Using social media data in diabetes care: bridging the conceptual gap between health providers and the network population

Abstract

Abstract Background Patients with diabetes who have poor health literacy about the disease may exhibit poor compliance and thus subsequently experience more complications. However, the conceptual gap of diabetes between health providers and the general population is still not well understood. Decoding concerns about diabetes on social media may help to close this gap. Methods Social media data were collected from the OpView social media platform. After checking the quality of the data, we analyzed the trends in people’s discussions on the internet using text mining. The natural language process includes word segmentation, word counting and counting the relationships between the words. A word cloud was developed, and clustering analyses were performed. Results There were 19,565 posts about diabetes collected from forums, community websites, and Q&A websites in the summer (June, July, and August) of 2017. The three most popular aspects of diabetes were diet (33.2%), life adjustment (21.2%), and avoiding complications (15.6%). Most discussions about diabetes were negative. The negative/positive ratios of the top three aspects were avoiding complications (7.60), problem solving (4.08), and exercise (3.97). In terms of diet, the most popular topics were Chinese medicine and special diet therapy. In terms of life adjustment, financial issues, weight reduction, and a less painful glucometer were discussed the most. Furthermore, sexual dysfunction, neuropathy, nephropathy, and retinopathy were the most worrisome issues in avoiding complications. Using text mining, we found that people care most about sexual dysfunction. Health providers care about the benefits of exercise in diabetes care, but people are mostly concerned about sexual functioning. Conclusion A conceptual gap between health providers and the network population existed in this real-world social media investigation. To spread healthy diabetic education concepts in the media, health providers might wish to provide more information related to the network population’s actual areas of concern, such as sexual function, Chinese medicine, and weight reduction.

Keywords

• Diabetes
• OpView
• Text mining
• Social media
• Sentiment analysis