2018 FALL ADVISORY BOARD MEETING October 11-12 | Mclean, Virginia

2018 FALL ADVISORY BOARD MEETING October 11-12 | McLean, Virginia

2 CHOT 2018 FALL IAB MEETING AGENDA

Thursday, October 11, 2018 11:00—11:30 Check in

11:30—12:10 Welcome & Meeting Overview Dr. Eva Lee, Site Director, Georgia Institute of Technology Opening Plenary Speaker Dr. Steven Luxenberg, MD, FACP, Associate Director for Health Informatics, Food and Drug Administration

12:10—1:30 CHOT 10 Year Celebratory Lunch 12:20—12:30 Dr. Alexander Schwarzkopf, Founder of Industry/University Collaborative Research Center (I/UCRC) Program Sharing Successful CHOT Projects 12:40—1:05 Dr. William Mahle, MD, Chief of Children’s Healthcare of Atlanta Sibley Heart Center, Marcus Professor of Pediatric Cardiology 1:05—1:30 Dr. Alexander Quarshie, MD, MS, Professor, Dept. of Community Health and Preventive Medicine, Director, Biomedical Informatics Program, and Director, Master of Science in Clinical Research Program

1:30—2:00 State of the Center Dr. Thomas Ferris, CHOT Center Director NSF LIFE Forms Review & NSF Update: Dr. Craig Scott, NSF I/UCRC Evaluator

2:00—2:45 Industry Research Themes/Needs Workshop Dr. Karan Uppal, IAB Chairperson

2:45—3:00 Break

Meeting Venue Location | Georgia Institute of Technology MITRE Facility | 7525 Colshire Drive | McLean, VA 22102

3 CHOT 2018 FALL IAB MEETING AGENDA

Thursday, October 11, 2018 3:00—3:30 Industry Research Themes/Needs Workshop Dr. Karan Uppal, IAB Chairperson

3:30—4:45 CHOT Research Impact and Insight Presentations Kristina Sheridan, MITRE Eva K. Lee, Georgia Institute of Technology Personalized Evidence-Based Approaches for Advancing Healthcare Delivery — Successful Implementation Farzan Sasangohar, Texas A&M University A Systems Approach to Investigating Unnecessary Admissions Ganisher K. Davlyatov, University of Alabama at Birmingham Using Socioeconomic and Sociodemographic Variables in the Electronic Health Record to Predict Risk of Repeated Hospitalizations and Emergency Department Use Sakthi Kumar Arul Prakash, Pennsylvania State University From Pixels to Biomarkers: How Computer Vision and Machine Learning are Transforming Healthcare

4:45—5:00 NSF LIFE Forms

5:00—6:15 Student Posters Presentation

6:15—8:00 Networking Dinner Banquet

Meeting Venue Location | Georgia Institute of Technology MITRE Facility | 7525 Colshire Drive | McLean, VA 22102

4 CHOT 2018 FALL IAB MEETING AGENDA

Friday, October 12, 2018 8:30 Breakfast Begins 8:30—8:45 Day 2 Overview and Poster Award Presentation 8:45—9:15 Education and Engaged Scholarship Workshop

9:15—10:30 RFP Development Guided by Industry Advisory Board Members

10:30—10:45 Break

10:45—11:45 Reporting Out of RFPs and Soliciting IAB Member Engagement in Research Theme Project Proposals

11:45—12:30 IAB Closed Door Meeting For IAB Members Only

11:45—12:30 Student/Faculty Open Door Meeting Jean Paul, Healthcare Fellowship Opportunities, MITRE

12:30—1:00 Lunch

1:00—1:30 IAB Open Door Meeting

1:30—1:50 Completion of NSF Required Forms For IAB Members, Faculty, and Students

1:50—2:00 Debrief and Adjourn

Meeting Venue Location | Georgia Institute of Technology MITRE Facility | 7525 Colshire Drive | McLean, VA 22102

6 2018 FALL INDUSTRY ADVISORY BOARD MEETING October 11-12 | McLean, Virginia Participant List as of October 1, 2018

NATIONAL SCIENCE FOUNDATION Prakash Balan [email protected] National Science Foundation Andre Marshall [email protected] National Science Foundation Craig Scott [email protected] National Science Foundation

INDUSTRY ADVISORY BOARD MEMBERS Jennifer Rock-Klotz [email protected] American Society of Anesthesiologists Matthew Hagen [email protected] Care Coordination Institute Children's Healthcare of Atlanta - Sibley William Mahle [email protected] Heart Center Neil Bhattarai [email protected] Children's National Health System Rebecca Cahill [email protected] Children's National Health System Evan Hochberg [email protected] Children's National Health System Tishelle Ogunfiditimi [email protected] Children's National Health System Rahul Shah [email protected] Children's National Health System Karan Uppal [email protected] Emory/Grady Jamey Gigliotti [email protected] Highmark stephen.houghland@passporthealthpl Stephen Houghland Passport Health Plan an.com Christopher Hall [email protected] Philips Healthcare Jerome Jourquin [email protected] Susan G. Komen Stephanie Reffey [email protected] Susan G. Komen

Jessie Conta [email protected] Seattle Children's Hospital Lab Medicine

Monica Wellner [email protected] Seattle Children's Hospital Lab Medicine

Main Line Health - Center for Population Justin Beaupre [email protected] Health Research at LIMR Alexander Quarshie [email protected] Morehouse School of Medicine

Rod Cantey [email protected] Opelousas General Health System [email protected] Kenneth Cochran Opelousas General Health System m thomas.giannantonio@siemens- Thomas Giannantonio Siemens Healthineers healthineers.com terrance.talbot@siemens- Terrance Talbot Siemens Healthineers healthineers.com Texas A&M University Health Science Steven Brown [email protected] Center

GUESTS

Kimberly Beer [email protected] Christopher and Dana Reeve Foundation

7 2018 FALL INDUSTRY ADVISORY BOARD MEETING October 11-12 | McLean, Virginia Participant List as of October 1, 2018

GUESTS Steven Luxenburg Food and Drug Administration Brandy Kelly Pryor [email protected] Humana Foundation Alexander Schwarzkopf weareasis@comcast.net Formerly National Science Foundation Shyrea Thompson [email protected] The IRIS Collaborative, LLC David Lacks [email protected] The Lacks Family Jeri Lacks Whye [email protected] The Lacks Family Craig Blakely [email protected] University of Louisville Johns Hopkins University Malone Center Greg Hager [email protected] for Engineering in Healthcare Jerry Folsom [email protected] Siemens Building Technologies Larry Gamm [email protected] Texas A&M University

UNIVERSITY FACULTY & RESEARCHERS Lauren Irlinger [email protected] CHOT Robert Weech-Maldonado [email protected] University of Alabama at Birmingham Eva Lee [email protected] Georgia Institute of Technology J'Aime Jennings [email protected] University of Louisville

Christopher Johnson [email protected] University of Louisville

Tiffany Robinson [email protected] University of Louisville Theo Edmonds [email protected] University of Louisville Stephen Timmons [email protected] University of Nottingham Andrea Sillner [email protected] Pennsylvania State University Chris DeFlitch [email protected] Pennsylvania State University Neal Thomas [email protected] Pennsylvania State University Prasenjit Mitra [email protected] Pennsylvania State University Conrad Tucker [email protected] Pennsylvania State University Hui Yang [email protected] Pennsylvania State University Amarnath Banerjee [email protected] Texas A&M University Alva Ferdinand [email protected] Texas A&M University Thomas Ferris [email protected] Texas A&M University Bita Kash [email protected] Texas A&M University Tony McDonald [email protected] Texas A&M University Arjun Rao [email protected] Texas A&M University Farzan Sasangohar [email protected] Texas A&M University Joseph Heim [email protected] University of Washington

