nicu length of stay initiative: lessons learned
TRANSCRIPT
NICU Length of Stay Initiative:Lessons Learned
Rick McClead MD MHA
Professor and Vice-chairman
Department of Pediatrics
Division of Neonatology
In the beginning….I have this
great idea!
C.O.P.E. can
lower LOS in
the NICU by 8
days! I only
need $10K!
What’s your
idea?
Dr. McClead,
Make it
happen!
You gotta be $#!^ing me!
(Spring 2009)Dr. Brilli,
Make it
happen!
Benefits of COPECost Savings / Decreased Length of NICU Stay for Infants
4 day shorter length of NICU stay for preterms 26-34 weeks 8 day shorter length of NICU stay for preterms under 32 week
Improved Parent Outcomes Less stress in the NICU Stronger beliefs/confidence in their ability to care for their perterm infants More developmentally sensitive interactions with their preterm infants Less depression and anxiety symptoms during and after NICU stay
Greater satisfaction with the NICU stay Higher readiness for their infant's discharge from the NICU
Evidence-Based Educational-Behavioral Program
How can I reduce NICU LOS with only one intervention?
You Can’t!
Lesson # 1
• You must understand your data!
• What does you population look like?
• What factors contribute to long length of stay, and which factors can you impact?
• What is a long length of stay?
• Where are the data to answer these questions?
Two Populations Contribute to LOS
What is a long LOS?What are the contributing factors to a long LOS?
What is the Average LOS?
Why are infants hospitalized longer than 37 weeks CGA?
Not Statistically Significant VON DataAdmit by DOL 3 or DOL 7RaceHispanic ethnicityNo PNCDelivery ModeMultiple GestationHFVRDSSteroids for BPDIndocin therapy for PDANECNEC SurgeryPVLFungal SepsisLate Bacterial SepsisIVHHome O2/monitor
Statistically Significant* VON DataLower EGAInbornNo Antenatal SteroidsMale GenderPDAPDA LigationSevere ROP/ ROP SurgeryOther SurgeryPneumothoraxGI PerforationCongenital Malformation (Gastroschisis)CONS SepsisOxygen at 36 weeks CGANAS
* Items in red could be influenced by QI team
Where do we start?
Start with engaged staff!
Lesson # 2
• Start with something that already is of interest to the staff, and has a physician and nurse champion!
• COPE
• Neonatal Abstinence Syndrome
• Gastroschisis
• Feeding failure
• Palliative Care
How do I organize the QI work?
Decide on a QI methodology
Lesson # 3
• Choose a QI infrastructure!
• Model for Improvement
• Six-sigma/Lean
• Toyota Production System
• FOCUS-PDSA
What are we trying toaccomplish?
How will we know that achange is an improvement?
What change can we make thatwill result in improvement?
Model for Improvement
Act Plan
Study Do
Repeated Use of the PDSA Cycle
Hunches Theories
Ideas
Changes That Result in
Improvement
A P
S D
A P
S D
Very Small
Scale Test
Follow-up
Tests
Wide-Scale
Tests of Change
Implementation
of Change
PDSA Cycles for TestingWhy Test?
• Increase your belief that the change will result in improvement
• Document how much improvement can be expected from the change
• Learn how to adapt the change to conditions in the local environment
• Evaluate costs and side-effects of the change
• Minimize resistance upon implementation
Lesson # 3
• Choose a QI infrastructure!
• Model for Improvement
• Six-sigma/Lean
• Toyota Production System
• FOCUS-PDSA
• Create a specific SMART aim and a key driver diagram
Specif ic
___What is the goal or intent. Precisely and concisely describe what is to be achieved. It MUST focus on achieving only ONE thing.
Measurable
___There is a direct relation between the increase and the decrease of a measure and the attainment or loss of the goal. Recommend: Start the Aim Statement with Increase/Decrease…then describe the object of what is to be measured
___There are means with which to measure and monitor progress over time (to take, collect, and record the measurement)
Actionable
___The team can take action to overcome any known barriers to achieving the proposed measurable results
___The ‘HOW’ of achieving this goal in NOT part of the Aim Statement (this would restrict other plausible solutions)
Realistic
___Given the resources available, it is within the teams ability to achieve, control, or influence the Aim’s attainment
___There is no significant that will compete with the time, attention or ability to achieve the goal
Timely
___The goal has a target date. If timeline is beyond 6 months there are interim Milestones. Recommend: …achieve intent by a specified date ___There is nothing that should compete with the time and attention
needed to achieve the goal
Worksheet for Creating a SMART Aim Statement
Key Driver Diagram
• This Learning Structure helps everyone on the team from the physician to the housekeeper understand the approaches, recommendations, and processes that have to be followed to reach the goal.
