Artificial intelligence enhances the treatment of stroke patients
23 October 2019
Every year, some 24 000 people in Finland suffer a stroke. The condition affects not only the patient but his or her relatives as well. A joint project by Kanta-Häme Hospital District and Tieto improves the understanding of the care process so that patients can be offered better, individual care.
It can often be challenging to prevent and treat a stroke, especially if it is not possible to address the factors that may lead to a stroke individually. A stroke can result from either a cerebral infarction or an intracranial haemorrhage, both of which damage brain tissue suddenly as a result of a disturbance in blood circulation. Many strokes could be prevented if the risk factors could be understood better and if the patient had the opportunity to influence his or her care better.
A joint project by Kanta-Häme Hospital District and Tieto has set out to improve the care of stroke patients by better understanding each individual’s background information. Thanks to new and more precise information, care practices and the hospital’s recording processes can be further developed in order to make the services more effective. The patient data utilized in the project comes from Tieto’s Lifecare system and is analysed on Tieto's Intelligent Wellbeing analysis platform.
”We are developing value-based healthcare extensively, and the data developed in the project will be used for this purpose. The care process of strokes has been developed as a co-operation project between first aid, acute medicine, neurology and radiology, and the results have also generated interest internationally,” says Seppo Ranta, Head of the Kanta-Häme Hospital District.
Resources for the development of care instead of data processing
Started in the spring of 2019, the project gives a deeper understanding of stoke patients’ way of living and environment by investigating patient data. The information needed for the analysis is often buried in medical reports, and going through them manually is a slow and challenging process. Instead, artificial intelligence can be used for running through extensive medical reports, quickly surfacing new and important information useful in decision-making.
By exploiting Natural Language Processing (NLP), artificial intelligence rapidly identifies essential background factors from text, such as how many of the patients smoke and their home environment. Artificial intelligence can identify different expressions of natural language from patient records and journals. A patient’s home environment may, for instance, be described using different expressions like “in his own home alone”, “home with the family” or “at a care centre”.
“We taught the application to understand sentences with a tool that is easy to use, which improved the precision of artificial intelligence. We were positively surprised by how well essential factors could be identified from text. The identified information can be used for developing care as well as in future research,” says Hanna Narsakka, Project Manager at the Development Unit of Kanta-Häme Hospital District.
“The project found an innovative way of processing large amounts of data efficiently. The Tieto Intelligent Wellbeing solution enables real-time analyses, which means that physicians will have access to the data needed for decision-making more rapidly than before,” says Hannu Pousi, Lead Software Developer at Tieto.
Kanta-Häme Hospital District: Hanna Narsakka, Project Manager, Development Unit, firstname.lastname@example.org, +358 3 629 3178.