Mental health disorders significantly impact people globally, regardless of background, education, or socioeconomic status. However, access to adequate care remains a challenge, particularly for underserved communities with limited resources. Text mining tools offer immense potential to support mental healthcare by assisting professionals in diagnosing and treating patients. This study addresses the scarcity of Arabic mental health resources for developing such tools. We introduce MentalQA, a novel Arabic dataset featuring questioning-answering (QA) style, including a total of 1000 annotations. To ensure data quality, we conducted a rigorous annotation process using a well-defined schema with quality control measures. Data was collected from a question-answering medical platform. The annotation schema for mental health questions and corresponding answers draws upon existing classification schemes with some modifications. Question types encompass six distinct categories: diagnosis, treatment, anatomy & physiology, epidemiology, healthy lifestyle, and provider choice. Answer strategies include information provision, direct guidance, and emotional support. Three experienced annotators collaboratively annotated the data to ensure consistency. Our findings demonstrate high inter-annotator agreement, with Fleiss’ Kappa of 0.61 for question types and 0.98 for answer strategies. Our in-depth analysis uncovered insightful patterns in patients behavior and doctor responses. We found a link between the types of questions asked and the strategies doctors use in their answers. Both genders focused on treatment-related questions, though with some variation in emphasis. Interestingly, the types of questions asked also differed by age group. Patients tended to express negative emotions in their questions, while doctors maintained a neutral tone in their responses. Finally, we observed differences in how long it took doctors to answer and the number of words they used depending on the type of question. MentalQA offers a valuable foundation for developing Arabic text mining tools capable of supporting mental health professionals and individuals seeking information.

Background: Web-based question and answer (Q&A) sites have emerged as an alternative source for serving individuals’ health information needs. Although a number of studies have analyzed user-generated content in web-based Q&A sites, there is insufficient understanding of the effect of disease complexity on information-seeking needs and the types of information shared, and little research has been devoted to the questions concerning multimorbidity. Objective: This study aims to investigate seeking of health information in Q&A sites at different levels of disease complexity. Specifically, this study investigates the effects of disease complexity on information-seeking needs, types of information shared, and stages of disease development. Methods: First, we selected a random sample of 400 questions separately from each of the Q&A sites: Yahoo Answers and WebMD Answers. The data cleaning resulted in a final set of 624 questions from the two sites. We used a mixed methods approach, including qualitative content analysis and quantitative statistical analysis. Results: The one-way results of ANOVA showed significant effects of disease complexity (single vs multimorbid disease questions) on two information-seeking needs: diagnosis (F1,622=5.08; P=.02) and treatment (F1,622=4.82; P=.02). There were also significant differences between the two levels of disease complexity in two stages of disease development: the general health stage (F1,622=48.02; P<.001) and the chronic stage (F1,622=54.01; P<.001). In addition, our results showed significant effects of disease complexity across all types of shared information: demographic information (F1,622=32.24; P<.001), medical diagnosis (F1,622=11.04; P<.001), and treatment and prevention (F1,622=14.55; P<.001). Conclusions: Our findings present implications for the design of web-based Q&A sites to better support health information seeking. Future studies should be conducted to validate the generality of these findings and apply them to improve the effectiveness of health information in Q&A sites.

COVID-19 has had broad disruptive effects on economies, healthcare systems, governments, societies, and individuals. Uncertainty concerning the scale of this crisis has given rise to countless rumors, hoaxes, and misinformation. Much of this type of conversation and misinformation about the pandemic now occurs online and in particular on social media platforms like Twitter. This study analysis incorporated a data-driven approach to map the contours of misinformation and contextualize the COVID-19 pandemic with regards to socio-religious-political information. This work consists of a combined system bridging quantitative and qualitative methodologies to assess how information-exchanging behaviors can be used to minimize the effects of emergent misinformation. The study revealed that the social media platforms detected the most significant source of rumors in transmitting information rapidly in the community. It showed that WhatsApp users made up about 46% of the source of rumors in online platforms, while, through Twitter, it demonstrated a declining trend of rumors by 41%. Moreover, the results indicate the second-most common type of misinformation was provided by pharmaceutical companies; however, a prevalent type of misinformation spreading in the world during this pandemic has to do with the biological war. In this combined retrospective analysis of the study, social media with varying approaches in public discourse contributes to efficient public health responses.

