Frontiers - Capacity Management

8/12/2019 Frontiers - Capacity Management

1/58

Volume 20

Number 4

Summer

2004

FRONTIERSO F H E A L T H S E R V I C E S M A N A G E M E N T

Carol Haraden and Roger Resar

Suzanne S. Horton

Diana Henderson, Christy Dempsey, an

Debra Appleby

Richard S. Zimmerman

Matthew Lambert III

Leo P. Brideau

Capacity Managemen

BreakthroughStrategies for

Improving Patient

Flow


2/58


3/58

F

R

O

N

T

I

E

R

S

Editorial

Audrey Kaufman

Lead Articles

3 Patient Flow in Hospitals: Understanding and Controlling It Better

Potential benefits of improved flow include better clinical outcomes, improved patientsafety, greater patient and staff satisfaction, and improved financial performance.Carol Haraden, Ph.D., and Roger Resar, M.D.

17 Increasing Capacity While Improving the Bottom Line

Overcrowding of the ED is not all about the ED but is a symptom of a broken system.

It is linked to the admission, discharge, and transition processes throughout the orga-

nization.Suzanne S. Horton, R.N.

25 A Case Study of Successful Patient Flow Methods: St. Johns Hospital

St. Johns keys to success for patient flow include having strong senior leadership sup-

port involving key stakeholders and multidisciplinary members on teams, having

teams meet weekly to discuss successes and challenges, having a physician cham-pion(s) leading each team, and good communication to all staff.Diana Henderson, Christy Dempsey, and Debra Appleby

Commentaries

31 The Commentaries: A Summary

33 Hospital Capacity, Productivity, and Patient SafetyIt All Flows Together

Creating successful change is more apt to occur and be sustained when there is align-

ment of goals linking healthcare institution and practitioner for the benefit of the pa-

tient.Richard S. Zimmerman, M.D.

39 Improvement and Innovation in Hospital Operations: A Key to Organizational Health

For an organization to be successful it has to do more than just improve operations;

substantial innovation must also take place.Matthew Lambert III , M.D., FACHE

47 Flow: Why Does It Matter?

IHI can convene, support, cajole, and lead. But in the end, real improvement requires

hard work on the front lines. Senior leaders must cause severe discomfort with the

status quo and give frontline caregivers the time, tools, permission, and support to

create improvement.Leo P. Brideau, FACHE

F

R

O

N

T

I

E

R

S

Volume 20

Number 4

Summer

2004

FRONTIERSO F H E A L T H S E R V I C E S M A N A G E M E N TFrontiers of Health Services Management is committed to providing our readers with compelling,

in-depth features and commentaries that are of current importance to the practice of health services

management by drawing on the expertise of the best practitioners and scholars.

1


4/58

Laurence M. Merlis, FACHE, ChairPresident and CEO, Greater Baltimore MedicalCenter, Baltimore, MD

Barbara A. Arrington, Ph.D., FACHEAssociate Professor, St. Louis University,St. Louis, MO

Raymond G. Brazier, FACHECEO, Mesa View Regional Hospital,Mesquite, NV

Joseph F. Damore, FACHEPresident and CEO, Sparrow Health System,Lansing, MI

Annette V. DrennanPresident, The Specialty Hospital,Meridian, MS

Cathy E. Duquette, Ph.D.Senior Vice President, Hospital Association of

Rhode Island, Providence, RI

Vivian A. EchavarriaDirector, Ancillary Support Services, AlaskaNative Medical Center, Anchorage, AK

Elizabeth J. Freeman, FACHEDirector, VAPAHCS, Palo Alto, CA

Earl G. Greenia, FACHEAssistant Administrator, Irvine RegionalMedical Center, Irvine, CA

MAJ Heather A. Kness, CHEOperations Officer, Pacific Regional MedicalCommand, Tamc, HI

Brenda Stevenson Marshall, Ph.D.Associate Professor, Cleveland State University,Cleveland, OH

Frederic P. Skinner, J.D.Watertown, NY

Rick L. StevensExecutive Director, Support Services, CARITASHealth Services, Louisville, KY

Frontiers of Health Services Management(ISSN 0748-8157) is published

quarterly by the Foundation of the American College of Healthcare

Executives, One North Franklin Street, Chicago, Illinois 60606-4425.

Subscription rate: $85.00 per year (four issues) in the United States,

$95.00 in other countries; $23.00/single issue; multiple-year sub-

scriptions are not available. Periodicals postage paid at Chicago,

Illinois, and additional mailing o;ces.

POSTMASTER: Send address changes to Frontiers of Health Ser-

vices Management, Subscription Services, One North Franklin, Suite

1700, Chicago, Illinois 60606-4425.

2004. Correspondence on editorialmatters shouldbe addressed

to the Managing Editor in care of the editorial o;ce at Health Admin-

istration Press, One North Franklin Street, Suite 1700, Chicago,

Illinois 60606, (312) 424-2800. Address subscription orders, notice

of change of address, and questions about missing or defective issues

to Foundation of the American College of Healthcare Executives,

Subscription Services, One North Franklin, Suite 1700, Chicago,

Illinois 60606-4425. Send both old and new addresses, including zip

codes, and allow six weeks for processing. Requests for back issues orsingle issues should also be addressed to Subscription Services.

Checks should be made payable to Frontiers of Health Services

Management.