STUDENTS Ganisher Davlyatov [email protected] University of Alabama at Birmingham Reena Joseph [email protected] University of Alabama at Birmingham Neeraj Puro [email protected] University of Alabama at Birmingham

8 2018 FALL INDUSTRY ADVISORY BOARD MEETING October 11-12 | McLean, Virginia Participant List as of October 1, 2018

STUDENTS Shaina Bolden [email protected] Georgia Institute of Technology Siawpeng Er [email protected] Georgia Institute of Technology Guanlin Chen [email protected] Georgia Institute of Technology Di Liu [email protected] Georgia Institute of Technology Joshua Morgan [email protected] Georgia Institute of Technology Pavan Thaker [email protected] Georgia Institute of Technology Hairong Wang [email protected] Georgia Institute of Technology Yuanbo Wang [email protected] Georgia Institute of Technology Peijue Zhang [email protected] Georgia Institute of Technology Mohamed Ahmed [email protected] University of Louisville Emmanuel Ezekekwu [email protected] University of Louisville Molly O'Keefe [email protected] University of Louisville Kelsey White [email protected] University of Louisville Fariha Azhar [email protected] Pennsylvania State University Rajeev Bhatt Ambati [email protected] Pennsylvania State University Farhad Imani [email protected] Pennsylvania State University Chonghan Lee [email protected] Pennsylvania State University Sakthi Kumar Arul Prakash [email protected] Pennsylvania State University Dillon Madrigal [email protected] Pennsylvania State University Mallory Peterson [email protected] Pennsylvania State University Pennsylvania State University Health David Morrell [email protected] Milton S. Hershey Medical Center Johnathan McKenzie [email protected] Texas A&M University Grace Ranft-Garcia [email protected] Texas A&M University Larissa Prates Guimaraes [email protected] University of Washington Petroianu

14 NSF CENTER FOR HEALTH ORGANIZATION TRANSFORMATION chotnsf.org

As a National Science Foundation industry-university cooperative research center (I/UCRC), CHOT follows a model of an industry- DFDGHPLFSDUWQHUVKLSWKDWKDVEHQHȴWHGLQGXVWU\IRFXVHGUHVHDUFKDFURVVPRUHWKDQGLVFLSOLQHV2IWKHΖ8&5&VZLWKLQ WKH8QLWHG6WDWHV&+27LVWKHRQO\RQHIRFXVHGRQLQQRYDWLRQVLQKHDOWKFDUHGHOLYHU\&+27UHVHDUFKHUVZRUNDORQJVLGHWKH Industry Advisory Board (IAB) to conduct research that supports the implementation of evidence-based transformational VWUDWHJLHVZLWKLQWKHKHDOWKFDUHVHFWRU&+27FUHDWHVDVDIHPXWXDOO\ EHQHȴFLDO FRRSHUDWLYH HQYLURQPHQW ZKHUH OHDGLQJ KHDOWKFDUHLQGXVWU\PHPEHUVFDQFRPHWRJHWKHUWRFROODERUDWHDQGWRLQQRYDWH

2XUUHVHDUFKPRGHOUHOLHVRQWKHNQRZOHGJHDQGH[SHULHQFHRIKHDOWKFDUHOHDGHUVWRJXLGHDFDGHPLFUHVHDUFK7KLV FRRSHUDWLYH PRGHO HQVXUHV WKDW WKH UHVHDUFK LV ERWK PHDQLQJIXO DQG DSSOLFDEOH WR WKH KHDOWKFDUH LQGXVWU\ DQG SURYLGHV LPPHGLDWHGHFLVLRQVXSSRUW

INDUSTRY ADVISORY BOARD (IAB)

INDUSTRY MEMBERSHIP CHOT UNIVERSITY SITES: = $50,000

Pooled Members $

CORE FUNDS Institutional NSF & Support Funds SUPPLEMENTAL FUNDS RESEARCH PROJECTS

Value $700,000 Created

Innovations in Healthcare Delivery

INVESTIGATE VALIDATE IMPLEMENT 5HVHDUFKLQIRUPHGVWUDWHJLFGHFLVLRQV Innovations and prototypes (YLGHQFHEDVHGLQQRYDWLRQDFURVVVHWWLQJV

CHOT’s research model relies on the knowledge and experience of healthcare leaders to guide academic research to ensure that it is meaningful and applicable to the healthcare industry and provides immediate decision support.

15 CHOTCHOT’S CURRENT’S CURRENT INDUSTRY INDUSTRY MEMBERS MEMBERS

Last Best Chance

16 NSF Center for Health Organization Transformation CHOT University Sites and Members

State of the Center – Fall 2018 October 11 & 12, 2018

Thomas Ferris, CHOT Center Director Bita Kash, CHOT Center Co-Director Karan Uppal, Incoming IAB Chair Lauren Irlinger, Managing Director 17

Welcome IAB Members! CURRENT CHOT INDUSTRY MEMBERS (34 members)

Health Systems/Providers (14) Associations/Government (8) Alacare Home Health and Hospice American Society of Anesthesiologists Care Coordination Institute Central Texas Veterans Health Care System Children’s Healthcare of Atlanta Sibley Center Centre Collaboration for Leadership in Applied Health Research and Care East Midlands Children’s National Health System East Midlands Patient Safety Collaborative Grady Health System Palm Health Foundation Lakeshore Foundation Main Line Health Primary Care Development Centre Morehouse School of Medicine Texas A&M University College of Medicine Opelousas General Health System Susan G Komen Foundation

Penn State Hershey Medical Center Retail/Tech/Vendors (9) Insurers/Payers (3) Seattle Children’s Hospital Primary Care AT&T Highmark Seattle Children’s Hospital Lab Medicine Avizia, Inc. Passport Health Plan UAB Health System GTech Procure York Risk Group University of Louisville Hospital Last Best Chance, LLC Philips Healthcare Quantum Innovation Restore Medical Solutions Sanofi Siemens NSF I/UCRC MODEL & CHOT’S VALUE PROPOSITION CHOT INDUSTRY ADVISORY BOARD (IAB) Industry Advisory Board Industry Members delegate representatives to serve on the CHOT IAB The IAB: Pooled Members •Contributes to the Center’s strategic direction $ •Advises on projects, new university partners, industry members, and project voting and selection Core & •Meets twice per year to conduct CHOT business University Investment of NSF Funds Center Indirects Supplemental •Participates in additional meetings and conference calls during the year if willing and able RESEARCH Funds Sites PROJECTS •Elects Chairperson & Vice-Chairperson for two year terms to represent the IAB as a whole ම Karan Uppal takes the place of IAB Chair Fall 2018 Value Created $

Innovations in Healthcare Delivery 18

ANNUAL RESEARCH CYCLE CHOT CENTER UPDATES

Publications per Year

Student & Faculty Engagement Fall CHOT Winter Spring Summer IAB meeting CHOT IAB meeting

Communicate Coordinate Collaborate Create

Research project Research Present research CHOT sites Presentation proposals proposals conduct research Singular projects vs. Collaborative Projects IAB members developed with IAB provides projects share research IAB input. CHOT feedback & ranks ideas and sites facilitate research proposals develop research collaborative themesme research 2018 – 2019 RESEARCH CYCLE 2018 – 2019 RESEARCH CYCLE

Population Health

Care Coordination Research Themes Projects Budget Population Health 3 $290,000 Care Coordination 2 $142,000 Analytics and Innovative Technologies Analytics & Innovative 2 $200,000 Technology Patient Experience 2 $150,000 Patient Experience Access to Care 2 $240,000