• The diagram identifies and spells out acceptable practice for providing care and creates a common language that everyone is expected to follow.
• Without these interventions, you are less likely to achieve the goal.
• To the left of the key drivers are the components of “how” you will get there.
• Using a diagram like this as a visual can help to focus staff and management on the important and necessary things for reaching a positive outcome.
• http://www.stanleyhealthcare.com/sites/www.stanleyhealthcare.com/files/pdf/Aug08.pdf
Global Aim & Key Drivers
Reduce ALOS for all infantsadmitted to Main Campus Neonatal Services from a baseline* of 42.4 days to 28.4 days by Sept 30, 2010.
Key Drivers
Design Changes/Interventions
Global Specific Aim
Team Communication
Care of ELBW Infants
Care of NAS patient
Care of Gastroschisis
Weekly meetingsMonthly meetingsQI Companion
Establish small baby teamDevelop sub-aim with KDD &
interventions
Establish NAS QI TeamDevelop KDD with interventions
Establish Gastroschisis TeamDevelop sub-aim with KDD &
interventions
* Average LOS from Jan-Sept 2009
Balancing measures
Mortality rate
Readmission rate
ED visit rate
Care of lethal anomaly Engage palliative care groupDevelop sub-aim with KDD &
interventions
Aim & Key Drivers for SBP
Reduce ALOS for infants < 27Weeks EGA admitted to Main Campus Neonatal Services from a baseline* of 95.5 days To 90.5 days by December 31, 2010.
Key Drivers
Design Changes/Interventions
Specific Strategic AimQI Team Communication
Parent Empowerment
Severity of illness on admission
Feeding Failure
Weekly meetingsMonthly meetingsQI Companion
Implement C.O.P.E. in SBPP.E.E.P. all main campus unitsParent education complete before DOD
Eliminate hypothermia (done)Standardize management of the PDA(MCSA and RMH)-doneMinimize surgical ligation
Establish Feeding Failure Team-6/10Standardize feeding regimen-Conduct weekly feeding rounds-6/10Monitor feeding milestonesSpread to C4 from J4- in progress
* Average LOS from Jan-Sept 2009
Balancing measures
Mortality rate
Readmission rate
ED visit rate
Inadequate nursing documentation
Survey doctors for expectationsStandardize Nursing documentation
That is great, but how do I know when I have actually improved?
Look at your data
Lesson # 4• Use charts and graphs to visualize your data
Run Chart Frequency Plot
Waiting Time for Clinic Visit
20
25
30
35
40
45
50
55
60
Avera
ge D
ays
Distribution of Wait Times
0
10
20
30
40
50
60
5 15 25 35 45 55 65 75 85 95 105
Wait time (days) for Visit
nu
mb
er
of
vis
its
Pareto ChartScatterplot
Lesson # 4
• Use charts and graphs to visualize your data
• Understand variability using annotated statistical process control charting
• Improved Visibility into Process & Outcomes• What kinds of problems are most prominent?
• How is our performance changing over time?
• What impact have our actions or interventions had?
All Processes Vary!
Common Cause Variation
Unavoidable variation inherent in the system. This will affect numbers of
medication errors, falls, BSIs, etc. (within limited ranges). The amount of variation depends upon the system’s current characteristics, but a certain amount will always exist.
The point-to-point ups and downs in common cause variation are RANDOM.
They do not have particular causes. Looking for cause is generally futile. And attributing a cause, however tempting that may be, is risky at best and can lead to misdirected efforts.
Common cause variation on a control chart is what generally appears between the control limits (with the exception of certain “special cause” patterns).
All Processes Vary!
Special Cause Variation
Variation that is not an inherent part of the system or process. It will generally
arise from specific circumstances. Examples of potential special causes are:
Faulty Materials or Equipment
Unusually High/Low Patient Volume or increase/decrease in the
proportion of high-risk patients.
Unusual Staff Shortage
Sudden Influx of New Staff
Non-standardized practices (e.g., differences across shifts).
Special cause is indicated on a control chart by points outside the control limits or by certain patterns, for which we have several “rules of thumb” for recognizing.
When do you have special cause variation?
