Flow Improves Everything

1
Flow Improves Everything

• Carol Haraden
• Kate Silvester

2
The Work of Institute for Healthcare Improvement
Innovation Team
3
How does Flow relate to current organizational
priorities?

• Patient satisfaction
• Effective ED operation
• Staff satisfaction
• Safety
• Finance

4
Aim of Improving Flow

• Increase throughput and minimize delays while
  assuring that high performance in flow is not at
  the expense of poor quality

5
The Myth of One-Hundred Percent Utilization

• Increasing average occupancy levels leaves less
  room for unscheduled admissions.
• Emergency admissions must be diverted, denied, or
  queued.
• Eventually, quality of care declines as all
  patients are increasingly placed into holding
  patterns
• New admissions managed off service
• More boarders in the ED or PACU
• Delayed elective medical or surgical admission.

6
We Can Control the Flow!

• Elective this is published everyday in advance
• How much does this vary?
• Unplanned use queuing theory to predict
• What is left?

7
Lessons from Queuing Theory Background

• A descriptive modeling tool
• Describes stead-state functioning of systems
  involving waiting lines
• Mathematics dependent upon the variability of
  arrival and service rates
• Three variables define the system
• Arrival rate
• Service time
• Number of servers

Michael McManus, Boston Childrens Hospital, 2001
8
Elective Surgical Requests vs Total Refusals
Michael L. McManus, M.D., M.P.H. Michael C.
Long, M.D. Abbot Cooper James Mandell, M.D.
Donald M. Berwick, MD Marcello Pagano, Ph.D.
Eugene Litvak, Ph.D. Impact of Variability in
Surgical Caseload on Access to Intensive Care
Services, Anesthesiology 2003 98 1491-1496.
9
Flow Work

• See your flow
• Develop your administration system for real-time
  flow management
• Prioritize your flow work
• Work on the changes that are most pressing in
  your organization
• Create demand and capacity predictions
• 80 reliable by area

10
Flow Work

• Create a pull system
• Discharge appointments that pull patients forward
  from ED, PACU, ICU
• Create partnerships
• Example Help nursing homes prevent unnecessary
  admissions
• Consider reducing variation in elective surgical
  scheduling
• Build the business case for flow with CFO

11
Seeing Flow Administrative System

• Bed Management Process
  Aim Efficiently transition
  patients through the system
• Some components of the process could be
• Bed Coordinator
• Bed Huddles
• Bed status (electronic bed tracking systems, Bed
  boards)
• Demand/capacity measures and planning
• Early Warning and Response Systems
  Aim
  Manage large fluctuations in demand or capacity
• Week ahead
• Day ahead
• Day of

12
Building a Real Time Demand Capacity System
Some Learning and Some Examples
13
The Sequence for Building a Real Time Demand
Capacity System

• Step 1 Predict your capacity (discharges plus
  available beds) for a single time interval
• Step 2 Predict your demand (admissions based on
  flow throughput considerations) for a single time
  interval
• Step 3 Determine whether you need a plan for the
  time interval by comparing demand and capacity
  predictions
• Step 4 Evaluate the success of the plan. Use the
  knowledge from successful and unsuccessful plans
  to implement standardized adjustments and
  identify system improvements.
• Step 5 Refine the plan to synchronize predicted
  admissions more closely with predicted capacity

14
Some learning

• 1. Initially, predict the current reality. Try to
  avoid taking on specific improvement projects
  until predictions are reliable.
• 2. Develop a standard process (who, what, when,
  where, how) on each unit to gather the
  information used in the prediction formula.
• 3. Start with one unit working on prediction to
  learn fundamentals then move quickly to include
  collaborating units or service lines.
• 4. The unit plan to match capacity to demand
  should be focused on the unit front line staff
  and the plan for specific patients
• 5. Create formal feedback loops for learning.
  Agree on the who, what, where and when for
  learning each day from both successful and
  unsuccessful predictions and plans.
  

15
Some learning

• In may be tempting to start using the information
  on prediction of discharges and admissions,
  gathered in Step 1 and Step 2, to develop a plan
  (Step 3) before you have reached 80 reliability.
  
