printing - university of michigan
TRANSCRIPT
Printing: This poster is 48” wide by 36” high. It’s designed to be printed on a large-format printer.
Customizing the Content: The placeholders in this poster are formatted for you. Type in the placeholders to add text, or click an icon to add a table, chart, SmartArt graphic, picture or multimedia file.
To add or remove bullet points from text, click the Bullets button on the Home tab.
If you need more placeholders for titles, content or body text, make a copy of what you need and drag it into place. PowerPoint’s Smart Guides will help you align it with everything else.
Want to use your own pictures instead of ours? No problem! Just click a picture, press the Delete key, then click the icon to add your picture.
Problem Statement Solution Approach
Inputs Decisions
Impact/Results
Optimizing Nurse Staffing with Absenteeism Boying Liu1, Kayse Lee Maass1, Zhehui Wang1, Rama Mwenesi1,2, Mary Duck1,2, Hannah Schapiro1, and Mark S. Daskin1
1Department of Industrial and Operations Engineering, University of Michigan, 2University of Michigan Health System
High Variability in Nursing Needs
• Day–to-day
• Among units
Unit and Pool Nurse Absenteeism
• Approximately 20% at UMHS
Why Nurse Staffing
6:1 3.2:1 10:3 40:1
US Economy: 15.7 Trillion
Healthcare: 2.7 Trillion
In-Hospital Nursing:
250 Billion
Hospitals: 847 Billion
U. Mich Budget
6.6 Billion
Effects of Nursing Levels
↑ cost
Already ~10% of HC cost
Increased N:P ratio
Key Issues
Rank Cause Deaths/year Relative
1 Heart disease 652,000 665
2 Cancer 559,000 570
3 Stroke 144,000 147
4 Chronic Lower Respiratory Disease 131,000 134
5 Accidents 118,000 120
6 Preventable Medical Errors 98,000 100
7 Diabetes 75,000 77
8 Alzheimer’s Disease 72,000 73
What is the proper staffing level?
↑ patient care ↑ nurse satisfaction ↓ medication errors
• 5 pools of patient census data
- date range July 2005 - June 2013
- pool size from 3 to 13 units
• Nurse absenteeism: UMHS ~20%; Nationally ~7-10%
Data Analysis Results
• Day of week: Positive correlation
-Monday census is closely correlated to Sunday census
• Monthly: Positive correlation among consecutive months
• Annual: Positive correlation between 07/08 and 10/11
• Unit-to-unit: Positive correlation among most units
• Distributional analysis
-Data does not follow truncated Poisson distribution
Artificially simulating the process will be difficult due to lack of a
distributional form and correlations
2012 Historical Data • 4 Pediatric Units
• Unit Cost: 1
• Pool Cost: 1.0667
• Temp Cost: 1.1555
Cost Savings compared
to 80% staffing: -3.6% savings
- ~$9 billion in savings nationally
UMHS Data STRATEGIC –
How many nurses to hire in each
unit and the pool
TACTICAL –
Which nurses should work each
shift during a week
OPERATIONAL –
How to allocate pool nurses to
units; How many temps to hire
Distribution of nursing
needs by unit
Nurse preferences for
shifts; work rules
Distribution of nursing
needs by unit
Model Description
Solve no absenteeism model (𝑼𝒋 ∀𝒋 ∈ 𝑼𝒏𝒊𝒕𝒔, 𝑷)
Initialize staffing level w/
𝑼 𝒋 = 𝑼𝒋/(𝟏 − 𝑨𝒃𝒔 𝒓𝒂𝒕𝒆) ∀𝒋 and
𝑷 = P/(𝟏 − 𝑨𝒃𝒔 𝒓𝒂𝒕𝒆)
Solve absenteeism model, Sample # absent from binomial
distribution using 𝑼 𝒋 & 𝑷
Converge? Stop Y
Revise staffing level
N
How is Nursing Organized at UMHS?
Spatial and temporal demand correlations
Unit 1 Unit 2 Unit |J|
Temp nurses
0
10
20
30
40
50
60
70
80
90
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
Nu
mb
er o
f D
ay
s E
ach
Ye
ar
Nurses Needed
Unit 2
0
10
20
30
40
50
60
70
80
90
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
Nu
mb
er o
f D
ay
s E
ach
Ye
ar
Nurses Needed
Unit 3
0
10
20
30
40
50
60
70
80
90
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
Nu
mb
er o
f D
ay
s E
ach
Ye
ar
Nurses Needed
Unit 4
Pool Nurses
$$$ Flexible
$$ Some Flex
$ Inflexible
Absenteeism
Absenteeism
Optimal Staffing Level Without Absenteeism
0
2
4
6
8
10
12
14
16
0% 10% 20% 30% 40% 50% 60% 70% 80% 90%100%
Nu
mb
er
of
nu
rse
s
Salvage cost as percent of unit cost
Nurse staffing vs. salvage cost
Unit 1
Unit 2
Unit 3
Unit 4
Pool
E(Temp)
E(Salvage)
Conclusions • Optimal staffing must account for variability in demand, differences
in the cost of various types of nurses, and nurse absenteeism rates • Small percentage, but significant amount, of cost savings possible
Future Work • Refine absenteeism rates, perhaps by day of week or time of year • Account for seasonality of nursing demand, perhaps through another
layer of nursing staff between unit and pool nurses • Optimize assignment of units to pools • Work with UMHS nursing staff to implement findings
120
130
140
150
160
170
180
190
200
210
07/0
1/1
2
09/3
0/1
2
12/3
0/1
2
03/3
1/1
3
06/3
0/1
3
Cen
su
s
Date
Mott Census Daily per Fiscal Year: FY13
FY13
30 per. Mov. Avg. (FY13)
7 per. Mov. Avg. (FY13)
Min Unit nurse cost + Pool nurse cost + Temp nurse cost – Benefit of extra nurses s. t. • Hire enough nurses to adequately cover patient demand in each unit each day • All pool nurses are assigned each day
Inputs Decisions
Current Focus
Current Focus
Future Work
Solution Algorithm
• Sample results: Jan 2012- June 2013 • Relative Costs: -Unit Nurse: 1 -Pool Nurse: 1.10
-Temp Nurse: 1.19 -Salvage: 0.45
• 22% increase in total cost with absenteeism
Unit 1 Unit 2 Unit 3 Unit 4 Pool
No Absenteeism 7 6 8 10 1
20% Absenteeism 8 7 9 12 3
No Absenteeism vs. Absenteeism
U: # of unit nurses P: # of pool nurses