1 data point > 3 δ from centerline
9 point in a row, above or below centerline
6 points in a row, all increasing or decreasing
2 of 3 points > 2 δ from center line
15 points in a row within 1 δ of centerline
Specific Aim: Reduce ALOS for infants < 27 Weeks EGA admitted to Main Campus Neonatal Services
from a baseline* of 95.5 days To 90.5 days by December 31, 2010
Apr-2011Jan-2011Oct-2010Jul-2010Apr-2010Jan-2010Oct-2009Jul-2009Apr-2009Jan-2009
300
150
0
DCMoYr
Sample Mean
__X=95.5
UCL=151.4
LCL=39.5
Apr-2011Jan-2011Oct-2010Jul-2010Apr-2010Jan-2010Oct-2009Jul-2009Apr-2009Jan-2009
150
100
50
0
DCMoYr
Sample StDev
_S=63.1
UCL=104.0
LCL=22.3
2
1
Xbar-S Chart of Main Campus Initial Length of Stay<= 27 Gestational Weeks - Jan 2009 to Date
Historical Mean and Control Limits from Jan-Aug 2009
Ave LOS 101 66 110.3 120 79.7 7.6 75.2 108.3 88.5 56.3 6.0 94 114.3 99.7 144.8 123.1 97.6 68 78.9 97.9 80.1 81.0 91.3 84.9 89.1 57.5 72.7 47.8D/C N = 10 4 10 19 13 9 11 10 14 8 1 17 14 12 11 15 9 5 11 6 13 11 12 12 13 14 7 12
SpecialCause Variation(9 points at or below centerline)
Lesson # 4
• Use charts and graphs to visualize your data
• Understand variability using annotated statistical process control charting.
• When you find special cause variation, investigate its cause.
0
50
100
150
200
250
300
350
400
450
500
5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90
Da
ys
Patients
SBP LOS 12/08/08-10/26/10 Survivors
De
c 0
8 (P
atie
nt 1
)
Ju
n 0
9 (P
atie
nt 9
)
Se
pt 0
9 (P
atie
nt 2
0)
De
c 0
9 (P
atie
nt 4
4)
Dates of Admission
COPE Starts
COPE Planning
Impact of COPE/PEEP on Neonates admitted to Small Baby Pod
COPE Training
No
v 0
9(P
atie
nt 3
6)
Ap
r 10
(Patie
nt 5
4)
Ju
n 1
0 (P
atie
nt 6
8)
PEEP Discharge Teaching
Feeding Runds
PEEP RN Documentation
Se
pt 1
0 (P
atie
nt 7
6)
Oc
t 10
(Patie
nt 8
1)
Ma
r 20
11
(Patie
nt 9
2)
Impact of Feeding Program on LOS for infants < 32 weeks PMA without Surgical Problems of Congenital Anomalies
Specific Aim: Reduce ALOS for all infants admitted to Main Campus Neonatal Services from a baseline of 42.4 days to 28.4 days by Sept 30, 2011.
Apr-2011Jan-2011Oct-2010Jul-2010Apr-2010Jan-2010Oct-2009Jul-2009Apr-2009Jan-2009
60
40
20
DCMoYr
Sample Mean
__X=42.47
UCL=62.26
LCL=22.67
Apr-2011Jan-2011Oct-2010Jul-2010Apr-2010Jan-2010Oct-2009Jul-2009Apr-2009Jan-2009
60
50
40
30
DCMoYr
Sample StDev
_S=50.04
UCL=64.13
LCL=35.95
6
6
11
11
1
1
1
1
Xbar-S Chart of Main Campus Initial Length of StayAll Gestational Weeks - Jan 2009 to Date
Historical Mean and Control Limits from Jan-Aug 2009
Ave LOS 44.6 34.9 48.4 51.0 41.2 33.9 36.7 47.4 39.0 29.0 31.8 41.9 43.9 46.3 45.9 44.2 42.6 28.3 34.3 35.2 34.3 29.3 45.3 37.3 44.7 29.3 30.6 28.1
D/C N = 52 53 56 75 66 66 65 70 75 69 79 49 70 59 62 67 81 60 73 75 69 69 54 73 55 56 61 58
Apr-11Jan-11Oct-10Jul-10Apr-10Jan-10Oct-09Jul-09Apr-09Jan-09
0.20
0.15
0.10
0.05
0.00
MoYr
Proportion
_P=0.0694
UCL=0.1694
LCL=0
P Chart of All Deaths
Tests performed with unequal sample sizes
Balancing Measure for Specific Aim: Reduce ALOS for all infants admitted to Main Campus Neonatal Services from a baseline of 42.4 days to 28.4 days by Sept 30, 2011.
