hospital flow case study: cincinnati children’s...

11/27/2016

1

Frederick C. Ryckman, MDProfessor of Surgery / TransplantationSr. Vice President – Medical Operations

Cincinnati Children’s HospitalCincinnati, Ohio

IHI Forum

Dec. 5, 2016

Hospital Flow Case Study: Cincinnati Children’s Hospital

This presenter has nothing to disclose.

James Anderson Center for Health Systems Excellence

James M. Anderson CenterFor

Health Systems Excellence

Cincinnati Children’s Hospital550 Bed Medical CenterAdmissions/Year – 30,848 Opt Visits 1.02 MSurgical Procedures – 32,000 cases28 OR’

’’

’s, 2 IR suites, Hybrid Cath lab8 OR Outpatient Surgery Center1.4 M sq. ft. Research Space 15,000 Employees

11/27/2016

2



ACCESS FLOWPATIENT SAFETY

CLINICALEXCELLENCE

REDUCE HASSLES

Health Care Delivery System TransformationStrategic Improvement Priorities and System Level Measures

TEAMWELLBEING

Risk Adjusted Cost per Discharge

Syst

em L

evel

Mea

sure

sFAMILY

CENTEREDCARE

3rd Next available

appointment

% of eligible patients

with delays

Discharge Prediction

and Execution

Growth

Prediction

Adverse drug events (ADE)

per 1,000 doses

Nosocomial infection rates: Bloodstream infection rate

Surgical site infection rate

infection rates:

VAP

Safe Practices

Serious Safety Events

Codes outside the ICU

rate/1000 days

Standardized PICU

Mortality Ratio –

Expected/ Actual

% use of

Evidence-Based Care for eligible

patients

Functional Health Status

Functional Health Status

Touch Time for Providers

Employee Satisfaction

Staffing

Effectiveness Staff

Physician Satisfaction

Voluntary

staff turnover rate

Accident rate for staff with Work days

lost

Overall Rating: Patient

Experience



What Do Patients “Hire” Us to ProvideWhat do they call “Value”

• Make the Right Diagnosis

• Deliver the Correct Therapy / Treatment

• Prevent Complications or Errors in Care

• Deliver Safe Care regardless of the Inherent Risks

• Get Me Home, Keep me at Home

• Respect my needs

• Give me my Money’s Worth

Outcomes

Safety

Patient / Family Experience

Value

This is all FLOW management –it is essential for SAFETY, PATIENT / FAMILY EXPERIENCE

and QUALITY DELIVERY

11/27/2016

3



“Flow” is a Safety Initiative

• Getting the “Rights” Right

• Right Diagnosis and Treatment

• Right Patient in Right Bed – Location

• Right Nursing Staff and Staffing Expertise

• Disease Specific Expertise

• Equipment Expertise

• Prediction Framework for Safety

• Requires ability to “Predict” future needs, and manage present capacity control variability

• Operations Management techniques to understand and manage variability are the key to success

BestCare

Model



Value Equation for Healthcare

Value = (Outcomes + Patient Experience) x Appropriateness

Cost + “Hassle Factor”

