SCIENTIFIC SESSION

With forecasts that the number of long-term breast cancer survivors will continue to increase, it is becoming increasingly clear that better follow-up and analysis of their needs are required. We present a new data mining methodology, based on unsupervised clustering and signal processing techniques. This methodology, once applied to the SURBCAN longitudinal cohort of care pathways, enables the identification of complex temporal patterns (clusters) in women’s healthcare service use over the six years of follow-up.

After this, the characteristics of the extracted clusters and the respective patients can be studied and analysed, in order to improve the care received and optimise the use of healthcare services.