@inproceedings{10.1145/3544793.3563429, author = {Tlachac, ML and Ogden, Samuel S.}, title = {Left on Read: Reply Latency for Anxiety \& Depression Screening}, year = {2023}, isbn = {9781450394239}, publisher = {Association for Computing Machinery}, address = {New York, NY, USA}, url = {https://doi.org/10.1145/3544793.3563429}, doi = {10.1145/3544793.3563429}, abstract = {Mental health is a critical societal issue and early screening is vital to enabling timely treatment. The rise of text-based communications provides new modalities that can be used to passively screen for mental illnesses. In this paper we present an approach to screen for anxiety and depression through reply latency of text messages. We demonstrate that by constructing machine learning models with reply latency features. Our models screen for anxiety with a balanced accuracy of 0.62 and F1 of 0.73, a notable improvement over prior approaches. With the same participants, our models likewise screen for depression with a balanced accuracy of 0.70 and F1 of 0.80. We additionally compare these results to those of models trained on data collected prior to the COVID-19 pandemic. Finally, we demonstrate generalizability for screening by combining datasets which results in comparable accuracy. Latency features could thus be useful in multimodal mobile mental illness screening.}, booktitle = {Adjunct Proceedings of the 2022 ACM International Joint Conference on Pervasive and Ubiquitous Computing and the 2022 ACM International Symposium on Wearable Computers}, pages = {500–502}, numpages = {3}, keywords = {digital phenotype, mental health screening, metadata, mobile health}, location = {Cambridge, United Kingdom}, series = {UbiComp/ISWC '22 Adjunct} }