An Evidence-based Practice Center in an Academic Medical Center Supports Continuing Quality Improvement Joel S. Betesh, MD; Matthew D. Mitchell, PhD; Craig A. Umscheid, MD, MSCE University of Pennsylvania Health System www.uphs.upenn.edu/cep The University of Pennsylvania Health System (UPHS) is a major US academic medical center. UPHS consists of three teaching hospitals, two primary care and specialty networks, a rehabilitative nursing facility, a hospice, and a home health agency. With so many institutions and so many providers, the poten- tial for variations in practice is high. Such variations can adversely affect the quality, safety, and value of care. Establishment of CEP The Center for Evidence-based Practice (CEP) opened in July 2006. Its mission is to gather scientific evidence and apply it to decisions regarding clinical practice, formularies, and purchases, and, in the process, reduce unnecessary variations in care and maximize the value of each dollar invested in patient care. The center’s primary financial support comes from the office of the Chief Medical Officer (CMO). Staffing and resources Two co-directors Physicians (internal medicine faculty with hospitalist specialization) Trained in clinical epidemiology Regular patient care responsibilities in our hospitals Three full-time research analysts Doctoral-level training Diverse backgrounds including health policy, nursing, and research Four physician and nurse liaisons Clinical leadership at the hospitals and outpatient practices in our system Bring topics of interest in their institutions to CEP’s attention Help disseminate report findings back to their institutions Two clinical liaison librarians Assist with searching for and acquiring evidence Teach evidence-based practice methods to library users Consulting partners Biostatistician Health economist with pharmacoeconomics background Administrative assistant Total staffing : 5.5 FTE Annual budget : $750,000 Clients served Academic years 2006 – 2009 More than half of the reports were commissioned by hospital and health system CMOs or purchasing committees. Most clients were from the UPHS flagship hospital. Academic years 2009 – 2012 Clients are from all health system hospitals and many outpatient practices. Clinical departments now represent the most frequent user of CEP reports. New opportunities emerging with the introduction of computerized clinical decision support systems across the health system. Source of request AY 2007-2009 AY 2010-2012 Clinical Department 9 (12%) 31 (30%) CMO 18 (23%) 19 (18%) Purchasing Committee 23 (29%) 8 (8%) Quality/Safety Committee 11 (14%) 15 (15%) Pharmacy Committee 5 (6%) 13 (13%) Administrative Department 2 (3%) 7 (7%) Nursing Administration 4 (5%) 2 (2%) Other 6 (8%) 8 (8%) Total number of reports 78 (100%) 103 (100%) Center for Evidence-based Practice Physicians and Nurses Pharmacy and Therapeutics Committees Supply Chain and Technology Committees Clinical Effectiveness and Quality Improvement Provides local utilization and outcomes data Uses CEP evidence to develop quality programs Clinical partners Request review of drugs, devices, procedures, and processes of care Provide clinical context for evidence reviews UPHS Executive Staff (Chief Medical Officer) Sets priorities for clinical guideline development Uses CEP evidence to develop quality improvement partnerships CEP information products Evidence Reviews: Systematic reviews of published clinical studies on a well-defined topic, with meta-analyses of data when appropriate. These are similar to evidence reports published by other comparative effectiveness organizations. Local utilization and cost data are incorporated so reports are tailored to our medical center’s needs. Evidence Advisories: Concise reports typically based on limited searches of guidelines and systematic reviews. An advisory can often answer a question more efficiently than a full evidence review. Evidence Inventories: Summaries of the quantity and type of published evidence for a particular topic. Inventories offer a snapshot of the research landscape and can help determine the need for a full review. Clinical Practice Guidelines: Developed in collaboration with a task force of key stakeholders and includes practice recommendations. An Evidence Review is part of the process. Sample topics R–Evidence Review A–Evidence Advisory I–Evidence Inventory G-Guideline Process of care •Guidelines for admission to long-term acute care hospitals (A) •One to one nursing care for patient safety (R) •Telemedicine in follow-up of chemotherapy patients (R) •Best practices for preventing aspiration