11/27/2016

4



Critical Flow Failure Recognition

0

20

40

60

80

100

120

140

0

20

40

60

80

100

120

1407/1

4/2

008 (

Wed-…

8/1

8/0

89/2

2/0

810

/27/0

812

/1/0

81/5

/09

2/9

/09

3/1

6/0

94/2

0/0

95/2

5/0

96/2

9/0

98/3

/09

9/7

/09

10

/12/0

911

/16/0

912

/21/0

91/2

5/1

03/1

/10

4/5

/10

5/1

0/1

06/1

4/1

07/1

9/1

08/2

3/1

09/2

7/1

011

/1/1

012

/6/1

01/1

0/1

12/1

4/1

13/2

1/1

14/2

5/1

15/3

0/1

17/4

/11

8/8

/11

9/1

2/1

110

/17/1

111

/21/1

112

/26/1

11/3

0/1

23/5

/12

4/9

/12

5/1

4/1

26/1

8/1

27/2

3/1

28/2

7/1

210

/1/1

211

/5/1

212

/10/1

21/1

4/1

32/1

8/1

33/2

5/1

34/2

9/1

36/3

/13

7/8

/13

8/1

2/1

39/1

6/1

310

/21/1

311

/25/1

312

/30/1

32/3

/14

3/1

0/1

44/1

4/1

45/1

9/1

46/2

3/1

47/2

8/1

49/1

/14

10

/6/1

411

/10/1

412

/15/1

41/1

9/1

52/2

3/1

53/3

0/1

55/4

/15

6/8

/15

7/1

3/1

58/1

7/1

59/2

1/1

510

/26/1

511

/30/1

51/4

/16

2/8

/16

3/1

4/1

64/1

8/1

65/2

3/1

66/2

7/1

68/1

/16

9/5

/16

To

tal

# o

f B

ed

Days

# o

f N

ew

Pa

tie

nt

Fa

ilu

res

Week Beginning

Weekly Critical Flow Failures

# of New Failures Total Failures (Bed Days)

Type of Control Chart: P Chart



Critical Flow Failure Recognition

0

20

40

60

80

100

120

140

160

0

20

40

60

80

100

120

140

160

10/5

/201

5

10/1

2/2

015

10/1

9/2

015

10/2

6/2

015

11/2

/201

5

11/9

/201

5

11/1

6/2

015

11/2

3/2

015

11/3

0/2

015

12/7

/201

5

12/1

4/2

015

12/2

1/2

015

12/2

8/2

015

1/4

/20

16

1/1

1/2

01

6

1/1

8/2

01

6

1/2

5/2

01

6

2/1

/20

16

2/8

/20

16

2/1

5/2

01

6

2/2

2/2

01

6

2/2

9/2

01

6

3/7

/20

16

3/1

4/2

01

6

3/2

1/2

01

6

3/2

8/2

01

6

4/4

/20

16

4/1

1/2

01

6

4/1

8/2

01

6

4/2

5/2

01

6

5/2

/20

16

5/9

/20

16

5/1

6/2

01

6

5/2

3/2

01

6

5/3

0/2

01

6

6/6

/20

16

6/1

3/2

01

6

6/2

0/2

01

6

6/2

7/2

01

6

7/4

/20

16

7/1

1/2

01

6

7/1

8/2

01

6

7/2

5/2

01

6

8/1

/20

16

8/8

/20

16

8/1

5/2

01

6

8/2

2/2

01

6

8/2

9/2

01

6

9/5

/20

16

9/1

2/2

01

6

9/1

9/2

01

6

9/2

6/2

01

6

10/3

/201

6

To

tal

# o

f B

ed

Days

# o

f N

ew

Pa

tie

nt

Fa

ilu

res

Week Beginning

Weekly Critical Flow FailuresOver the last 52 weeks

# of New Failures Total Failures (Bed Days)

Type of Control Chart: P ChartType of Control Chart: P ChartType of Control Chart: P ChartType of Control Chart: P Chart

11/27/2016

5



0

50

100

150

200

250

300

JU

L

AU

G

SE

P

OC

T

NO

V

DE

C

JA

N

FE

B

MA

R

AP

R

MA

Y

JU

N

# o

f P

ati

en

ts w

ith

a N

ew

Failu

re

Month of Fiscal Year

Critical Patient Flow Failures by Month

2009 2010 2011 2012 2013 2014 2015 2016

Last Update: 3/4/2016 by M. Ponti-Zins, James M. Anderson Center for Health Systems Excellence Data Source: MPS

Growth Requires Constant Efficiency Improvement



Key Drivers for Capacity Management

Management of Variability

Predictable CareDelivery

CapacityPrediction

CapacityManagement

Optimization ofFlow Delivery

Flow:SafetyMatching

Identify Patient Streams – Inpatient/Outpatient/ORManage System Variation

Best Practices, Analysis of ALOS and outliers,Standardize then Customize, Eliminate unnecessary care

Integration of simulation modeling and planning“Environmental Impact” Reports for growth programs

Simulation for design and patient placement“Environments Impact” Planning

Placement initiatives – D:C Matching plansDischarge prediction and planning,

Home Care, Parent Initiatives

Flow Failure Analysis, Predictive Risk Analysis

IHI Drivers CCHMC Initiative Operations Possibilities

D/C

Match

System

Re-Design

Shape /

Reduce

Demand

11/27/2016

6



• Decrease overutilization of hospital services

• Optimize patient placement to insure the right care, in the right place, at the right time

• Increase clinician and staff satisfaction

• Demonstrate a ROI for the systems moving to bundled payment arrangements

1. Redesign surgical schedules to improve throughput and to improve

smooth flow of patients to downstream ICUs and inpatient units

2. Separate scheduled and unscheduled flows in the OR

3. ED efficiency changes to decrease LOS

4. Decrease LOS in ICUs (timely consults, tests and procedures)

5. Decrease LOS on Med/Surg Units (case management for patients with

complex medical and social needs)

