Personalised medicine involves the use of individual information to determine the best medical treatment. Such information include the historical health records of the patient. In this thesis, the records used are part of the Finnish Hospital Discharge Register. This information is utilized to identify disease trajectories for individuals for the FINRISK cohorts.
The techniques usually implemented to analyse longitudinal register data use Markov chains because of their capability to capture temporal relations. In this thesis a first order Markov chain is used to feed the MCL algorithm that identifies disease trajectories.
These trajectories highlight the most prevalent diseases in the Finnish population: circulatory diseases, neoplasms and musculoskeletal disorders. Also, they defined high level interactions between other diseases, some of them showing an agreement with physiological interactions widely studied. For example, circulatory diseases and their thoroughly studied association with symptoms from the metabolic syndrome.