Authorization to photocopy items for internal or personal use, or

the internal or personal use for speci>c clients, is granted by Health

Administration Press for libraries and other users registered with the

Copyright Clearance Center (CCC) transactional reporting services,

provided that the base fee of $3.50 per article is paid directly to CCC,

27 Congress Street, Salem, Massachusetts 01970. ISSN 0748-8157/

90/$3.50. Frontiers of Health Services Management is available on

micro>

lm from University Micro>

lms, Inc., 300 North Zeeb Road,Ann Arbor, Michigan 48106. All rights reserved.

Claims for undelivered copies must be made no later than eight

months following month of publication. The publisher will supply

missing copies when losses have been sustained in transit and when

the reserve stock will permit. Opinions and views expressed in the

articles are those of the author and do not necessarily reflect those

of the Foundation of the American College of Healthcare Executives.

The paper in this publication meets the requirement of the ANSI

Standard z39.48-1984 (Permanence of Paper), effective with Volume

8, Number 1.

Frontiers of Health Services ManagementMaureen C. Glass, CHE, CAE

Publisher, American College of

Healthcare Executives

Chicago, IL

Joyce A. Sherman

Managing Editor

Health Administration Press

Chicago, IL

Audrey Kaufman, Editor

Janet Davis, Associate Editor

Health Administration Press

Chicago, IL


5/58

F

R

O

N

T

I

E

R

S

audrey kaufman 1

Editorial

editorial

What comes to mind when you have to go to a hospital emergency room, either as

the patient or as a companion to the patient? Other than the immediate illness,

most people think of crowded rooms and hours of waiting. Overcrowded and bot-

tlenecked emergency departments are a chronic problem that has now reached cri-

sis proportions. Emergency departments are so overcrowded that patients are

being parked in hallways for hours waiting for a hospital bed or are turned away

altogether and sent to another institution. This is not just frustrating for patients

and their companions, it is also unsafe, non-patient-centered care.

With the advent of an uninsured and underinsured middle class, and other

such trends, the number of patients coming into the hospital through the emer-

gency department has increased dramatically and continues to increase. Somehospitals are tackling this problem by building more spaceenlarging the emer-

gency department, adding more beds on the floors, building another wing. But is

this really a long-term solution?

This issue of Frontiers takes a look at the patient flow/capacity management

problem from a systems perspective. We have asked Carol Haraden, Ph.D., vice

president at the Institute for Healthcare Improvement (IHI) in Boston, and Roger

Resar, M.D., IHI fellow at the Mayo Health System in Eau Claire, Wisconsin, to

draw from their work with more than 60 hospitals, where they evaluated what fac-

tors are involved in achieving the smooth, timely flow of patients through hospital

departments and helped develop methods for improving flow. These authors sug-gest that reducing delays and bottlenecks in the emergency department depends

on assessing and improving flow between and among departments. Hospitals

must view the problem in terms of an interdependent system rather than indi-

vidual departments. Improving the flow in one area alone, increasing nurse

staffing ratios, and placing patients off service (in the hallways) have not solved

the bottleneck problem and could significantly increase the risk of harm to the pa-

tient. However, they argue, by managing the flow of elective surgeries, achieving

timely and efficient transfer of patients from the intensive care units to

medical/surgical units, and improving the flow of inpatients to long-term-care fa-

cilities, the emergency department will be able to more efficiently move patientsonto floors and into beds where they can get appropriate care.

Complementing the article by Haraden and Resar are two case studies. The

first is written by Suzanne S. Horton, R.N., director of nursing at Baptist Memorial

audrey kaufman, assistant director and acquisitions manager

at health administration press in chicago, has worked in

healthcare publishing for more than 25 years.


6/58

2 frontiers of health services management 20:4

Hospital in Memphis, Tennessee. The second comes from Diana Henderson, ex-

ecutive director for quality; Christy Dempsy, director of perioperative services; and

Debra Appleby, supervisor of quality resources at St. Johns Hospital in

Springfield, Missouri. Both hospitals sought to solve their patient flow problems

by joining IHIs IMPACT collaborative to improve patient flow and care delivery.Each case study describes the particular issues that these institutions were facing

and the specific strategies used to reduce delays and increase patient flow.

Quality and patient-focused care are at the heart of the initiatives described in

this issue of Frontiers. Keeping patients in the emergency department or its corri-

dors for hours, keeping patients in the intensive care unit longer than necessary,

and keeping surgical patients waiting for an unoccupied surgery suite is no longer

acceptablein each case the patient is put at risk. Furthermore, these are not sep-

arate problems: departments in the hospitals cannot be viewed as single units.

Each is an interdependent part of the system as a whole. The authors emphasize

this fundamental idea throughout their discussions.

ON A PERSONAL NOTE

This is the first issue of Frontiers that will be managed in-house at Health

Administration Press. For the past 20 years, the journal has had an academic edi-

tor, most recently Leonard Friedman, Ph.D., who leaves very big shoes to fill. As

the new editor from within, I and my colleague, Janet Davis, will do our best to

maintain the high standards for which this journal has become known and to de-

velop issues that are both thought provoking and useful to the busy executive.

Audrey Kaufman

E R R A T U M In the Spring 2004 issue of Frontiers, the amount that Caterpillar Inc. spent on

healthcare costs for employees, retirees, and their dependents in 2003 was incorrectly

stated in the Editorial as $500 billion. The correct amount spent was $500 million.


