{"paper":{"title":"PriHA: A RAG-Enhanced LLM Framework for Primary Healthcare Assistant in Hong Kong","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"PriHA uses query optimization and dual retrieval to give more accurate answers to Hong Kong primary healthcare questions than standard LLMs.","cross_cats":["cs.AI"],"primary_cat":"cs.IR","authors_text":"Hao Chen, Liangjun Jiang, Richard Wai Cheung Chan, Shanru Lin, Wenqi Fan, Ya-nan Ma","submitted_at":"2026-04-10T06:47:43Z","abstract_excerpt":"To address the unsustainable rise in public health expenditures, the Hong Kong SAR Government is shifting its strategic focus to primary healthcare and encouraging citizens to use community resources to self-manage their health. However, official clinical guidelines are fragmented across disparate departments and formats, creating significant access barriers. While general-purpose Large Language Models (LLMs) such as ChatGPT and DeepSeek offer potential solutions for information accessibility, they are prone to generating factually inaccurate content due to a lack of localized and domain-speci"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Comprehensive experiments and a detailed case study demonstrate that our proposed method can outperform both ablations and baseline in terms of accuracy and clarity.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That the mixed-source retrieval in DRAG will consistently surface accurate, up-to-date, and complete guideline fragments without introducing new factual errors or omissions for medical queries.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"PriHA combines query optimization with a Dual Retrieval Augmented Generation pipeline to improve accuracy and clarity of LLM responses on fragmented Hong Kong primary care guidelines.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"PriHA uses query optimization and dual retrieval to give more accurate answers to Hong Kong primary healthcare questions than standard LLMs.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"f94aceea7084c8cec53f223ddba4f95be38b729855c780c8ba1bcb6961577992"},"source":{"id":"2604.14215","kind":"arxiv","version":2},"verdict":{"id":"4e6a6443-9c8c-49f6-9914-ead55a343abd","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-10T17:06:38.192160Z","strongest_claim":"Comprehensive experiments and a detailed case study demonstrate that our proposed method can outperform both ablations and baseline in terms of accuracy and clarity.","one_line_summary":"PriHA combines query optimization with a Dual Retrieval Augmented Generation pipeline to improve accuracy and clarity of LLM responses on fragmented Hong Kong primary care guidelines.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That the mixed-source retrieval in DRAG will consistently surface accurate, up-to-date, and complete guideline fragments without introducing new factual errors or omissions for medical queries.","pith_extraction_headline":"PriHA uses query optimization and dual retrieval to give more accurate answers to Hong Kong primary healthcare questions than standard LLMs."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2604.14215/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","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"}