Access to Care 2018-2019 Research Budget $1,022,000 19

2018 – 2019 PROJECT UPDATES 2018 – 2019 PROJECT UPDATES

Population Health Projects Budget Spent ComprehensiveCompre Analysis on Impact of Social Determinants to $15,000 Patient Experience Projects Budget Spent ImproveImprov Care Across Populations CareCare Coordination Activities for Individuals with Spinal Cord $10,000 ParticipatingParticip in a Community Health Improvement Network $13,916 InjuryInjyju EmbeddingEmbedding RoutineRo Informal, Family Caregiver Assessment of $20,000 TheThe EffectivenessEff of Substance Abuse Treatment Services in $20,000 Delirium SupeSuperimposed on Dementia into Acute Care CombatingComba Opioid Crisis

Care Coordination Projects Budget Spent CareCare Coordination and Patient Experience across the $10,000 Access to Care Projects Budget Spent ContinuumCon of Care: A Value Based Reimbursement PerspectivePerspective TelemedicineTelemedicine in Practice: MultidisciplinaryMulti Utilization of $20,000 TeTelehealthlehealth anandd Remote Patient Monitoring Systems DevelopingDeveloping a RiskRisk PredictionPredictio Model for Hospital Acquired $10,000 Clostridium Difficile Infection An IIntervention to Address Health Literacy Barriers, Increase $10,000 PatPatienti Engagement, and Improve Patient Experience and Outcomes Analytics and Innovative Technologies Budget Spent Projects LeveragingLeveraging TechnologyTechnology to EnhanceEnhance CommunicationCommunication in $30,000 Healthcare Data-drivenData-driven AnalyticsAnalytics andand MachineMachine LearningLearning forfor ImprovingImpro $10,000 Healthcare Outcomes CHOT WEBSITE – chotnsf.org CHOT WEBSITE – chotnsf.org 20

FALL 2018 MEETING ATTENDANCE SPRING IAB MEETING – MARCH/APRIL

Total Registered Attendees as of 10.01.18 Hosted by Siemens & Pennsylvania State University Total Attendees: 88

Type Number of Attendees University Site Administrator 6 Faculty / Researcher 13 NSF Representative 1 IAB Member 21 NSF/IUCRC Fall Meeting Expectations • 1.5 days minimum Guests 18 • Current project report updates Graduate & Undergraduate Student 29 • LIFE Forms • Social event • Closed door IAB business meeting “If you want to go FAST, go alone.

If you want to go FAR, go together.” - African Proverb 21

22 RESEARCH IMPACTS & INSIGHTS

Personalized Evidence-Based Approaches for Advancing Healthcare Delivery — Successful Implementation Eva K. Lee, Georgia Institute of Technology

A Systems Approach to Investigating Unnecessary Admissions Farzan Sasangohar, Texas A&M University

Using Socioeconomic and Sociodemographic Variables in the Electronic Health Record to Predict Risk of Repeated Hospitalizations and Emergency Department Use Ganisher K. Davlyatov, University of Alabama at Birmingham

From Pixels to Biomarkers: How Computer Vision and Machine Learning are Transforming Healthcare Sakthi Kumar Arul Prakash, Pennsylvania State University

24 Grady + 40 hospitals: Transforming ED Personalized Evidence-based 1. Patients 2. Data / 3. Machine 4. Simulation- Approaches for Advancing providers Process / NLP / Learning optimization Healthcare Delivery hospital mining

Eva K Lee, R Zhou, S Lahlou, A Widmaier, J Phillips, C Allen, F Yuan, Y Cao, J Lee GEORGIA INSTITUTE OF TECHNOLOGY

Observe IAB: CHILDREN’S HEALTHCARE OF ATLANTA, GRADY, ? MOREHOUSE, EMORY, NORTHSIDE, CARE COORDINATION INSTITUTE, RESTORE

Predict: 1) Disease, treatment characteristic 2) Readmission status

CHOT FALL 2018 INDUSTRY ADVISORY BOARD MEETING 25

Quality, Timeliness, Efficiency, Personalized Treatment Effectiveness, Safety, Cost, Saving Lives ª Improve Timeliness & Efficiency Improve Quality, Reduce Mortality & Disability Transforming Cancer Care Incorporating Biological Tumor Control º « n Annual Reduction Improve outcome; save lives, reduce « » Average length of stay reduced -33% (-3.35 hrs) mortality, reduce disability; timely diagnosis « )( etS )( tdb » - Trauma % stroke -20% (-90 mins) TCP t «1)( t » and treatment Annual Increase Reduced « )( tdb dt ' » Average wait-time reduced ED Service ¬ )(1 etbS tdb ')( volume ³ )'( etS - 369,295 hours -70% (-3.12 hrs) 0 ¼» Airlift trauma 3,395 417 (+14%) Readmission reduced -28% (-602) 15,992 1665 (+8.4%) Death ~ 56 Left without being seen reduced -30% (-5,553) Trauma patients n = initial number (at time t = 0) of tumor cells Disability ~ 160 Annual Volume S(t) = the survival probability at time t of tumor cells, Comprehensive Care 39,059 2001 (+5.52%) Disability ~ 390 b = birth rates of tumor cells = 0.693/T ED* (treat more patients) +7.8% (+8,114) pot Extended Care 29,645 902 (+2.9%) Disability ~ 296 d = death rates of tumor cells Trauma +8.4% (+1,664) T pot = potential doubling time; tumor cell loss factor = b/d Severe Trauma +14.0% (+417) Minimize max t ¦¦ 20 Impacts & Significance i Blue/Red/Pace: -50% Outcome-driven personalized approach Length of Stay Quality of care: $168.85 mil Maximize \ PP EDvv PP 15 Trauma: -18% P BTVPTV P OAR Reduce LOS (-30%), wait-time (-70%), LWBS (-30%) Maximize TCP 10 Detention: -36% ¦ Reduce readmission (-28%) $7.7 mil penalties Subject to 5 , Original p Timeliness of care: saving lives (trauma/blue patients) iP tD i PrDose p Ly , P)1(07.0 t PTV \ BTV 0 ¦i Blue Red Trauma Detention PACe Walk-in Now Efficiency and effectiveness: Zone Zone ,iP - pPi )1( d UvNtD p, P OARs, PTV \ BTV 25.0% i 2010 72-hour Increase ED throughput (+19%) $96.6 mil ¦ return BTV || y P t PTV ||95.0 20.0% 2011 72-hour Reduce/redirect non-urgent patients (-32%) $21.6 mil ¦ P \ BTVPTV return tD escDose, 15.0% New business for alternative care $12.45 mil ,iP i t P BTV 2012 72-hour i return Expand trauma care 10.0% 2013 72-hour d xTt iii % revisits of % return Increased throughput (+14%) $30.5 mil ¦ x i d Maxseeds 5.0% 2010 30-day i return 90 minute reduction in treatment time (saving lives, +500) p ,, t},{xvy t 010, 0.0% 2011 30-day p i i 12345Acuity Level return Sustained improvement Pediatric Congenital Heart Diseases: A Novel Challenges and Impact Analytic Collaborative Challenges Impact & Broad Dissemination OR-analytic computer engine Objectives are complex, competing Treatment outcome & quality-of-life - Practice variance & inter- dependencies Biological and clinical objectives are Cost savings ($1/2 billion on prostate - Uncover patterns of outcome not readily available nor easily cancer alone) - Machine learning to pinpoint key incorporated Quality control and training factors - Influence network to prioritize Establish Computational advances are needed Drastic reduction in side-effect - System modeling to gauge change Assess Impact Consensus CPG Elements of uncertainty is high Licensed and became national (and impact - Quality and Multi-source data Impleme care Implementation High translational value… but must be international) standards nt -EMR, lab/images strategies -Outcome patient and persistent…. - Personnel, changed -Cost resources Data reporting practice - Compliance Reduce100 Side Effects Improve Tumor Control Probability -Observation guidelines 80 Pre -Planne d TPI -Time-motion study Monitoring Intraoperative 3D TPI PET-image/PTV Ratio Category Small Medium Large - Interviews Planning Target Volume (PTV in cc) 82.80 137.47 92.47 Collaborative learning protocol 60 PET-identified volume (BTV in cc) 8.60 27.63 29.74 - Multi-team site visits Active & Control Ratio: BTV/PTV 10.4% 20.1% 32.2% 40 - Deep learning groups Treatment Planning Model Tumor Control Probability - Develop consensus Standard Clinical Plan 0.48 0.53 0.63 20 - Establish Clinical Practice PET-guided Dose Escalated Personalized Plan 0.99 0.91 0.95 Persistent Grade 2 GU symptoms GU 2 Grade Persistent Guideline 0 0 1224364860728496 - Determine suitability Months