7/58

F

R

O

N

T

I

E

R

S

c ar ol h ar a de n an d ro ge r re sa r 3

Patient Flow in Hospitals:

Understanding and Controlling It Better

C A R O L H A R A D E N , P H . D . , A N D

R O G E R R E S A R , M . D .

Summary Because waits, delays, and cancellations are so common in

healthcare, patients and providers assume that waiting is an inevitable,

but regrettable, part of the care process. For years, hospitals responded

to delays by adding resourcesmore beds and buildings or more staff

as the only way to deal with an increasingly needy population.

Furthermore, as long as payment for services covered the costs, more

construction and more staff allowed for continued inefficiencies in the

system. Today, few organizations can afford this solution. Moreover, re-

cent work on assessing the reasons for delays suggests that adding re-

sources is not the answer. In many cases, delays are not a resource prob-

lem; they are a flow problem. The Institute for Healthcare Improvement

has worked with more than 60 hospitals in the United States and the

United Kingdom to evaluate what influences the smooth and timely flow

of patients through hospital departments and to develop and implementmethods for improving flow. Specific areas of focus include smoothing

the flow of elective surgery, reducing waits for inpatient admission

through emergency departments, achieving timely and efficient transfer

of patients from the intensive care unit to medical/surgical units, and

improving flow from the inpatient setting to long-term-care facilities.

leadarticle

carol haraden, ph.d., is vice president at the institute for

healthcare improvement (ihi) in boston, and roger resar, m.d., is

an ihi fellow at luther midelfort, mayo health system, in eau

claire, wisconsin.


8/58

4 frontiers of health services management 20 :4

Patients and providers regard

waits, delays, and cancellations as an

inevitable consequence of receiving

care. That the hospital is a complex or-

ganization often serves as an easy ex-planation, but one that results in de-

layed and unstructured improvements

regarding flow. With more hospitals

unable to take care of their commu-

nity healthcare needs, problems with

flow in particular have taken center

stage in boardrooms and newsrooms.

Although no individual hospital area

(now sometimes known as a mi-

crosystem) is designed to achieve op-

timal flow of patients, the emergencydepartment, intensive care unit, and

operating rooms and their related pre-

and postcare areas tend to be major

bottlenecks because they are noninter-

changeable resources and conduits for

much of care received in the hospital.

The Institute for Healthcare Im-

provement (IHI) has worked with

more than 60 hospitals in the United

States and the United Kingdom to

evaluate what influences the smoothand timely flow of patients through

hospital departments and to develop

and implement methods for improv-

ing flow. Reducing delays and unclog-

ging bottlenecks depend on assessing

and improving flow between and

among these departments. To im-

prove flow, hospitals must view the

problem in terms of an interdepen-

dent system rather than individual de-

partments. Any individual depart-ment that improves flow in its area

alone often in fact exacerbates the

problem for other dependent depart-

ments. For example, consider the

emergency department (ED) that tries

to improve flow through the ED by

moving patients into a hallway or

lounge while they wait for a bed on a

medical or surgical unit to become

available. Although this might seem

like a good solution, it ultimatelyworsens the situation for both the pa-

tient and the receiving floor. Studies

have shown that increased nurse

staffing ratios (Aiken et al. 2002) and

placing patients off service (D. Brailer,

CareScience, personal communica-

tion) have significantly added to the

risk for mortality and morbidity.

THE KEY TO UNDERSTANDING

FLOWOne of the first steps necessary for un-

derstanding flow is to accept that flow

depends on the inherent variation

found in the healthcare delivery sys-

tem. It is a common but an incorrect

assumption that healthcare flow is a

result of what appears to be the ran-

domness and complexity of disease

presentation: Who could possibly pre-

dict broken legs, heart attacks, or

strokes and then always have the rightresources ready for quick, timely, and

safe care? Providers and consumers

alike believe that waits, delays, and

cancellations are to be expected be-

cause of the presumed and logical in-

ability to predict and manage unsched-

uled and emergent demand. This

belief has been challenged by studies

on the effects of variability on the

healthcare delivery system. When ana-

lyzed, the variation introduced by thevery structure of the delivery system it-

self far outweighs the variation caused

by the randomness of patient arrivals

in the ED as well as the disease state

with which they present. Variation

from the randomness of disease,

Reducing delaysand unclogging

bottlenecks

depend on

assessing and

improving flow

between and

among

departments.


9/58

F

R

O

N

T

I

E

R

S


termed natural variation, can be ac-

commodated by managing demand

based on historical data and queuing

methods. Similarly, the natural varia-

tion introduced by differing levels ofstaff competency and clinical abilities

can and must also be managed. All or-

ganizations strive to maintain the

competency of staff to minimize this

type of variation. Generally, this varia-

tion is handled reasonably well by

state and federal regulatory agencies

or by processes within a hospital to re-

view the competency of staff and to

provide training. Natural variation can

affect flow but commonly plays asmaller role and tends to become rela-

tively constant in the system.

Variation introduced into the care

system from personal preferences

and beliefs of individual clinicians,

called artificial variation, cannot be

similarly managed and needs to be

eliminated. The population of surgi-

cal schedules or basis of decisions

made about when to admit or dis-

charge a patient is largely dependenton a wide variety of human partici-

pants, with many unknown inputs.

