Analyzing Technostress in Pregnancy Applications through User Sentiment: A Case Study in the Indonesian Digital Health Ecosystem

Pregnancy applications are increasingly used by expecting mothers in Indonesia to access health information and monitor their pregnancy. While these apps offer valuable features, their growing complexity may lead to technostress, which refers to psychological strain caused by overwhelming technology use. This study investigates indicators of technostress by analyzing 48,733 user reviews collected from 12 popular Indonesian pregnancy applications between 2016 and 2025. Reviews were preprocessed using standard NLP techniques, and latent themes were extracted through Latent Dirichlet Allocation (LDA) topic modeling. Five dominant aspects associated with technostress were identified including technical issues, features, education and community content, monitoring functions, and complementary feeding. Sub-topic analysis of negative reviews revealed specific stressors such as login failures, unclear content, and inconsistent monitoring features, which align with technostress dimensions like techno-overload, techno-complexity, and techno-uncertainty. To classify user sentiment, a Support Vector Machine (SVM) model with a linear kernel and TF-IDF-transformed features was applied using an 80:20 train-test split and showed acceptable results in recognizing sentiment patterns. These findings highlight common stressors in maternal app usage and provide insights for building more emotionally supportive, stress-aware digital health tools.