Can online discussion sites generate quality data for research purposes?

• Main Authors: Smith, Helen, Bulbul, Alpaslan, Jones, Christina J.
• Other Authors: Lee Kong Chian School of Medicine (LKCMedicine)
• Format: Article
• Language: English
• Published: 2018
• Subjects: Social Media; Health Information Seeking
• Online Access: https://hdl.handle.net/10356/86046; http://hdl.handle.net/10220/45376
• Institution: Nanyang Technological University

We have recently presented orally the preliminary findings of some research where we derived our data from online discussion sites. Our research described parents’ and caregivers’ experiences of administering medications to young children (1, 2). Different aspects of these data have been presented at four different conferences (each in a different country, and across two continents). On each occasion, our audience of clinicians and researchers have been divided about the credibility and quality of our results. In response to these divergent opinions encountered, we have reflected on whether online discussion sites can contribute to quality research in the arena of health. Our own foray into the analysis of the content of online discussion sites was unplanned; it was precipitated by a delay in the approval and recruitment processes for a conventional qualitative research study. Our original plan was to use one-to-one interviews with the parents of young atopic children to explore their experiences of administering health-care treatments. The atopic child may suffer from eczema, asthma, rhinoconjunctivitis, and food allergy and therefore may require many different treatments. The formulations of these treatments are diverse, including inhalers, nasal sprays, liquid medicines, eye drops, and topical preparations. It was the complexity of some children’s treatment regimens that stimulated our interest in understanding how parents coped with sustained, complex treatment administration. For our analysis of online data, we first identified websites where parents discuss the care of their children. We then searched on these, using such terms as “how to eye drops toddler,” substituting in turn other medication delivery methods for the term “eye drops.” Data were plentiful; for example, the transcripts of online discussions relating to the administration of eye drops extended to 39 pages. Caregivers described in detail their own experiences and responded to requests for help with pragmatic suggestions. Presentation was sometimes informal, with errors of literacy and spelling but the message was always clear. A thematic analysis (3) identified three major themes: children’s negative reactions to having eye drops, physical restraint techniques adopted by parents to administer eye drops, facilitators and bribes used to ensure adherence. The accounts we identified were very powerful, engaging, and alerted us to previously unrecognized challenges and the parental distress these difficulties generated. We fully recognized that this approach had limitations and was not a substitute for our original research plan, but it did form a useful adjunct. First, analysis of online discussion lacked the specificity that we were seeking; for example, the discussions relating to eye drops included any ophthalmological conditions, not just the atopic eye conditions of interest to us. Second, this research method did not allow us to explore the impact of administering multiple medications, but rather focused on individual medications. Nonetheless, it has provided us with novel data that will impact on clinical practice, and it has informed protocol refinement for our definitive face-to-face interviews with the parents of atopic children. Our audiences recognized the richness of the data, particularly when it raised issues encountered by parents, that they, as health-care professionals or clinical scientists, were previously unaware. However, the critics in our audiences challenged our data on issues of its quality; they queried its truthfulness, highlighted our lack of knowledge of our respondents, and were concerned about bias and ethics. Below we take some of the questions posed of our data and discuss to what extent they are justified or can be rebutted.