The effect of artificial variation on

flow far exceeds the effect of variation

resulting from random, highly com-

plex disease presentations. For this

reason, emphasis needs to be placed

on change concepts related to reduc-

ing the artificial variation in health-

care delivery. Litvak and colleagues

found that typically 50 percent of theadmissions to a hospital are from the

ED and 30 percent are for elective

surgery (McManus et al. 2003). The

elective surgical admissions have

more effect on flow than the random

admissions from the ED. How is it

possible that an elective set of admis-

sions can have more effect than auto-

mobile accidents or heart attacks? The

answer is related to the exceedingly

arbitrary nature of elective schedulingdecisions.

If a surgical schedule is viewed

several weeks ahead, the placement of

patients in the schedule has little rela-

tionship to the resources that might

be demanded when the patients are

actually admitted. Surgeons with no

information about the impact of ad-

missions, projected intensive care

unit (ICU) use, or other resource de-

mands tend to schedule cases by theslots available. When the slots are

filled, the hospital responds three

weeks hence in a reactive mode,

rather than predict and try to smooth

the demand on the resources.

One vital key to improving flow

lies in reducing variation in processes

related to flow. While some variability,

such as the types of patients coming

into the ED, is normal and expected,

other types of variation should not beexpected or tolerated; they require

elimination before additional beds are

built and staffed. Hospitals in partner-

ship with IHI have tested a broad

range of changes to reduce process

variation and improve flow. These

changes and measurements as well as

eventual goals are discussed in this ar-

ticle.

THE ED ONLY LOOKS LIKETHE PROBLEM

The emergency department increas-

ingly has faced waits for care, with

consequent delays in treatment, to a

point where diversions have become a

common part of ED language. Patients

leadarticle


10/58


who are increasingly frustrated with

the length of these delays and waits

are leaving EDs untreated. In the

United States, EDs experienced a 20

percent increase in patient visits overthe past decade (CDC 2003). Not sur-

prisingly, ED waiting times have also

increased. According to the Centers

for Disease Control and Prevention,

the average wait time for nonurgent

visits increased between 1997 and

2000 by 33 percent, from 51 minutes

to 68 minutes (CDC 2002).

Of those individuals who leave the

ED because of the lengthy wait times,

how many return to the hospital evensicker and in need of additional re-

sources is unknown and of concern.

Increasingly, hospitals are realizing

the community obligations of their

EDs and have decreed a no-diversion

status policy. Although this policy is

commendable, it does not solve the

problem; it only forces the problem

downstream in the hospital. Patient

safety and patient-centered care con-

tinue to be illusions.Diverting ambulances away from

hospitals that are at capacity does still

happen and is, in fact, on the rise in

most areas, although some communi-

ties have banded together to make

this practice a thing of the past. An

October 2001 U.S. government study

shows that ambulance diversions

have impeded access to emergency

services in metropolitan areas in at

least 22 states since January 1, 2000.More than 75 million Americans re-

side in the areas affected by these am-

bulance diversions (U.S. House of

Representatives 2001, i).

Examples abound, according to the

study.

In Tucson, Arizona, so many hospitals

diverted ambulances that paramedics

had to struggle to find any place to

bring patients. In the Boston area, am-

bulance diversions last year ran asmuch as ten times higher than in pre-

vious years. On some days in Atlanta,

eight to ten hospitals diverted ambu-

lances at the same time. In Los Ange-

les, two dozen emergency rooms at the

heart of the areas emergency system

were closed to ambulances almost one-

third of the time in June 2001. (U.S.

House of Representatives 2001, i)

The so-called ED problem, how-ever, is actually a systemwide prob-

lem. EDs do not exist in isolation, but

are part of a system of care through

which patients flow. Increasing capac-

ity in the ED to accommodate more

patients, a solution chosen by many

hospitals, is like broadening the large

end of a funnel without increasing the

capacity at the neck or constriction

point.

In a recent report on ED crowding,the U.S. General Accounting Office

(GAO 2003, 1) noted the connection

between the ED and the rest of the

hospital system:

While no single factor stands out as

the reason why crowding occurs, GAO

found the factor most commonly as-

sociated with crowding was the inabil-

ity to transfer emergency patients to

inpatient beds once a decision hadbeen made to admit them as hospital

patients rather than to treat and re-

lease them. When patients board in

the emergency department due to the

inability to transfer them elsewhere,

the space, staff, and other resources

Flow problemscannot be solved

simply by working

harder or by

adding beds and

staff.


11/58

F

R

O

N

T

I

E

R

S


available to treat new emergency pa-

tients are diminished.

Another recent study of ED over-

crowding also showed that the pri-mary reason hospitals go on diversion

is the lack of available critical care

beds (The Lewin Group 2002). An in-

teresting way to look at the dimin-

ished capacity is to measure for a

week the number of patients waiting

to transfer to a hospital bed. The aver-

age number of patients waiting repre-

sents the percentage of ED capacity

that is taken away by waiting.

Poor ED flow is not a trivial prob-lem; at a minimum, it has the follow-

ing costs:

Patient safety. The ED becomes an

inappropriate and expensive hold-

ing area when patients are not

transferred to an inpatient unit in a

timely manner. The patients can

best be described as off service and

parked (i.e., keeping or placing

admitted patients in a holding loca-tionsometimes in the ED, some-

times simply in a hallwaywhen

they cannot be moved immediately

to their intended bed or location).