6. Advance planning for transfers to community-based care settings

7. Cooperative agreements with rehab facilities, SNFs and nursing homes

1. Proactive advanced illness planning

2. Development of palliative care programs (hospital-based and

community-based)

3. Reduce readmissions for high risk populations

4. Extended hours in primary care practices

5. Urgent Care and Retail Clinics

6. Enroll patients in community-based mental health services

7. Paramedics & EMTs triaging & treating patients at home

8. Greater use of clinical pathways and evidence-based medicine

9. Care management for vulnerable/high risk patient populations

10. Decrease complications/harm (HAPU, CAUTI, SSI, falls with harm) and

subsequent LOS

11. Redesign surgical schedules to create an predictable flow of patients to

downstream ICUs and inpatient units

Redesign the

System

Shape or Reduce Demand

Match

Capacity

and

Demand

Outcomes Primary Drivers Secondary Drivers Specific Change Ideas

1. Assess seasonal variations and changes in demand patterns and

proactively plan for variations

2. Daily flow planning huddles (improve predictions to synchronize

admissions, discharges and discharges)

3. Real-time demand and capacity problem-solving (managing constraints

and bottlenecks)

4. Planning capacity to meet predicted demand patterns

5. High census protocols to expedite admissions from the ED and manage

surgical schedules.

Relocate care in ICUs in accordance with

patients EOL wishes

Relocate care in Med/Surg Units to

community-based care settings

Relocate low-acuity care in EDs to

community-based care settings

Decrease demand for hospital beds

through delivering appropriate care

Oversight system for hospital-wide

operations to optimize patient flow

Flex capacity to meet hourly, daily and

seasonal variations in demand

Real-time demand and capacity

management processes

Reducing unnecessary variations in care

and managing LOS “outliers”

Service Line Optimization (frail elders, SNF

residents, stroke patients, etc.)

Improve efficiencies and throughput in the

OR, ED, ICUs and Med/Surg Units

Decrease demand for hospital beds by

reducing hospital acquired conditions

Decrease variation in surgical scheduling

Early recognition for high census and surge

planning

IHI Theory on Flow






CapacityPrediction

CapacityManagement


Flow:SafetyMatching


Evidence Based Best Practices, Analysis of ALOS/outliers,Standardize then Customize, Eliminate unnecessary care







D/C

Match

System

Re-Design

Shape /

Reduce

Demand

11/27/2016

7



Evidence Based Care

• Evidence Based Care Guidelines serve as an interface between rapidly evolving scientific information and busy clinical practices

• Developed by Inter-disciplinary teams – experts

• Implementation

• Awareness of recommendation to facilitate change

• Easy access to the Evidence

• Feedback on Outcomes

• Feedback on further improvements

• Culture of Improvement / Evidence Based Care



Integration – Priority – Practice - Plan

Prioritization – Goal = Exceptional, Safe, Affordable Care Every Child

Owner – Executive Leadership

Practice – “What we Do” – Essential Steps, Decisions and Actions

Owner – Clinical Leadership Teams – Departments / Divisions

Processes – “How we Do It” – Processes to execute to the goal

Owner – Operational Leaders – Sites of Care

Plan – “Implement the Processes - plan through application of process steps

Owner – Sites of Care leaders and clinical staff (MD / RN)

Front Line Implementation – “Just Do It” – Every day for every child

11/27/2016

8



Cytomegalovirus Prophylaxis

• 75% Decrease in CMV infection – liver/intestine transplants

• Decreased IV-IGG expense

Danziger-Isakov, Lara et al. CCHMC Integrated Solid Organ Transplant



Yearly SSI Patients - CCHMC

0

10

20

30

40

50

60

70

80

90

100

2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016

86

39

52 51 50

23

29

43 41

34

23 21

To

tal

SS

I P

ae

nts

Fiscal Year

1032SSI’’’’s

492SSI’’’’s

12 Years540

SSI’’’’sPrevented

11/27/2016

9



Yearly SSI Patients - CCHMC

0

10

20

30

40

50

60

70

80

90

100

2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016

86

39

52 51 50

23

29

43 41

34

23 21

To

tal

SS

I P

ae

nts

Fiscal Year

1032SSI’’’’s

492SSI’’’’s

12 Years540

SSI’’’’sPrevented

Case Average10 days LOS$27,000.00

Business Case5400 days LOS$14.58 million



Focus on Excellence – SSI Spines

Standardization for OutcomesMerging Evidence Based Care and Practice

11/27/2016

10



SSI Accomplishments

• Baseline rate: 4.4 SSIs/100 procedures, Current Rate: 1.7 SSIs/100 procedures – 60% reduction, 32% reduction in past 3 years - $17.4 Million from SSI alone

