These presenters have nothing to disclose. Applying Critical ED Improvement Principles Jody Crane, MD, MBA Kevin Nolan, MStat, MA April 28, 2015 Cambridge, MA Session Objectives After this session, participants will be able to: Understand flow as it pertains to Lean Healthcare. Understand the basic characteristics of a queue. Understand the impact of variation in healthcare. Understand the impact of server utilization on service times. 1
Ruraltownville ED Part 1 20,000 annual visits 11% LWOBS rate, peaks at 25% in the evenings Poor reputation in the community 2
Ruraltownville ED Pat Sat/Value 2015, X32 Healthcare, LLC, All Rights Reserved 5 Ruraltownville ED Part 1 20,000 annual visits 11% LWOBS rate, peaks at 25% in the evenings Poor reputation in the community and your own leadership, You have the worst performing ED in the Province! 3
Ruraltownville ED Part 1 20,000 annual visits 11% LWOBS rate, peaks at 25% in the evenings Poor reputation in the community and your own leadership, You have the worst performing ED in the Province! You have a boarding problem, only 8 treatment spaces and a very poor physical layout. 1 doc, 1 NP, 3 nurses and 1 triage nurse at peak times Ruraltownville ED Geography 2012, X32 Healthcare, LLC, All Rights Reserved 8 4
Ruraltownville ED Part 1 20,000 annual visits 11% LWOBS rate, peaks at 25% in the evenings Poor reputation in the community and your own leadership, You have the worst performing ED in the Province! You have a boarding problem, only 8 treatment spaces and a very poor physical layout. 1 doc, 1 NP, 2 nurses and 1 triage nurse at peak times Exercise: How would YOUapproach this problem if it were YOURED? Overview Queuing Theory Lean Concepts and Tools Hospital-wide flow and the Theory of Constraints 5
Queuing Theory - A Simple Queue Customer Arrivals Queue (waiting line) Server Customer Departures Queuing Analysis Arrival Rate (λ ) and Arrival Distribution Avg Wait in Queue (W q ) Avg Number in Queue (L q ) Service Rate (µ ) and Service Distribution Avg Time in System (W ) Avg Number in System (L ) 6
Sigma ED Triage - Assumptions 1. The nurse is always on time for the start of her shift. 2. On average, 6 patients arrive per hour. 3. The time it takes the nurse to triage a patient averages 12 minutes (can triage 5 per hour). Will there be waiting??? YES! Sigma ED Triage - Assumptions 1. The nurse is always on time for the start of her shift. 4 patients 2. On average, 6 patients arrive per hour. 3. The time it takes the nurse to triage a patient averages 12 minutes (can triage 5 per hour). Now will there be waiting??? It Depends! 7
Sigma ED Triage - Assumptions 1. The nurse is always on time for the start of her shift. 2. On average, 4 patients arrive per hour. Assume 1 patient arrives every 15 minutes. 3. The time it takes the nurse to triage a patient averages 12 minutes (can triage 5 per hour). Assume exactly 12 minutes per patient. Sigma ED Triage - Capacity & Performance arrive1 arrive2 arrive3 arrive4 triage1 triage2 triage3 triage4 4:00 4:15 4:30 4:45 The nurse and patient arrive at 4pm The first triage encounter lasts exactly 12 min The nurse has exactly 3 min of idle time The next patient arrives at exactly 4:15 And so on...... 8
Sigma ED - Ideal Triage QueueCalc Sigma ED Triage Add Variation 1. The nurse is always on time for the start of her shift. 2. On average, 4 patients arrive per hour. Assume 1 patient arrives every 15 minutes. random arrival process. 3. The time it takes the nurse to triage a patient averages 12 minutes (can triage 5 per hour). Assume exactly 12 minutes per patient. 9
Arrival Data from a Real ED Arrival data from a California hospital. Mondays, 2pm-6pm. Sigma ED Triage Add Variation 1. The nurse is always on time for the start of her shift. 2. On average, 4 patients arrive per hour. Assume 1 patient arrives every 15 minutes. random arrival process. 3. The time it takes the nurse to triage a patient averages 12 minutes (can triage 5 per hour). Assume exactly 12 minutes per patient. variation around service times 10
