Recommendation Systems for healthcare
Scroll to see more
Jesús Manzanares
Given the high capacities of data gathering that service providers have achieved lately, Recommender systems are one of the technologies to have gained great growth in its implantation. When the data being gathered consists of user preferences and tastes, collaborative filtering has proved as one of the most used techniques.
On the one hand, it is a filtering as it permits to distinguish relevant information from a set of information large enough. On the other, it is collaborative as it relies on the principle that when several users have expressed their relish for the same things, they will also in the future.
Incorporating into recommending techniques objective sources of information (medical related) is one of the challenges we strive for in Gradiant’s EHealth department; employing user’s medical records to establish similarities between patients and compiling biomedical signals processing them into ratings.
Different scenarios may employ this technology among the health sciences: Occupational therapy, best suited individual selection in medical trials and even, in the near future, assistance to healthcare professionals in the election of medical treatments, therapies and procedures.
Currently, these capacities are being successfully applied to the music therapy field in the frame of the SMEC project. In this particular use case we have been able to use one of the most acknowledged fields of recommendation (music) as a reference to restate our determination in applying Information Technologies in the Health sector.
Music therapy is the process in which music activities and elements are used to promote the relation, communication, learning and expression (among others) among people. These activities may include interaction with musical instruments, involvement in group dynamic of musical expression or performance listening.
Taking as a reference users’ clinical data (patients in this occasion), we have stablished mechanisms by which we are able to enrich existing similarity algorithms using its information. Analogously, by gathering biomedical signals regardless of their nature (Heart or respiratory frequency, but also more complex data such as eeg or egg) we may reflect not only subjective ratings from patients, but an objective point of view in order to augment the preferences dimension.
It is undeniable how the integration of IT in so many aspects of life has increased the quality of services. Despite the challenges being faced in the Health sector, it may arise as the one in which we can strike the most, not only making the most out of the things we enjoy, but savouring them for longer.
