Transcript T-A-0945-1045 Tools for Data-Driven Decision Making (Mayo

Tools for Data Driven
Decisions
Agenda
•
•
•
The Toolkit
Designing for Outputs
Understanding Tool Inputs
•
•
•
•
•
•
Data sources
Data collection
Tool parameters
Spreadsheet Tools
Analytical Models
Simulation Models
FOR OFFICIAL USE ONLY
2
Context, Purpose, Outcome
1) Context: Navy Medicine faces key operational decisions
that can be made easier with the appropriate use of data
through tools
2) Purpose: Guidance will be provided in choosing the right
tool for the problem at hand, and examples of successful
tools used for performance improvement will be discussed
3) Outcome: Identification of opportunities for decision tools
and the appropriate application of these tools will lead to
more confidence in future decision making
FOR OFFICIAL USE ONLY
3
Tool Choice
"When the only tool you have is a
hammer, every problem begins to
resemble a nail."
– Abraham Maslow
FOR OFFICIAL USE ONLY
4
The Toolkit
•
•
•
•
Spreadsheet Models
Analytical Models
Simulation
Other Support Tools
FOR OFFICIAL USE ONLY
5
Designing for the Outputs
•
What does the customer want out of the tool?
•
•
What is a usable tool for the customer?
•
•
•
IT skill level
Angle of interest
What decisions will be made with this tool?
•
•
•
•
What questions are to be answered?
Day-to-day operations?
System-wide planning?
Facilities/Construction?
Beware of tendency to overdesign!
FOR OFFICIAL USE ONLY
6
Understanding Tool Inputs
•
•
•
Data inputs must match tool choice
Will the tool be “live” or based on a snapshot of
available data
Availability of input data impacts fidelity of output
•
•
•
•
Quantity of data
Granularity of data
Quality of data
Examples of data sources:
•
•
•
Less specific: M2
More specific: Local data pulls (CHCS, Operating Room
systems, Essentris)
FOR OFFICIAL USE ONLY
Custom to needs: Manual
data collection
7
Electronic Data Mining
•
Reasons for electronic data mining
•
•
•
Existing processes can be described by the data they generate
Allows for massive sample sizes
Examples of electronic data sources
•
•
•
Coarse information: M2
Detailed information: CHCS, Essentris, S3, etc.
Custom information: data from specially built applications
such as a discharge board or triage board
FOR OFFICIAL USE ONLY
8
Manual Data Collection
•
Reasons for manual data collection
•
•
•
Examples of manual data needs
•
•
•
Validation of system data
Needed data not available in system
Process times
Hourly volume counts
Processes for manual data collection
FOR OFFICIAL USE ONLY
9
Tool Parameters
•
Queue types
•
•
Levers
•
•
•
•
Resources such as staff or rooms
Volumes
Process changes including sequencing and scheduling
Process types
•
•
What are processes for serving customers?
Can be statistically described
Acceptable service levels
•
•
What is acceptable wait time?
What is acceptable turn away rate?
FOR OFFICIAL USE ONLY
10
Tool Parameter Examples
•
Scheduling modalities
•
•
•
•
•
•
•
Number of queues
Number of servers
Rooms
Staff
Service Levels
•
•
•
Process types
•
Capacity
•
Queuing
•
•
Templates
Random
Shortest case first
•
•
Wait time
ALOS
Resource utilization
Patients deferred
Sequencing and location
of process steps
FOR OFFICIAL USE ONLY
11
Spreadsheet Tools
•
•
•
•
•
Software tools widely available
Users familiar with interface
Distributable
Ability for modification
Less robust when representing dynamic systems
FOR OFFICIAL USE ONLY
12
Birth Volume Tool
•
What is the primary question?
•
•
Who are expected tool users?
•
•
“How many births should be expected in catchment area X?”
High level decision makers, not “in the weeds”
What decisions will be made with this tool?
•
•
System-wide planning
What-if ? analysis
FOR OFFICIAL USE ONLY
13
Birth Volume Tool
•
Solution implemented: Spreadsheet Model
Select whether
you want to use
the baseline for
% direct care,
population
change and
fertility rate. If
not, select “No”
and enter data in
blue cells
Learn more at
Population Based
Healthcare
Delivery Design
(II) Session:
Wednesday
1100-1200
Forecast
Results
14
FOR OFFICIAL USE ONLY
OR Volume Tool
•
What is the primary question?
•
•
Who are expected tool users?
•
•
“How do changes in volume and capacity impact OR
utilization”
Experienced analysts with report outs to decision makers
What decisions will be made with this tool?
