Evaluation of a Large Language Model Discharge Summary Hospital Course Tool: Improved Quality but Longer Documentation Time

{"BACKGROUND:":[0],"No":[1],"studies":[2],"have":[3],"examined":[4],"the":[5,21,150,181,189,247,266],"effect":[6,22],"of":[7,23,121,149,276],"artificial":[8],"intelligence":[9],"(AI)-generated":[10],"hospital":[11,26,65,77,97],"courses":[12],"on":[13,29,85],"discharge":[14,30,82,104,161],"summary":[15,31,83,105,162,253,281],"documentation":[16,290],"time.":[17,291],"OBJECTIVES:":[18],"To":[19],"evaluate":[20],"an":[24],"AI-generated":[25,230],"course":[27,66,78],"tool":[28,67,152,191,207,248,268,279],"edit":[32,106,163,184,215,254],"time,":[33,107],"quality,":[34],"and":[35,52,133,145,235,283],"clinician":[36],"perceptions.":[37],"METHODS:":[38],"Observational":[39],"study":[40,166,260],"between":[41,165,259],"a":[42,53,69,75,139,157,211],"\\"Pre-AI":[43],"period\\"":[44,55],"(March":[45],"16,":[46],"2025":[47,58],"to":[48,59,73,114],"November":[49,86],"19,":[50],"2025)":[51],"\\"Post-AI":[54],"(November":[56],"20,":[57,87],"February":[60],"11,":[61],"2026).":[62],"An":[63],"AI":[64,151,190,206,267],"using":[68],"GPT":[70],"4.1":[71],"model":[72],"generate":[74],"draft":[76],"integrated":[79],"into":[80],"mandatory":[81],"templates":[84],"2025.":[88],"We":[89],"included":[90],"8,298":[91],"hospitalized":[92],"adults":[93],"cared":[94],"for":[95,126,131],"by":[96],"medicine":[98],"clinicians.":[99],"The":[100,278],"primary":[101],"outcome":[102],"was":[103,124,153,186,192,208,263,269,273],"defined":[108],"as":[109],"time":[110,164,185,216,255],"from":[111],"note":[112],"creation":[113],"signature":[115],"across":[116],"all":[117],"sessions.":[118],"A":[119],"subset":[120],"27":[122],"encounters":[123,179],"reviewed":[125],"cohesiveness,":[127],"comprehensiveness,":[128],"conciseness,":[129],"potential":[130,285],"harm,":[132],"overall":[134],"quality.":[135,146],"Thirty-six":[136],"clinicians":[137,245],"completed":[138],"survey":[140],"assessing":[141],"usability,":[142],"efficiency,":[143],"trustworthiness,":[144],"RESULTS:":[147],"Implementation":[148],"not":[154,257],"associated":[155,209],"with":[156,210],"significant":[158],"change":[159],"in":[160,180,214],"periods":[167,261],"(6.60":[168],"[IQR":[169,173,195,199],"3.4-13.5]":[170],"vs":[171,197],"6.28":[172],"3.2-13.0]":[174],"minutes,":[175,201],"p=0.11).":[176],"Among":[177],"2,206":[178],"Post-AI":[182],"period,":[183],"longer":[187,264],"when":[188,265],"used":[193],"(9.20":[194],"4.9-16.3]":[196],"4.93":[198],"2.5-9.7]":[200],"p<0.001).":[202],"In":[203],"multivariable":[204],"analysis,":[205],"32%":[212],"increase":[213,289],"(95%":[217],"CI":[218],"24%-41%,":[219],"p":[220],"<":[221],"0.001;":[222],"adjusted":[223],"R²":[224],"=":[225],"0.34).":[226],"Faculty":[227],"review":[228],"found":[229],"summaries":[231],"had":[232],"higher":[233],"quality":[234,282],"lower":[236],"harm":[237,286],"scores":[238],"but":[239,262,287],"were":[240],"less":[241],"concise.":[242],"31":[243],"(86.1%)":[244],"agreed":[246],"improved":[249,280],"efficiency.":[250,277],"CONCLUSIONS:":[251],"Discharge":[252],"did":[256],"differ":[258],"used.":[270],"However,":[271],"there":[272],"increased":[274],"perception":[275],"reduced":[284],"may":[288]}

Authors

Institutions

Publication Details

Journal
Applied Clinical Informatics
Published
2026-09-18
DOI
https://doi.org/10.1055/a-2962-8004
Primary Topic
Artificial Intelligence in Healthcare and Education
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
article

Evaluation of a Large Language Model Discharge Summary Hospital Course Tool: Improved Quality but Longer Documentation Time

Sebastián Suárez, Alexander Dobek, Sayari Patel, Maritza Suarez et al.
Applied Clinical Informatics
Artificial Intelligence in Healthcare and Education
article

Evaluation of a Large Language Model Discharge Summary Hospital Course Tool: Improved Quality but Longer Documentation Time

Sebastián Suárez, Alexander Dobek, Sayari Patel, Maritza Suarez, Christina DeBenedictus, Fawaz Naeem
article en

Abstract

BACKGROUND: No studies have examined the effect of artificial intelligence (AI)-generated hospital courses on discharge summary documentation time. OBJECTIVES: To evaluate the effect of an AI-generated hospital course tool on discharge summary edit time, quality, and clinician perceptions. METHODS: Observational study between a "Pre-AI period" (March 16, 2025 to November 19, 2025) and a "Post-AI period" (November 20, 2025 to February 11, 2026). An AI hospital course tool using a GPT 4.1 model to generate a draft hospital course integrated into mandatory discharge summary templates on November 20, 2025. We included 8,298 hospitalized adults cared for by hospital medicine clinicians. The primary outcome was discharge summary edit time, defined as time from note creation to signature across all sessions. A subset of 27 encounters was reviewed for cohesiveness, comprehensiveness, conciseness, potential for harm, and overall quality. Thirty-six clinicians completed a survey assessing usability, efficiency, trustworthiness, and quality. RESULTS: Implementation of the AI tool was not associated with a significant change in discharge summary edit time between study periods (6.60 [IQR 3.4-13.5] vs 6.28 [IQR 3.2-13.0] minutes, p=0.11). Among 2,206 encounters in the Post-AI period, edit time was longer when the AI tool was used (9.20 [IQR 4.9-16.3] vs 4.93 [IQR 2.5-9.7] minutes, p<0.001). In multivariable analysis, AI tool was associated with a 32% increase in edit time (95% CI 24%-41%, p < 0.001; adjusted R² = 0.34). Faculty review found AI-generated summaries had higher quality and lower harm scores but were less concise. 31 (86.1%) clinicians agreed the tool improved efficiency. CONCLUSIONS: Discharge summary edit time did not differ between study periods but was longer when the AI tool was used. However, there was increased perception of efficiency. The tool improved summary quality and reduced potential harm but may increase documentation time.

Applied Clinical Informatics
University of Miami (US)
Openalex Percentile: Top 15%
Artificial Intelligence in Healthcare and Education
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.