D/C N = 52 53 56 75 6 66 65 70 75 69 79 49 70 59 62 67 81 60 73 75 63 69 54 73 55 56 61 58
Montly Discharges from the NCH NICU (Jan 2009-April 2011)
0
10
20
30
40
50
60
70
80
90
100
J09 F M A M J J A S O N D J10 F M A M J J A S O N D J11
Month
Nu
mb
er
of
Pa
tie
nts
APR-DRG Mortality Risk for Neonates admitted to Main Campus Neonatal Services
Sep10Jul10May10Mar10Jan10Nov09Sep09Jul09May09Mar09Jan09
1.65
1.50
1.35
1.20
Discharge Month
Sample Mean
__X=1.4755
UCL=1.6522
LCL=1.2988
Sep10Jul10May10Mar10Jan10Nov09Sep09Jul09May09Mar09Jan09
1.0
0.9
0.8
0.7
Month-Yr
Sample StDev
_S=0.8279
UCL=0.9531
LCL=0.7027
1
Tests performed with unequal sample sizes
Xbar-S Chart of Risk of Mortality
Lesson # 4
• Use charts and graphs to visualize your data
• Understand variability using annotated statistical process control charting.
• When you find special cause variation, investigate its cause.
• Track a balancing measure to make sure something untoward is not happening.
More on Lesson # 4
Use the right control chart to track your data
Nov-2010
Sep-2010
Jul-2010
May-2010
Mar-2010
Jan-2010
Nov-2009
Sep-2009
Jul-2009
May-2009
Mar-2009
Jan-2009
0.06
0.05
0.04
0.03
0.02
0.01
0.00
DCMoYr
Proportion
_P=0.00926
UCL=0.05162
LCL=0
P Chart of Mortality at RMH
Tests performed with unequal sample sizes
Mortality for all infants admitted to the NICU at RMH
Mortality for all infants admitted to the NICU at RMH
111098765432
300
250
200
150
100
50
0
Observation
Days Between Events
LCL= 0.000
UCL=282.695
CL= 70.300
G Chart of RMH Deaths
Lesson # 5
• When you find something that works to improve a process, spread the learning to another area!
Aim & Key Drivers for NAS
Reduce LOS of main campus NAS patientsfrom 31 to 24 days
by December 31, 2010
Specific AimNursing Assessment
Nursing Documentation
Weaning Protocol
Maternal Management
Education re patient assessment and scoring
Compliance Monitoring
Develop alternative Weaning protocol
Collaborate with OBGYNs
Key Drivers
Design Changes / Interventions
0
50
100
150
200
250
Mar April May JuneAugust Sept Oct NovDec Jan Feb Mar April MayJun JulyAugust Oct Nov Jan Feb Mar
LO
S (
Da
ys
)
Date
NAS Patients Discharged from Main Campus NICUs 3/2/09-7/12/10
X chart
LOS Data for Main Campus NAS Patients
Methadone Weaning Protocol
NAS Task Force
Convened
Morphine weaning protocol
24.1 days
Failed Morphine weaning
30.94 days
58 days
IUGR infant required only 6
days of morphine but prolonged W/U for hypoglycemia Surgical infant
required opioid for pain control; began
weaning DOL13
Specific Aim: Reduce LOS of main campus NAS patients from 24.1 to 18 days by Sept 30, 2011
Infant with CHD
0
20
40
60
80
100
120
140
160
180
200
Oct Dec Jan Feb Mar Apr May June July Aug Sept Oct Nov Dec Jan'2011Feb Mar Apr
LO
S (
Day
s)
Date
NAS Patients Discharged from GMC October 2009-January 2011
Morphine Protocol
Methadone Protocol
20102009
Black circle = Methadone wean
Blue triangle = Morphine wean
LOS Data for GMC NAS Patients
2011
Lesson # 6
• Be careful of “before” and “after” comparison
Aim & Key Drivers for Gastroschisis
Reduce LOS of uncomplicated Gastroschisis patients from
44 to 34 days by Sept 30, 2010
Specific AimGuideline
Feeding Delays
Discharge Preparation
Revise GuidelineObtain surgical supportMonitor guideline use& explainVariancesSurgical attending round 2x/wk
Develop Feeding Plan per Jadcherla et al
Develop & Implement Parent Road Map(Similar to COPE & PEEP)
Key Drivers
Design Changes / Interventions
Pain ManagementStandardized post-op protocolImplement protocol
0
20
40
60
80
100
120
Jan Feb MarApril MayJul AugSept Oct DecJanMar AprilJulAug Sept OctNov Dec Feb
Le
ng
th o
f S
tay
(D
ay
