Hospital Flow Case Study: Cincinnati Children s Hospital

Transcription

1 Hospital Flow Case Study: Cincinnati Children s Hospital Frederick C. Ryckman, MD Professor of Surgery / Transplantation Sr. Vice President Medical Operations Cincinnati Children s Hospital Cincinnati, Ohio This presenter has nothing to disclose. IHI um Dec. 5, 2016 James Anderson Center for Health Systems Excellence Cincinnati Children s Hospital 550 Bed Medical Center Admissions/Year 30,848 Opt Visits 1.02 M Surgical Procedures 32,000 cases 28 OR s, 2 IR suites, Hybrid Cath lab 8 OR Outpatient Surgery Center 1.4 M sq. ft. Research Space 15,000 Employees 1

2 Health Care Delivery System Transformation Strategic Improvement Priorities and System Level Measures ACCESS FLOW PATIENT SAFETY CLINICAL EXCELLENCE REDUCE HASSLES TEAM WELLBEING FAMILY CENTERED CARE System Level Measures 3 rd Next available appointment % of eligible patients with delays Discharge Prediction and Execution Growth Prediction Adverse drug events (ADE) per 1,000 doses Nosocomial infection rates: Bloodstream infection rate Surgical site infection rate infection rates: VAP Safe Practices Serious Safety Events Codes outside the ICU rate/1000 days Standardized PICU Mortality Ratio Expected/ Actual % use of Evidence- Based Care for eligible patients Functional Health Status Touch Time for Providers Employee Satisfaction Staffing Effectiveness Staff Physician Satisfaction Voluntary staff turnover rate Accident rate for staff with Work days lost Overall Rating: Patient Experience Functional Health Status Risk Adjusted Cost per Discharge What Do Patients Hire Us to Provide What do they call Value Make the Right Diagnosis Deliver the Correct Therapy / Treatment Outcomes Prevent Complications or Errors in Care Deliver Safe Care regardless of the Inherent Risks Safety Get Me Home, Keep me at Home Respect my needs Give me my Money s Worth Patient / Family Experience Value This is all FLOW management it is essential for SAFETY, PATIENT / FAMILY EXPERIENCE and QUALITY DELIVERY 2

3 Flow is a Safety Initiative Getting the Rights Right Right Diagnosis and Treatment Right Patient in Right Bed Location Right Nursing Staff and Staffing Expertise Disease Specific Expertise Equipment Expertise Prediction Framework for Safety Best Care Model Requires ability to Predict future needs, and manage present capacity control variability Operations Management techniques to understand and manage variability are the key to success Value Equation for Healthcare Value = (Outcomes + Patient Experience) x Appropriateness Cost + Hassle Factor 3

4 140 Critical Flow Failure Recognition Weekly Critical Flow Failures # of New Patient Failures Total # of Bed Days /14/2008 (Wed- 8/18/08 9/22/08 10/27/08 12/1/08 1/5/09 2/9/09 3/16/09 4/20/09 5/25/09 6/29/09 8/3/09 9/7/09 10/12/09 11/16/09 12/21/09 1/25/10 3/1/10 4/5/10 5/10/10 6/14/10 7/19/10 8/23/10 9/27/10 11/1/10 12/6/10 1/10/11 2/14/11 3/21/11 4/25/11 5/30/11 7/4/11 8/8/11 9/12/11 10/17/11 11/21/11 12/26/11 1/30/12 3/5/12 4/9/12 5/14/12 6/18/12 7/23/12 8/27/12 10/1/12 11/5/12 12/10/12 1/14/13 2/18/13 3/25/13 4/29/13 6/3/13 7/8/13 Week Beginning # of New Failures Total Failures (Bed Days) 8/12/13 9/16/13 10/21/13 11/25/13 12/30/13 2/3/14 3/10/14 4/14/14 5/19/14 6/23/14 7/28/14 9/1/14 10/6/14 11/10/14 12/15/14 1/19/15 2/23/15 3/30/15 5/4/15 6/8/15 7/13/15 8/17/15 9/21/15 10/26/15 11/30/15 1/4/16 2/8/16 3/14/16 4/18/16 5/23/16 6/27/16 8/1/16 9/5/16 0 Type of Control Chart: P Chart Critical Flow Failure Recognition # of New Patient Failures Weekly Critical Flow Failures Over the last 52 weeks Total # of Bed Days 0 10/5/ /12/ /19/ /26/ /2/ /9/ /16/ /23/ /30/ /7/ /14/ /21/ /28/2015 1/4/2016 1/11/2016 1/18/2016 1/25/2016 2/1/2016 2/8/2016 2/15/2016 2/22/2016 2/29/2016 3/7/2016 3/14/2016 3/21/2016 3/28/2016 4/4/2016 4/11/2016 4/18/2016 4/25/2016 5/2/2016 5/9/2016 5/16/2016 5/23/2016 5/30/2016 6/6/2016 6/13/2016 6/20/2016 6/27/2016 7/4/2016 7/11/2016 7/18/2016 7/25/2016 8/1/2016 8/8/2016 8/15/2016 8/22/2016 8/29/2016 9/5/2016 9/12/2016 9/19/2016 9/26/ /3/2016 Week Beginning # of New Failures Total Failures (Bed Days) 0 Type of Control Chart: P Chart 4