The safety level of ventilated pa-

tients in an ED is without question

more tenuous than in an ICU; the

ED just is not the best place for this

type of care. Solving the ED prob-

lem by parking patients in hallways

on the receiving unit to await trans-fer to a bed in that unit is de-

plorable, but in todays world, park-

ing is part of the plan in many

hospitals. Who among us, espe-

cially when sick or vulnerable,

would like to be placed in a hallway?

Care not given. When the ED is

overcrowded because patients can-

not be transferred quickly to care

units or operating rooms, incoming

patients can experience harmful de-lays in receiving care, and many

leave without being treated. One

hospital described those patients as

only 1.3 percent of all patients regis-

tered. When the math was done in

terms of number of patients, how-

ever, such elopements involved

more than 1,000 patients. What

more complicated diseases did the

ED treat in the long run because of

the failure to treat initially? Reduced institutional revenues. Any

patient waiting for an inpatient bed

in the ED basically robs the hospi-

tal of ED capacity and, subse-

quently, robs the community of

service. The ED is a source of rev-

enue and an important care site for

patients; when it is not accessible

and patients are diverted, both the

patients future care and potential

revenue are lost.

Waits and delays, bottlenecks, and

backlogs are not the result of lack of

effort or commitment on the part of

staff. Their source lies in what

Berwick (1996) calls the first law of

improvement: ...[E]very system is

perfectly designed to achieve the re-

sults it achieves. The answer to im-

proving flow of patients lies in re-

designing the overall systemwidework processes that create the flow

problems. These problems cannot be

solved simply by working harder or

by adding beds and staff.

Optimal care can only be delivered

when the right patient is in the right

leadarticle


12/58


place with the right provider and the

right information at the right time.

Improvement efforts by hospitals in

the United States are showing that it

is possible to reduce waits and delaysin the hospital environment by en-

hancing the flow of patients and in-

formation throughout the care sys-

tem. Better flow will increase access;

reduce waiting times; lower costs;

and, ultimately, improve outcomes in

safety, throughput, and financial sta-

tus. The section that follows describes

one such program.

THE CHALLENGE FOR

HOSPITALS

The Institute for Healthcare Improve-

ment has developed a process and

methodology for hospitals to use in

evaluating and improving patient flow

in acute care settings. As part of its ef-

fort to foster improvements through-

out the healthcare system, IHI invites

hospitals to engage in this process,

using the methods described in the

following sections.

1. Evaluate Flow: How Much of the

Time Do You Get It Right?

The first step in evaluating the flow

of patients through your acute care

setting(s) is to find out, on average,

how much of the time your hospital

gets it right in moving patients

through the system in a timely and

efficient manner. In considering thisquestion, your hospital needs to look

at both the frequency of parking pa-

tients and hospital occupancy as key

indicators.

Two key questions help bring these

issues into focus:

1. Do you park more than 2 percent

of your admitted patients at some

time during the day at least 50 per-

cent of the time?

Example: In a hospital with amidnight census of 500 pa-

tients, 10 patients (2 percent)

were parked during the day,

waiting for admission to the

final destination bed. This oc-

curs more than half the time

during the sample period.

2. Does your hospital have a mid-

night census of 90 percent or

more of your bed capacity more

than 50 percent of the time?Example: A 500-bed hospital

had more than 450 patients in

the hospital at midnight (90

percent of capacity) more

than half the time during the

sample period.

If you answer yes to one or both

of these questions, your hospital is

likely struggling with flow problems

on a regular basis. Parking patients isa clear indication that the system is

inhibiting the smooth forward move-

ment of patients to their appropriate

destination. And if your midnight

census typically is high, you probably

experience capacity problems be-

cause your hospital is virtually full at

the start of the day, leaving little ca-

pacity for new admissions. To ad-

dress these issues, you will have two

tasks: (1) work to reduce flow varia-tion and (2) extend the chainthat

is, work with others along the contin-

uum of care, including those outside

your hospital, to smooth the flow of

patients into and out of your organi-

zation.

To addresscapacity issues,

you will have two

tasks: (1) work to

reduce flow

variation and (2)

extend the

chain.


13/58

F

R

O

N

T

I

E

R

S


Even if you answer no to both of

these questions but still feel that pa-

tients do not consistently move

smoothly through the system, this

may indicate a need to reduce flowvariation.

2. Measure and Understand Flow

Variation

Variation is intrinsic in healthcare. It

is the result of clinical variability(number of patients presenting with

certain clinical conditions),flow vari-

ability(the ebb and flow of patients ar-

riving throughout the day), andprofes-

sional variability(the variation in skilllevels and techniques among pro-

viders). Eugene Litvak (IHI 2003a)

suggests that only the following sce-

nario would (in theory) eliminate vari-

ability entirely:

1. All patients have the same disease

with the same severity.

2. Patients arrive at the same rate

every hour.

3. All providers (physicians andnurses) are equal in their ability to

provide quality care.

Some kinds of variability (so-called

random variability) cannot be elimi-

nated, or even reduced; they must be

managed. This is true of patient vari-

ability. We cannot eliminate the many

types of problems from which pa-

tients suffer, nor can we control when

they arrive in the ED.Nonrandom variability, on the

other hand, is often driven by individ-

ual priorities, resulting, for example,

in surgical schedules that are heavy on

Wednesdays but light on Fridays be-

cause of surgeons preferences rather

than actual demand. Nonrandom vari-

ability should not be managed; it

should be eliminated.