Distribution of Actual ED Triage Times Average = 5.06 Std.Dev. = 4.97 Time study data from MWH. Sigma ED - Ideal Triage QueueCalc 11
Sigma ED - Ideal Triage Simulation Queue Behavior as a Foundation of Server Utilization High Utilization Contributes to Queuing 12
An example from queuing theory: MICU Utilization and the Patient Rejection Rate Michael McManus, Boston Children s Hospital, 2001 25 Arrival Rate of 10/hour, Service Rate of 12/hour, Server Utilization of 83.33% High Variation Contributes to Queuing No server variation Maximal server variation 13
Achieving Lean Flow Principle - To reduce flow time through an individual queue, you must do one of the following: Reduce average rate of arrivals (rationalize, offload) Reduce variation in time between arrivals (standardize) Reduce average service times (eliminate waste) Reduce variation in service times (standardize) Add server capacity or change the timing of server capacity (align) 2012, Jody Crane, MD, MBA, Charles E. Noon, Ph.D. Achieving Lean Flow Principle - To reduce flow time through an individual queue, you must do one of the following: Reduce average rate of arrivals (rationalize, offload) Reduce variation in time between arrivals (standardize) Reduce average service times (eliminate waste) Reduce variation in service times (standardize) Add server capacity or change the timing of server capacity (align) 2012, Jody Crane, MD, MBA, Charles E. Noon, Ph.D. 14
Affecting the Arrival Rate 2015, Jody Crane, MD, MBA, Kevin Nolan, 2012, Jody MStat, Crane, MA MD, MBA, Charles E. Noon, Ph.D. Affecting Variation in Arrivals 2012, Jody Crane, MD, MBA, Charles E. Noon, Ph.D. 15
Achieving Lean Flow Principle - To reduce flow time through an individual queue, you must do one of the following: Reduce average rate of arrivals (rationalize, offload) Reduce variation in time between arrivals (standardize) Reduce average service times (eliminate waste) Reduce variation in service times (standardize) Add server capacity or change the timing of server capacity (align) 2012, Jody Crane, MD, MBA, Charles E. Noon, Ph.D. Overview Queuing Theory Lean Concepts and Tools Hospital-wide flow and the Theory of Constraints 16
Lean Healthcare An operations and process management strategy Derived from the Toyota Production System Focuses on creating patient value, eliminating waste, promoting flow and continuous improvement Key Principles Focus on Processes that deliver Customer Value Value-added An activity in the process that moves the patient closer to wellness Waste Any other activity in the process 2014 Crane, Noon 17
Waste TIM WOOD Transportation Unnecessary patient movement Inventory Having more supplies than are necessary (gloves) Movement Walking to various locations to get supplies Covering beds in remote areas Waiting Waiting to be seen Waiting for biopsy or stress test results Over-processing Multiple providers asking the same questions Ordering too many tests Overproduction Monitoring a patient that doesn t need to be monitored Unnecessary ICU admission Defects Rework of lab tests (hemolysis) Repeat visits 2014 Crane, Noon Key Principles Focus on Processes that deliver Customer Value Eliminate waste Promote flow Align capacity with demand Establish clear signals and handoffs Reduce variation Eliminate Queues 2014 Crane, Noon 18
Key Principles Focus on Processes that deliver Customer Value Eliminate waste Promote flow Continuously improve the processes Mastering change management Creating a community of scientists Willing to try, ok to fail, but always learning PDCA, Rapid Cycle Testing (RCT) and Rapid Performance Improvement Events (RIEs) 2014 Crane, Noon Improvement Tools Process Mapping Standard Work, Demand/Capacity Management Workplace Organization Inventory Management (Pull systems), visual controls Setup/changeover reduction (Rapid Changeover) Mistake proofing, root cause analysis 2014 Crane, Noon 19
Process Mapping Low Tech Value, Waste, and Tools System Improvements 2014 Crane, Noon 20