5 Growth Requires Constant Efficiency Improvement 300 Critical Patient Flow Failures by Month 250 # of Patients with a New Failure JUL AUG SEP OCT NOV DEC JAN FEB MAR APR Month of Fiscal Year MAY JUN Last Update: 3/4/2016 by M. Ponti-Zins, for Data Source: MPS Key Drivers for Capacity Management IHI Drivers CCHMC Initiative Operations Possibilities Shape / Reduce Demand Predictable Care Delivery Management of Variability Best Practices, Analysis of ALOS and outliers, Standardize then Customize, Eliminate unnecessary care Identify Patient Streams Inpatient/Outpatient/OR Manage System Variation D/C Match Optimization of Flow Delivery Capacity Prediction Placement initiatives D:C Matching plans Discharge prediction and planning, Home Care, Parent Initiatives Integration of simulation modeling and planning Environmental Impact Reports for growth programs System Re-Design Capacity Management Flow:Safety Matching Simulation for design and patient placement Environments Impact Planning Flow Failure Analysis, Predictive Risk Analysis 5

6 Decrease overutilization of hospital services Shape or Reduce Demand IHI Theory on Flow Outcomes Primary Drivers Secondary Drivers Specific Change Ideas Relocate care in ICUs in accordance with patients EOL wishes Relocate care in Med/Surg Units to community-based care settings Relocate low-acuity care in EDs to community-based care settings Decrease demand for hospital beds through delivering appropriate care Decrease demand for hospital beds by reducing hospital acquired conditions 1. Proactive advanced illness planning 2. Development of palliative care programs (hospital-based and community-based) 3. Reduce readmissions for high risk populations 4. Extended hours in primary care practices 5. Urgent Care and Retail Clinics 6. Enroll patients in community-based mental health services 7. Paramedics & EMTs triaging & treating patients at home 8. Greater use of clinical pathways and evidence-based medicine 9. Care management for vulnerable/high risk patient populations 10. Decrease complications/harm (HAPU, CAUTI, SSI, falls with harm) and subsequent LOS 11. Redesign surgical schedules to create an predictable flow of patients to downstream ICUs and inpatient units Optimize patient placement to insure the right care, in the right place, at the right time Increase clinician and staff satisfaction Demonstrate a ROI for the systems moving to bundled payment arrangements Match Capacity and Demand Redesign the System Decrease variation in surgical scheduling Oversight system for hospital-wide operations to optimize patient flow Real-time demand and capacity management processes Flex capacity to meet hourly, daily and seasonal variations in demand Early recognition for high census and surge planning Improve efficiencies and throughput in the OR, ED, ICUs and Med/Surg Units Service Line Optimization (frail elders, SNF residents, stroke patients, etc.) Reducing unnecessary variations in care James and managing M. Anderson LOS outliers Center 1. Assess seasonal variations and changes in demand patterns and proactively plan for variations 2. Daily flow planning huddles (improve predictions to synchronize admissions, discharges and discharges) 3. Real-time demand and capacity problem-solving (managing constraints and bottlenecks) 4. Planning capacity to meet predicted demand patterns 5. High census protocols to expedite admissions from the ED and manage surgical schedules. 1. Redesign surgical schedules to improve throughput and to improve smooth flow of patients to downstream ICUs and inpatient units 2. Separate scheduled and unscheduled flows in the OR 3. ED efficiency changes to decrease LOS 4. Decrease LOS in ICUs (timely consults, tests and procedures) 5. Decrease LOS on Med/Surg Units (case management for patients with complex medical and social needs) 6. Advance planning for transfers to community-based care settings 7. Cooperative agreements with rehab facilities, SNFs and nursing homes Key Drivers for Capacity Management IHI Drivers CCHMC Initiative Operations Possibilities Shape / Reduce Demand Predictable Care Delivery Management of Variability Evidence Based Best Practices, Analysis of ALOS/outliers, Standardize then Customize, Eliminate unnecessary care Identify Patient Streams Inpatient/Outpatient/OR Manage System Variation D/C Match Optimization of Flow Delivery Capacity Prediction Placement initiatives D:C Matching plans Discharge prediction and planning, Home Care, Parent Initiatives Integration of simulation modeling and planning Environmental Impact Reports for growth programs System Re-Design Capacity Management Flow:Safety Matching Simulation for design and patient placement Environments Impact Planning Flow Failure Analysis, Predictive Risk Analysis 6