• Overall SPS - Estimated 6,686 fewer children harmed

• Since October 2009 - $121.4 Million saved in SPS Network

Toltzis P, O’Riordan M, Cunningham DJ, Ryckman FC, Bracke TM, Olivea J, Lyren A. A statewide collaborative to reduce pediatric surgical site infections. Pediatrics 2014. 134:1174-80.

http://www.solutionsforpatientsafety.org/our-results/



Getting to “Better”

• “Clinical Practice Guidelines” –systematically developed statements

that guide decision making…

• Widely developed, limited effectiveness

• There are times when it is advisable and appropriate to deviate from standard guidelines

Parikh K, Agrawal S. JAMA Ped 2015;169;301-302

Recommended = 0

11/27/2016

11



Getting to “Better”

• “Achievable Benchmarks of Care” (ABC’s)

• Actual performance at health care sites performing in the top 10%

Parikh K, Agrawal S. JAMAPed 2015;169;301-302



Getting to “Better” – ABC’s

• Hospital Site Performance

• Individual Performance

11/27/2016

12






CapacityPrediction

CapacityManagement


Flow:SafetyMatching


Evidence Based Best Practices, Analysis of ALOS and outliers,Standardize then Customize, Eliminate unnecessary care







D/C

Match

System

Re-Design

Shape /

Reduce

Demand



Pre-Smoothing ICU Elective Census

0

2

4

6

8

10

12

06

/01

/07

06

/03

/07

06

/05

/07

06

/07

/07

06

/09

/07

06

/11

/07

06

/13

/07

06

/15

/07

06

/17

/07

06

/19

/07

06

/21

/07

06

/23

/07

06

/25

/07

06

/27

/07

06

/29

/07

07

/01

/07

07

/03

/07

07

/05

/07

07

/07

/07

07

/09

/07

07

/11

/07

07

/13

/07

07

/15

/07

07

/17

/07

07

/19

/07

07

/21

/07

07

/23

/07

07

/25

/07

07

/27

/07

07

/29

/07

07

/31

/07

08

/02

/07

08

/04

/07

08

/06

/07

08

/08

/07

08

/10

/07

08

/12

/07

08

/14

/07

08

/16

/07

08

/18

/07

08

/20

/07

08

/22

/07

Nu

mb

er

of

Pa

tie

nts

in

IC

U B

ed

s

ICU Daily Elective Census Prior to ICU Model for Smoothing

ICU Daily Elective Patient Census Center Line - Mean Control Limits

11/27/2016

13



ICU Bed Availability – ICU Scheduling

Case Statistics by Category

CategoryTotal PICU

Days Case Count ALOS

Short 224.47 177 (61%) 1.27 (27%)

Medium 304.74 82 (28%) 3.72 (37%)

Long 302.56 31 (11%) 9.76 (36%)

