Gosta Labs reduces the workload of healthcare professionals with AI
Europe’s healthcare systems face mounting pressure from aging populations and growing staff shortages. Maintaining high-quality care requires innovative solutions that optimize workflows and empower healthcare professionals. Finnish startup Gosta Labs is tackling this challenge head-on, developing specialized machine learning models to drive improved efficiency and outcomes in healthcare.
In Finland, it has been estimated that physicians may spend up to half their workday on administrative tasks, including patient record documentation. Streamlining this process could free more time for patient care.
Henri Viertolahti and Lauri Sippola have worked with health technology for over a decade. In 2012, they founded the Kaiku Health application, which, among other things, automates symptom monitoring for cancer patients. The application has been adopted in hospitals globally, and in 2020, company called Elekta acquired Kaiku Health’s business.
After working with Kaiku Health for the following years, in 2023, they felt that the time was right – also in terms of technological advancements – to tackle new challenges in healthcare.
“We had long been fascinated by the potential of large language models,” explains Viertolahti. “In 2023, we conducted extensive interviews with stakeholders across healthcare, medtech, and pharma to understand key challenges in these sectors.”
The interviews revealed the significant time drain in healthcare caused by administrative tasks, sparking the idea to automate these processes. Gosta Labs was founded in the fall of 2023 to address this need.
Complying with a wide range of requirements with Microsoft Azure
Leveraging generative AI for healthcare documentation presents complex regulatory hurdles. “Data privacy, medical device legislation, GDPR, and the EU AI Act all play a crucial role,” Viertolahti notes. Choosing the right technological foundation was paramount. After careful consideration, Gosta Labs selected Microsoft Azure for its robust security, compliance features, and seamless integration with existing healthcare systems.
“Microsoft Azure is widely used by Finnish healthcare providers,” says Viertolahti. “It offers a secure, compliant environment for running and hosting digital health solutions. Plus, Azure’s interoperability simplifies integration with existing healthcare infrastructure.” Access to Azure Machine Learning Studio and Microsoft’s EU Data Boundary operating model further solidified their decision, enabling rapid development and ensuring data sovereignty.
Expanding all over the world with the support of Microsoft’s expertise
Gosta Labs’ flagship solution is an AI assistant that automates patient record-keeping for physicians. The service listens to patient consultations, generating structured records that adhere to local standards such as the guidelines of the Finnish Institute for Health and Welfare, THL. Physicians then review and finalize these records, significantly reducing documentation time.
In the summer of 2024, the service was first piloted at a health center in the municipality of Nummela, part of the Western Uusimaa Wellbeing Services County. Initial feedback has been promising: users reported an improved ability to focus on client and patient care, enhanced interaction with patients, a reduction in documentation time, and increased job satisfaction.
Consequently, the pilot was expanded to three additional health centers in the Western Uusimaa region. Further implementations are underway with public and private healthcare providers in Finland and in Europe, demonstrating the solution’s worldwide potential.
Viertolahti and Sippola credit the Microsoft for Startups program for accelerating their growth. “Despite starting small, we’ve gained access to a global network of Microsoft experts with significant healthcare industry knowledge. The support from both the global and Finnish Microsoft teams has been invaluable for us.”
Extending the service beyond automated record-keeping
Gosta Labs’ vision extends beyond automated record-keeping. The company is developing a comprehensive AI assistant to provide real-time decision-making support during consultations.
To further enhance the service, Gosta Labs is also developing its own proprietary language models. The development is happening with the LUMI supercomputer, the most powerful high-performance computing resource in Europe.
In addition, the company is actively engaged in research, with their first scientific results on using large language models to interpret lung cancer guidelines published at the 2024 ASCO Annual Meeting.
Gosta Labs’ future goals also include internationalization and establishing the technology.
“The future offers very interesting opportunities, as there are challenges related to reducing the workload of professionals and improving the quality of treatment all over the world. We already have the experience and understanding of what it takes to make our service to become a certified medical device.”