7 Evidence Based Care Evidence Based Care Guidelines serve as an interface between rapidly evolving scientific information and busy clinical practices Developed by Inter-disciplinary teams experts Implementation Awareness of recommendation to facilitate change Easy access to the Evidence Feedback on Outcomes Feedback on further improvements Culture of Improvement / Evidence Based Care Integration Priority Practice - Plan Prioritization Goal = Exceptional, Safe, Affordable Care Every Child Owner Executive Leadership Practice What we Do Essential Steps, Decisions and Actions Owner Clinical Leadership Teams Departments / Divisions Processes How we Do It Processes to execute to the goal Owner Operational Leaders Sites of Care Plan Implement the Processes - plan through application of process steps Owner Sites of Care leaders and clinical staff (MD / RN) Front Line Implementation Just Do It Every day for every child 7

8 Cytomegalovirus Prophylaxis 75% Decrease in CMV infection liver/intestine transplants Decreased IV-IGG expense Danziger-Isakov, Lara et al. CCHMC Integrated Solid Organ Transplant Yearly SSI Patients - CCHMC Total SSI Pa ents Fiscal Year SSI s 12 Years 540 SSI s Prevented 492 SSI s 8

9 Yearly SSI Patients - CCHMC Total SSI Pa ents Case Average 10 days LOS $27, Fiscal Year 29 Business Case 5400 days LOS $14.58 million SSI s 12 Years 540 SSI s Prevented 492 SSI s Standardization for Outcomes Merging Evidence Based Care and Practice Focus on Excellence SSI Spines 9

10 SSI Accomplishments Baseline rate: 4.4 SSIs/100 procedures, Current Rate: 1.7 SSIs/100 procedures 60% reduction, 32% reduction in past 3 years - $17.4 Million from SSI alone Overall SPS - Estimated 6,686 fewer children harmed Since October $121.4 Million saved in SPS Network Toltzis P, O Riordan M, Cunningham DJ, Ryckman FC, Bracke TM, Olivea J, Lyren A. A statewide collaborative to reduce pediatric surgical site infections. Pediatrics : Getting to Better Clinical Practice Guidelines systematically developed statements that guide decision making Widely developed, limited effectiveness There are times when it is advisable and appropriate to deviate from standard guidelines Parikh K, Agrawal S. JAMA Ped 2015;169; Recommended = 0 10

11 Getting to Better Achievable Benchmarks of Care (ABC s) Actual performance at health care sites performing in the top 10% Parikh K, Agrawal S. JAMAPed 2015;169; Getting to Better ABC s Hospital Site Performance Individual Performance 11