Grand Total 831.78 290 2.87



ICU Admission Model – Elective CasesShort Stay Cases – Access Cap

# Cases on Schedule / DayLong Stay Cases

Fixed # Beds

11/27/2016

14



Predicting ICU Discharge



Critical Flow Failures

0

1

2

3

4

5

6

7

8

9

7/1

6/2

00

81

0/1

4/2

008

1/1

2/2

00

94

/12

/200

97

/11

/200

91

0/9

/200

91

/7/2

010

4/7

/20

10

7/6

/20

10

10/4

/201

01

/2/2

011

4/2

/20

11

7/1

/20

11

9/2

9/2

01

11

2/2

8/2

011

3/2

7/2

01

26

/25

/201

29

/23

/201

21

2/2

2/2

012

3/2

2/2

01

36

/20

/201

39

/18

/201

31

2/1

7/2

013

3/1

7/2

01

46

/15

/201

49

/13

/201

41

2/1

2/2

014

3/1

2/2

01

56

/10

/201

59

/8/2

015

12/7

/201

53

/6/2

016

6/4

/20

16

9/2

/20

16

12/1

/201

63

/1/2

017

5/3

0/2

01

78

/28

/201

71

1/2

6/2

017

2/2

4/2

01

8# o

f P

ati

en

ts w

ith

a N

ew

Failu

re

Daily Failures

Delayed or Canceled Surgery Due to Bed Capacity

0

1

2

3

4

5

6

7

8

9

7/1

6/2

00

81

0/1

4/2

008

1/1

2/2

00

94

/12

/200

97

/11

/200

91

0/9

/200

91

/7/2

010

4/7

/20

10

7/6

/20

10

10/4

/201

01

/2/2

011

4/2

/20

11

7/1

/20

11

9/2

9/2

01

11

2/2

8/2

011

3/2

7/2

01

26

/25

/201

29

/23

/201

21

2/2

2/2

012

3/2

2/2

01

36

/20

/201

39

/18

/201

31

2/1

7/2

013

3/1

7/2

01

46

/15

/201

49

/13

/201

41

2/1

2/2

014

3/1

2/2

01

56

/10

/201

59

/8/2

015

12/7

/201

53

/6/2

016

6/4

/20

16

9/2

/20

16

12/1

/201

63

/1/2

017

5/3

0/2

01

78

/28

/201

71

1/2

6/2

017

2/2

4/2

01

8# o

f P

ati

en

ts w

ith

a N

ew

Failu

re

PICU Bed Not Available for Urgent Use

Daily Failures

0

1

2

3

4

5

6

7

8

9

7/1

6/2

00

89

/14

/200

81

1/1

3/2

008

1/1

2/2

00

93

/13

/200

95

/12

/200

97

/11

/200

99

/9/2

009

11/8

/200

91

/7/2

010

3/8

/20

10

5/7

/20

10

7/6

/20

10

9/4

/20

10

11/3

/201

01

/2/2

011

3/3

/20

11

5/2

/20

11

7/1

/20

11

8/3

0/2

01

11

0/2

9/2

011

12/2

8/2

011

2/2

6/2

01

24

/26

/201

26

/25

/201

28

/24

/201

21

0/2

3/2

012

12/2

2/2

012

2/2

0/2

01

34

/21

/201

36

/20

/201

38

/19

/201

31

0/1

8/2

013

12/1

7/2

013

2/1

5/2

01

44

/16

/201

46

/15

/201

48

/14

/201

41

0/1

3/2

014

12/1

2/2

014

2/1

0/2

01

54

/11

/201

56

/10

/201

58

/9/2

015

10/8

/201

51

2/7

/201

52

/5/2

016

4/5

/20

16

6/4

/20

16

8/3

/20

16

10/2

/201

61

2/1

/201

61

/30

/201

73

/31

/201

75

/30

/201

77

/29

/201

79

/27

/201

71

1/2

6/2

017

1/2

5/2

01

83

/26

/201

8

# o

f P

ati

en

ts w

ith

a N

ew

Failu

re

Patients who Utilize an ICU bed b/c an Appropriate Bed is Not Available

0

2

4

6

8

10

12

7/1

6/2

00

89

/14

/200

81

1/1

3/2

008

1/1

2/2

00

93

/13

/200

95

/12

/200

97

/11

/200

99

/9/2

009

11/8

/200

91

/7/2

010

3/8

/20

10

5/7

/20

10

7/6

/20

10

9/4

/20

10

11/3

/201

01

/2/2

011

3/3

/20

11

5/2

/20

11

7/1

/20

11

8/3

0/2

01

11

0/2

9/2

011

12/2

8/2

011

2/2

6/2

01

24

/26

/201

26

/25

/201

28

/24

/201

21

0/2

3/2

012

12/2

2/2

012

2/2

0/2

01

34

/21

/201

36

/20

/201

38

/19

/201

31

0/1

8/2

013

12/1

7/2

013

2/1

5/2

01

44

/16

/201

46

/15

/201

48

/14

/201

41

0/1

3/2

014

12/1

2/2

014

2/1

0/2

01

54

/11

/201

56

/10

/201

58

/9/2

015

10/8

/201

51

2/7

/201

52

/5/2

016

4/5

/20

16

6/4

/20

16

8/3

/20

16

10/2

/201

61

2/1

/201

61

/30

/201

73

/31

/201

75

/30

/201

77

/29

/201

79

/27

/201

71

1/2

6/2

017

1/2

5/2

01

83

/26

/201

8# o

f P

ati

en

ts w

ith

a N

ew

Failu

re

Daily Failures

Psychiatry Patients Placed Outside of their Primary Unit

11/27/2016

15



Mental Health Impact on Flow






CapacityPrediction

CapacityManagement


Flow:SafetyMatching









D/C

Match

System

Re-Design

Shape /

Reduce

Demand

11/27/2016

16



Early Day BedsPICU, CICUCritical Units

Later Day BedsAll units

Specific BedD:C Match

Unit BedAwareness

Strategies for Patient Placement

