{"id":"https://openalex.org/W4415138906","doi":"https://doi.org/10.3390/bdcc9100256","title":"Robust Clinical Querying with Local LLMs: Lexical Challenges in NL2SQL and Retrieval-Augmented QA on EHRs","display_name":"Robust Clinical Querying with Local LLMs: Lexical Challenges in NL2SQL and Retrieval-Augmented QA on EHRs","publication_year":2025,"publication_date":"2025-10-11","ids":{"openalex":"https://openalex.org/W4415138906","doi":"https://doi.org/10.3390/bdcc9100256"},"language":"en","primary_location":{"id":"doi:10.3390/bdcc9100256","is_oa":true,"landing_page_url":"https://doi.org/10.3390/bdcc9100256","pdf_url":"https://www.mdpi.com/2504-2289/9/10/256/pdf?version=1760347130","source":{"id":"https://openalex.org/S4210238752","display_name":"Big Data and Cognitive Computing","issn_l":"2504-2289","issn":["2504-2289"],"is_oa":true,"is_in_doaj":true,"is_core":true,"host_organization":"https://openalex.org/P4310310987","host_organization_name":"Multidisciplinary Digital Publishing Institute","host_organization_lineage":["https://openalex.org/P4310310987"],"host_organization_lineage_names":["Multidisciplinary Digital Publishing Institute"],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Big Data and Cognitive Computing","raw_type":"journal-article"},"type":"article","indexed_in":["crossref","doaj"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://www.mdpi.com/2504-2289/9/10/256/pdf?version=1760347130","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5045596935","display_name":"Luka Bla\u0161kovi\u0107","orcid":"https://orcid.org/0000-0002-8942-8585"},"institutions":[{"id":"https://openalex.org/I25607154","display_name":"Juraj Dobrila University of Pula","ror":"https://ror.org/006ks2460","country_code":"HR","type":"education","lineage":["https://openalex.org/I25607154"]}],"countries":["HR"],"is_corresponding":false,"raw_author_name":"Luka Bla\u0161kovi\u0107","raw_affiliation_strings":["Faculty of Informatics, Juraj Dobrila University of Pula, 52100 Pula, Croatia"],"affiliations":[{"raw_affiliation_string":"Faculty of Informatics, Juraj Dobrila University of Pula, 52100 Pula, Croatia","institution_ids":["https://openalex.org/I25607154"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5065852136","display_name":"Nikola Tankovi\u0107","orcid":"https://orcid.org/0000-0003-3936-1244"},"institutions":[{"id":"https://openalex.org/I25607154","display_name":"Juraj Dobrila University of Pula","ror":"https://ror.org/006ks2460","country_code":"HR","type":"education","lineage":["https://openalex.org/I25607154"]}],"countries":["HR"],"is_corresponding":true,"raw_author_name":"Nikola Tankovi\u0107","raw_affiliation_strings":["Faculty of Informatics, Juraj Dobrila University of Pula, 52100 Pula, Croatia"],"affiliations":[{"raw_affiliation_string":"Faculty of Informatics, Juraj Dobrila University of Pula, 52100 Pula, Croatia","institution_ids":["https://openalex.org/I25607154"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5078677385","display_name":"Ivan Lorencin","orcid":"https://orcid.org/0000-0002-5964-245X"},"institutions":[{"id":"https://openalex.org/I25607154","display_name":"Juraj Dobrila University of Pula","ror":"https://ror.org/006ks2460","country_code":"HR","type":"education","lineage":["https://openalex.org/I25607154"]}],"countries":["HR"],"is_corresponding":true,"raw_author_name":"Ivan Lorencin","raw_affiliation_strings":["Faculty of Informatics, Juraj Dobrila University of Pula, 52100 Pula, Croatia"],"affiliations":[{"raw_affiliation_string":"Faculty of Informatics, Juraj Dobrila University of Pula, 52100 Pula, Croatia","institution_ids":["https://openalex.org/I25607154"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5054098756","display_name":"Sandi Baressi \u0160egota","orcid":"https://orcid.org/0000-0002-3015-1024"},"institutions":[{"id":"https://openalex.org/I154347574","display_name":"University of Rijeka","ror":"https://ror.org/05r8dqr10","country_code":"HR","type":"education","lineage":["https://openalex.org/I154347574"]}],"countries":["HR"],"is_corresponding":true,"raw_author_name":"Sandi Baressi \u0160egota","raw_affiliation_strings":["Department of Automation and Electronics, Faculty of Engineering, University of Rijeka, 51000 Rijeka, Croatia"],"affiliations":[{"raw_affiliation_string":"Department of Automation and Electronics, Faculty of Engineering, University of Rijeka, 51000 Rijeka, Croatia","institution_ids":["https://openalex.org/I154347574"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":4,"corresponding_author_ids":["https://openalex.org/A5054098756","https://openalex.org/A5065852136","https://openalex.org/A5078677385"],"corresponding_institution_ids":["https://openalex.org/I154347574","https://openalex.org/I25607154"],"apc_list":{"value":1400,"currency":"CHF","value_usd":1515},"apc_paid":{"value":1400,"currency":"CHF","value_usd":1515},"fwci":0.7244,"has_fulltext":true,"cited_by_count":1,"citation_normalized_percentile":{"value":0.75492398,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":96,"max":98},"biblio":{"volume":"9","issue":"10","first_page":"256","last_page":"256"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11710","display_name":"Biomedical Text Mining and Ontologies","score":0.9995999932289124,"subfield":{"id":"https://openalex.org/subfields/1312","display_name":"Molecular Biology"},"field":{"id":"https://openalex.org/fields/13","display_name":"Biochemistry, Genetics and Molecular Biology"},"domain":{"id":"https://openalex.org/domains/1","display_name":"Life Sciences"}},"topics":[{"id":"https://openalex.org/T11710","display_name":"Biomedical Text Mining and Ontologies","score":0.9995999932289124,"subfield":{"id":"https://openalex.org/subfields/1312","display_name":"Molecular Biology"},"field":{"id":"https://openalex.org/fields/13","display_name":"Biochemistry, Genetics and Molecular Biology"},"domain":{"id":"https://openalex.org/domains/1","display_name":"Life Sciences"}},{"id":"https://openalex.org/T10181","display_name":"Natural Language Processing Techniques","score":0.9987999796867371,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.9976000189781189,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/sql","display_name":"SQL","score":0.5289999842643738},{"id":"https://openalex.org/keywords/relational-database","display_name":"Relational database","score":0.504800021648407},{"id":"https://openalex.org/keywords/benchmark","display_name":"Benchmark (surveying)","score":0.4936999976634979},{"id":"https://openalex.org/keywords/bottleneck","display_name":"Bottleneck","score":0.4925999939441681},{"id":"https://openalex.org/keywords/security-token","display_name":"Security token","score":0.4586000144481659},{"id":"https://openalex.org/keywords/question-answering","display_name":"Question answering","score":0.4408000111579895},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.4187000095844269},{"id":"https://openalex.org/keywords/pairwise-comparison","display_name":"Pairwise comparison","score":0.4058000147342682}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7889000177383423},{"id":"https://openalex.org/C510870499","wikidata":"https://www.wikidata.org/wiki/Q47607","display_name":"SQL","level":2,"score":0.5289999842643738},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.5073000192642212},{"id":"https://openalex.org/C5655090","wikidata":"https://www.wikidata.org/wiki/Q192588","display_name":"Relational