12 Key Drivers for Capacity Management IHI Drivers CCHMC Initiative Operations Possibilities Shape / Reduce Demand Predictable Care Delivery Management of Variability Evidence Based Best Practices, Analysis of ALOS and outliers, Standardize then Customize, Eliminate unnecessary care Identify Patient Streams Inpatient/Outpatient/OR Manage System Variation D/C Match Optimization of Flow Delivery Capacity Prediction Placement initiatives D:C Matching plans Discharge prediction and planning, Home Care, Parent Initiatives Integration of simulation modeling and planning Environmental Impact Reports for growth programs System Re-Design Capacity Management Flow:Safety Matching Simulation for design and patient placement Environments Impact Planning Flow Failure Analysis, Predictive Risk Analysis Pre-Smoothing ICU Elective Census 12 ICU Daily Elective Census Prior to ICU Model for Smoothing N u m b e r o f P a tie n ts in IC U B ed s /01 / /03 / /05 / /07 / /09 / /11 / /13 / /15 / /17 / /19 / /21 / /23 / /25 / /27 / /29 / /01 / /03 / /05 / /07 / /09 / /11 / /13 / /15 / /17 / /19 / /21 / /23 / /25 / /27 / /29 / /31 / /02 / /04 / /06 / /08 / /10 / /12 / /14 / /16 / /18 / /20 / /22 /0 7 ICU Daily Elective Patient Census Center Line - Mean Control Limits 12

13 ICU Bed Availability ICU Scheduling Category Case Statistics by Category Total PICU Days Case Count ALOS Short (61%) 1.27 (27%) Medium (28%) 3.72 (37%) Long (11%) 9.76 (36%) Grand Total ICU Admission Model Elective Cases Short Stay Cases Access Cap # Cases on Schedule / Day Long Stay Cases Fixed # Beds 13

14 # of Patients with a New Failure Patients who Utilize an ICU bed b/c an Appropriate Bed is Not Available 9 7/16/ /14/2008 1/12/2009 4/12/2009 7/11/ /9/2009 1/7/2010 4/7/2010 7/6/ /4/2010 1/2/2011 4/2/2011 7/1/2011 9/29/ /28/2011 3/27/2012 6/25/2012 9/23/ /22/2012 3/22/2013 6/20/2013 9/18/ /17/2013 3/17/2014 6/15/2014 9/13/ /12/2014 3/12/2015 6/10/2015 9/8/ /7/2015 3/6/2016 6/4/2016 9/2/ /1/2016 3/1/2017 5/30/2017 8/28/ /26/2017 2/24/2018 # of Patients with a New Failure 6 7/16/2008 9/14/ /13/2008 1/12/2009 3/13/2009 5/12/2009 7/11/2009 9/9/ /8/2009 1/7/2010 3/8/2010 5/7/2010 7/6/2010 9/4/ /3/2010 1/2/2011 3/3/2011 5/2/2011 7/1/2011 8/30/ /29/ /28/2011 2/26/2012 4/26/2012 6/25/2012 8/24/ /23/ /22/2012 2/20/2013 4/21/2013 6/20/2013 8/19/ /18/ /17/2013 2/15/2014 4/16/2014 6/15/2014 8/14/ /13/ /12/2014 2/10/2015 4/11/2015 6/10/2015 8/9/ /8/ /7/2015 2/5/2016 4/5/2016 6/4/2016 8/3/ /2/ /1/2016 1/30/2017 3/31/2017 5/30/2017 7/29/2017 9/27/ /26/2017 1/25/2018 3/26/2018 # of Patients with a New Failure 7/16/ /14/2008 1/12/2009 4/12/2009 7/11/ /9/2009 1/7/2010 4/7/2010 7/6/ /4/2010 1/2/2011 4/2/2011 7/1/2011 9/29/ /28/2011 3/27/2012 6/25/2012 9/23/ /22/2012 3/22/2013 6/20/2013 9/18/ /17/2013 3/17/2014 6/15/2014 9/13/ /12/2014 3/12/2015 6/10/2015 9/8/ /7/2015 3/6/2016 6/4/2016 9/2/ /1/2016 3/1/2017 5/30/2017 8/28/ /26/2017 2/24/2018 Delayed or Canceled Surgery Due to Bed Capacity # of Patients with a New Failure /16/2008 9/14/ /13/2008 1/12/2009 3/13/2009 5/12/2009 7/11/2009 9/9/ /8/2009 1/7/2010 3/8/2010 5/7/2010 7/6/2010 9/4/ /3/2010 1/2/2011 3/3/2011 5/2/2011 7/1/2011 8/30/ /29/ /28/2011 2/26/2012 4/26/2012 6/25/2012 8/24/ /23/ /22/2012 2/20/2013 4/21/2013 6/20/2013 8/19/ /18/ /17/2013 2/15/2014 4/16/2014 6/15/2014 8/14/ /13/ /12/2014 2/10/2015 4/11/2015 6/10/2015 8/9/ /8/ /7/2015 2/5/2016 4/5/2016 6/4/2016 8/3/ /2/ /1/2016 1/30/2017 3/31/2017 5/30/2017 7/29/2017 9/27/ /26/2017 1/25/2018 3/26/ /27/2016 Predicting ICU Discharge Critical Flow Failures 9 8 PICU Bed Not Available for Urgent Use Daily Failures Psychiatry Patients Placed Outside of their Primary Unit Daily Failures Daily Failures 14

