Kevin S. Hughes, MD, FACS Co-Director, Avon Comprehensive Breast Evaluation Center Massachusetts General Hospital Associate Professor of Surgery Harvard Medical school Creating the EHR of the future Using Computers to Decrease Workload and Improve Quality Israeli Association for Medical Informatics
The EHR has tremendous promise as a means of decreasing workload, decreasing cost and improving quality of care
EHR HIT has tremendous promise as a means of decreasing workload, decreasing cost and improving quality of care
EHR and productivity varies by specialty 100 internists, pediatricians and family practitioners 25 to 33 percent drop in MD productivity Over time –Internists slightly above original productivity –Pediatricians /family practitioners never recovered Hemant Bhargava, UC Davis Graduate School of Management
“Computerization hasn't saved a dime, nor has it improved administrative efficiency” 4,000 hospitals 2003 to 2007 Computerization –Weak correlation Quality for MI –No correlation Cost savings Improvements in administrative efficiency Quality for pneumonia Quality for heart failure Overall quality (MI, heart failure, pneumonia) Himmelstein, The American Journal of Medicine (2010) 123, 40-46
Quote from a surgeon beginning an EHR …our productivity is down 28% I am the highest paid transcriptionist in the state Each cancer patient chart takes me apprx 1 hour For the first time in my career, I turned down an add- on patient
EHR Generic Interface Mostly Filing Cabinet Or Document Management System Small Database Meds Allergies Minimal Generic CDS
EHR Database Breast MedOnc Interface Breast Surgery Interface Mammography Interface Pathology Interface RT Interface Specialty Specific CDS Specialty Specific CDS Specialty Specific CDS Specialty Specific CDS Specialty Specific CDS
Current EHRFuture EHR Decrease productivity or neutralIncrease productivity Mostly document repositoryDatabase Mostly free textStructured data Data entered by staff or providerData entered by patient, staff or provider Generic interfaceSpecialty specific interfaces Rudimentary CDS/Drug-Drug interactions Effective CDS for multiple specialties View isolated transactionsView consolidated information about a given problem Proprietary hidden informationOpen access to patient data Monolithic, barely intraoperableInteroperable with multiple ‘best of breed’ systems
Clinical Decision Support (CDS) Apply Algorithms/Guidelines to patient data Identify best course of action Results displayed as intuitive Visualizations BRCAPRO Mutation Risk 25% Suggest Genetic Testing Facilitates best action as part of workflow
CDS Genetic Testing and Risk
Click open 4 screens BRCA1+ Genetic Testing and Risk
Programs exist that can fill this gap HughesRiskApps Progeny Jameslink CancerGene BRCAPRO Current EHR Future EHR Monolithic Interoperable with multiple ‘Specialty Specific’ systems
Current EHR Future EHR Monolithic Interoperable with multiple ‘Specialty Specific’ systems
Current EHR Future EHR Monolithic Interoperable with multiple ‘Specialty Specific’ systems
Current EHR Future EHR Monolithic Interoperable with multiple ‘Specialty Specific’ systems
Current EHR Future EHR Mostly free text Structured data Using free text for CDS will require Natural Language Processing and is not trivial
Natural Language Processing of breast pathology reports DiagnosisPhrases Invasive Ductal Carcinoma124 Invasive Lobular Carcinoma95 Ductal Carcinoma in situ52 Lobular Carcinoma in situ53 Atypical Lobular Hyperplasia17 Atypical Ductal Hyperplasia14
Natural Language Processing of breast pathology reports DiagnosisPhrasesNegateTotal Patterns Invasive Ductal Carcinoma Invasive Lobular Carcinoma Ductal Carcinoma in situ Lobular Carcinoma in situ Atypical Lobular Hyperplasia Atypical Ductal Hyperplasia negations before diagnosis ( No evidence of …) 12 negations after diagnosis ( … was not seen)
Current EHR Future EHR Mostly free text Structured data Specialty Specific Systems ask for specialty specific data and in return provide specialty specific CDS Providers will enter Structured Data if they get value in return or if it saves them time
EHR: Paper + extra work + memory Patient completes paper form Reviews data using memory of guidelines Staff enters data into the EHR Current EHR Future EHR Data entered by staff or provider Data entered by patient, staff or provider Documents and Orders
Reviews Report & Pedigree Reviews suggested management Documents and Orders Patient educational materials Clinical Decision Support Patient enters data Tablet PC iPad Website Clinical Decision Support Current EHR Future EHR Data entered by staff or provider Data entered by patient, staff or provider
Breast Cancer Patient Print 3 Op Reports and 3 Path Reports Accession Number: Report Status: Updated Type: Surgical Pathology Pathology Report: CASE: PATIENT: Date Taken: 12/11/2008 Source Care Unit: Path Subspecialty Service: Results To: Signed Out by. ********** Addended Report *********** Accession Number: Report Status: Updated Type: Surgical Pathology Pathology Report: CASE: PATIENT: Date Taken: 12/11/2008 Source Care Unit: Path Subspecialty Service: Results To: Signed Out by. ********** Addended Report *********** CLINICAL DATA: Accession Number: Report Status: Updated Type: Surgical Pathology Pathology Report: CASE: PATIENT: Date Taken: 12/11/2008 Source Care Unit: Path Subspecialty Service: Results To: Current EHR Future EHR View isolated transactions View consolidated information about a given problem
Current EHR Future EHR View isolated transactions View consolidated information about a given problem
EHR Specialty Interfaces/CDS Will allow us to increase quality while decreasing cost Specialty Interfaces/CDS
HughesRiskApps Breast Surgery Module Free software available at: HughesRiskApps.net
Reviews Report & Pedigree Reviews suggested management Documents and Orders Patient educational materials Clinical Decision Support Patient enters data Tablet PC iPad Website Clinical Decision Support EHR
Demo An explanation of HughesRiskApps
HughesRiskApps Breast Surgery Module Free software available at: HughesRiskApps.net