s)
Patients
LOS of Uncomplicated Gastroschisis Patients 1/2009-12/2010
43.8 days
BG Reeder: Silo
Feeding intolerance
Cholestasis, Malro
BB Hayes: Silo
Feeding intolerance
Cholestasis
New Protocol
Specific Aim: Reduce LOS of uncomplicated* Gastroschisis patients from 44 to 34 days by Sept 30, 2011
* Excludes premature infants < 34 weeks PMA and any infant with multiple surgeries (e.g. stenosis or atresia repairs)
Good
33.7 days
0
20
40
60
80
100
120
Jan Feb MarApril MayJul AugSept Oct DecJanMar AprilJulAug Sept OctNov Dec Feb Mar
Le
ng
th o
f S
tay
(D
ay
s)
Patients
LOS of Uncomplicated Gastroschisis Patients 1/2009-12/2010
43.8 days
BG Reeder: Silo
Feeding intolerance
Cholestasis, Malro
BB Hayes: Silo
Feeding intolerance
Cholestasis
New Protocol
Specific Aim: Reduce LOS of uncomplicated* Gastroschisis patients from 44 to 34 days by Sept 30, 2011
* Excludes premature infants < 34 weeks PMA and any infant with multiple surgeries (e.g. stenosis or atresia repairs)
Good
Lesson # 7
• When you introduce a “test of change,” monitor compliance.
Gastroschisis Timeline ReviewFirst Infant on Protocol
LOS=
44 days
Gastroschisis Review
Lesson # 8• Track the performance of the teams
1.0 Team has been formed; target population identified; baseline measurement begun.
1.5 Team is meeting, discussion is occurring. Plans for the project have been made.
2.0 Team actively engaged in development and discussion, but no changes have been tested.
2.5 Components of Change Package being tested but no improvement in measures. Data on measures are being reported.
3.0 Initial test cycles have been completed and implementation begun for several components. Evidence of moderate improvement in process measures.
3.5 Some improvement in outcome measures, process measures continuing to improve, PDSA cycles on multiple components of the Change Package.
4.0 Evidence of sustained improvement in outcome measures, halfway toward accomplishing all of the goals. Plans for spreading the improvements are in place.
4.5 Sustained improvement in outcome measures, 75% of goals achieved, spread to a larger population has begun.
5.0 Major change in all areas of the Change Package has occurred, all goals of the AIM have been accomplished, outcome measures at national benchmark levels, and spread to another department is underway.
Lesson # 9
• For big QI projects, you need senior leadership on board.
NICU LOS Reduction ProjectResults to Date
1. Overall ALOS for main campus NICU has decreased 7 days from 42 to 35 days
2. ALOS for infants < 27 weeks EGA admitted to main campus NICU has decreased 10 days from 95-85 days.
3. ALOS for main campus NAS patient has decreased 7 days from 31 to 24 days. ALOS for NAS patients at offsite nurseries has decreased to < 20 days.
4. ALOS for infants with uncomplicated gastroschisis has decreased ~ 9 days from 44 to 35 days.
5. Over 4,000 unnecessary patient days have been avoided since June 2010.
6. The estimated financial impact of the project to date is in excess of $6,000,000.
Aim & Key Drivers for Intestinal Injury
Reduce the incidence of NEC in infants
< 1500 g Birth weight from 7.4% to 3.8%* by
December 31,2012
Strategic AimFeeding Protocol
Transfusion Policy
Restrict feeds
Engagement by referring Hospital NICU/SCNs
Feed according to protocolFeed mothers own milk or
Donor breast milk exclusivelyPerform oral care with EBM
Avoid routine transfusions for Hb > 7 g/dL
Hold feeds during transfusion
Meet with medical directors of allReferring NICU/SCNs to obtain
Agreement to feeding protocol and Transfusion guidelines
Key Drivers
Design Changes / Interventions
* NEC @ NCH, 7.4%
CHCA Lower quartile, 3.8%