Mental health disorders represent a substantial global health concern, impacting millions and placing a significant burden on public health systems. Natural Language Processing (NLP) has emerged as a promising tool for analyzing large textual datasets to identify and predict mental health challenges. The aim of this scoping review is to identify the Arabic NLP techniques employed in mental health research, the specific mental health conditions addressed, and the effectiveness of these techniques in detecting and predicting such conditions. This scoping review was conducted according to the PRISMA-ScR (Preferred Reporting Items for Systematic reviews and Meta-Analyses extension for Scoping Reviews) framework. Studies were included if they focused on the application of NLP techniques, addressed mental health issues (e.g., depression, anxiety, suicidal ideation) within Arabic text data, were published in peer-reviewed journals or conference proceedings, and were written in English or Arabic. The relevant literature was identified through a systematic search of four databases: PubMed, ScienceDirect, IEEE Xplore, and Google Scholar. The results of the included studies revealed a variety of NLP techniques used to address specific mental health issues among Arabic-speaking populations. Commonly utilized techniques included Support Vector Machine (SVM), Random Forest (RF), Decision Tree (DT), Recurrent Neural Network (RNN), and advanced transformer-based models such as AraBERT and MARBERT. The studies predominantly focused on detecting and predicting symptoms of depression and suicidality from Arabic social media data. The effectiveness of these techniques varied, with trans-former-based models like AraBERT and MARBERT demonstrating superior performance, achieving accuracy rates of up to 99.3% and 98.3%, respectively. Traditional machine learning models and RNNs also showed promise but generally lagged in accuracy and depth of insight compared to transformer models. This scoping review highlights the significant potential of NLP techniques, particularly advanced transformer-based models, in addressing mental health issues among Arabic-speaking populations. Ongoing research is essential to keep pace with the rapidly evolving field and to validate current findings.

Background: Pre-Trained Language Models hold significant promise for revolutionizing mental health care by delivering accessible and culturally sensitive resources. Despite this potential, their efficacy in mental health applications, particularly in the Arabic language, remains largely unexplored. To the best of our knowledge, comprehensive studies specifically evaluating the performance of PLMs on diverse Arabic mental health tasks are still scarce. This study aims to bridge this gap by evaluating the performance of pre-trained language models in classifying questions and answers within the mental health care domain. Methods: We used the MentalQA dataset, which comprises Arabic Questions and Answers interactions related to mental health. Our experiments involved four distinct learning strategies: traditional feature extraction, using PLMs as feature extractors, fine-tuning PLMs, and employing prompt-based techniques with models, such as GPT-3.5 and GPT-4 in zero-shot and few-shot learning scenarios. Arabic-specific PLMs, including AraBERT, CAMelBERT, and MARBERT, were evaluated. Results: Traditional feature-extraction methods paired with Support Vector Machines (SVM) showed competitive performance, but PLMs outperformed them due to their superior ability to capture semantic nuances. In particular, MARBERT achieved the highest performance, with Jaccard scores of 0.80 for the question classification and 0.86 for the answer classification. Further analysis revealed that fine-tuning PLMs enhances their performance, and the size of the training dataset plays a critical role in model effectiveness. Prompt-based techniques, particularly few-shot learning with GPT-3.5, demonstrated significant improvements, increasing the accuracy of question classification by 12% and the accuracy of answer classification by 45%. Conclusions: The study demonstrates the potential of PLMs and prompt-based approaches to provide mental health support to Arabic-speaking populations, providing valuable tools for individuals seeking assistance in this field. This research advances the understanding of PLMs in mental health care and emphasizes their potential to improve accessibility and effectiveness in Arabic-speaking contexts.