15 Mental Health Impact on Flow Key Drivers for Capacity Management IHI Drivers CCHMC Initiative Operations Possibilities Shape / Reduce Demand Predictable Care Delivery Management of Variability Evidence Based Best Practices, Analysis of ALOS and outliers, Standardize then Customize, Eliminate unnecessary care Identify Patient Streams Inpatient/Outpatient/OR Manage System Variation D/C Match Optimization of Flow Delivery Capacity Prediction Placement initiatives D:C Matching plans Discharge prediction and planning, Home Care, Parent Initiatives Integration of simulation modeling and planning Environmental Impact Reports for growth programs System Re-Design Capacity Management Flow:Safety Matching Simulation for design and patient placement Environments Impact Planning Flow Failure Analysis, Predictive Risk Analysis 15

16 Strategies for Patient Placement Early Day Beds PICU, CICU Critical Units Later Day Beds All units Demand : Capacity Match Opportunistic with ethought Specific Bed D:C Match Unit Bed Awareness Discharge Prediction Various approaches to Discharge Management 1980 s Keep it a Secret 1990 s 2000 s Discharge goals Reactive AM before 11 > 30-40% Shift goals 4 hour time block goals with prediction of window Not Patient Centered Prediction Discharge when Medically Ready 16

17 Discharge Prediction Buckets Discharge Prediction Various approaches to Discharge Management 1980 s Keep it a Secret 1990 s 2000 s Discharge goals Reactive AM before 11 > 30-40% Shift goals 4 hour time block goals with prediction of window Not Patient Centered Prediction Discharge when Medically Ready 17

18 Timeline for DC when Medically Ready Admission to Floor Discharge Criteria Set Treatment Protocol Followed Discharge Criteria Met Nurse Notifies Staff 2 Hrs Discharge Home Re-Adm & LOS Tracked Standardized Criteria Buy-In by Staff Standardized Protocols for most Tx Evaluation Criteria Criteria established at admission Nurse at bedside notifies service when Medical discharge criteria are met Discharge from floor in < 2 hours Review Length of Stay and Re-Admissions as balancing measures Not about Speed Now about Efficiency Modify Rounding Clear Discharge Criteria Communication Family Discharge When Medically Ready Karen Tucker, Angela Statile, Diane Herzog, and Christy White Increase percentage of all HM patients who have met* Medically ready criteria who will be discharged within two hours of reaching that goal* on A6S, A6N, LA1W from 75%to 80% by June 30, 2014 Productivity: Optimize use of facilities and staff and improve patient flow to achieve 20% greater utilization of existing assets by June 30, 2015 Criteria for Medically Ready Defined at Admission Shared Ownership/ Accountability and Buy-In Among Physicians and Nurses Communication regarding prediction of discharge and defined goals is ongoing through the hospital stay Potential Barriers to Discharge are Clearly Articulated and Mitigation Plans Established Performance by team is transparent Evidence of Preoccupation with Failure Clear expectations for Parents/ Families Agreement among HM attendings and nursing staff of discharge criteria for order set diagnoses and general admissions (LOR 2) 1) 8 pm Huddle discussion re: early discharges (LOR 2) 2) 0630 notification of patients ready for discharge (LOR 1) Performance Management (LOR 1) Standardized and modifiable order sets (LOR 2) Identify and Mitigate Plans: 1) Transportation- census based (LOR 1) 2) Pharmacy-priority fills (LOR 2), Outpt delivery to patient room (LOR 1) 3) Consults- proactive evaluation (LOR 2) 4) RT- process in PICU (LOR 1) 5) Home Health Care Daily Feedback reports to RNs and MD s with ID and mitigation of process and outcome measure failures (LOR 2) Feedback of data by HM team In conference room and by (LOR 1) Auto notification to resident team that patient has met all criteria (LOR 2) 18

