Managing Data in Help4Mood

EAI Endorsed Transactions Research Article on Ambient Systems Managing Data in Help4Mood Maria K. Wolters1,∗ , Juan Martínez-Miranda2 , Soraya Estevez3 , Helen F. Hastie4 , Colin Matheson1 1 School of Informatics, 10 Crichton Street, Edinburgh EH8 9aB, University of Edinburgh Edinburgh, UK 2 IBIME, Universitat Politécnica de Valencia, Valencia, Spain 3 Fundaciò i2CAT, Barcelona, Spain 4 School of Mathematical and Computer Sciences, Heriot-Watt University, Edinburgh, UK Abstract Help4Mood is a system that supports the treatment of people with depression in the community. It collects rich cognitive, psychomotor, and motor data through a Personal Monitoring System and a Virtual Agent, which is then analysed by a Decision Support System; analysis results are fed back to patients and their treating clinicians. In this paper, we describe how the complex data is managed and discuss ethical issues. Data is stored in functional units that correspond to treatment relevant entities. Custom XML DTDs are defined for each unit, which are used to exchange information between system components. As far as possible, observations and findings are coded using SNOMED CT to ensure interoperability with other applications such as Electronic Health Records. Keywords: XML, depression, SNOMED CT, decision support Received on 22 May 2012; accepted on 25 September 2012; published on 19 March 2013 Copyright © 2013 Wolters et al., licensed to ICST. This is an open access article distributed under the terms of the Creative Commons Attribution license (http://creativecommons.org/licenses/by/3.0/), which permits unlimited use, distribution and reproduction in any medium so long as the original work is properly cited. doi:10.4108/trans.amsys.01-06.2013.e2 1. Help4Mood—Supporting People with Depression can be unreliable, especially if the patient is not keeping regular notes or a diary. Depression is the main cause of disability worldwide [1]. It is characterised by a persistent and intense change of mood which affects behaviour, cognition, and physiology. Two types of depression can be distinguished, a melancholic form where patients’ movements are significantly slowed down, and a nonmelancholic form, where movements are not affected or agitated. Slowed movements are reflected in both gross motor function, such as gait, and fine motor function, such as movement initiation and reaction times [2, 3]. They also contribute to slowed speech and a flat intonation [4, 5]. Sleep duration can be either severely reduced (insomnia) or significantly increased (hypersomnia). At the moment, recovery is monitored infrequently through self-reported patient questionnaires that require the person with depression to remember their symptoms over a period of time that can be as long as two weeks (e.g., PHQ-9 [6]). Those self-reports Help4Mood enables patients to monitor selected cognitive, behavioural, and physiological aspects of their depression. Patients interact with the system every day to share how they are feeling and complete a few tasks that are informed by cognitive behaviour therapy, such as tracking and challenging negative thoughts. Help4Mood also collects activity and sleep data through a personal monitoring system. ∗ Corresponding author. Email: EAI European Alliance for Innovation Once every 1–2 weeks, Help4Mood generates a summary of the data that patients will then discuss with the clinician who treats them. This can be a family physician, a psychologist, a psychiatrist, or another health professional with mental health training. The summary includes overall mood, sleep and activity trends, and a list of frequent intrusive negative thoughts, which can be treated using cognitive behaviour therapy. The data generated by Help4Mood will also be useful for reviewing the effectiveness of medication. This design is based on extensive consultation with clinicians and patients in Spain, Romania, and the UK [7] 1 EAI Endorsed Transactions on Ambient Systems January-June 2013 | Volume 12 | Issue 1-6 | e2 M. Wolters et al. In this paper, we describe our approach to data management in Help4Mood. We focus on the high-level data structures that form the basis for communicating with clinicians, patients, and other stakeholders. In Section 2, we give an overview of the Help4Mood system. The basic high-level Help4Mood data structures are described in Section 3. Ethical issues are discussed in Section 4, and provisions for interoperability with Electronic Health Records are outlined in Section 5. Future work plans are summarised in Section 6. 2. Overview of Help4Mood Help4Mood consists of a Personal Monitoring System, a Virtual agent, and a Decision Support System. Help4Mood is structured around patients’ sessions with the Virtual Agent. Ideally, patients interact with their Virtual Agent daily. The Virtual Agent asks questions, sets tasks, and summarises the results of each session. Sessions can include summaries of activity and sleep patterns. Some of these tasks will yield cognitive data, such as relevant negative automatic thoughts, others are designed to capture relevant neuropsychomotor symptoms of depression. The sensors of the Personal Monitoring System assess sleep and activity patterns using sleep sensors and a wrist actigraph. While sleep data is collected every night, the wrist actigraph will only be worn for 72 hours at a time. Sessions with the Virtual Agent can include summaries of activity and sleep patterns. The Decision Support System plans and controls sessions with the Virtual Agent and converts data about the patient’s sleep, motor, speech, and other psychomotor patterns into graphical, textual, and conceptual summaries that can be communicated to clinicians, patients, and electronic health records. As yet, there are very few rules for adjusting medications and interpreting data that could be implemented in a traditional decision support framework [7, 8]. Therefore, the decision support system focuses on trend analysis and planning the interactions between Help4Mood and the patient. Figure 1 shows the internal structure of Help4Mood. On the left of the graph, we see the sensing / monitoring components, the Virtual Agent and the Monitoring System. The (Personal) Monitoring System includes the sensors and wireless communication infrastructure. The structure of the Virtual Agent is more complex. The Virtual Agent consists of a Graphical User Interface (GUI) and a talking head (the “Agent”). Verbal messages are generated by the Natural Language Generation (NLG) component. These messages are displayed by the GUI and spoken by the Agent. Spoken messages are passed to a text-to-speech synthesis engine (TTS) that creates speech with annotations that help synchronise the Virtual Agent’s head and facial movements. The EAI European Alliance for Innovation Figure 1. System Structure Dialogue System controls the flow of messages, using scripts to ensure the correct wording where clinically re (...truncated)