This paper provides an overview of Task 2 from the Social Media Mining for Health and Health Real-World Data (#SMM4H-HeaRD) 2026 Workshop and Shared Tasks, which focused on the detection of insomnia in clinical notes derived from the MIMIC-III dataset. The task consisted of two subtasks: binary text classification to determine whether a patient is likely experiencing insomnia (Subtask 1), and multi-label classification combined with character-level evidence extraction to identify supporting evidence for specific insomnia crite- ria (Subtask 2). Eight teams participated, using approaches ranging from large language model (LLM) prompting and fine-tuned encoder mod- els to hybrid rule-based pipelines. Results demonstrated that structured LLM pipelines with deterministic post-processing achieved the strongest overall performance, while character-level span extraction remained substantially harder than classification across all systems. These findings highlight both the promise of NLP for identifying underdiagnosed conditions in electronic health records and the ongoing difficulty of producing interpretable, evidence-grounded clinical predictions.