuable time.

Real-World Clinical Settings

SURI is built to function in real-world clinical settings. It integrates with EHR systems and other hospital management platforms, ensuring that structured reports are securely stored and accessible. The system relies on Amazon Web Services (AWS) for data management, ensuring HIPAA (Health Insurance and Portability Accountability Act) compliance and scalability.

“[…] these technologies support SURI’s efficient, scalable deployment across various healthcare settings, making it adaptable to the documentation needs of global, multilingual healthcare environments,” the researchers said

“Undeniable Benefits”

Healthcare organizations using SURI have reported a 60% reduction in documentation errors and a 70% decrease in reporting time compared to traditional methods. The economic benefits are also “undeniable,” according to the researchers, with potential cost-saving of up to 88% in documentation efforts.

“SURI not only provides a practical solution to a pressing issue in healthcare but also sets a benchmark for integrating AI into medical communication workflows”, they said.

Dr. Hamid Abbasi, Chief Medical Officer at Inspired Spine, wrote in a LinkedIn post, “for two years we have been actively developing and using our own artificial intelligence system, not only for medical documentation, but billing, coding, and hopefully one day, an AI-based medical record system that surgeons will love.”

Despite its strengths, SURI faces challenges, particularly in noisy clinical environments where background sounds can impact accuracy and in accommodating the variability of regional accents and lesser-represented languages. Additionally, user feedback has highlighted the need for improvements in the user interface to maximize adoption and usability.

For those interested in seeing SURI in action, a tutorial is available here.

Authors: Jiawen Zhan, Dominic Moore, Yuanzhe Lu, and Hamid Abbasi