Standard Work Standard Work, Demand/Capacity Management Defining the current work sequence and making sure it functions in accordance with patient demand Demand Analysis Quantitative analysis of arrival rates/patterns Calculating takt time Capacity Analysis Work sequencing Appropriate staffing Designed to meet takt time Demand Capacity Matching Applying these principles in complex service industries where demand varies over time 2014 Crane, Noon Hourly Demand 2014 Crane, Noon 21
Stafford ED RN Demand/Capacity Aggregate Demand/RN Capacity 9.00 8.00 7.00 6.00 5.00 4.00 3.00 2.00 1.00 0.00 Projected Total RN Demand Total RN Staffing 2014 Crane, Noon Stafford ED RN Demand/Capacity Aggregate Demand/RN Capacity 9.00 8.00 7.00 6.00 5.00 4.00 3.00 2.00 1.00 0.00 Projected Total RN Demand Total RN Staffing 2014 Crane, Noon 22
Workplace Organization A tool for redefining the organization and performance of your workplace to better promote effective patient care Approach Sort Simplify Sweep Standardize Sustain A place for everything, everything in it s place 2014 Crane, Noon MWH Workplace Organization Project 2014 Crane, Noon 23
MWH Workplace Organization Project 2014 Crane, Noon Visual Management 48 24
4/29/2015 Visual Nursing Server 2014 Crane, Noon Visual Controls -Inventory Pictures Compliments of Karl Kraber, Fanciscan Health System 2014 Crane, Noon 25
Anyone gotten a flat recently? 2014 Crane, Noon Formula One Pit Stop 2014 Crane, Noon 26
How quickly can you get your bed back online? 2014 Crane, Noon Achieving Lean Flow Principle - To reduce flow time through an individual queue, you must do one of the following: Reduce average rate of arrivals (rationalize, offload) Reduce variation in time between arrivals (standardize) Reduce average service times (eliminate waste) Reduce variation in service times (standardize) Add server capacity or change the timing of server capacity (align) 2012, Jody Crane, MD, MBA, Charles E. Noon, Ph.D. 27
Ruraltownville ED Part 2 55 CTAS Distribution 56 28
Ruraltownville ED Acuity by HOD 2011, Jody Crane, MD, MBA 57 Key Metrics Ruraltownville ED 2012, X32 Healthcare, LLC, All Rights Reserved 58 29
Ruraltownville ED Capacity 59 Exercise Ruraltownville ED P60 During peak hours 1. What is the triage nurse utilization if each patient takes 7.5 minutes to triage? (assume all patients are triaged) 2. What is the Nurse Practitioner Productivity in patients per hour if the nurse practitioner is able to see all of the low acuity patients (CTAS 4 and 5)? Give some reasons for the actual productivity of the NP being only 1.4 patients per hr. 3. Calculate the NP utilization in #2 above if he/she spends 20 minutes working on each patient. 4. If the door to bed time was 20 minutes and all patients occupied a bed until they were discharged, calculate the bed utilization based on the current T&R LOS of 200 minutes (assume all patients are discharged) 30
1. Triage Nurse Utilization 1 RN Arrival rate (avg peak) = 3.5 pts/hr Service time = 7.5min Service rate = 60 / 7.5 = 8 pts/hr Utilization = (arrival rate / service rate) = 3.5 / 8 = 44% 2. Nurse Practitioner Productivity CTAS 4 = 42.7% and CTAS 5 = 23.0% Total = 65.7% of total volume NP Productivity if able to see all lower acuity patients =.657 x 3.5 pts/hr = 2.2 pts/hr 31
3. NP Utilization Arrival rate (avg peak) = 2.2 pts/hr (CTAS 4 and 5) Service time = 20 min Service rate = 60 / 20= 3 pts/hr Utilization = (arrival rate / service rate) = 2.2 / 3 = 73% 4. Bed Utilization Arrival rate (avg peak) = 3.5 pts/hr Service time (in bed) = 180 min (3 hours) Service rate = 1/3 pt per hour Total beds (in ED) = 8 Utilization = Arrival rate / (total beds) x service rate = 3.5 / 8 x (1/3) = 131% 32
Nurse, MLP and Physician 2012, X32 Healthcare, LLC, All Rights Reserved 65 2012, X32 Healthcare, LLC, All Rights Reserved 66 33
Any Thoughts? Would you change any of your recommendations from your group exercise? Overview Queuing Theory Lean Concepts and Tools ED and Hospital-wide flow and the Theory of Constraints (TOC) 34