Volume, census, or occupancy

rates are often calculated and dis-played as means or averages. How-

ever, it is the variation in these metrics

that causes most of the flow problems

in our hospital systems. Consider this

example: The mean elective surgical

volume for two hospitals for one week

is 125 patient cases each. Hospital A

has a steady flow of surgical cases

throughout the week, allowing for op-

timal scheduling and predictable de-

mand for staffing and patient beds.Hospital B, which also has a mean of

125 cases, schedules 50 percent of its

cases on Mondays and Wednesdays

and 50 percent on the remaining days.

Because the caseload is so high on

Mondays and Wednesdays, there is no

room for the seemingly random but

historically predictable surgical com-

plications and added cases. The de-

mand for staff, beds, and equipment

is at a maximum. Any added volumeor decrease in capacity is felt quickly

as waits, delays, and cancellations.

A helpful exercise is to look at the

variation in census between each day

of the week and the variation in cen-

sus within each day. These measures

can point to different problems and

solutions. To assess within-day varia-

tion, an organization can measure the

midnight census and the midday cen-

sus. The difference between these twomeasurements gives an idea of the

variation within that day. A common

way to express this is the standard de-

viation between the midnight census

and the midday census for two weeks

(count only on weekdays). The cause

leadarticle


14/58


for the within-day variation is mostly

related to a lack of synchronization

between admissions and discharges.

For between-day variation, an orga-

nization can measure the variationbetween elective surgical admissions

and ED admissions expressed either

as a standard deviation or as residual

differences between days. The cause

for this variation is commonly due to

wide variations in elective surgical ad-

missions.

3. Test Changes to Improve Flow

Hospitals that want to improve flow

should consider testing two maingroups of changes:

1. Changes that can be made within

the hospital

2. Changes that result in cooperative

relationships with other healthcare

providers outside of the hospital

Changes Within the Hospital

Smooth the surgical schedule. The

elective surgical schedule is the majorsource of artificial variation in flow.

Several concepts, such as the follow-

ing, are being tested by several of the

60 hospitals working in the flow do-

main of the IMPACT collaborative1

and show early promise in smoothing

the elective surgical schedule by re-

ducing the demands on hospital re-

sources:

Smooth the number of elective sched-

uled cases. Scheduling the maximumnumber of elective surgeries without

regard to eventual downstream use of

resourceswhether it is a surgical-

bed or an ICU-bed caseleaves little

flexibility for the demands from a

fairly predictable ED admission

schedule or emergency surgical de-

mand. By smoothing the number of

elective surgical admissions, the de-

mands on downstream resources will

be known and predictable.Separate emergent surgery from elec-

tive surgery. Because the vast majority

of surgery is scheduled, most of the

operating room space should be as-

signed as such. Utilization of the

scheduled rooms then becomes pre-

dictable and controllable, and wait

times for unscheduled surgery be-

come manageable. Setting aside

rooms for emergent surgery as a sepa-

rate flow stream then does not resultin cancellations of elective surgery.

Additional benefits can be achieved by

designating a specific surgeon to do

the emergent cases if the volume will

warrant this approach. In most cases,

the separation of emergent and ur-

gent cases within the current call sys-

tem will be the first significant and

helpful step in improving flow.

Schedule the discharge. Admissionbottlenecks are often created because

discharges are not managed effi-

ciently. Creating a more consistent

and predictable discharge schedule by

taking the following steps can help

improve flow:

Provide a process for scheduling the

time that patients will be discharged. Al-

though the date and time of discharge

may be known a day in advance for

some patients, hospitals have histori-cally tried to build all discharge strate-

gies on predicting discharges. Data

suggest that discharges from surgical

floors can be predicted with consider-

able accuracy because of the pre-

dictability of the surgical procedures

In most cases, theseparation of

emergent and

urgent surgery

cases within the

current call

system will be the

first significant

and helpful step

in improving flow.


15/58

F

R

O

N

T

I

E

R

S

carol haraden an d roger resar 11

themselves. In fact, early data indicate

that nurses, doctors, and other health-

care providers can usually predict one

day in advance, with more than 80

percent accuracy, which patients on asurgical unit will be discharged the

following day. A process to schedule

discharge times and dates with these

patients is therefore relatively easy.

The opposite is true on medical

units. Frequently, the discharge is de-

pendent on tests to be run that day. In

those units the ability to schedule the

time for discharge depends less on

predicting the day before and more

on the ability of the unit to respond toa discharge order and give a patient a

reasonable discharge time. Regard-

less of whether discharge time is pre-

dicted the day before or the same day,

scheduling the discharge time should

lead to optimizing and planning the

discharge, and, eventually, the dis-

charge times for transferred or admit-

ted patients from other parts of the

hospital will also be synchronized to

that schedule. This work will mostlikely require a centralized planning

and scheduling function. Planners

can record data about the ability of

the system to comply with the sched-

ule and document reasons for non-

compliance to identify bottlenecks

and processes that need improve-

ment.

Orchestrate the discharge. A series of

tasks must be performed in a specific

order prior to discharging a patient.The ancillary services (e.g., lab, phar-

macy, physical therapy), however,

work best without rigid schedules.

