| Popis: |
Digitalised patient records represent a large potential source of real-world data. Nevertheless, confidentiality and data protection has made big data extraction from patient records impossible in the past. Future options for artificial intelligence in free text reading might enable data extraction while maintaining confidentiality. In turn this could enable improvement in risk prediction for several disease groups. Still, it is not known if free text record data provides an appropriate data source for this purpose. In this project we have analysed a pilot dataset of patients with carotid stenosis in order to estimate the stroke risk in relation to treatment, and by that to assess the individual risk profile. The pilot dataset is applied to design a statistical model suitable for patient record data analysis. In this article a detailed description of the data set, and the methods behind our model choice will be presented. |