database","level":2,"score":0.504800021648407},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4970000088214874},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.4936999976634979},{"id":"https://openalex.org/C2780513914","wikidata":"https://www.wikidata.org/wiki/Q18210350","display_name":"Bottleneck","level":2,"score":0.4925999939441681},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.4586000144481659},{"id":"https://openalex.org/C44291984","wikidata":"https://www.wikidata.org/wiki/Q1074173","display_name":"Question answering","level":2,"score":0.4408000111579895},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.4187000095844269},{"id":"https://openalex.org/C184898388","wikidata":"https://www.wikidata.org/wiki/Q1435712","display_name":"Pairwise comparison","level":2,"score":0.4058000147342682},{"id":"https://openalex.org/C192209626","wikidata":"https://www.wikidata.org/wiki/Q190909","display_name":"Focus (optics)","level":2,"score":0.3946000039577484},{"id":"https://openalex.org/C157486923","wikidata":"https://www.wikidata.org/wiki/Q1376436","display_name":"String (physics)","level":2,"score":0.39149999618530273},{"id":"https://openalex.org/C23123220","wikidata":"https://www.wikidata.org/wiki/Q816826","display_name":"Information retrieval","level":1,"score":0.37450000643730164},{"id":"https://openalex.org/C195324797","wikidata":"https://www.wikidata.org/wiki/Q33742","display_name":"Natural language","level":2,"score":0.34299999475479126},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.33660000562667847},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.2847999930381775},{"id":"https://openalex.org/C3019952477","wikidata":"https://www.wikidata.org/wiki/Q1324077","display_name":"Health records","level":3,"score":0.2793000042438507},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.27730000019073486},{"id":"https://openalex.org/C47487241","wikidata":"https://www.wikidata.org/wiki/Q5227230","display_name":"Data access","level":2,"score":0.27379998564720154},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.2662000060081482},{"id":"https://openalex.org/C177148314","wikidata":"https://www.wikidata.org/wiki/Q170084","display_name":"Generalization","level":2,"score":0.26570001244544983},{"id":"https://openalex.org/C67186912","wikidata":"https://www.wikidata.org/wiki/Q367664","display_name":"Data modeling","level":2,"score":0.2603999972343445},{"id":"https://openalex.org/C148524875","wikidata":"https://www.wikidata.org/wiki/Q6975395","display_name":"F1 score","level":2,"score":0.2531000077724457}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.3390/bdcc9100256","is_oa":true,"landing_page_url":"https://doi.org/10.3390/bdcc9100256","pdf_url":"https://www.mdpi.com/2504-2289/9/10/256/pdf?version=1760347130","source":{"id":"https://openalex.org/S4210238752","display_name":"Big Data and Cognitive Computing","issn_l":"2504-2289","issn":["2504-2289"],"is_oa":true,"is_in_doaj":true,"is_core":true,"host_organization":"https://openalex.org/P4310310987","host_organization_name":"Multidisciplinary Digital Publishing Institute","host_organization_lineage":["https://openalex.org/P4310310987"],"host_organization_lineage_names":["Multidisciplinary Digital Publishing Institute"],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Big Data and Cognitive Computing","raw_type":"journal-article"},{"id":"pmh:oai:doaj.org/article:c0ea565179224ed28f5d62d109815891","is_oa":true,"landing_page_url":"https://doaj.org/article/c0ea565179224ed28f5d62d109815891","pdf_url":null,"source":{"id":"https://openalex.org/S112646816","display_name":"SHILAP Revista de lepidopterolog\u00eda","issn_l":"0300-5267","issn":["0300-5267","2340-4078"],"is_oa":true,"is_in_doaj":true,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"Big Data and Cognitive Computing, Vol 9, Iss 10, p 256 (2025)","raw_type":"article"}],"best_oa_location":{"id":"doi:10.3390/bdcc9100256","is_oa":true,"landing_page_url":"https://doi.org/10.3390/bdcc9100256","pdf_url":"https://www.mdpi.com/2504-2289/9/10/256/pdf?version=1760347130","source":{"id":"https://openalex.org/S4210238752","display_name":"Big Data and Cognitive Computing","issn_l":"2504-2289","issn":["2504-2289"],"is_oa":true,"is_in_doaj":true,"is_core":true,"host_organization":"https://openalex.org/P4310310987","host_organization_name":"Multidisciplinary Digital Publishing Institute","host_organization_lineage":["https://openalex.org/P4310310987"],"host_organization_lineage_names":["Multidisciplinary Digital Publishing Institute"],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Big Data and Cognitive Computing","raw_type":"journal-article"},"sustainable_development_goals":[],"awards":[],"funders":[{"id":"https://openalex.org/F4320320300","display_name":"European Commission","ror":"https://ror.org/00k4n6c32"},{"id":"https://openalex.org/F4320321697","display_name":"LOEWE Zentrum AdRIA","ror":"https://ror.org/05n911h24"}],"has_content":{"pdf":true,"grobid_xml":false},"content_urls":{"pdf":"https://content.openalex.org/works/W4415138906.pdf"},"referenced_works_count":66,"referenced_works":["https://openalex.org/W86247481","https://openalex.org/W147952491","https://openalex.org/W1992358349","https://openalex.org/W2101105183","https://openalex.org/W2159583324","https://openalex.org/W2164765923","https://openalex.org/W2781002470","https://openalex.org/W2785625179","https://openalex.org/W2799503098","https://openalex.org/W2890431379","https://openalex.org/W2911489562","https://openalex.org/W2945102109","https://openalex.org/W3022196370","https://openalex.org/W3027879771","https://openalex.org/W3156770071","https://openalex.org/W3209235691","https://openalex.org/W4214860116","https://openalex.org/W4317790708","https://openalex.org/W4360885664","https://openalex.org/W4375869868","https://openalex.org/W4382753509","https://openalex.org/W4384071683","https://openalex.org/W4386448382","https://openalex.org/W4387316166","https://openalex.org/W4388033286","https://openalex.org/W4389524197","https://openalex.org/W4391227238","https://openalex.org/W4391940656","https://openalex.org/W4392186815","https://openalex.org/W4392348090","https://openalex.org/W4392353733","https://openalex.org/W4392544551","https://openalex.org/W4396700714","https://openalex.org/W4399628104","https://openalex.org/W4399660853","https://openalex.org/W4399810434","https://openalex.org/W4400837480","https://openalex.org/W4400909566","https://openalex.org/W4401043171","https://openalex.org/W4401103265","https://openalex.org/W4402369854","https://openalex.org/W4402670290","https://openalex.org/W4406152279","https://openalex.org/W4406379614","https://openalex.org/W4406421570","https://openalex.org/W4406642447","https://openalex.org/W4406771321","https://openalex.org/W4406818618","https://openalex.org/W4407051322","https://openalex.org/W4407840751","https://openalex.org/W4408197148","https://openalex.org/W4409060418","https://openalex.org/W4409657176","https://openalex.org/W4409690035","https://openalex.org/W4410030662","https://openalex.org/W4411203672","https://openalex.org/W4411231479","https://openalex.org/W4411969741","https://openalex.org/W4411985537","https://openalex.org/W4412809772","https://openalex.org/W4412875459","https://openalex.org/W4412877147","https://openalex.org/W4412889134","https://openalex.org/W4412889782","https://openalex.org/W4413293038","https://openalex.org/W7081978817"],"related_works":[],"abstract_inverted_index":{"Electronic":[0],"health":[1],"records":[2,198],"(EHRs)":[3],"are":[4],"typically":[5],"stored":[6],"in":[7,187],"relational":[8],"databases,":[9],"making":[10],"them":[11],"difficult":[12],"to":[13,31,80,143,256],"query":[14],"for":[15,34,40,174,185,303],"nontechnical":[16],"users,":[17],"especially":[18],"under":[19,289],"privacy