6 26

Postoperative Clinical Outcomes Contact Information Exceeded Expectation! Percentage changed -10.00% -30.00% -50.00% -70.00% -90.00% Eva K Lee, [email protected] Time to extubation (hr) -78.77% Time to continuous… -55.73% Posters Time to first introduction of… -37.46% Cumulative opioid (mg/kg) -37.50% Challenges in Telemedicine - a Systematic Review and Cumulative dose of… -46.30% Engagement with Rural Communities Cumulative benzodiazepine… -33.33% Postoperative ICU LOS (days) -27.59% Machine Learning for Evidence-Based Practice, Risk Postoperative LOS (days) -16.67% Prediction, and Optimal Care Coordination

Mahle et al. 2016, Lee, et al 2018

7 Background Readmissions Problem

• Estimated total hospital costs at $44 billion per year for re- Systems Approach to Investigate hospitalizations within 30 days of hospital discharge (Jencks, 2010) Unnecessary Admissions • Factors which might lead to readmissions: • Medicaid and Medicare, black, Hispanic, divorced or “other” marital RESEARCHERS: status, presenting with major disease severity, or 34 years of age or Farzan Sasangohar, PhD older Bita Kash, PhD, MBA • Large variability between EDs

IAB MEMBERS: •Need: holistic methods that uncover context-dependent factors Texas A&M University contributing to the admission decision-making

CHOT FALL 2018 INDUSTRY ADVISORY BOARD MEETING 2 27

Background Project Timeline Methodology: System Modeling

Research Objectives: To understand systemic contributors to unnecessary admissions at MLH EDs. • Iterative modeling using SysML (Systems Modeling Language) to: • Phase 1 (Quantitative Research) 1. Identify the levels of interest within the MLH organizational • Literature review hierarchy accounting for depth and breadth, • Data analysis of admissions for determinants of readmissions in ED 2. Identify the relevant stakeholders within each level, • Phase 2 (Qualitative Research) 3. Provide insights about which relevant stakeholder perspectives have • Interviews: March 2018 (sharp-end) been collected and which potential stakeholder perspectives must • Interviews & Focus group: July 2018 (blunt-end) be collected in the next stage of the project, and 4. Facilitate communication of the current status of the data collection Secondary process between the different stakeholders involved in this research Literature Focus data Interviews endeavor. review groups analysis

3 4 Main Line Health (MLH) Stakeholders Interviewed: Senior Leadership President and CEO Participant 1

Executive Vice Participant 2 President and Chief Methodology: Qualitative Data Analysis Financial Officer Participant 3

Senior Vice Senior Vice Senior Senior Vice Senior Vice Participant 4 Senior Vice President President, Vice President President Chief President, and Facilities President, and Chief and Chief Medical Participant 5 (Strauss & Corbin, 1998) Developme Understanding inductive and deductive results from the enquiry General Design and Human Information Nursing Officer nt Counsel Construction Resources Officer Officer Participant 6

President, Potential President, President, Main Line President, Segment-Direct Lankenau Mirmont Health HomeCare & Lankenau Interviewees quote Institute of Treatment Center Hospice Medical Center Medical Research for second President, President, President, President, President, round Bryn Mawr Bryn Mawr Riddle Main Line Paoli Hospital Hospital Rehab Hospital Hospital HealthCare DVACO Code-Initial code Oversees day-to-day operations, strategic planning, clinical program Manager of development, physician recruitment, fundraising and financial operations. Care sub-code-focus Coordination coding Hospital In-patient Ambulatory Regional VP Chief of Emergency Chief Department Chief Department Operations Provider DVACO Chief Nursing Regional System Administrator Administrator Clinical Oversees Oversees VPMA Officer Administrator Oversees Oversees Chief of Emergency DVACO Care Nursing Manager Oversees Department Medical Staff Coordinators Provide direction for In-patient Ambulatory and Educator enhancement of all aspects Functions Physicians Physicians Oversees day-to- of provision of medical care. Oversees day operations

Emergency Department Focused Suppo coding: Interpretation- Nurses rt Physicians Initial coding: Care Coordinators *Oversee identify Mapping: associations, Expectations of the Data collection and Identifying a patient care that must relationship search for concepts, condition transcription thematic Perform receive. between patterns explanation *Administer Provide information of patient framework Social Case medications initial status and severity. different for the data Workers Managers *Perform themes healthcare Triage procedures • Software: MAXQDA 12 • Double coders, high intercoder reliability

*Clinical/physiological status *Coordinates assured follow-up *Evaluates patient condition. *First contact with the ED. care after discharge. *Provide healthcare and treatment *Manage challenges and *Make patient admission decisions difficulties to healthcare access. 6 Patients 5 28

Work System

Technology and Tools Environment

1. Meaningful metrics and tools to categorize avoidable admissions are Processes Outcomes Methodology: SEIPS Modeling unavailable. 1. Space constraints in the ED. 2. Standardized guidelines and protocols are limited. 3. Need decision-support tools to implement #2 above. Patient-level • Healthcare Work System Processes Outcomes • Patient admission Person decision-making • Quality of ER care • Patient follow-up • Follow-up care Provider Patient • Home-nursing care • Satisfaction with • Unscheduled care clinical encounter • Duration of stay Technology and 1. Demographics - Age Environment 1. Risk-averse admission decisions to avoid litigation 2. Demographics – Social Status Tools processes. 3. Gaps in understanding of condition. 2. Resistance to standardized guidelines. 4. Cultural expectations of the community. Employee and Patient Outcomes: 3. Pressure to meet throughput goals. 5. Not having a PCP and relying on ED care. Organizational 4. Inadequate understanding of the Home Nursing 6. Patient preference/push to be admitted. -quality of care capabilities. 7. Conditions with higher readmission rate. Outcomes 5. Physicians being pulled at different ends. -patient safety 6. Depending on the circumstances, physicians opt for the path of least resistance. Management/Leaders Processes: 7. Lack of a health system perspective. *care process 1. No shared understanding of the problem. Person • Short- and long-term *other processes financial performance Organization Employee and • Cultural change • Meet staffing or capacity organizational Care Coordination Information Management Resources Allocation requirements outcomes • Legal consequences 1. Lack of resources available in the ED to setup follow-up 1. Lack of care coordination with local clinics and PCPs care (social workers and case managers). • Hospital bed turnover in the area. 2. Uneven assignment of resources within the ED’s. 2. A non-standardized approach to care delivery among 1. Not providing information/feedback to the • ER utilization the hospitals. physicians about the status of the unnecessary 3. Lack of care coordination between the ACO and the admissions and readmissions. -missing info • Providers pressure Organization 2. Potential lack of information in the emergency Tasks hospitals. Infrastructure 4. Inadequate/Lack of involvement of the PCPs in the room. unnecessary admissions assessment process. 3. Insufficient/Lack of information sharing between 5. Lack of 24-hours care coordinators. ACO and MLH. 6. Care managers time is not well used. 4. Lack of data collection about unnecessary 7. Lack of timely and effectual care management. admissions and readmissions at a granular level. - missing info 5. Lack of a health information exchange pipe with 1. Lack of well-designed unscheduled care vehicles. PCPs and specialists. 2. Inadequate/lack of infrastructure for patient education Culture 6. Inefficient communication between Retirement and advocacy. Centers, the Emergency Department, and HMS. 3. Lack of ACO involvement in providing/educating about other alternatives. 4. Absence of robust community behavioral health services. 5. Lack of efforts on identifying and addressing social 1. ED physicians are independent practice oriented, determinants of health. aligned to their campuses and programs. 6. Lack of assured and timely follow-up for the patient. 2. Incentives to keep hospital beds full to avoid loss of Laws and Regulations 7. Unavailability of alternative pathways for healthcare, revenue. other than the ED’s. 3. Difficulties in creating a home-preferred mindset. 8. Inadequate/lack of outpatient support home care systems. 4. Medicare lack of control on the approval of patients’ 9. Need for better mechanisms to measure total costs and 1. Lack of federal regulations and penalties to admissions to hospitals. patients outcomes over long periods of time. control unnecessary admission rates. 5. Litigious environment. 10. Higher admission rates at nigh due to lack of resources at 2. Lack of legal protection to physicians’ decisions 6. No testing and test reading over the weekend. that time of the day. made by applying guidelines and standard protocols 7. Social admissions. 11. Need for more involvement from Utilization Management for admission. 8. Time-based decisions. (UM) 3. CMS 72-hours admission rule. 9. Complex decision-making and bureaucracy. 4. Need of a transition from a fee-for-service to a pay-for-value incentive system. 7 8 Findings: Care Coordination Summary Summary of Findings

