Utilization of a HIPAA-compliant large language model chatbot in an academic pediatric medical center

{"Despite":[0],"growing":[1],"interest":[2],"in":[3,12,50,212,244,262],"LLM":[4,26,192,242],"chatbots,":[5],"little":[6],"evidence":[7],"exists":[8],"on":[9,77,82,249,255],"usage":[10],"patterns":[11],"hospital":[13,56,186],"systems.":[14],"This":[15],"paper":[16],"presents":[17],"a":[18,24,139,151,156,163,194],"mixed-methods":[19],"case":[20],"study":[21],"of":[22,32,40,110,114,142,167,176,210],"\\"InternalGPT,\\"":[23],"HIPAA-compliant":[25,191],"chatbot":[27],"made":[28],"available":[29],"to":[30,118,155,170,178,235],"employees":[31,143,187],"the":[33,51,105,190,207],"institution.":[34,52],"Survey":[35],"responses":[36],"and":[37,80,84,99,128,214,220,229,265],"14":[38],"months":[39],"system":[41,257],"utilization":[42],"data":[43],"were":[44,223],"analyzed":[45],"across":[46],"diverse":[47],"professional":[48,68],"roles":[49],"Of":[53],"approximately":[54],"15,800":[55],"employees,":[57],"2,149":[58],"(13.6%)":[59],"requested":[60],"InternalGPT":[61,87,146],"access.":[62],"Motivations":[63],"for":[64,125,216],"sign-up":[65],"aligned":[66],"with":[67,144],"roles:":[69],"operations":[70],"professionals":[71],"prioritized":[72],"administrative":[73],"automation,":[74],"clinicians":[75],"focused":[76],"clinical":[78],"augmentation,":[79],"scientists":[81],"research":[83],"coding.":[85],"Among":[86,138],"users,":[88],"52.8%":[89],"recorded":[90],"at":[91],"least":[92],"one":[93],"token":[94],"use;":[95],"33.6%":[96],"never":[97,103,188],"logged-in,":[98],"13.7%":[100],"logged-in":[101],"but":[102,253],"used":[104,189],"tool.":[106],"The":[107],"top":[108],"20%":[109],"users":[111],"consumed":[112],"69.4%":[113],"tokens.":[115],"Key":[116],"barriers":[117,234],"engagement":[119],"among":[120],"non-users":[121],"included":[122],"limited":[123],"time":[124],"exploration":[126],"(51.4%)":[127],"difficulty":[129],"using":[130],"(21.8%),":[131],"integrating":[132],"(15.5%),":[133],"or":[134],"accessing":[135],"(9.2%)":[136],"InternalGPT.":[137],"small":[140,195],"group":[141],"sustained":[145],"use":[147],"(n":[148,159],"=":[149,160],"461),":[150],"subset":[152],"who":[153,197],"responded":[154],"self-report":[157],"survey":[158],"92)":[161],"indicated":[162],"mean":[164],"productivity":[165,174,203],"gain":[166],"30%,":[168],"corresponding":[169],"an":[171],"exploratory":[172],"perceived":[173],"value":[175,208],"$6.3M":[177],"$18.9M":[179],"under":[180],"varying":[181],"extrapolation":[182],"assumptions.":[183],"Although":[184],"most":[185],"chatbot,":[193],"minority":[196],"utilized":[198],"it":[199],"regularly":[200],"reported":[201],"substantial":[202],"improvements,":[204],"highlighting":[205],"both":[206],"proposition":[209],"LLMs":[211],"healthcare":[213,245],"potential":[215],"increased":[217],"engagement.":[218],"Implementation":[219],"knowledge":[221],"challenges":[222],"more":[224],"frequently":[225],"cited":[226],"than":[227],"AI-feature":[228],"performance":[230],"concerns":[231],"as":[232],"key":[233],"adoption.":[236],"These":[237],"findings":[238],"underscore":[239],"that":[240],"successful":[241],"implementation":[243],"hinges":[246],"not":[247],"only":[248],"advanced":[250],"AI":[251],"technology,":[252],"also":[254],"robust":[256],"capacity":[258],"building":[259],"including":[260],"investments":[261],"user":[263],"education":[264],"technical":[266],"integration.":[267]}

Authors

Institutions

Publication Details

Journal
PLOS Digital Health
Published
2026-09-17
DOI
https://doi.org/10.1371/journal.pdig.0001684
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

Utilization of a HIPAA-compliant large language model chatbot in an academic pediatric medical center

Benjamin Rader, Peter C. Laussen, Sampath Chennuri, John S. Brownstein et al.
PLOS Digital Health
Artificial Intelligence in Healthcare and Education
article

Utilization of a HIPAA-compliant large language model chatbot in an academic pediatric medical center

Benjamin Rader, Peter C. Laussen, Sampath Chennuri, John S. Brownstein, Melissa L. Stewart, Max H. Witwer, Sarah Lejfer, Dinesh K. Rai, James Hartwell, Victor Moyano, Heather L. Nelson
article en

Abstract

Despite growing interest in LLM chatbots, little evidence exists on usage patterns in hospital systems. This paper presents a mixed-methods case study of "InternalGPT," a HIPAA-compliant LLM chatbot made available to employees of the institution. Survey responses and 14 months of system utilization data were analyzed across diverse professional roles in the institution. Of approximately 15,800 hospital employees, 2,149 (13.6%) requested InternalGPT access. Motivations for sign-up aligned with professional roles: operations professionals prioritized administrative automation, clinicians focused on clinical augmentation, and scientists on research and coding. Among InternalGPT users, 52.8% recorded at least one token use; 33.6% never logged-in, and 13.7% logged-in but never used the tool. The top 20% of users consumed 69.4% of tokens. Key barriers to engagement among non-users included limited time for exploration (51.4%) and difficulty using (21.8%), integrating (15.5%), or accessing (9.2%) InternalGPT. Among a small group of employees with sustained InternalGPT use (n = 461), a subset who responded to a self-report survey (n = 92) indicated a mean productivity gain of 30%, corresponding to an exploratory perceived productivity value of $6.3M to $18.9M under varying extrapolation assumptions. Although most hospital employees never used the HIPAA-compliant LLM chatbot, a small minority who utilized it regularly reported substantial productivity improvements, highlighting both the value proposition of LLMs in healthcare and potential for increased engagement. Implementation and knowledge challenges were more frequently cited than AI-feature and performance concerns as key barriers to adoption. These findings underscore that successful LLM implementation in healthcare hinges not only on advanced AI technology, but also on robust system capacity building including investments in user education and technical integration.

PLOS Digital HealthVol. 5(9)
Boston Children's Hospital (US), Harvard University (US), Boston Children's Museum (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.