Voice and AI Could Assist Nurses, Researchers Find
Worldwide, artificial intelligence delivered via a voice interface could help healthcare organizations deal with enormous cost-cutting pressures, labor shortages, and growing demand for their services, according to research by VTT Technical Research Centre of Finland, Turku University of Applied Sciences, the University of Eastern Finland, the Northern Savo and Southwest Finland well-being services counties, Mediconsult, Oiva Health, and Solita.
According to Senior Scientist Jouni Kaartinen from VTT, artificial intelligence could, for instance, automatically and illustratively provide relevant patient information to professionals through a graphical or voice interface. AI could also be an excellent assistant in instructing patients when they are searching for help, during their hospital stays and afterwards. This could facilitate access to the right service, reduce repeat visits, and improve treatment monitoring.
Various digital tasks, such as patient registration, reviewing patient information, familiarizing with instructions and giving them, take a significant part of nurses' daily working hours. Reducing the time consumed on these tasks with the help of better technological solutions could free up working time for patient encounters and lighten the nurses' workload, the researchers assert.
"For instance, the German consultancy TLGG has estimated that if used properly, artificial intelligence could save more than 30 percent of nurses' working hours, which they could spend on more valuable tasks, like meeting customers and providing care," Kaartinen adds.
Ville Salaspuro, medical director of Finnish Mediconsult, believes that AI will be an important part of the future work of social welfare and healthcare professionals. "These tools will make professionals' work more efficient. They will have more time to focus on customers instead of consuming time on information systems," Salaspuro says.
A goal of the research project is to craft solutions to many of the problems that have prevented AI's expansion in the healthcare sector.
"A traditional problem in the healthcare sector has been that nursing professionals have not been involved in the technology development. As a result, the functionality or usability of the systems doesn't always support nursing work. Additionally, research evidence on the impact of different solutions on the operation of organisations is yet scarce," says Associate Professor Laura-Maria Peltonen from the University of Eastern Finland.
"Nurse-centricity is at the core of the project. The developed technology will be brought to our nursing students for testing. This way, we will learn how development should be done to promote end-user acceptability of new technology and prevent it from being perceived as a threat or as clumsy in terms of usability," says Senior Lecturer Paula Savolainen from Turku University of Applied Sciences.
Kaartinen says that the difficulty of obtaining necessary data has significantly hampered the development of AI solutions in healthcare. Healthcare involves so much professional terminology that it is paramount to train a truly useful conversational AI with comprehensive material: authentic anonymised speech data, patient data, and other healthcare documentation.
"As a lot of health data is already digitised, obtaining it securely for research and product development creates significant advantages. However, this has been complicated from a legal perspective. Anonymised and consent-based data would ultimately benefit all parties," Kaartinen says.
"Software companies must protect their customers' personal information and data by legislation. The PROFIT project is developing technical solutions and operating methods that would enable data use in accordance with the law," says Manu Setälä, head of research at Solita.
Several ethical issues also must be resolved. Despite the use of AI, decision-making and responsibility remain with professionals. According to Kaartinen, AI could not, for instance, make direct entries in patient or customer records without the professional's separate approval. Instead, AI could make an entry suggestion that the professional can either accept as is or with modifications. The project examines the acceptability and ethics of AI-assisted solutions from the perspectives of both nurses and customers.