• Synchronizing admissions to discharges at
  specific times is an important refinement to the
  system.
• Gross mismatches in capacity and demand need to
  be addressed from a more central leadership
  perspective.
• To attract people to the effort, demonstrate this
  work leads to improved performance. Outcome
  measures need to be tracked including, but not
  only, staff satisfaction.

16
Improving Flow through the Bottleneck

• Focus on the specific stream of patients
  identified
• Develop a flowchart for the flow of patients in
  that stream
• Walk the route of the patients in the stream to
  learn more about where/when patients are backing
  up (You may have to ask why multiple
  times)
• Based on what is learned, identify and test
  interventions to alleviate the bottleneck.
  Consider where flow high leverage changes
  (Synchronization, MDR, reducing variation,
  extending the chain, etc. ) might apply.

17
Use of Data
Testing Real Time Demand Capacity

• Process Reliability of predictions
• Intermediate Outcome Transition time between
  units, Holds, .
• System Outcome (for example) ED LOS for admitted
  pts

Real Time D/C
Specific Bottlenecks
Discharge efficiencies
Transfer slots with NH

18
UPMC Shadyside

• Real Time Demand Capacity Management Update
• January, 2008

19
Where We Are

• Real Time Demand Capacity Sequencing has be
  rolled out to all inpatient units
• All units are now involved in Step 5 refining
  the plan
• Identification of barriers and bottlenecks to
  successful plans implementation of tests of
  change to eliminate bottlenecks
• Observation of how Real Time Demand Capacity
  functions when our High Census Alert is fired

20
Identifying Bottlenecks and
Creating New Work Processes

• Orthopedic Unit
• Bottleneck Sequencing of getting patients out
  of the PACU problematic.patients delayed in
  PACU, potentially backs up the OR
• All ortho patients leave on some type of
  anti-coagulation regimen. Bottleneck appeared to
  be getting medication delivered on day of
  discharge
• TOC Pharmacy to deliver meds on post op day 1
• TOC Outcome 43/43 patients received medication
  on post-op day 1 with process change

21
Identifying Bottlenecks and
Creating New Work Processes

• Orthopedic Unit
• 21/43 patients still had gt3hrs between written
  discharge actually leaving
• 12 of them due to need for 2nd PT session
• Current TOC
• PT identifies patients for dc in AM that would
  only need one PT session on day of discharge with
  a PT magnet on white board
• These patients are scheduled first in the AM and
  ready for discharge by 10am

22
Identifying Bottlenecks and
Creating New Work Processes

• Surgical Units
• Bottleneck delays in getting patients out of
  the PACU due to the lack of beds on the surgical
  specialty units
• Right-sizing exercise completed on all surgical
  units. None of the surgical specialty units had a
  great enough volume to support its specialty
  keep the unit productive one surgical unit had
  numerous specialties of overnight cases
• Each surgical unit now has a primary specialty
  and a secondary lower volume specialty of
  overnight cases to help fill the surgical beds
  with surgical patients provide early dcs
• This resulted also in additional medicine beds
  all on one unit vs scattered among the various
  surgical units

23
(No Transcript)
24
(No Transcript)
25
(No Transcript)
26
(No Transcript)
27
4M efforts moved to get the right number of
patients out by 11am so that PACUs can be
accommodated when ready..not the most out by
11am
28
Median ED Admit LOS CY07
29
(No Transcript)
30
Synchronization What Does it Mean in
Healthcare?

• Admissions are linked to planned discharges based
  on a historical data
• Assumes there is a repetitive character to the
  flow stream.
• Assumes that expectations of discharge slot
  reliability will be tied to expectations for the
  upstream department to orchestrate to use the
  linked slots
• Allows the start of consideration for down stream
  resource

31
Creating Partnerships Why?

• Hospitals are dramatically affected by the flow
  of patients in and out of their facility
• They feel buffeted by forces beyond their control
• They own the problem but not the geography

32
Creating Partnerships The Solution

• Partner with organizations that impact your flow
  whether you own them or not
• Develop cooperative agreements that are mutually
  beneficial
• Develop a strong business case for both parties