•
•
System-wide planning
What-if ? analysis
FOR OFFICIAL USE ONLY
15
OR Volume Tool
Click to
Include
Solution implemented: Spreadsheet Model
Procedures Click to Adjust
Have Same %
Procedures
a
a
a
a
a
a
a
a
a
a
a
a
a
a
MOR Service Category
Your
Considered
Volume
Average Case
Time (hours)
Percentage
Accepted
Accepted
Volume
Total Case
Room Hours
Percent of Hours
q
Cardiothoracic
494
4.80
100%
494
2373
IIIIII
q
General Surgery
3314
2.99
50%
1641
4910
IIIIIIIIIIIII
60,000
q
GYN Surgery
1332
2.58
50%
666
1720
IIII
55,000
q
Neurosurgery
741
3.57
100%
741
2643
IIIIIII
50,000
q
Obstetrics
607
2.18
97%
592
1291
III
45,000
Average Historic Volume
53018.5(Step 1) 36144.0
q
Ophthalmology
885
1.76
50%
439
773
II
40,000
Considered Volume
53018.5
(Step 2)
36144.0
q
Orthopedic Surgery
2693
3.24
50%
1337
4328
IIIIIIIIIII
35,000
Accepted Volume 36116.3
(Step 4)
36144.0
q
Other Diagnostic_Therapeutic
453
1.93
34%
154
298
q
Otolaryngology
1184
2.68
50%
587
1572
IIII
q
Plastic_Recon Surgery
831
2.70
50%
414
1118
III
q
Unmatched
70
2.35
14%
10
22
q
Urology_GU Surgery
1265
3.26
50%
637
2077
IIIII
q
Vascular
607
3.67
65%
392
1439
III
q
VIP and WW
3042
3.80
100%
3042
11552
IIIIIIIIIIIIIIIIIIIIIIIIIIIIIII
q
Room Case Time in Hours
•
Chart data
Case Room
Hours
36,144
30,000
25,000
20,000
15,000
10,000
5,000
0
Average Historic
Volume (Step 1)
q
Target
Room
Considered Volume Accepted Volume
(Step 2)
(Step 4)
q
Total Accepted
Volume (count)
11,145
Target Utilization
(in Hours)
36,144
Total Accepted Case
Room Hours
36,116
Current scenario enables all case volume to be scheduled. There are 27
hours avilable to schedule more cases.
16
FOR OFFICIAL USE ONLY
OB Bed Sizing
•
What is the primary question?
•
•
Who are expected tool users?
•
•
“What is the facility capacity?”
Experienced analysts with report outs to decision makers
What decisions will be made with this tool?
•
•
Setting facility volume targets
Construction needs
FOR OFFICIAL USE ONLY
17
OB Bed Sizing
•
Solution implemented:
•
•
•
•
Spreadsheet model
Spreadsheet model has many assumptions
Assumptions reduce data needs while providing
reasonable answer
Model can highlight the need for additional analysis if
results are borderline
FOR OFFICIAL USE ONLY
18
OB Procedure Scheduling Model
•
What is the primary question?
•
•
Who are expected tool users?
•
•
“How should inductions and c-sections be scheduled to level
demand”
Experienced analysts with report outs to decision makers
What decisions will be made with this tool?
•
A recommended schedule template for c-sections and
inductions
FOR OFFICIAL USE ONLY
19
OB Procedure Scheduling Model
•
Solutions implemented
•
Analytical model produces a new template for scheduled
procedures that levels day of week peaks
FOR OFFICIAL USE ONLY
20
Simulation Tools
•
•
•
•
•
Software tools not widely available
Limited familiarity with tools
Very robust representation of dynamic systems
Can require extensive data as inputs
Can provide very granular results
FOR OFFICIAL USE ONLY
21
OB Simulation Model
•
What is the primary question?
•
•
Who are expected tool users?
•
•
“What is the capacity of our facility?”
Experienced analysts with report outs to decision makers
What decisions will be made with this tool?
•
•
Setting facility targets
Construction needs
FOR OFFICIAL USE ONLY
22
OB Simulation Model
•
Solution implemented:
•
•
Additional analysis:
•
•
Simulation identified the bottleneck was MBU beds
“Assuming process and facility reconfigurations that alleviate
the bottleneck, what is the new capacity”
Solution recommended:
•
Reclaim non-used LDRP rooms and keep vaginally delivered
patients in the same room for entire stay
FOR OFFICIAL USE ONLY
23
Pharmacy Simulation Model
•
What is the primary question?
•
•
Who are expected tool users?
•
•
“What processes/equipment are required to meet
performance criteria at a given volume”
Experienced analysts with report outs to decision makers
What decisions will be made with this tool?
•
•
•
Staffing decisions
Facility decisions
Equipment decisions
FOR OFFICIAL USE ONLY
24
Pharmacy Simulation Model
•
Solution implemented:
•
•
•
Simulation model
Complex input needs
Complex output processing needs
FOR OFFICIAL USE ONLY
25
Simulation User Interface
•
Run all possible scenarios and present results in
spreadsheet model
•
•
Method for reducing
demands on users
Limit input requirements
for front line
FOR OFFICIAL USE ONLY
26
Dynamic Decision Support Tools
•
OB Forecasting Tool
•
•
•
Answers the question “Should I accept or defer this patient”
Used by front line staff
Dynamically updated with live data
FOR OFFICIAL USE ONLY
27