• Lack of patient education regarding behavioral health Qualitative SEIPS Element Analysis Evidence determinants Lack of care coordination with local clinics and PCPs in 2 the area “… each campus do it their • Approach to admissions within MLH based on cultural issues own way, they figure out what A non-standardized approach to care delivery among the 6 hospitals -standardization works for them. They all suffer from a lack of the • Environment and locality-based issues (litigation, dominant socio- Lack of care coordination between the ACO and the 1 mindset of standard economic status of the localities) hospitals pathways” Inadequate/Lack of involvement of the PCPs in the 1 • Standardized practice (campuses operate in silos, lack of unnecessary admissions assessment process “You need a pitcher and a catcher. We’re missing a guidelines, lack of personnel for efficient post-discharge care) catcher.” Lack of resources e.g., 24-hour care coordinators 1 • Lack of care coordination resources Care managers time is not utilized well 1 • Low patient engagement Lack of timely and effectual care management 2

9 10 29

Summary Conclusion Next steps: connections • Issues mainly pertain to lack of resources, approach to admissions and culture, missing information to make informed decisions, etc. between • Important recommendations based on this analysis include: issues and solutions 1. standardization of care across campuses; starting with physician, nurse and patient education 2. pro-active approach to admissions, creation of transition care and an ambulatory catch system 3. creation of strong communication networks between physicians and care coordinators 4. creation of clinical accountability units to measure success and recommend improvements to the system

11 12 Contact Information

Dr. Farzan Sasangohar [email protected] http://acelab.tamu.edu 30 Using Socioeconomic and Sociodemographic Problem Background variables in the Electronic Health Record to The cycle of crisis care – repeated ED visits and Predict Risk of Repeated Hospitalizations and hospitalizations for preventable conditions Emergency Department Use (ambulatory sensitive conditions)

RESEARCHERS: 33 out of 67 counties in Alabama have avoidable Allyson Hall, PhD, Sue Feldman, RN, MEd, PhD, Tapan Mehta, PhD, hospital stays in the 3rd or 4th quartile. MSEE, Ganisher Davlyatov, PhD Alabama ranks in the 4th quartile overall in terms of IAB MEMBERS: avoidable hospital use and associated cost Lakeshore Foundation & UAB Health System

CHOT UNIVERSITY SITES: University of Alabama at Birmingham

CHOT FALL 2018 INDUSTRY ADVISORY BOARD MEETING 31

Description Methodology Socio-demographic factors are associated with the Patient-level data in the UAB Enterprise Data Warehouse was health and well-being of individuals (Healthy People used to analyze the likelihood of readmission for ambulatory care 2020) sensitive conditions (N=29,530, visit=50,873) The purpose of this project is to use social Random effects logistic regression was performed to determine determinants of health variables to assess the the effects of socio-demographic and community deprivation on likelihood of a 30, 60, and 90 day repeat risk of being re-hospitalized following discharge from the hospital hospitalizations; to develop and test a predictive model for preventable readmissions 30,000 25,000 20,000 The project will help hospitals in developing targeted 15,000 interventions aimed at helping the patients avoid 10,000 5,000 preventable readmissions patients Number of 0 123456789 Number of visits Research Findings Random effects logistic regression (OR) Next Steps 30-day 60-day 90-day Variables readmission readmission readmission Gender (ref=Male) Female 0.8974*** 0.9058*** 0.9061*** Obtain additional variables Race (ref=White) Black 0.8798*** 0.8845*** 0.8775*** Hispanic 1.2427 1.1546 1.0678 Perform time to event analysis Discharge disposition (ref=Home) Home health 1.2740*** 1.2331*** 1.1822*** Nursing home 1.2412*** 1.1332** 1.0671 Disseminate the findings I/P rehab 18.8178*** 13.4842*** 11.2858*** Hospice home 0.3677*** 0.2854*** 0.2642*** Insurance (ref=Private) Self-pay 0.8428** 0.8017*** 0.7669*** Medicare 0.9043** 0.9398* 0.9397* Medicaid 0.9887 1.0095 1.0040 Length of stay 1.0021 1.0025** 1.0029** Number of comorbidities 1.2304*** 1.2951*** 1.3394*** Pain score 1.0171*** 1.0149*** 1.0124** Area deprivation index (ref=Low) Medium 0.9759 0.9796 0.9827 High 0.9427 0.9614 0.9754 32

Contact Information

Ganisher K. Davlyatov, Ph.D., M.S.

University of Alabama at Birmingham

Phone: 205-975-8701

Email: [email protected] Description From Pixels to Biomarkers: How Computer Vision and Machine Learning Objective: are Transforming Healthcare • Develop a non-contact, video based heart rate estimation algorithm. RESEARCHERS: Partners / Relevance: Conrad Tucker, PhD; Christopher Deflitch, MD; Nilam Ram, PhD; Sakthi Kumar Arul Prakash • Remote and affordable health care diagnosis, and non contact biometric IAB MEMBERS: Siemens, Highmark, Hershey Medical Center, AT&T sensing.

CHOT FALL 2018 INDUSTRY ADVISORY BOARD MEETING 33

Academia to Industry Problem Background What is the problem your research project is trying to address? Reliable heart rate estimation even when lighting and skin tone changes

Reliable heart rate estimation even when someone is moving Approach Experimental Approach Experimental Approach

• As blood absorbs incident light, camera captures pixel variations from skin regions.

Time-Frequency Transformation

ECG Monitor displaying pulse rate and other vitals

Amplification of pixel intensity changes from face

Approach Experimental Approach Experimental Approach • Kalman filter coupled with a bounded extrapolator was used to track the head movements. This enabled the algorithm to achieve a high signal to noise ratio in the presence of motion artifacts.

Pixels to Pulse rate

34 Research Findings Next Steps • Benchmark results show the performance of the method We were recently awarded $100K in funding to expand the capabilities of our approach globally.