Staff who perform these services can

self-organize and schedule the work

to achieve a common goal: setting the

prearranged discharge time, whether

it is known the day before or the same

day. Orchestrating the discharge al-

lows all care providers to schedule

their work accordingly so that a con-tinuous flow of patients can be dis-

charged throughout the day and into

the early evening.

Provide a process and a team for dis-

charging patients with more complex is-

sues, using data from discharge coordina-

tors. Not all patients can be easily

discharged. Because of the complexity

of their disease, lack of support at

home, or psychosocial problems,

some patients are difficult to place inappropriate settings after discharge.

Although the time and date of dis-

charge should be scheduled for these

patients as for others, the orchestra-

tion of the discharge may need to be

handled separately and differently.

This may require a special person or

team capable of crafting customized

solutions to meet the discharge needs

of these patients. The complexity of

the patient should not allow regres-sion to a chaotic and unplanned dis-

charge.

Synchronize other movements to the

discharge schedule. Once an organized,

preplanned set of discharge times is

in place on each hospital unit, internal

transfers of patients from an ICU to a

step-down unit, admissions from the

ED, or transfers from other hospitals

can be synchronized to that schedule.

Individual units can create schedulesbased on their own resources and de-

mands but coordinate on a larger sys-

tem level. This synchronization allows

local, unit-level control and sys-

temwide optimization to occur simul-

taneously.

leadarticle


16/58


Decrease demand within the hospital.

Use a rapid-response team. The assump-

tion that all demand for hospital ser-

vices originates outside of the hospital

must be challenged. Recent work inboth the United States and Australia

has shown the effect of rapid-response

teams in decreasing demand for ICU

beds. For example, a study published

recently shows a reduction of cardiac

arrests and a marked decrease in the

use of ICU beds if symptoms are rec-

ognized before the actual code occurs.

The time of maximal impact from a

rapid-response team is during the pe-

riod of time commonly known asfail-ure to rescue. This is a term used to

connote an approximately four-hour

period that often precedes a patients

cardiac arrest, respiratory arrest, or

stroke. In the failure-to-rescue period,

there is a change in the patients

symptoms, which, although continu-

ous, is often subtle. While the care-

giver is concerned and often attempts

to communicate the changes, the

signs and symptoms are often seen assoft signs, such as decreased ap-

petite, decreased respiratory rate, or

increased tiredness, and more moni-

toring is ordered. The patient contin-

ues to deteriorate until an event oc-

curs, such as an arrest, a stroke, or

other failure, that raises the alarm in

the system. The implementation of ap-

propriate rescue care has shown a re-

duction of up to 90 percent in patient

days that result from return to theICU after a cardiac arrest. Avoiding

placement in or a return to the ICU of

hospitalized patients can have a large

impact on ICU flow (GAO 2003).

Decrease length of stay in the ICU.

Capacity in the intensive care unit is

directly dependent on the length of

stay in the ICU. Teams in IHIs ICU

IMPACT collaborative who have been

working with ventilated patients have

recently demonstrated marked de-creases in ventilator hours and, con-

sequently, decreases in ICU lengths

of stay. Using a process whereby all

evidence-based care that is necessary

and sufficient for ventilator-depen-

dent patients is combined, or bun-

dled, patients have a shorter stay on

the ventilator and, in some cases, sig-

nificant decreases in ventilator-associ-

ated pneumonias. The effect of an 8-

to 12-hour decrease in average lengthof stay for a given ICU is equivalent

to the addition of three ICU beds.

Changes Involving Providers Outside of

the Hospital

The issues affecting patient flow in

the hospital do not all occur within

the walls of the hospital. The raw

facts suggest that unless hospitals ex-

pand their partnerships with other

providers in the communitycreat-ing working relationships with new

partners and developing creative new

care systems with existing partners

the problems that the community

brings to the hospital will never be

solved. It is crucial for hospitals to

work with the community to find so-

lutions to the flow problem. The hos-

pitals participating in IHIs collabora-

tive project on flow call this

extending the chain.Responsibility driven by geogra-

phythat is, addressing only those

problems in ones own areais the

source of much variation in hospi-

tals. Particularly for hospitals with

patient flow problems, working more

The effect of an 8-to 12-hour

decrease in

average length of

stay for a given

ICU is equivalent

to the addition of

three ICU beds.


17/58

F

R

O

N

T

I

E

R

S


intensively with physicians and long-

term-care facilities in the commu-

nitythose with the power to affect

both admissions and dischargesis

an effective strategy to improve flow.

Set up specialized units outside of the

hospital. A common and particularly

frustrating bottleneck in the ICU or a

step-down unit, for example, is the in-

ability to transfer chronic ventilator

patients off the unit because of a lack

of chronic ventilator beds in other set-

tings. Multiple examples of hospitals

using their own resources to help de-

velop chronic ventilator beds areknown. One such example is paying a

knowledgeable physician to set up a

unit in a nursing home. If a hospital

feels that this type of financial assis-

tance for an unaffiliated nursing

home is not appropriate, it does not

understand the overall effect that a

few chronic ventilator patients can

have on the throughput of the hospi-

tal. In addition to helping with flow in

the hospital, such ventilator units,when set up properly, can and will

lead to better patient outcomes at

lower cost. Similar principles have

been used in creating neurology and

cardiac rehabilitation units outsourced

to long-term-care facilities. In all of

these examples, the business case is

easily demonstrated.

