Abstract

Hospitalshaverapidlyadoptedtheuseofelectronicmedicalrecords(EMR)forroutinemanagementandreportingof patienthealthcareutilization.InspiteofthecomprehensivedatacollectedinEMRs,theyhavenotrealizedtheir potentialforconductingroutinesurveillanceofqualitymeasures,formeasuringhospitalperformance,orfor surveillanceofpatientsafety.TheuseofEMRsforpatientsafetysurveillanceandforpredictiveanalyticshasbeen underutilizedespeciallyforacutemyocardialinfarction(AMI).Reasonsforthisunderuseincludefragmentationofdata entryandstorage,poorcomplianceincompletingstructuredfieldsforqualityreporting,andtheabundanceof unstructuredinformationdescribedinnarrativenotes.Weproposetodeveloparobustautomatedsurveillancetoolkit builtintwoindependentEMRswithexternalvalidationinmultipleEMRs.Wewillcombinetherichinformationlocked inclinicalnoteswithstructureddatatoquantifytheriskforreadmissionafteranAMIdirectlyfromtheEMR,validate, anddemonstrateitsportabilityacrossinstitutionstootherEMRs.Ouroverallhypothesisisthataddingstructured variablesfromtheEMRwithNLP-derivedvariableswillimproveourabilitytopredict30-dayreadmissionfromAMI. Wewillevaluatethishypothesisbymappingrelevantvariablestocommoninformationmodels,developingand validatingpredictionmodelsforAMI,andcreatingandvalidatingaportabletoolkitforgeneratingpredictivemodels frommultipleEMRsinthefollowingspecificaims:1)ToevaluatepotentialAMIriskfactorsfor30-dayreadmission fromAMItoacommoninformationmodelusingstructuredEMRvariablesandnovelNLPvariablesextractedfrom EMRtext;?2)Todevelopanoptimalpredictionmodelfor30-dayreadmissionfromAMIateachsiteusingregistrydata, structuredEMR,andnovelsocialNLPvariablesextractedfromunstructuredEMRtextandtocross-validateeach modelatanotherinstitution;?3)Tovalidateanautomatedsurveillancetoolkit(ReX)forportabilitytothreeotherEMRs. ThisresearchissignificantinthatitwillimproveourabilitytoidentifyAMIpatientsatriskof30-dayreadmission, identifyriskforcausesofreadmissionforactionableinterventionbeforereadmissionoccurs,andforthefirsttime provideavalidatedportablesurveillancetoolkit.Ourresearchisinnovative,becauseitexpandstheuseofNLPtools tonovelvariablespreviouslyonlyobtainedthroughmanualextraction(e.g.,socialriskfactors)anddevelopsa generalizableandportabletoolkitbuiltinparallelontwoindependentEMRswithexternalvalidationinmultipleEMRs. Wewillshifttheparadigmfromcurrentsingle-centerapproachestoa2-centerparalleldevelopmentandcross- validationmethodallowingfornovelinformationevaluationandsystematicdifferencesindatarepresentationbetween thetwoinstitutionsandadaptingourportabletoolkitaccordingly.Wewillsignificantlyadvancebiomedicalinformatics tooldevelopmentandourabilitytoperformriskassessmentforAMIpatients,enablingimprovedclinicalcareand improvedpatientoutcomes.

Public Health Relevance

Theproposedresearchwilldevelopaportableandexternallyvalidatedtoolkitforextractingcomplexinformationfrom theelectronicmedicalrecordaboutpatienthealthandhealthcarefactorsthatpredicttheriskofreturningtothe hospitalforanotherstay.Thetoolsdevelopedinthisresearchwilladvancethefieldsofhealthinformationtechnology, computerscienceandhealthcaredeliverysciencewhileassistinghealthcareprofessionalswithtoolstoimprove patientcareandsafety.Thesemethodswillbetranslatedintootherreasonsforhospitalstaystohelphealthcare workersmaximizepatientsafetyandhealthafterpatientsleavethehospitalwhilereducinghealthcarecostsin preventingunnecessaryhospitalstays.

  Funding Agency

Agency
National Institute of Health (NIH)
Institute
National Heart, Lung, and Blood Institute (NHLBI)
Type
Research Project (R01)
Project #
5R01HL130828-04
Application #
9692759
Study Section
Biomedical Computing and Health Informatics Study Section (BCHI)
Program Officer
Hsu, Lucy L
Project Start
2016-07-01
Project End
2021-04-30
Budget Start
2019-05-01
Budget End
2021-04-30
Support Year
4
Fiscal Year
2019
Total Cost
Indirect Cost

  Institution

Name
Dartmouth College
Department
Type
Schools of Medicine
DUNS #
041027822
City
Hanover
State
NH
Country
United States
Zip Code
03755

  Publications

Rosenbloom, S T; Carroll, R J; Warner, J L et al. (2017) Representing Knowledge Consistently Across Health Systems. Yearb Med Inform 26:139-147
Caracciolo, Chris; Parker, Devin; Marshall, Emily et al. (2017) Excess Readmission vs Excess Penalties: Maximum Readmission Penalties as a Function of Socioeconomics and Geography. J Hosp Med 12:610-617
Brown, Jeremiah R; Chang, Chiang-Hua; Zhou, Weiping et al. (2014) Health system characteristics and rates of readmission after acute myocardial infarction in the United States. J Am Heart Assoc 3:e000714
Brown, Jeremiah R; Conley, Sheila M; Niles 2nd, Nathaniel W (2013) Predicting readmission or death after acute ST-elevation myocardial infarction. Clin Cardiol 36:570-5