The Reducing Readmissions Through Improving Care Transitions (RRTICT) Quality Improvement Program Ashley Ketterer Gruszkowski, MHA Health System Specialist Veterans Engineering Resource Center (VERC) - Pittsburgh Co-Authors: Michael W. Kennedy, PhD; Emily Rentschler Drobek, MSPPM; Lior Turgeman, PhD; Aleksandra Milicevic, PhD; Terrence L. Hubert, PhD; Larissa Myaskovsky, PhD; Jerrold May, PhD, Robert Monte, RPh, MBA; Kathryn Sapnas, PhD, RN-BC, CNOR
VA Pilot Sites & Collaborators • • Albany VA Medical Center Office of Strategic Integration | Veterans Engineering Resource • Bay Pines VA Healthcare System Center (OSI|VERC) • Gainesville - Malcom Randall • VA National Program Office of VAMC, NF/SGVHS Patient Care Services (PCS) • Miami VA Healthcare System • Center for Health Equity Research • Omaha VA Medical Center and Promotion (CHERP) • Pittsburgh - VA Pittsburgh • University of Pittsburgh Katz Healthcare System Graduate School of Business • San Juan - VA Caribbean Healthcare System • Sioux Falls - Royal C Johnson Veterans Memorial Medical Center VETERANS HEALTH ADMINISTRATION
Agenda • History of Big Data in the VA • Why readmissions are important • RRTICT program overview VETERANS HEALTH ADMINISTRATION 3
History of Big Data in the Veterans Administration (VA) • VHA is the nation’s largest integrated healthcare system • Serves 9 million Veterans in 1,700 healthcare facilities • Access to > 1 billion Veteran health data points • Ability to connect data analysts with clinical expertise – Improve clinical quality metrics – Prospectively identify patients for care needs VETERANS HEALTH ADMINISTRATION
Corporate Data Warehouse 1,700 VISTA/CPRS = EHR (CDW) hospitals, clinics, CLCs 6.6 billion 9 million lab tests Veterans 2.1 billion 3.8 billion inpatient & clinical outpatient orders encounters
Importance of Reducing Readmissions (Private vs VA) Private Hospitals Veterans Administration • Not driven by economic • Affordable Care Act Hospital Readmissions incentives Reduction Program • Data-driven performance • Economic incentives and measurement penalties from Center for • Reduce burden on Medicare and Medicaid patients, families and Service (CMS) hospital resources – Quality of Care • Increase inpatient access – Outcome Measures • Quality and safety VETERANS HEALTH ADMINISTRATION
30 Day Readmission Rate • 30 DRR defined by CMS – Readmissions after a longer time period may not correlate with care received in the hospital – Financial penalties and reimbursement threshold • VA readmission rate data = CMS definition – Data sources Inpatient Evaluation Center (IPEC) and Strategic Analytics for Improvement and Learning (SAIL) VETERANS HEALTH ADMINISTRATION
Reducing Readmissions Through Improving Care Transitions (RRTICT) Pilot • Goal: Determine if a bundled strategy approach reduces 30DRR • Background: – VERC/Patient Care Services (PCS)/VA facilities – Risk predictive model – decision support tool • Guided clinicians to target those most at risk – Evidence-based best practices in bundled options to reduce readmission • Provider- and system-level strategies to address major patient transition failure points VETERANS HEALTH ADMINISTRATION
Risk-Stratified Predictive Model • VERC University of Pittsburgh • Used at time of admission • Addresses limitation of previous models that focused on prediction at discharge or post-discharge • Allows clinicians to proactively plan for discharge • Tailor specific interventions to patients at greatest risk for readmission VETERANS HEALTH ADMINISTRATION
Model Variables Demographics • Age, sex, marital status, next of kin • # recent ED visits, # recent inpatient Patient History admissions, time since last inpatient discharge, number & type of medications Current • Source (reason) of admission, lab values, Hospitalization systolic / diastolic, pulse, respiration, BUN / creatinine, WBC, albumen Information • Anemia, asthma, chronic renal failure, cerebrovascular accident stroke, COPD, diabetes, Co-morbidities depression, dementia, hypertension, ischemic heart 12/2 VETERANS HEALTH ADMINISTRATION 1/20 disease, obesity, PTSD, peripheral vascular disease 16 10
Risk Prediction Output Example Site Name SSN Admit Date Risk Bay Pines Pt 1 P0001 2/19/2016 Low Bay Pines Pt 2 P0002 2/22/2016 Moderate Bay Pines Pt 3 P0003 2/24/2016 High VETERANS HEALTH ADMINISTRATION 11
Table of Bundled Strategies In-Hospital Strategies Post-Discharge Strategies Appointment Scheduling C-TraC Program Patient Education Hospital In Home Discharge Instructions Home Based Primary Care Discharge Team Telehealth Pharmacist-Led Med-Rec PACT Team Rounding Community Residential Care Daily Huddles PILL Program Discharge Coordinator VETERANS HEALTH ADMINISTRATION
Aligning Model with Strategies Suggested Site Name SSN Admit Date Risk Strategy Bundle Attending Bay Pines Pt 1 P0001 2/19/2016 Low Preference Bay Pines Pt 2 P0002 2/22/2016 Moderate In-Hospital In-Hospital & Bay Pines Pt 3 P0003 2/24/2016 High Post-Discharge VETERANS HEALTH ADMINISTRATION 13
RRTICT Pilot Procedures • 7 Facilities across 3 VISNs volunteered to participate • 6 Month pilot (4/1/2015 – 9/30/2015) • Required to collaborate on monthly virtual learning sessions • Implement RRTICT VETERANS HEALTH ADMINISTRATION
Analyses • Qualitative feedback from all sites • Quantitative comparison of pre/post 30DRR • Data from VA’s IPEC all -cause 30DRR – Percent of patient discharges with readmission within 30 days – Derived from CMS definition (exclusion criteria) • Two-tailed z-test, with p < 0.01 significance VETERANS HEALTH ADMINISTRATION
Feedback from Pilot Sites • “Pairing evidenced based interventions with cost effective interventions” • “Developed a strong culture that embraces change” • “Use of the predictive model helped develop a well established discharge planning team and meeting format” VETERANS HEALTH ADMINISTRATION
Readmission Rate and Percent Difference (FY14 vs FY15) 30.0% 25.4% 25.0% -28% 19.8% 20.0% 16.2% 15.9% 15.8% 15.2% 15.2% 13.9% 15.0% 12.1% 11.7% 11.4% 10.0% 9.9% 10.0% 8.5% 5.0% 0.0% Overall Albany *Bay Pines (+) Gainesville Omaha *San Juan Sioux Falls FY14 (Historical) FY15 (Intervention) *P ≤ 0.01; (+) 2 pilot sites VETERANS HEALTH ADMINISTRATION 17
Limitations • QI based study with limited data points • Predictive model limitations: – Excluded patients new to the VA – Excluded length of stay and admitting diagnosis codes • Comparison made with historical group • Variation of implementation of bundle VETERANS HEALTH ADMINISTRATION
Next Steps • VA Leadership interested in OSI|VERC leading a prospective controlled trial – HSR&D Merit or QUERI – VA Cooperative Study • Monitoring RRTICT’s impact on other variables – Mortality – Length of Stay • Promote RRTICT through presentations and publications VETERANS HEALTH ADMINISTRATION
Thank You! Ashley Ketterer Gruszkowski, MHA Ashley.Gruszkowski@va.gov VETERANS HEALTH ADMINISTRATION
Recommend
More recommend
