Collaborating to Fund and Advance ResearchThe OHSU Experience Jeffrey A. Gold, M.D. Professor of Medicine December 17th, 2014 Disclosures • Received funding from AHRQ, AAMC/Donaghue Foundation Statement of Clinical Problem • Electronic Health Records (EHRs) are at the center of all data communication, clinical decision making and care implementation in the health care system • Numerous studies document poor use of EHR can lead to increased patient errors and undermine communication • The ICU appears to be a highly vulnerable environment to these issues • Massive amounts of data (>1800 data points/day) • Most physicians feel that current EHR training is inadequate Solution To use simulation with high fidelity, contextually relevant cases to improve EHR utilization and help focus EHR redesign What resources are required? • High fidelity (realistic) cases for training. Need to be created by content experts Cases need to be designed specifically to test EHR usability and safety (the controlled lie) • Cases need to be built into the EHR • Need a cloned version of EHR which looks identical to clinical system but doesn’t have actual patient data • Must maintain user customization • Problem- I am a basic scientist/fellowship director with background in ICU administration Step #1-Reach out to Informatics • Chair of Informatics-Bill Hersh • Introduced me to Fran Biagioli who user EHR simulated cases for Medical Students • Used resources from ARRA funds to help fund EHR educator (Gretchen Scholl) to help build simulated cases, create separate simulation environment and trouble shoot problems Leveraged support from hospital Human Resources to allow for shared position Human Resources controls EHR training Engagement of hospital CMO Barriers to Overcome • Cases need to be created • • Cases need to be imported into EHR • • Manually done-no autoimport Cases must be temporally contextually correct • • Need to test meaningful use of EHR • Recognition of patient safety issues/errors-not just charting They must exist in realtime –cloned forward Can’t use actual production EHR environment • • Impacts billing, pharmacy, meaningful use etc…. Need to maintain user customization and workflow Methods 5 day real life ICU stay created in EPIC simulation environment Cases originally created in EXCEL Every data point created and entered by hand in relative real time (no way to download data into system)-Patient “cloned” forward to day of testing so can be used in real time Case contains clinical decompensation with 15 built- in patient safety issues • Vitals trends, medication misdosing, lack of best practices Methods Cont.. Trainees given written history, relevant clinical info for last 5 days, Bld Cx results and PE • No radiology in sim and residents told not to look for it Trainees given 10 min to gather data in EPIC • Done in ICU to recapitulate effect of environment (lights etc…) Subjects told to present case as if giving daily plan and signout for weekend Residents allowed to use own login for customization Subjects could be tested again, at least 1 week later • Repeat testing with different case-random order Types of Patient Safety Issues Issues of Cognition • Recognizing trends in vitals • Recognize high Pplat Familiarity Issues • Do they even know where things are located Medical knowledge Issues • Do they even know appropriate VT Fragmentation issues • How many screens are used Step #2-See if it works Run 40 housestaff through simulations and document that average clinicians miss >50% of safety items within a case 40 # S u b je c ts • C o rre c t Inc o rre c t 30 20 10 R ecurrent S epsis B P /H R P lat Fever V an c troug h V an co D ose Zosyn D ose D5 G lucose > 200 N eed for In sulin MASS D aily A w akening TV Fluid B alan ce 0 March et al BMJ Open 2013 C linic al C hang e M e d ic atio n F ailure to E rro r A d he re to B e s t p rac tic e Step #2-See if it works • Run 40 housestaff through simulations and document that average clinicians miss >50% of safety items within a case • Data used to obtain R18 from AHRQ for using simulation to improve patient safety • Funds from this grant allow for creation of additional cases to document training effect • Allow to increase N>150 Trainees Fail to Recognize Patient Safety Issues Stephenson et al BMC Med Ed 2014 Can Simulation Improve Performance? B. C. Case 1 First, then Case 2 Pre Post p<0.0001 40 60 40 20 20 0 0 p=0.0003 st 0 60 Po 20 80 % Correct 40 80 Po st 60 100 Pr e % Correct % Correct 80 Case 2 First, then 1 100 e All Subjects 100 Pr A. p=0.001 Stephenson et al BMC Med Ed 2014 Next Level of Clinical Problem • In the ICU we round as an Interprofessional team (RN, MD, Pharmacist, RT) • Effective clinical decision making on rounds is dependent upon Everyone accurately retrieving and reporting data Everyone effectively using the EHR Best Practice for ICU Rounds • Interprofessional Rounds, including RN, pharmacy, and RT • Multiple studies document improved cost, improved morbidity and patient satisfaction with interprofessional rounds • Multiple barriers, including information retrieval and EHR Both increase time and decrease communication • Little data in controlled settings to determine whether improved error recognition by the group Swiss cheese or Cheese cloth EHRs Differentially Impact RN and MD Workflow • RNs like EHRs more often than MDs • EHR has more dramatic affects on efficiency for MDs (Poissant) • Only 46% of handoff items overlap in data transmitted during handoff (Collins) • RNs unaware of abnormal vitals in 43% of ward patients (Fuhrman 2012) • 25% of goals stated in rounds are not present in EHR (collins 2009) Solution-ICU Round Simulation • Obtained funding from the AAMC/Donaghue foundation to allow for entire Interprofessional team to participate in simulation activities. • Engaged hospital Nursing, Pharmacy and Medical leadership to facilitating testing of all groups. • Have now tested all of ICU RNs and Pharmacists on same case as MD. • Beginning full ICU rounds simulation this winter. • Project is now an Incubator Project for National Center for Inteprofessional Education and Practice Allowed leveraging of resources from OHSU IPE committee including nursing and pharmacy Creation of High-Fidelity EHR Simulation for all members of IP Team. • Engaged RN and pharmacy leaders to help in scenario and case design • Initial simulation cases modified to ensure contain all relevant information for all professions (almost an extra 400 data pts/day for RNs) • Needed to frame scenario for workflow • Pharmacists review chart during mock presentation • RNs get a mock signout from another RN Proclivity for Different Professions in Identifying Safety Issues Fluid Balance TV Recurrent Sepsis 100 90 80 70 Vanc trough Vanco Dose 60 50 40 30 Plat Zosyn Dose 20 10 0 Pharm Housetsaff RN Daily Awakening Fever maas BP/HR D5 Glucose Insulin Differential Patterns in EHR Utilization Amongst Professional Groups Total of 135 Different Screens Used Overlap in Screen Utilization Housetaff 3 1 3 3 Pharmacist 4 0 RN 6 What Are Differences in Workflow? 1 135 100 2 90 128 3 80 70 113 5 60 50 40 111 6 30 20 10 110 Pharmacist 9 0 Housestaff RN 85 10 82 14 36 17 23 19 20 Next Steps • Leveraging these data for redesign of ICU rounds • Focusing on data veracity and patient safety • OHSU created an ICU Change Management group focused on standardization of EHR utilization and redesign. • Allows dissemination across all ICUs • RO1 submitted to AHRQ. Our Team Now Division of Pulmonary Critical Care • Jeffrey A. Gold, PCCM/CCM Program Director Department of Medical Informatics • Vishnu Mohan, Bill Hersh School of Medicine • Gretchen Scholl School of Nursing • Judith Baggs School of Pharmacy • Dave Bearden OHSU Hospital • Jesse Bierman (Pharmacy), Ashley Mulanax (ICU RN), Adrienne McDougal (ICU RN) • OHSU ICU Change Management Group