People use language to express their thoughts and feelings, unveiling important aspects of their psychological traits and social interactions. Although there are several studies describing methodologies to create a collection of words in English related to depression and other conditions, in most of them the selection of words is not clinical or expert based. The objective of this study is twofold: firstly, to introduce a comprehensive collection of Spanish words commonly used by patients suffering from depression, which will be available as a free open source for research purposes (GitHub), and secondly, to study the usefulness of this collection of words in identifying social media posts that could be indicative of patients suffering from depression. The level of agreement among medical doctors to determine the best words that should be used to select tweets related to depression was low. This finding may be due to the complexity of depression and the extraordinary diversity in the way people express themselves when describing their illness. It is critical to perform a thorough analysis of the specific language used in each condition, before deciding the best words to be used for filtering the tweets in each disease. As our study shows, the words supposedly more linked to depression are very common words used in other contexts, and consequently less specific for detecting depressive users. In addition, grammatical gender forms should be considered when analysing some languages such as Spanish.
Keywords: Depression; social media; surveys and questionnaires; terminology.