Using Social Media to Help Understand Patient-Reported Health Outcomes of Post-COVID-19 Condition: Natural Language Processing Approach

Portrait of mental health identified by people with the post-covid syndrome

OBJECTIVES

This study aimed to produce a patient-centered understanding of mental health symptoms of people with the post-COVID-19 syndrome (PCS).

METHODS

A cross-sectional analysis of 414 participants in a longitudinal study was carried out involving people who self-identified as having symptoms of PCS. People were asked to name their most frequent and most bothersome mental health symptoms affected by PCS using the structure of the Patient Generated Index (PGI). The text threads from the PGI were grouped into topics using BERTopic analysis.

RESULTS

20 topics were identified from 818 text threads referring to PCS mental health symptoms. 35% of threads were identified as relating to anxiety, discussed in terms of five topics: generalized/social anxiety, fear/worry, post-traumatic stress, panic, and nervous. 29% of threads were identified as relating to low mood, represented by five topics: depression, discouragement, emotional distress, sadness, and loneliness. A cognitive domain (22% of threads) was covered by four topics referring to concentration, memory, brain fog, and mental fatigue. Topics related to frustration, anger, irritability. and mood swings (7%) were considered as one domain and there were separate topics related to motivation, insomnia, and isolation.

CONCLUSIONS

This novel method of digital transformation of unstructured text data uncovered different ways in which people think about classical mental health domains. This information could be used to evaluate whether existing measures cover the content identified by people with PCS, to initiate a clinical conversation, or to justify the development of a new measure of the mental health impact of PCS.

Toward the novel AI tasks in infection biology

Machine learning and artificial intelligence (AI) are becoming more common in infection biology laboratories around the world. Yet, as they gain traction in research, novel frontiers arise. Novel artificial intelligence algorithms are capable of addressing advanced tasks like image generation and question answering. However, similar algorithms can prove useful in addressing advanced questions in infection biology like prediction of host-pathogen interactions or inferring virus protein conformations. Addressing such tasks requires large annotated data sets, which are often scarce in biomedical research. In this review, I bring together several successful examples where such tasks were addressed. I underline the importance of formulating novel AI tasks in infection biology accompanied by freely available benchmark data sets to address these tasks. Furthermore, I discuss the current state of the field and potential future trends. I argue that one such trend involves AI tools becoming more versatile.