Demand : Capacity Match Opportunistic with Forethought




• Various approaches to Discharge Management

• 1980’s – Keep it a Secret

• 1990’s – 2000’s Discharge goals• AM before 11 > 30-40%

• “Shift” goals

• 4 hour time block goals with prediction of “window”

• Reactive

• “Not Patient Centered”

• 2008 - Prediction

• 2013-14 – Discharge when Medically Ready

11/27/2016

17



Discharge Prediction Buckets




• Various approaches to Discharge Management

• 1980’s – Keep it a Secret

• 1990’s – 2000’s Discharge goals• AM before 11 > 30-40%

• “Shift” goals

• 4 hour time block goals with prediction of “window”

• Reactive

• “Not Patient Centered”

• 2008 - Prediction

• 2013-14 – Discharge when Medically Ready

11/27/2016

18



Timeline for DC when Medically Ready

• Criteria established at admission

• Nurse at bedside notifies service when Medical discharge criteria are met

• Discharge from floor in < 2 hours

• Review Length of Stay and Re-Admissions as balancing measures

Criteria Set

Admission to

Floor

Discharge

Criteria Set Tracked

Discharge

Home

Re-Adm & LOS

TrackedStaff

Discharge

Criteria Met

Nurse Notifies

Staff

Treatment

Protocol

Followed

Standardized

Criteria

Buy-In by Staff

Modify Rounding

Clear Discharge Criteria

Communication Family

Standardized

• Protocols for most Tx

• Evaluation Criteria

2 Hrs

Not about Speed – Now about Efficiency



Discharge When Medically ReadyKaren Tucker, Angela Statile, Diane Herzog, and Christy White

Increase percentage of

all HM patients who

have met* Medically

ready criteria who will

be discharged within

two hours of reaching

that goal* on A6S,

A6N, LA1W from

75%to 80% by June 30,

2014

Productivity: Optimize

use of facilities and

staff and improve

patient flow to achieve

20% greater utilization

of existing assets by

June 30, 2015

Criteria for Medically Ready Defined at

Admission

Agreement among HM attendings and nursing

staff

of discharge criteria for order set diagnoses and

general admissions (LOR 2)

1) 8 pm Huddle discussion re: early discharges

(LOR 2)

2) 0630 notification of patients ready for

discharge (LOR 1)

Standardized and modifiable order sets (LOR 2)

Feedback of data by HM team

In conference room and by email (LOR 1)

Identify and Mitigate Plans:

1) Transportation- census based (LOR 1)

2) Pharmacy- priority fills (LOR 2), Outpt

delivery to patient room (LOR 1)

3) Consults- proactive evaluation (LOR 2)

4) RT- process in PICU (LOR 1)

5) Home Health Care

Daily Feedback reports to RNs and MD’s with ID

and mitigation of process and outcome measure

failures (LOR 2)

Shared Ownership/ Accountability and Buy-In

Among Physicians and Nurses

Communication regarding prediction of

discharge and defined goals is ongoing through

the hospital stay

Evidence of Preoccupation with Failure

Performance by team is transparent

Potential Barriers to Discharge are Clearly

Articulated and Mitigation Plans Established

Performance Management (LOR 1)

Auto notification to resident team that patient

has met all criteria (LOR 2)

Clear expectations for Parents/ Families

11/27/2016

19



EMR Discharge Criteria: Physician View

© 2013 Epic Systems Corporation. Used with permission.



Poster in Resident Conference Room

11/27/2016

20



Discharge when Medically Ready

All Units



Service Level DC when Medically Ready

11/27/2016

21



Discharge Failure Reasons Comparison



Balancing Measures – Length of Stay

2.12

2.42

1.14

2.12.11.97

1.4

1.87

0

0.5

1

1.5

2

2.5

3

Unit 1 Unit 2 Unit 3 All

Ave

rag

e L

en

gth

of

Sta

y (

Da

ys)

Hospital Medicine Average Length of Stay

For patients with selected diagnosis

FY11 FY13

11/27/2016

22



Balancing Measures – Readmission Rate






CapacityPrediction

CapacityManagement


Flow:SafetyMatching









D/C

Match

System

Re-Design

Shape /

Reduce

Demand

11/27/2016

23



Prediction – Model for the Future

• Static Analytics

• Performing a ONE TIME analysis of processes with historical data in order to PREDICT what’s going to happen under certain circumstances.

