The computational turn in online mental health research: A systematic review

Abstract

Digital trace data and computational methods are increasingly being used by researchers to study mental health phenomena (i.e. psychopathology and well-being) in social media. Computer-assisted mental health research is not simply a continuation of previous studies, but rather raises ethical, conceptual and methodological issues that are critical to behavioural science but have not yet been systematically explored. Based on a systematic review of n = 147 studies, we reveal a multidisciplinary field of research that has grown immensely since 2010, spanning the humanities, social sciences, and engineering. We find that a substantial majority of studies in our sample lack a standardized form of ethical consideration, focus on specific constructs and have a rather narrow focus on specific social media platforms. Based on our findings, we discuss how computational elements have influenced mental health research, highlight academic gaps and suggest promising directions for future studies.