19 EMR Discharge Criteria: Physician View 2013 Epic Systems Corporation. Used with permission. Poster in Resident Conference Room 19

20 Discharge when Medically Ready All Units Service Level DC when Medically Ready 20

21 Discharge Failure Reasons Comparison Balancing Measures Length of Stay Hospital Medicine Average Length of Stay patients with selected diagnosis Average Length of Stay (Days) Unit 1 Unit 2 Unit 3 All FY11 FY

22 Balancing Measures Readmission Rate Key Drivers for Capacity Management IHI Drivers CCHMC Initiative Operations Possibilities Shape / Reduce Demand Predictable Care Delivery Management of Variability Evidence Based Best Practices, Analysis of ALOS and outliers, Standardize then Customize, Eliminate unnecessary care Identify Patient Streams Inpatient/Outpatient/OR Manage System Variation D/C Match Optimization of Flow Delivery Capacity Prediction Placement initiatives D:C Matching plans Discharge prediction and planning, Home Care, Parent Initiatives Integration of simulation modeling and planning Environmental Impact Reports for growth programs System Re-Design Capacity Management Flow:Safety Matching Simulation for design and patient placement Environments Impact Planning Flow Failure Analysis, Predictive Risk Analysis 22

23 Prediction Model for the Future Static Analytics Performing a ONE TIME analysis of processes with historical data in order to PREDICT what s going to happen under certain circumstances. Critical care Bed Modeling for Growth Real-Time Prediction Performing ONGOING analysis of processes with latest available data in order to continuously PREDICT what s going to happen under certain circumstances. RN Bedside Nurse Staffing Model Critical Care Bed Predictions 23

24 Critical Care Bed Growth Analysis ecasted Bed Needs - Advantages of Efficiency Estimated number of beds required for given probability of the unit being full. ecast Time Frame Probability of Full Unit PICU Beds CICU Beds ICU Bed Needs Combined ICUs Estimated Savings Year 2 10% % % % Year 5 10% % % % Year7 10% % % % POPULATION: Unscheduled Medical/Surgical, BMT, ENT Airway ICU Elective Cases, Heart Institute Patients 24

25 Predicting Future Programmatic Needs Environmental Impact Studies LUNG TRANSPLANTATION ICU Bed Needs Number of beds needed based on probability of having a full unit (5%, 2%, 1%, 0%) and the growth estimate. Low/Conservative Mid-Range/Most Likely High/Aggressive Growth 1 Yr 3 Yr 5 Yr 7 Yr 10 Yr 1 Yr 3 Yr 5 Yr 7 Yr 10 Yr 1 Yr 3 Yr 5 Yr 7 Yr 5% % % % Yr 25

26 Predicted to Actual Pre-Op Bed Demand Predicted to Actual Pulmonary Bed Demand Prediction Pulmonary demand appears to be in line with original projections Actual Pulmonary Mid Level Conservative Aggressive Year Year Year Year Year Year Predicted to Actual ICU Bed Demand Predicted to Actual ICU Bed Demand ICU Prediction Mid Level Conservative Aggressive Actual Year Year Year Year Year Year ICU demand appears to be dependent on pre-transplant severity and much higher than originally anticipated 26

27 Predicted to Actual Post-Op Floor Demand Predicted to Actual Cardiology Bed Demand Heart Prediction Mid Level Conservative Aggressive Year Year Year Year Year Year Cardiology demand appears to be lower than original projections Actual Outpatient / Procedural Demand 27