• Critical care Bed Modeling for Growth

• Real-Time Prediction

• Performing ONGOING analysis of processes with latest available data in order to continuously PREDICT what’s going to happen under certain circumstances.

• RN Bedside Nurse Staffing Model



Critical Care Bed Predictions

11/27/2016

24



Critical Care Bed Growth Analysis



Forecasted Bed Needs - Advantages of Efficiency

Forecast

Time

Frame

Probability

of Full

Unit

PICU

Beds

CICU

Beds

ICU Bed

Needs

Combined

ICUs

Estimated

Savings

Year 2 10% 34 27 61 56 5

5% 36 29 65 58 7

3% 38 30 68 59 9

1% 40 33 73 62 11

Year 5 10% 35 31 66 60 6

5% 37 32 69 64 5

3% 38 34 72 66 6

1% 41 37 78 71 7

Year7 10% 36 33 69 64 5

5% 38 35 73 66 7

3% 39 37 76 68 8

1% 42 39 81 71 10

Estimated number of beds required for given probability of the unit being full.

POPULATION: Unscheduled Medical/Surgical, BMT, ENT Airway ICU Elective Cases, Heart Institute Patients

11/27/2016

25



Environmental Impact Studies

LUNG TRANSPLANTATION

Predicting Future Programmatic Needs



ICU Bed NeedsNumber of beds needed based on probability of having a full unit (5%, 2%, 1%, 0%) and the growth estimate.

Low/Conservative Mid-Range/Most Likely High/Aggressive Growth

1 Yr 3 Yr 5 Yr 7 Yr 10

Yr

1 Yr 3 Yr 5 Yr 7 Yr 10 Yr 1 Yr 3 Yr 5 Yr 7

Yr

10 Yr

5% 1 2 2 2 2 2 2 2 2 2 2 2 2 2 3

2% 2 2 2 2 2 2 2 2 2 3 2 2 3 3 3

1% 2 2 2 2 3 2 2 2 3 3 2 2 3 3 3

0% 2 2 3 3 3 3 3 3 3 4 3 3 3 3 4

11/27/2016

26



Predicted to Actual Pulmonary Bed Demand

Pulmonary demand appears to be in line with original

projections

PulmonaryPrediction

ActualMid Level Conservative Aggressive

Year 1 72.7 38.3 102.1 6.2

Year 2 80.4 58.4 139.4 148.4

Year 3 88.0 78.4 176.6

Year 5 230.7 121.8 473.5

Year 7 475.0 236.3 867.8

Year 10 874.8 461.2 1505.4

Predicted to Actual Pre-Op Bed Demand



Predicted to Actual ICU Bed Demand

Predicted to Actual ICU Bed Demand

ICU demand appears to be dependent on pre-transplant

severity and much higher than originally anticipated

ICU PredictionActual

Mid Level Conservative Aggressive

Year 1 23.9 12.5 54.4 348.5

Year 2 39.4 19.4 65.6 275.7

Year 3 54.9 26.3 76.7

Year 5 73.4 37.5 108.3

Year 7 107.4 70.8 155.3

Year 10 166.9 95.0 183.1

11/27/2016

27



Predicted to Actual Cardiology Bed Demand

Cardiology demand appears to be lower than original

projections

HeartPrediction

ActualMid Level Conservative Aggressive

Year 1 48.2 25.2 125.9 55.6

Year 2 68.9 41.1 154.4 40.4

Year 3 89.6 56.9 182.8

Year 5 157.6 78.1 249.4

Year 7 265.1 155.1 353.5

Year 10 373.5 207.8 356.6

Predicted to Actual Post-Op Floor Demand



Outpatient / Procedural Demand

11/27/2016

28



Outcomes and Observations

- Environmental Impact Assessment provided valuable information allowing for assessment and agreement across the hospital before program was initiated

- Answer questions about patient flow and placement

- Assess potential stress on existing resources

- Quantify demand and capacity needs (staffing, beds,

outpatient clinic rooms, PFT demand, OR demand)

