Ethical issues in using Twitter for population-level depression monitoring: a qualitative study.

Jude Mikal,Mike Conway,Samantha Hurst

doi:10.1186/s12910-016-0105-5

Jude Mikal, Mike Conway + Show 1 more

Open Access

https://doi.org/10.1186/s12910-016-0105-5

Copy DOI

Abstract

BackgroundRecently, significant research effort has focused on using Twitter (and other social media) to investigate mental health at the population-level. While there has been influential work in developing ethical guidelines for Internet discussion forum-based research in public health, there is currently limited work focused on addressing ethical problems in Twitter-based public health research, and less still that considers these issues from users’ own perspectives. In this work, we aim to investigate public attitudes towards utilizing public domain Twitter data for population-level mental health monitoring using a qualitative methodology.MethodsThe study explores user perspectives in a series of five, 2-h focus group interviews. Following a semi-structured protocol, 26 Twitter users with and without a diagnosed history of depression discussed general Twitter use, along with privacy expectations, and ethical issues in using social media for health monitoring, with a particular focus on mental health monitoring. Transcripts were then transcribed, redacted, and coded using a constant comparative approach.ResultsWhile participants expressed a wide range of opinions, there was an overall trend towards a relatively positive view of using public domain Twitter data as a resource for population level mental health monitoring, provided that results are appropriately aggregated. Results are divided into five sections: (1) a profile of respondents’ Twitter use patterns and use variability; (2) users’ privacy expectations, including expectations regarding data reach and permanence; (3) attitudes towards social media based population-level health monitoring in general, and attitudes towards mental health monitoring in particular; (4) attitudes towards individual versus population-level health monitoring; and (5) users’ own recommendations for the appropriate regulation of population-level mental health monitoring.ConclusionsFocus group data reveal a wide range of attitudes towards the use of public-domain social media “big data” in population health research, from enthusiasm, through acceptance, to opposition. Study results highlight new perspectives in the discussion of ethical use of public data, particularly with respect to consent, privacy, and oversight.

Highlights

Significant research effort has focused on using Twitter to investigate mental health at the population-level
Significant research has focused on using Twitter to investigate mental health at the population-level [19], including work on correlating suicide-related keywords with United States suicide rates [20] and automatically identifying depression symptoms [21,22,23]
Despite the potential of publicly available social media data, in combination with computationally efficient Natural Language Processing (NLP) techniques [24], to augment current telephone-based public health monitoring efforts, significant doubt remains among regulatory authorities and research ethics committees regarding ethically appropriate uses for these new data sources

Summary

Introduction

Significant research effort has focused on using Twitter (and other social media) to investigate mental health at the population-level. Despite the potential of publicly available social media data, in combination with computationally efficient Natural Language Processing (NLP) techniques [24], to augment current telephone-based public health monitoring efforts (e.g., in the United States, the Behavioral Risk Factor Surveillance System [25]), significant doubt remains among regulatory authorities and research ethics committees regarding ethically appropriate uses for these new data sources This is true in the wake of Facebook’s 2014 “emotional contagion” intervention study [26], and concerns expressed regarding Samaritans Radar, a Twitter app designed by Samaritans UK – a suicide prevention charity – to monitor the tweets of a user’s contacts for evidence of suicidal ideation [27].

Objectives

Methods

Results

Discussion

Conclusion