28 Outcomes and Observations - Environmental Impact Assessment provided valuable information allowing for assessment and agreement across the hospital before program was initiated - Answer questions about patient flow and placement - Assess potential stress on existing resources - Quantify demand and capacity needs (staffing, beds, outpatient clinic rooms, PFT demand, OR demand) - Requires assumptions and research for new programs - As always your results are only accurate if your assumptions are correct Staffing and Environment - Mortality Nurse Staffing and Hospital Mortality Tertiary Medical Center 197,691 patients, 176,696 RN shifts, 43 hospital units Relationship between nurse staffing and patient turnover Risk of Death 2-3 % for each below target shift Risk of Death 4-7 % for every high turnover shift Admissions, discharges, and transfers Risk of Death 12 % for each below target shift Risk of Death 15 % for every high turnover shift Independent Variables when considering risks Needleman J. et al. N Engl J Med 2011;364: ICU Patient Non-ICU Patient 1 st 5 days LOS 28

29 Census Calculation Tomorrow s census = Today s census + Today s admissions Today s discharges Easy enough, right? Developed more than 200 models: 17 units, 7 days a week, admissions, and discharges, multiple input and output streams Model Outputs -Available bed capacity -Midnight census -Scheduled admissions -Predicted admissions -Predicted discharges -Predicted demand (census + adm disch) -Predicted overflow placement 29

30 Staffing Prediction Proactive Planning Actual vs. Predicted Results 30

31 Impact of Analytics Bed demand predictions facilitate staffing and overflow planning right patient right team ED admit predictions improved from 40% to 70% accuracy resource allocation Encourages staff to more consistently predict and document estimated discharge date, which helps guide care system efficiency Uncovers scheduling issues efficiency and access One-stop source to determine where there is capacity (or lack thereof) to add services (infusions, etc.) efficiency and utilization Key Drivers for Capacity Management IHI Drivers CCHMC Initiative Operations Possibilities Shape / Reduce Demand Predictable Care Delivery Management of Variability Evidence Based Best Practices, Analysis of ALOS and outliers, Standardize then Customize, Eliminate unnecessary care Identify Patient Streams Inpatient/Outpatient/OR Manage System Variation D/C Match Optimization of Flow Delivery Capacity Prediction Placement initiatives D:C Matching plans Discharge prediction and planning, Home Care, Parent Initiatives Integration of simulation modeling and planning Environmental Impact Reports for growth programs System Re-Design Capacity Management Flow:Safety Matching Simulation for design and patient placement Environments Impact Planning Flow Failure Analysis, Predictive Risk Analysis 31

32 GARDiANS Hospital Wide System for Safety 3 Times - Every Day Individual Room / Floor / System Predictions Capacity and Safety Floor Huddles PeriOp Huddle Outpt, Home, Psych ED Huddle ICU Huddles Institutional Wide Bed Huddle Capacity Management Pharmacy Pt. Transport Facilities Institutional Daily Operations Brief System Prediction Mitigation Strategy Security Housekeeping P.F.E. 32

33 11/27/2016 Flow Dashboard Sites of Care Patient Satisfaction Only 3-4% of 1 Million outpatient visitors rank our care in the lower half (0-6 of 10 pts) 35,000 patient per year Great American Ballpark 42,319 Paul Brown Stadium 65,535 33

34 Understanding Outliers I thought: If I can get 80-85% of this under good control, that will solve at least 85% of the problem N=297 cases < 62 days (85.6%) Total N=347 N=50 >62 Days N=25 >100 Days N=14 >200 Days N=6 >365 Days 34

35 7,595 Bed Days 50 Patients 342 Acute Care Beds 22.2 Days Total Hospital Census 6.08% Yearly Hospital Census 1,795 bed days 1,339 bed days 2,131 bed days 2,340 bed days Observations on Outliers In these predictive models, it is important to be right. What is really important is the cumulative magnitude of your errors. The errors are often the result of big surprises, not multiple small issues. Failure to meet the predictive model leads to progressive and increasing cumulative error, things rarely get better fast (the more you are off, the more you are off.) It is hard to offset the surprise errors with great prediction of the expected. 35

36 Lessons Learned Building Will to work on Flow is a challenge When it works, it is not on anyone s radar If it works for me, your problem is not my problem. When it does not work, somebody else should solve it Linkage Safety and Flow Speed vs Efficiency Work Backwards not just ward Embrace Mathematics and Analytics Standardize processes and work flows Make it Personal Don t let the Data Drown out the Dream Stories not Statistics Names and Faces Accountability is Personal & Group Responsibility Collective Mission/Vision Cincinnati Children s Hospital Medical Center

37 11/27/2016 Thanks! 37