{"paper":{"title":"An Agentic LLM-Based Framework for Population-Scale Mental Health Screening","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"An agentic LLM framework builds stable pipelines for population-scale mental health screening by locking validated stages after proxy evaluation.","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Donald Cowan, Giuliano Lorenzoni, Paulo Alencar","submitted_at":"2026-05-13T06:08:43Z","abstract_excerpt":"Mental health disorders affect millions worldwide, and healthcare systems are increasingly overwhelmed by the volume of clinical data generated from electronic records, telemedicine platforms, and population-level screening programs. At the same time, the emergence of novel AI-based approaches in healthcare calls for intelligent frameworks capable of processing domain-specific unstructured clinical information while adapting to patient-specific needs. This paper proposes an agentic framework for building robust LLM-based pipelines, where each stage is encapsulated as a LangChain agent governed"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"The proposed framework evolves from feature-level exploration, through proxy-based tuning and freeze/rollback mechanisms, to full orchestration by an Orchestrator Agent that coordinates preprocessing, retrieval, selection, diversity, threshold optimization, and decoding. A proof-of-concept in transcript-based depression detection demonstrates that the framework converges to stable configurations, such as cosine similarity, dynamic Top-k, and threshold 0.75, while controlling evaluation costs and avoiding regressions.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That proxy-guided evaluation metrics reliably predict actual clinical performance and that locking validated stages will prevent regressions without blocking necessary future adaptations to new patient data or clinical contexts.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"An agentic framework orchestrates LLM agents for transcript-based depression detection and converges on stable configurations including cosine similarity, dynamic Top-k, and a 0.75 threshold.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"An agentic LLM framework builds stable pipelines for population-scale mental health screening by locking validated stages after proxy evaluation.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"9fe0471152253b93da57989b2915bb5367155fa567b9a3d8274e9699c033eeb6"},"source":{"id":"2605.13046","kind":"arxiv","version":1},"verdict":{"id":"2a0c1f00-06a6-4c45-943d-a46fa345d601","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-14T19:37:59.538593Z","strongest_claim":"The proposed framework evolves from feature-level exploration, through proxy-based tuning and freeze/rollback mechanisms, to full orchestration by an Orchestrator Agent that coordinates preprocessing, retrieval, selection, diversity, threshold optimization, and decoding. A proof-of-concept in transcript-based depression detection demonstrates that the framework converges to stable configurations, such as cosine similarity, dynamic Top-k, and threshold 0.75, while controlling evaluation costs and avoiding regressions.","one_line_summary":"An agentic framework orchestrates LLM agents for transcript-based depression detection and converges on stable configurations including cosine similarity, dynamic Top-k, and a 0.75 threshold.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That proxy-guided evaluation metrics reliably predict actual clinical performance and that locking validated stages will prevent regressions without blocking necessary future adaptations to new patient data or clinical contexts.","pith_extraction_headline":"An agentic LLM framework builds stable pipelines for population-scale mental health screening by locking validated stages after proxy evaluation."},"references":{"count":31,"sample":[{"doi":"","year":2024,"title":"A survey on RAG meeting LLMs: Towards retrieval-augmented large language models,","work_id":"73e9ecf1-2826-4237-8efc-d4fc5a057dc3","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2024,"title":"A survey on RAG with LLMs,","work_id":"eb521129-fa06-40d1-bf72-43486b2ff04c","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2024,"title":"Efficiency- driven custom chatbot development: Unleashing LangChain, RAG, and performance-optimized LLM fusion,","work_id":"fd92c87e-5d93-4976-b7ff-f68795d4a991","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":1985,"title":"Dynamic configuration for distributed sys- tems,","work_id":"138af8a2-f179-4943-bf28-5e9aba010157","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2004,"title":"Dynamic configuration of resource-aware services,","work_id":"ca507b62-83d7-4024-a784-fa1ed305097d","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":31,"snapshot_sha256":"b2de58da20938792af5645d03d49ea3b254dfd7b37b288d6754c6e51efc2e03f","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}