Principal Investigator Britt Elin Øiestad and Bjørnar Berg

Collaborators Margreth Grotle (Oslo metropolitan University), May Arna Risberg (Oslo University Hospital and Norwegian School of Sports Sciences), Karin Magnusson (Norwegian Institute of Public Health), Jon Fiva (Norwegian Buisness School), Marie Pedersen (Volvat NIMI), Thor-Einar Holmgaard (Norwegian Back Association)

External Funding Norwegian PhysioFund

There exist little data on the use of physiotherapy consultations in the private market. Patient consultations in the public health care are reported in health registries, including the services the public sector buys from the private sector. With no registry data on the use of private health care, we lack important information on the group of patients who seek private physiotherapy services with self-pay only. Consequently, we do not know patients’ needs with respect to physiotherapy services in Norway. In this innovative project, we aim to utilize machine learning to develop and test a framework for extracting and analyzing data from physiotherapy records in the private care and compare patient characteristics with public registry data to increase knowledge on patient populations.

This project will have two objectives resulting in two scientific papers: 1) to develop a framework for extracting and organizing text from physiotherapy records in private physiotherapy service, convert it to data and describe the patients’ characteristics, and 2) to compare patients using a private physiotherapy service to patients from the same geographic area registered in health registries with respect to sex, age, diagnosis, comorbidities, educational level, and work status. We will apply AI and natural language processing (NPL) methods for the data management. Volvat Nimi will provide us with physiotherapy records from 14 private physiotherapists. We will include data on first consultations only. The project has 3 phases: First, we will prepare text files from journals, then apply NPL to structure and organize the data in an efficient way, and finally analyze the data with traditional statistical methods. The output of the project will be 2 scientific papers.