Consumer-Generated Discourse on Cannabis as a Medicine: Scoping Review of Techniques.

This study is the first to review techniques used by research on consumer-generated text for understanding cannabis as a medicine. It is increasingly evident that consumer-generated data offer opportunities for a greater understanding of individual behavior and population health outcomes. However, research using these data has some limitations that include difficulties in establishing sample representativeness and a lack of methodological best practices. To address these limitations, deidentified annotated data sources should be made publicly available, researchers should determine the origins of posts (organizations, bots, power users, or ordinary individuals), and powerful analytical techniques should be used.