The algorithm shows no statistically significant difference when compared with existing contact based FDA approved devices. 35

36 RESEARCH UPDATES

POPULATION HEALTH 01-06181.UAB-TAM-UW Comprehensive Analysis on Impact of Social Determinants to Improve Care Across Populations 02-06181.UL Participating in a Community Health Improvement Network (CHIN) 03-06181.PSU The Effectiveness of Substance Abuse Treatment Services (SATS) in Combating Opioid Crisis

CARE COORDINATION 04-06181.FAU-TAM-UAB Care Coordination & Patient Experience Across the Care Continuum: A Value-Based Reimbursement Perspective 05-06181.UAB Developing a Risk Prediction Model for Hospital Acquired Clostridium Difficile Infection (CDI)

ANALYTICS & INNOVATIVE TECHNOLOGY 06-06181.GIT Leveraging Technology to Enhance Communication in Healthcare 07-06181.GIT-PSU-UW Data-Driven Analytics and Machine Learning for Improving Health Outcomes

PATIENT EXPERIENCE 08-06181.UAB Care Coordination Activities for Individuals with Spinal Cord Injury (SCI) 09-06181.FAU-PSU Embedding Routine Informal, Family Caregiver Assessment of Delirium Superimposed on Dementia into Acute Care

ACCESS TO CARE 10-06181.GIT-UW-PSU Telemedicine in Primary Care & in Management of Chronic -UAB-UL-FAU Conditions: Exploring Patient & Provider Perspectives 11-06181.FAU-PSU Ask Me 3®: A Home Health Intervention to Address Health Literacy Barriers, Increase Patient Engagement, and Improve Patient Experience & Outcomes

37 Theme NSF Number Project Name Principal Investigators Universities IAB Members

Borkowski Comprehensive Analysis on Impact of UAB Population Health Christina Mastrangelo Central Texas Veterans Health Care System, 01-06181.UAB-TAM-UW Social Determinants to Improve Care TAM (Pop 1) Tom Ferris Seattle Children’s Hospital Across Populations UW Bita Kash Population Health Participating in a Community Health Passport Health Plan, Sanofi, University of 02-06181.UL Judah Thornewill UL (Pop 2) Improvement Network (CHIN) Louisville Hospital The Effectiveness of Substance Abuse Population Health Siemens, Hershey Medical Center, Highmark, 03-06181.PSU Treatment Services (SATS) in Combating Hui Zhao PSU (Pop 3) AT&T Opioid Crisis Ankur Agarwal FAU Care Coordination Care Coordination & Patient Experience Texas A&M University College of Medicine, 04-06181.FAU-TAM-UAB Tom Ferris TAM (Care 1) Across the Continuum of Care UAB Health System Robert Weech-Maldonado UAB Developing a Risk Prediction Model for Care Coordination Midge Ray 05-06181.UAB Hospital Acquired Clostridium Difficile UAB Opelousas General Health System (Care 2) Ferhat Zengul Infection (CDI) Analytics & Innovation Grady Health System, Children’s Healthcare of Leveraging Technology to Enhance Technology 06-06181.GIT Eva Lee GIT Atlanta, Morehouse School of Medicine, Communication in Healthcare (Tech 1) Restore Medical Solutions Analytics & Innovation Data-Driven Analytics and Machine Eva Lee GIT Philips Healthcare, Susan G Komen 38 Technology 07-06181.GIT-PSU-UW Learning for Improving Healthcare Christina Mastrangelo PSU Foundation (Tech 2) Outcomes Conrad Tucker UW Patient Experience Care Coordination Activities for Individuals Tapan Mehta 08-06181.UAB UAB Lakeshore Foundation, UAB Health System (Patient 1) with Spinal Cord Injury (SCI) Allyson Hall Embedding Routine Informal, Family Patient Experience Caregiver Assessment of Delirium FAU 09-06181.FAU-PSU Andrea Sillner Hershey Medical Center, Highmark (Patient 2) Superimposed on Dementia into Acute PSU Care Eva Lee GIT Cynthia LeRouge Telemedicine in Primary Care & in UW Access to Care 10-06181.GIT-UW-PSU- Conrad Tucker Avizia, University of Louisville Hospital, UAB Management of Chronic Conditions: PSU (Access 1) UAB-UL-FAU Nancy Borkowski Health System Exploring Patient & Provider Perspectives UAB Christopher Johnson UL Ankur Agarwal Ask Me 3®: A Home Health Intervention to Robert Weech-Maldonado Access to Care Address Health Literacy Barriers, Increase 11-06181.UAB Nancy Borkowski UAB Alacare Home & Hospice (Access 2) Patient Engagement, and Improve Patient Justin Lord Experience and Outcomes

39 40 41 42 43 I/UCRC EXECUTIVE SUMMARY | PROJECT UPDATE UPDATED: SEPTEMBER 24, 2018 PROJECT TITLE: Leveraging Technology to Enhance Communication in Healthcare

PROJECT ID:06-06181.GIT PI: Eva Lee

RESEARCH THEME: Analytics & Innovative Technologies BUDGET: $50,000

MULTI-UNIVERSITY PROJECT: No PROJECT START DATE: 9/1/2018 DESCRIPTION: With continuous advancements in technology, care providers have access to more tools than ever to combat breakdowns in communication with referring physicians and to ultimately play a greater role in improved patient care. As care providers are often overwhelmed with heavy workloads, care communication may suffer. For example, radiologists may be hesitant to assume additional responsibilities related to conveying test results and ensuring proper follow-up with patients. Certain symptoms discovered during surgical procedures by surgeons may be conveyed ineffectively to intensivists and bed-side teams. Yet those activities can play an important role in not only carefully interpreting images or making recommendations but also acting as a safe, patient-centered back-up system and ensuring that actionable results are not overlooked. In a similar manner, non-English speaking patients may require enhanced care coordination plans to ensure that they understand the discharged and home care process.

PROJECT OBJECTIVES: SCOPE: Improve communication, compliance, & quality of care through Utilize machine learning, text mining, and deep learning automated machine translation. Initial focus will be the discharge techniques for hospital discharge notes to build an accurate procedure. Design chatbot & virtual messaging to enhance family automatic translation system which will facilitate discharge and engagement & knowledge dissemination (e.g. feeding plan, home care process design, particularly for non-English speaking compliance & awareness of hospital acquired infections). patients.

HOW THIS IS DIFFERENT THAN RELATED RESEARCH: This study is the first study which utilizes machine learning, text mining, and deep learning techniques for hospital discharge notes to build an accurate automatic translation system which will facilitate discharge and home care process design, particularly for non-English speaking patients. In addition, it incorporates system design and human-device interaction technologies to offer real-time decision support providers.

MILESTONES TARGETED START DATE TARGETED END DATE

Systematic Literature Review 6/1/2018 8/1/2018 Develop a Machine Learning Framework 8/1/2018 11/1/2018 Evaluation & Refinement of System 11/1/2018 2/1/2019 Design & Implementation of Chatbot 2/1/2019 5/31/2019

PERCENT COMPLETED OVERALL: 60% NEXT STEPS: This project focuses on designing a chatbot that serves as a communication tool for patient family members. We have thus far completed a thorough literature review on chatbot state-of-the-art systems. We have begun interviewing providers on frequently asked questions from parents and knowledge of certain procedures that parents need to understand. We have also begun designing the interface of the chatbot. We intend to link this communication system with our previous findings on hospital acquired infections, which will help educate parents and in turn help mitigate potential HAI risks.

BENEFITS TO INDUSTRY: EXPECTED DELIVERABLES: Enhance care coordination plan. Reduce workforce requirements Improve communication, compliance, and quality of care through for translation experts. Improve discharge and home care process. automated machine translation learning. Designed chatbox and Enhance family engagement. Improve patient compliance and virtual messaging to enhance family engagement and knowledge treatment outcomes. Reduce staff time and cost for treatment. dissemination.