pneumonia (A) •Management of acute lower GI bleeding (G) •Symptom-triggered vs. fixed-schedule treatment in alcohol withdrawal syndrome (R) Device •Antimicrobial sutures for preventing surgical site infections (R) •Portable intermittent compression devices to prevent VTE (A) •Indications for robotic-assisted surgery (I) Drug •Hyperthermic intraperitoneal chemotherapy (A) •Gastrointestinal bleeding risks with celecoxib (R) •Safety and effectiveness of rhBMP-2 for spinal fusion (R) Diagnostic Test •Brief screening tests for depression in ED patients (R) •Use of ultrasound for diagnosis of DVT in asymptomatic patients (G) Other •Physician compensation and medical professionalism (R) •Defining preventable readmission (Special Report) Type of topic AY 2007-2009 AY 2010-2013 Drug 19 (24%) 25 (24%) Device 26 (33%) 20 (19%) Diagnostic 3 (4%) 8 (8%) Process of Care 27 (35%) 46 (45%) Other 3 (4%) 4 (4%) TOTAL 78 (100%) 103 (100%) Using evidence to transform care Example: Transfusion thresholds Evidence Advisory protocol Review of published guidelines Indications for transfusion in adults Evidence Advisory findings Table of recommended indications for transfusion, stratified by three hemoglobin levels Implementation of evidence-based practice Clinical decision support tool presented those indications in the inpatient EMR screen for ordering RBCs Physician must check one of those indications or explain clinical rationale for transfusion Clinical results 15% fewer patients given RBCs 21% reduction in total RBC orders Example: Systems for early detection of sepsis Two CEP evidence reports New technologies for prediction or early detection of sepsis Clinical risk scores for prediction or early detection of sepsis Evidence Advisory findings Low quality, indirect evidence for aPTT waveform analysis No evidence on the predictive ability of biomarker tests Low-quality evidence for clinical risk scores to predict sepsis Moderate evidence for clinical risk scores to detect sepsis No evidence for commercial risk scoring products Implementation of evidence-based practice Clinical prediction rules used to develop automated alert in inpatient EMR Care team including physician text paged and told to gather at the bedside within 30 minutes to evaluate patient Clinical results Automated alert implemented September 2012 Survey of caregivers following the first two weeks of operation 6 of 63 patients with sepsis or other critical illness (10%) were detected first by the automated alert 33 of 103 patients (32%) were transferred to the ICU 41 of 103 survey respondents (40%) thought the alarm had value. Introducing CEP Dissemination of findings All CEP reports are posted to a searchable internal web site avail- able to all health system faculty and staff. Most reports can be made available to outside users on request. We disseminate our findings most often through the CRD Health Technology Assess- ment database, which is searchable through the Cochrane Library. CEP has developed two guidelines published by the Centers for Disease Control and Prevention, and three more are in progress. Other guidelines are submitted to AHRQ’s National Guideline Clearinghouse. Beginning in October 2012, CEP will be participating in the AHRQ Evidence-based Practice program, in partnership with the ECRI Institute. The AHRQ-designated Evidence-based Practice Centers perform systematic reviews to inform major health policy decisions and translate research findings into practice. Education Initiatives Faculty and staff education •“Academic detailing” thru distribution of InfoPOEM and PROVE (Penn Reviews of Value and Effectiveness) e-mails •CME credit for CEP task force participants •Local and national conferences and workshops Fellow education •Direct and teach systematic review and meta-analysis course Resident education •CEP elective on evidence-based medicine •Participate in Clinical Investigator Toolbox and Healthcare Systems Leadership & Quality Improvement Track programs Medical student education •Direct and teach in clinical decision making courses •Small group instructors in epidemiology & health policy courses Scientific contributions •41 articles published in peer-reviewed journals •31 presentations at international scientific conferences •Participation in methods development groups like GRADE UPHS internal web site 181 (100%) Integrated into clinical decision support 35 (19%) Peer-reviewed publications 19 (10%) National Guideline Clearinghouse 5 (3%) Health Technology Assessment database (Searchable through Cochrane Library) 136 (75%) Other means of dissemination 9 (5%) AHRQ EPC program Beginning 2012