Fast Track Scenario You are the ED manager and you are trying to optimize your fast track because your length of stay is too long (3 hours) On average, your fast track sees 4 ESI 4,5 patients per hour during peak times and you consistently have 5-10 patients waiting to be seen Your fast track is staffed by a physician and nurse, and 100% of your patients get x-rays, no lab testing What do you do? Fast Track Scenario Ok, you study the system and your IE returns the following report The process is : Nurse assess Doc assess X-ray Doc d/c Nurse d/c 10 min 8 min 7 min 2 min 5 min Now what? 35
Fast Track Scenario Ok, you study the system and your IE returns the following report The process is : Nurse assess Doc assess X-ray Doc d/c Nurse d/c 10 min 8 min 7 min 2 min 5 min Utilizations: Nurse = (10 + 5) = 15 x 4 pts/hr = 60min/hr = 100% Physician = (8 + 2) = 10 x 4 pts/hr = 40min/hr = 66% X-ray = 7 x 4pts/hr = 28min/hr = 47% Fast Track Scenario Utilizations: Nurse = (10 + 5) = 15 x 4 pts/hr = 60min/hr = 100% Physician = (8 + 2) = 10 x 4 pts/hr = 40min/hr = 66% X-ray = 7 x 4pts/hr = 28min/hr = 47% Because the nurse is the constraint, adding more docs or working on x-ray will not improve throughput or performance of the system, throughput can only be enhanced by offloading the nurse How might this be done? 36
TOC: The Theory of Constraints Goldratt: The goal of a business is to make money now and in the future. A system s constraints limit its performance or progression toward its goal A resource is either a bottleneck or a nonbottleneck Bottleneck- A resource that has the capacity equal to or less than the demand placed upon it Non-bottleneck- A resource that has a capacity that is greater than the demand placed upon it TOC: The Theory of Constraints An hour lost at a bottleneck is an hour lost for the whole system Time saved at a non-bottleneck is a mirage Efforts spent improving a non-critical bottleneck will not improve the overall performance of your process or system In highly variable systems (i.e. the ED), the bottlenecks can appear to jump around 74 37
The Throughput World: The Five Step Focusing Process of TOC Step 1: Identify the System s Constraint(s) Step 2: Decide how to Exploit the System s Constraint(s) Step 3: Subordinate Everything Else to that Decision Step 4: Elevate the System s Constraints Step 5: If a Constraint Was Broken in Previous Steps, Go to Step 1 2014, Crane, Noon 75 Ruraltownville ED The Fix 2012, X32 Healthcare, LLC, All Rights Reserved 76 38
Kaizen 3 Day Event Day 1 Education and Current State Analysis Day 2 Completion of Future State and Working Subgroups Day 3 Rapid Cycle Test 2015, X32 Healthcare, LLC, All Rights Reserved 77 Current State 39
Ruraltownville Mapping LWOBS Patient & Staff Satisfaction Waste No Test Lab Only X-ray only Both Treatment 107 minutes 182 minutes 172 minutes 182 minutes 147 minutes 2015, X32 Healthcare, LLC, All Rights Reserved Targets A - Visual Management 1 - Chart Flow 2 - Lab 3 - Wayfinding B - Nurse First Triage C - Low Acuity Flow 1 - Super Track 2 - Supplies 3 - Medications 4 - Bed or No Bed Criteria 2015, X32 Healthcare, LLC, All Rights Reserved 40
Before 81 Visual Actions 41
Lab Visual Patient Board Wayfinding 42
Nurse First Triage Triage and Registration 43
Waiting Room d to Super Track 44
45
All in One Day!? 46
Just About There 47
3 days Permanent Change 95 Summary You must have a firm understanding of Queuing theory and Lean in order to respond to the ever-changing landscape and operational demands of your ED. Approaching ED issues with an eye on process improvement over adding resources will insure you begin developing your community of scientists. TOC can provide the framework for targeting improvement and flow opportunities. 48
2010, Jody Crane, MD, MBA, Charles E. Noon, Ph.D. Suggested Reading The Toyota Way, Jeffery Liker Lean Hospitals, Mark Graben The Goal Eli Goldratt Going Lean in Healthcare, IHI White Paper. 49