- Requires assumptions and research for new programs

- As always – your results are only accurate if your assumptions are correct



Nurse Staffing and Hospital Mortality

• Tertiary Medical Center – 197,691 patients, 176,696 RN shifts, 43 hospital units

• Relationship between nurse staffing and patient turnover

• Risk of Death 2-3 % for each below target shift

• Risk of Death 4-7 % for every high turnover shift • Admissions, discharges, and transfers

• Risk of Death 12 % for each below target shift

• Risk of Death 15 % for every high turnover shift

• Independent Variables when considering risksNeedleman J. et al. N Engl J Med 2011;364:1037-45.

ICU

Patient

Non-ICU

Patient

1st 5 days

LOS

Staffing and Environment - Mortality

11/27/2016

29



Census Calculation

Tomorrow’s census =

Today’s census + Today’s admissions – Today’s discharges

Easy enough, right?

Developed more than 200 models:

17 units, 7 days a week, admissions, and discharges,

multiple input and output streams



Model Outputs

�-Available bed capacity

�-Midnight census

�-Scheduled admissions

�-Predicted admissions

�-Predicted discharges

�-Predicted demand (census + adm – disch)

�-Predicted overflow placement

11/27/2016

30



Staffing Prediction – Proactive Planning



Actual vs. Predicted Results

11/27/2016

31



Impact of Analytics

• Bed demand predictions facilitate staffing and overflow planning – right patient – right team

• ED admit predictions improved from 40% to 70% accuracy – resource allocation

• Encourages staff to more consistently predict and document estimated discharge date, which helps guide care – system efficiency

• Uncovers scheduling issues – efficiency and access

• One-stop source to determine where there is capacity (or lack thereof) to add services (infusions, etc.) –efficiency and utilization






CapacityPrediction

CapacityManagement


Flow:SafetyMatching









D/C

Match

System

Re-Design

Shape /

Reduce

Demand

11/27/2016

32



GARDiANS



Hospital Wide System for Safety

Floor Huddles ICU HuddlesED HuddlePeriOp Huddle

Institutional Wide Bed Huddle – Capacity Management

Individual Room / Floor / System Predictions – Capacity and Safety

Institutional Daily Operations Brief

System Prediction – Mitigation Strategy

Pharmacy Security

Pt. Transport Housekeeping

3 Times - Every Day

Facilities P.F.E.

Outpt, Home, Psych

11/27/2016

33



Flow Dashboard – Sites of Care



Patient Satisfaction

• Only 3-4% of 1 Million outpatient visitors rank our care in the lower half (0-6 of 10 pts)

• 35,000 patient per year

Great American Ballpark Paul Brown Stadium

42,319 65,535

11/27/2016

34



Understanding Outliers

I thought:

If I can get 80-85% of thisunder good control, that will

solve at least 85% of theproblem



N=14 >200 Days

N=25 >100 Days

N=50 >62 Days

N=6 >365 Days

N=297 cases < 62 days (85.6%)

Total N=347

11/27/2016

35



1,7

95 b

ed

days

1,3

39 b

ed

days

2,1

31 b

ed

days

2,3

40 b

ed

days

7,595 Bed Days – 50 Patients342 Acute Care Beds

22.2 Days – Total Hospital Census6.08% Yearly Hospital Census



Observations on Outliers

• In these predictive models, it is important to be right. What is really important is the cumulative magnitude of your errors.

• The “errors” are often the result of “big surprises”, not multiple small issues.

• Failure to meet the predictive model leads to progressive and increasing cumulative error, things rarely get better fast (the more you are off, the more you are off.)

• It is hard to offset the “surprise” errors with great prediction of the expected.

11/27/2016

36



Lessons Learned

• Building “Will” to work on Flow is a challenge

• When it works, it is not on anyone’s radar

• If it works for me, your problem is not my problem….

• When it does not work, somebody else should solve it

• Linkage Safety and Flow

• Speed vs Efficiency

• Work Backwards not just Forward

• Embrace Mathematics and Analytics

• Standardize processes and work flows



Make it Personal

• Don’t let the Data

Drown out the Dream

• Stories not Statistics

• Names and Faces

• Accountability is Personal & Group Responsibility

• Collective Mission/Vision

© Cincinnati Children’s Hospital Medical Center 2013

11/27/2016

37



Thanks !

hospital flow case study: cincinnati children’s...

Documents