44 CHOT CONFIDENTIAL I/UCRC EXECUTIVE SUMMARY | PROJECT UPDATE UPDATED: SEPTEMBER 24, 2018 PROJECT TITLE: Data-Driven Analytics and Machine Learning for Improving Healthcare Outcomes

PROJECT ID:07-06181.GIT-PSU-UW PI: Lee (GT), Mastrangelo (UW), Tucker (PSU)

RESEARCH THEME: Patient Experience BUDGET: $150,000

MULTI-UNIVERSITY PROJECT: Yes PROJECT START DATE: 1/1/2018 DESCRIPTION: Data-driven healthcare has the potential to revolutionize care delivery and trim costs. A major challenge is that providers must sift through and analyze mountains of disparate data to materialize the substantial gain. We continue our healthcare innovation through systems and data analytics. Utilizing EMR and various procedural and personal health data, along with social and behavioral information, we will address all aims - with specific regard to radiologic exam variability. This also has implications in utilizing predictive models to use at the point-of-care when treating infectious disease.

PROJECT OBJECTIVES: SCOPE: Use MRI log file data to identify variability and "wasted" time Increase utilization of large amounts of disparate medical data opportunities and develop predictive models of exam duration, idle including MRI imaging logs, treatment procedures, demographics, time, and repeated scans. Leverage the size & availability of and social and behavioral information. Optimize usage of hospital population health data to model and predict machine utilization resources, treatment process, and outcomes. Address individual efficiency. Apply machine learning to electronic health records. patient conditions and design personalized treatment.

HOW THIS IS DIFFERENT THAN RELATED RESEARCH: This is the first study where 1) large amounts of patient data are extracted unbiased and globally analyzed, 2) automated encryption of PHI and data integration through terminology mapping is achieved using natural language processing, 3) time series clustering is done with consideration of disease progression despite sparse and missing data, 4) discriminatory factors that inform key decisions are systematically selected using machine learning, 5) individual patient conditions are addressed with the design of personalized evidence- based treatment methods, and 6) this research will be able to be replicated to other cases and sites to improve their process.

MILESTONES TARGETED START DATE TARGETED END DATE

Conduct & Benchmark Literature Review 6/1/2018 7/31/2018 Data Collection, Extraction, & Encryption 7/1/2018 11/30/2018 Data Cleaning & Integration Across Records 8/1/2018 12/31/2018 Build Predictive Models 11/1/2018 5/31/2018

Run Pilot Projects & Analyze Findings 1/1/2019 5/31/2019

PERCENT COMPLETED OVERALL: 50% NEXT STEPS: We have extracted about 1.2 million patient data, each includes patient information, billing, laboratory and results, diagnosis codes, resource utilization, etc. The data is de-identified and encrypted to eliminate PHI information. Natural language processing is designed to extract key information effectively to sub-type patient conditions and the illness and resources they use. In particular, data fields related to laboratory usage, medication, and blood transfusion are kept, as well as crucial data containing patients' disease progress and health conditions. We have thus far designed a preliminary predictive model to uncover factors that influence the outcome and resource usage.

BENEFITS TO INDUSTRY: EXPECTED DELIVERABLES: Industry practitioners can make informed decisions and achieve Develop predictive models for KPI's (exam duration, idle time, ratio care that is personalized, timely, evidence-based and appropriate. of repeated scans), using logfile variables. Define better sequence Optimize the usage of hospital resources, treatment processes, and of scans per exam (exam cards). Design pilot project improving outcomes. Reduce waste, risk and cost. This research can be exam cards. Develop models that predict machine utilization replicated to other cases and sites to improve process. efficiency. Design support system and optimize treatment plans. 45 CHOT CONFIDENTIAL I/UCRC EXECUTIVE SUMMARY | PROJECT UPDATE UPDATED: SEPTEMBER 24, 2018 PROJECT TITLE: Care Coordination Activities for Individuals with Spinal Cord Injury (SCI)

PROJECT ID:08-06181.UAB PI: Tapan Mehta, Allyson Hall

RESEARCH THEME: Patient Experience BUDGET: $50,000

MULTI-UNIVERSITY PROJECT: No PROJECT START DATE: 9/1/2018 DESCRIPTION: This is Phase 2 of a project designed to develop and pilot test a care coordination program for people with newly diagnosed spinal cord injury (SCI). Phase 1 activities focused on developing the care coordination program. This phase includes (1) a review of relevant literature and (2) in-depth interviews with patients with SCI, their caregivers, physicians, other healthcare workers who specialize in SCI, and staff at the Lakeshore Foundation. Based on findings from these two activities, a pilot intervention will be developed. Phase 2 will implement the care coordination program developed in Phase 1 and assess the extent to which it improves the quality of life of participants.

PROJECT OBJECTIVES: SCOPE: Develop a care coordination model for individuals newly Phase 2 will be a mixed methods study with the following discharged from an inpatient setting. components: (1) participant self-reported assessments of their quality of life using validated instruments, (2) comparison of hospital use between intervention and control groups, and (3) in- depth interviews with both groups to focus on experiences.

HOW THIS IS DIFFERENT THAN RELATED RESEARCH: Several studies have documented the effectiveness of care coordination/transitions of care activities. Most of these studies focus on the general population and do not address the specific and unique needs of individuals newly diagnosed with SCI. The proposed project aims at addressing the needs of individuals with significant mobility limitations. In addition, the project specifically addresses how a local disability focused community organization can partner with an academic medical center to improve the quality of life of individuals with SCI.

MILESTONES TARGETED START DATE TARGETED END DATE

Hospital Use Data Acquisition & Analysis 9/1/2018 9/30/2018 Interviews with Lakeshore & UABHS 10/1/2018 12/10/2018 Interviews with UAB Rehab Patients 11/1/2018 12/10/2018 Analysis and Thematic Development 12/15/2018 4/1/2019

Final Report on Developed Intervention 5/1/2019 6/30/2019

PERCENT COMPLETED OVERALL: 5% NEXT STEPS: After conversation with our industry partner, UAB Health System, we revised the project guidelines and deliverables to better align with their current needs. We discovered that the ongoing initiative at UABHS focused on transitional care for patients with SCI.

BENEFITS TO INDUSTRY: EXPECTED DELIVERABLES: Evidence of the effectiveness of a care coordination program on Final report documenting study findings and submitted manuscript improving the quality of life of SCI patients. for publication.

46 CHOT CONFIDENTIAL I/UCRC EXECUTIVE SUMMARY | PROJECT UPDATE UPDATED: SEPTEMBER 24, 2018 PROJECT TITLE: Embedding Routine, Informal Family Caregiver Assessment of Delirium Superimposed on Dementia into Acute Care

PROJECT ID:09-06181.FAU-PSU PI: Andrea Sillner

RESEARCH THEME: Patient Experience BUDGET: $100,000

MULTI-UNIVERSITY PROJECT: Yes PROJECT START DATE: 6/1/2018 DESCRIPTION: The purpose of this pilot study is to assess initial accuracy and feasibility of communication of observed symptoms of delirium in older adults with complex multiple chronic conditions dementia by family caregivers utilizing app-based delivery of the Family Confusion Assessment Method (FAM-CAM) in the acute care setting.

PROJECT OBJECTIVES: SCOPE: Explore feasibility and accuracy of using data from family caregivers To assess the initial agreement of communication of observed to establish baseline cognitive function in the acute care setting for symptoms of delirium in older adults with complex multiple older adults with complex multiple chronic conditions and chronic conditions by family caregivers utilizing app-based delivery dementia to increase the likelihood of recognizing symptoms of of FAM-CAM compared to trained observers. To determine delirium. feasibility of embedding FAM-CAM within EMR.

