Lian Van der Krieke*, University of Groningen, University Medical Center Groningen, Groningen, Netherlands Ando C Emerencia, University of Groningen, Johann Bernoulli Institute for Mathematics and Computer Science, Groningen, Netherlands Elisabeth H Bos, University of Groningen, University Medical Center Groningen, Groningen, Netherlands Judith Gm Rosmalen, University of Groningen, University Medical Center Groningen, Groningen, Netherlands Harriëtte Riese, University of Groningen, University Medical Center Groningen, Groningen, Netherlands Marco Aiello, University of Groningen, Johann Bernoulli Institute for Mathematics and Computer Science, Groningen, Netherlands Sjoerd Sytema, University of Groningen, University Medical Center Groningen, Groningen, Netherlands Peter de Jonge, University of Groningen, University Medical Center Groningen, Groningen, Netherlands Track: Practice Presentation Topic: Participatory health care Presentation Type: Rapid-Fire Presentation Submission Type: Single Presentation Building: Mermaid Room: Room 4 - Queenshithe Date: 2013-09-24 02:00 PM – 03:30 PM Last modified: 2013-09-25

If you are the presenter of this abstract (or if you cite this abstract in a talk or on a poster), please show the QR code in your slide or poster (QR code contains this URL).

Background
Evidence-based treatment guidelines in psychiatry are predominantly based on research conducted on a group level (nomothetic approach). Samples of people are investigated to find general laws of symptomatology and functioning which are then generalized to all individual members of the investigated population. This approach has been criticized for leading to knowledge that is ‘true on average’. Although group-based research is useful to study variability between individuals in a sample, the results do not necessarily generalize to individual patients. An alternative research approach focuses on unique patterns within individuals over time (idiographic approach). In the last two decades, the idiographic research approach has taken off, due to the development of new quantitative methods to perform within-person research combined with the development of innovative mobile health techniques. The rise of self-tracking apps enables individuals to monitor themselves real-time and on a daily basis. The data collected with these apps can be analyzed by time series techniques such as Vector Autoregressive Modeling (VAR). VAR is a statistical technique which allows for analysis of time series data which can elucidate dynamic relationships between two or more variables, providing an impression of putative causal associations. For instance, VAR analysis can show that in patient x inactivity is likely to cause a depressive mood, whereas in patient y depression is likely to cause inactivity. This research approach has much potential for a personalized healthcare. However, there still is a significant gap between the research context in which daily mobile assessments combined with VAR analysis is experimented with and clinical practice in which individual patients may profit from its results. The main reason for this gap is that analysis of time series data requires advanced statistical expertise, including extensive knowledge of the statistical procedures and a high level of experience. To put it differently: we have the tools, we can collect the data, we know how to analyze the data, but we have not been able to put it to use.

Objective
Our objective was to bring the idiographic approach closer to clinical psychiatric practice, by automating the data analysis process and generate output in an intuitive way, so that it is interpretable by non-experts.

Method
We developed a web-based open source application, called AutoVAR, which automates time series analysis of self-tracking data. AutoVAR was developed to take over expert modeling work, otherwise conducted by an experienced statistician. We validated the output of AutoVAR by comparing it to the output of a manual analysis procedure.

Results
The AutoVAR application generates output in a few seconds and presents the information in a way that is interpretable by clinicians and patients. In our presentation, we will demonstrate AutoVAR and show how the application can significantly contribute to a patient-tailored treatment in clinical practice.