":[20],"constraints.":[21],"We":[22,51,146],"evaluate":[23],"two":[24],"practical":[25],"clinical":[26,41,285,305],"NLP":[27],"workflows,":[28],"natural":[29],"language":[30,55],"SQL":[32,151],"(NL2SQL)":[33],"EHR":[35],"querying":[36],"and":[37,68,70,74,91,133,208,224,268,300],"retrieval-augmented":[38],"generation":[39],"question":[42],"answering":[43],"(RAG-QA),":[44],"with":[45,129,154,221],"a":[46,82,139,149,271],"focus":[47],"on":[48,194,219,244],"privacy-preserving":[49],"deployment.":[50],"benchmark":[52],"nine":[53,98],"large":[54],"models,":[56,118,214],"spanning":[57,85],"open-weight":[58,117,265],"options":[59],"(DeepSeek":[60],"V3/V3.1,":[61],"Llama-3.3-70B,":[62],"Qwen2.5-32B,":[63],"Mixtral-8":[64],"\u00d7":[65,100],"22B,":[66],"BioMistral-7B,":[67],"GPT-OSS-20B)":[69],"proprietary":[71,92],"APIs":[72],"(GPT-4o":[73],"GPT-5).":[75],"The":[76],"models":[77,99],"were":[78,236],"chosen":[79],"represent":[81],"diverse":[83],"cross-section":[84],"sparse":[86],"MoE,":[87],"dense":[88],"general-purpose,":[89],"domain-adapted,":[90],"LLMs.":[93,240],"On":[94],"MIMICSQL":[95],"(27,000":[96],"generations;":[97],"three":[101,205],"runs),":[102,206],"the":[103,160,182,188,227,239,263,276,284],"best":[104,264],"NL2SQL":[105,186,286],"execution":[106],"accuracy":[107],"(EX)":[108],"is":[109],"66.1%":[110],"(GPT-4o),":[111],"followed":[112,164],"by":[113,165],"64.6%":[114],"(GPT-5).":[115],"Among":[116],"DeepSeek":[119,125,222,260],"V3.1":[120,223,261],"reaches":[121,127],"59.8%":[122],"EX,":[123],"while":[124,169],"V3":[126],"58.8%,":[128],"Llama-3.3-70B":[130],"at":[131],"54.5%":[132],"BioMistral-7B":[134],"achieving":[135],"only":[136,175],"11.8%,":[137],"underscoring":[138],"persistent":[140],"gap":[141],"relative":[142],"general-domain":[144],"benchmarks.":[145],"introduce":[147],"SQL-EC,":[148],"deterministic":[150],"error-classification":[152],"framework":[153],"adjudication,":[155],"revealing":[156],"string":[157],"mismatches":[158],"as":[159,181],"dominant":[161],"failure":[162,295],"(86.3%),":[163],"query-join":[166],"misinterpretations":[167],"(49.7%),":[168],"incorrect":[170],"aggregation-function":[171],"usage":[172,247],"accounts":[173],"6.7%.":[176],"This":[177],"highlights":[178],"lexical/ontology":[179],"grounding":[180],"key":[183],"bottleneck":[184],"biomedical":[189],"domain.":[190],"For":[191],"RAG-QA,":[192],"evaluated":[193],"100":[195],"synthetic":[196],"patient":[197],"across":[199,213],"20":[200],"questions":[201],"(54,000":[202],"reference\u2013generation":[203],"pairs;":[204],"BLEU":[207],"ROUGE-L":[209],"fluctuate":[210],"more":[211],"strongly":[212],"whereas":[215,283],"BERTScore":[216],"remains":[217,287],"high":[218],"most,":[220],"GPT-4o":[225],"among":[226,238],"top":[228],"performers;":[229],"pairwise":[230],"t-tests":[231],"confirm":[232],"that":[233],"significant":[234],"differences":[235],"observed":[237],"Cost\u2013performance":[241],"analysis":[242],"based":[243],"measured":[245],"token":[246],"shows":[248],"per-query":[249],"costs":[250],"ranging":[251],"from":[252],"USD":[253,257],"0.000285":[254],"(GPT-OSS-20B)":[255],"0.005918":[258],"(GPT-4o);":[259],"offers":[262],"cost\u2013accuracy":[266],"trade-off,":[267],"GPT-5":[269],"provides":[270],"balanced":[272],"API":[273],"alternative.":[274],"Overall,":[275],"privacy-conscious":[277],"RAG-QA":[278],"attains":[279],"strong":[280],"semantic":[281],"fidelity,":[282],"brittle":[288],"lexical":[290],"variation.":[291],"SQL-EC":[292],"pinpoints":[293],"actionable":[294],"modes,":[296],"motivating":[297],"ontology-aware":[298],"normalization":[299],"schema-linked":[301],"prompting":[302],"robust":[304],"querying.":[306]},"counts_by_year":[{"year":2026,"cited_by_count":1}],"updated_date":"2026-04-21T08:09:41.155169","created_date":"2025-10-14T00:00:00"}