HOW THIS IS DIFFERENT THAN RELATED RESEARCH: Current standards in diagnosing delirium rely on diagnostic criteria of the Diagnostic and Statistical Manual of Mental Disorders, 5th Edition (DSM-5) and the International Statistical Classification of Disease and Related Health Problems, 10th Revision (ICD-10); however, there are no specific diagnostic criteria for delirium in persons with preexisting dementia. Recommended assessment tools for delirium, such as the Confusion Assessment Method (CAM), take into account changes from normal, but often this is unknown to formal healthcare providers. For older adults with complex multiple chronic conditions and cognitive impairment, the person who may be best able to assess baseline cognitive function is the family caregiver.

MILESTONES TARGETED START DATE TARGETED END DATE

Development of FAM-CAM for Acute Care 6/1/2018 11/15/2018 Assessment of Accuracy & Feasibility of Tool 11/15/2018 2/1/2019 Determination of FAM-CAM Sync Within EMR 2/1/2019 4/15/2019

PERCENT COMPLETED OVERALL: 20% NEXT STEPS: The research team is in the process of investigating a software package that is able to capture survey data that is provided by family care givers. This will enable family care providers to complete surveys about their loved ones so that this data can be sued to get the patient family perspective on assessment of delirium superimposed on dementia into acute care.

BENEFITS TO INDUSTRY: EXPECTED DELIVERABLES: Understanding how we can allow informal family caregivers at the Dissemination of findings by publications and presentations to bedside to routinely communicate observed signs and symptoms stakeholders at all levels of care, including, but not limited to of common hospital adverse events to medical staff using app- healthcare providers, patients, informal caregivers, and industry based technology and standardized screening instruments. partners.

47 CHOT CONFIDENTIAL I/UCRC EXECUTIVE SUMMARY | PROJECT UPDATE UPDATED: SEPTEMBER 24, 2018 PROJECT TITLE: Telemedicine in Primary Care & in Management of Chronic Conditions: Exploring Patient and Provider Perspectives

PROJECT ID:10-06181.GIT-UW-PSU-UAB-UL-FAU PI: Lee, Lerouge, Tucker, Borkowski, Johnson, Agarwal

RESEARCH THEME: Access to Care BUDGET: $200,000

MULTI-UNIVERSITY PROJECT: Yes PROJECT START DATE: 6/1/2018 DESCRIPTION: Timely access to quality healthcare service is a real challenge—as outlined in the 2015 IOM report—and misalignment of resources and demands result in long delays for care. Telehealth can offer alternative and timely care to rural area patients who lack sufficient healthcare options. Telehealth can also help to improve health conditions and promote active patient engagement, which is particularly important for chronic disease management. This project identifies drivers and barriers of patient engagement by population groups and chronic conditions and provides recommendations for implementing appropriate telehealth/telemedicine interventions through multiple care settings given governmental policies, reimbursement payments, and delivery of care.

PROJECT OBJECTIVES: SCOPE: Explore & define iterations of telemedicine services and models in Literature review of existing primary care literature. Secondary data primary care for the care of chronic diseases. Identify drivers and analysis of former CHOT landscape project. Create a low-level barriers to patient engagement in telemedicine. Identify provider prototype showcasing the design of a personalized remote patient drivers and barriers to integrating telemedicine. Investigate and monitoring system. design a personalized remote patient monitoring system.

HOW THIS IS DIFFERENT THAN RELATED RESEARCH: The adoption of telemedicine and the level of patient engagement and services provided across healthcare facilities remain uneven and far from optimal. Little evidence, particularly in the form of understanding from the viewpoint and situation of providers, is available to guide stakeholder organizations as they consider introducing telemedicine into primary care practice. This study examines issues including point-of-access, administrative logistics, timely primary care, monitoring of chronic disease and mental health, and providing of equal and affordable care to poor and rural areas. We also investigate and design a personalized remote patient monitoring system to connect patients and providers.

MILESTONES TARGETED START DATE TARGETED END DATE

Perform Systematic Literature Review 6/1/2018 7/31/2018 Identify Gaps in Care Through Gap Analysis 6/1/2018 11/30/2018 Data & System Modeling 8/1/2018 12/31/2018 Develop & Administer Survey Instruments 1/1/2019 4/30/2019

Collect Feedback & Develop/Pilot Prototype 3/1/2019 5/31/2018

PERCENT COMPLETED OVERALL: 50% NEXT STEPS: Though extensive literature review, we identified service gaps across different regions and demographics. We also analyzed the potential return of access and improvement to rural health quality by expanding telemedicine practices to under-served populations, including inner cities and rural regions. Utilizing the rural Georgia data and the disease burden from the CDC, we are designing a community-based model that optimizes point-of-access for this rural population. We are currently exploring what networks are available for small rural providers join and what can be done to facilitate the expansion of telemedicine sites based on our recommended locations.

BENEFITS TO INDUSTRY: EXPECTED DELIVERABLES: Understanding the adoption and diffusion of telemedicine in Technology readiness model for telemedicine in primary care primary care can inform decision making regarding service design, settings. Conference presentations. Piloted technology in Primary implementation, operations, and provider engagement. An Care Assessment Survey. Optimized point-of-access for study sites. assessment tool based on these forces can help assess individual A low cost personalized prototype remote patient monitoring primary care organizational readiness for telemedicine innovation. device. 48 CHOT CONFIDENTIAL I/UCRC EXECUTIVE SUMMARY | PROJECT UPDATE UPDATED: SEPTEMBER 24, 2018 PROJECT TITLE: Intervention to Address Health Literacy Barriers, Increase Patient Engagement, & Improve Patient Experience & Outcomes

PROJECT ID:11-06181.FAU-PSU PI: Weech-Maldonado, Borkowski, Lord

RESEARCH THEME: Access to Care BUDGET: $40,000

MULTI-UNIVERSITY PROJECT: Yes PROJECT START DATE: 6/1/2018 DESCRIPTION: Value-based reimbursement in health care has resulted in an increasing focus on patient engagement as a mechanism to improve post- acute care outcomes, particularly in reducing readmissions. However, health system strategies aimed at increasing patient engagement should account for health literacy and generational differences. Strategies that may work with a high literacy population may not be as effective among a population with low literacy.

PROJECT OBJECTIVES: SCOPE: Pilot test a targeted intervention in a home health environment to Addressing health literacy barriers to care can improve patient reduce health literacy barriers to care and increase patient engagement, which can in turn improve health care outcomes, engagement. such as lower readmissions and improved patient experience.

HOW THIS IS DIFFERENT THAN RELATED RESEARCH: There has been limited research on the effectiveness of health literacy interventions in improving patient engagement and health outcomes, particularly in the home health context. During the current first phase, we are conducting a literature review to identify best practices/strategies in addressing health literacy barriers in a home health environment with the ultimate goals of improving patient engagement and reducing hospital readmissions. We are proposing a second phase, which will consist of a pilot intervention Ask Me 3® in a home health setting. Ask Me 3® is an educational program that encourages patients and families to ask three specific questions of their providers to better understand their health conditions & what they need to do to stay healthy.

MILESTONES TARGETED START DATE TARGETED END DATE

Obtain IRB Approval & Finalize Intervention 6/1/2018 9/1/2018 Collect Data & Implement Intervention 9/1/2018 1/1/2019 Collect Post-Intervention 1/1/2019 3/1/2019 Data Analysis 3/1/2019 5/1/2019

Final Report 5/1/2019 5/31/2019

PERCENT COMPLETED OVERALL: 5% NEXT STEPS: The intervention material has been collected. The next steps involve discussions with industry partners about the outcomes variables of the project.

BENEFITS TO INDUSTRY: EXPECTED DELIVERABLES: This pilot project will provide the foundation for future A final report outlining the findings of the pilot project. interventions of health system strategies to address barriers related to health literacy, increase patient engagement, and improve patient outcomes.

49 CHOT CONFIDENTIAL

50 THANK YOU!

2018 FALL INDUSTRY ADVISORY BOARD MEETING SPONSORS