Chapter Preview
TopFrom Computerized Therapy To Machine Learning
The efficacy of a computerized cognitive behavioural therapy were demonstrated (McCrone, 2004), and online interventions are proven effective for a variety of different mental diseases such as depression (De Graaf et al., 2009), anxiety (Andrews, Cuijpers, Craske, McEvoy & Titov, 2010), eating disorders (Dölemeyer, Tietjen, Kersting & Wagner 2013) and they can be used to improve the medication adherence (Linn, Vervloet, Dijk, Smit & Van Weert 2011).
In the beginning of e-mental health the quality and success of the e-mental health applications were questionable, because this field was lacking of development and style guides, which led to a variety of different applications with low user adherence and the results were hardly comparable. This was mostly due to the lack of user involvement in the design of e-mental health applications and missing collaboration between software developers and health service researchers (Pagliari, 2007).
Another major concern was the usability and safety of the clients. The potential of medical errors within the applications has to be minimized, to increase the client’s safety. This goal goes hand in hand with the usability of the software and adherence in therapy, because a user unfriendly software can lead to rejection by the customer and is prone to possible medical mistakes (Karsh, 2004). Nowadays still privacy concerns and doubt about the effectiveness of online treatment remain barriers for many patients (Musiat et al., 2014).
Key Terms in this Chapter
Data Mining: Basically the same as machine learning.
Framework: Describes how the software has to be constructed.
Outcome Prediction: Data from previous clients and machine learning is used, to predict the outcome of the therapy for new clients.
Gamification: Treatments are modified and programmed as a game. The game aims on alleviating the symptoms of the client as the regular intervention would do.
Machine Learning: Algorithms that allow to identify patterns in data to make predictions on new data based on old data.
Activity Recognition: Uses methods from machine learning to identify a person's current activity.
Sentiment Detection: Subfield of machine learning and natural language processing that tries to identify the emotions within a text or speech sample.