Make the nursing home reservation.

Most hospitals use a push system todischarge a patient to a nursing home

facility. Once a patient is determined

to need a nursing home bed, a search

is started by the hospitals social ser-

vices staff. A better, more efficient sys-

tem might be to synchronize hospital

and nursing home needs by establish-

ing a reservation system whereby

hospitals can reserve beds in nursing

homes once a patient in need is iden-

tified, and vice versa. The nursinghome or the hospital still receives pay-

ment if the reservation is not can-

celled and the bed goes unused, much

like the system used in the hotel in-

dustry.

Extend the chain on the front end of

the hospital. Use midlevel providers to

decrease demand for hospital services.

Acute changes that occur in patients

in nursing homes or at home com-monly cause these patients to seek

care in the emergency department.

Using on-site, midlevel providers such

as nurse practitioners and physician

assistants to give tests and treatments

in these settings can avoid ED visits

and, in many cases, admissions to the

hospital. An extension of this idea has

been used in the San Diego school

system with excellent results in treat-

ing asthma patients. Here, the hospi-tal partnered with the schools to pro-

vide treatment to patients at their

school, thus improving their care and

health and reducing ED visits and in-

patient admissions.

SUMMARY

The frontline units (microsystems)

that actually deliver care in hospitals

are tightly linked. Understanding pa-

tient flow therefore requires looking atthe whole system of care, not just the

isolated units. A key element in this

understanding is the concept of nat-

ural and artificial variation, from

sources both inside and outside the

hospital. Variation can and should be

leadarticle


18/58


measured as a first step to dealing

with the problems of patient flow in

hospitals.

Interventions that can smooth flow

include rethinking the scheduling ofelective surgical procedures; creating

a system for planned patient dis-

charge; and extending the chain of

care, both prior to and following dis-

charge. Potential benefits of better

flow include better clinical outcomes,

improved patient safety, greater pa-

tient and staff satisfaction, and im-

proved financial performance.

NOTE

1. The Institute for Healthcare Im-

provement has been working for the

last two years with more than 60 hos-

pitals in the United States and the

United Kingdom in a collaborative

project to improve flow through acute

care settings. This work is becoming

embedded in a group of more than

100 hospitals that are addressing the

issue of healthcare improvement as

part of IHIs IMPACT network. TheIMPACT network is a group of

change-oriented healthcare organiza-

tions committed to ambitious levels of

improvement on a broad scale. Seven

hospitals participating in IMPACT

have agreed to become true innova-

tion sites and test the new flow-

change concepts in almost a Skunk

Workssm

type environment (IHI

2003b). Leadership commitment, in-

tense faculty involvement, and heavilyfocused resources contribute to a

rapid growth of knowledge in how to

improve flow.

Through IMPACT and the work in

the innovation sites, hospitals have

been testing the theory that the key to

improving flow throughout the acute

care setting lies in understanding the

variability throughout the hospital

system, as discussed in this article

(IHI 2003b).

REFERENCES

Aiken, L. H., S. P. Clarke, D. M. Sloane, J.

Sochalski, and J. H. Silber. 2002. Hospital

Nurse Staffing and Patient Mortality, Nurse

Burnout, and Job Dissatisfaction.Journal

of the American Medical Association 288

(16): 198793.

Berwick, D. M. 1996. A Primer on Leading

the Improvement of Systems. British Med-

ical Journal 312 (March 9): 61922.

Centers for Disease Control and Prevention

(CDC), National Center for Health Statis-

tics. 2003. More Americans Seek Medical

Care in Hospital Emergency Rooms: In-

juries Cause One in Three Visits. National

Hospital Ambulatory Medical Care Survey:

2001 Emergency Department Summary. Ad-

vance Data No. 335. (PHS) 2003-1250. [On-

line report; retrieved 3/11/04.] www.cdc.gov

/nchs/releases/03news/ervisits.htm.

. 2002. Visits to the Emergency De-

partment Increase Nationwide. National

Hospital Ambulatory Medical Care Survey:

2000 Emergency Department Summary. Ad-

vance Data No. 326. (PHS) 2002-1250.

[Online press release; retrieved 3/11/04.]

http://www.cdc.gov/nchs/releases/02news

/emergency.htm.

Institute for Healthcare Improvement (IHI).

2003a. Maximizing Hospital Flow for Effi-

cient and Effective Care: Manifestations of

Variability Within the ED and OR. Slide

presentation, March 11.

. 2003b. Optimizing Patient Flow,

Moving Patients Smoothly Through Acute

Care Settings. Innovation Series 2003

paper. [Online article; retrieved 2/18/04.]http://www.ihi.org/newsandpublications

/whitepapers/flowfinal.pdf.

The Lewin Group. 2002. Analysis of AHA ED

and Hospital Capacity Survey. [Online re-

port; retrieved 2/2/04.] www

.hospitalconnect/aha/press_room-info.

McManus, M. L., M. C. Ling, A. B. Cooper, J.


19/58

F

R

O

N

T

I

E

R

S


Mandell, D. M. Berwick, M. Pagano, and E.

Litvak. 2003. Variability in Surgical Case-

load and Access to Intensive Care Ser-

vices. Anesthesiology98 (6): 149196.

U.S. General Accounting Office (GAO). 2003.

Hospital Emergency Departments: CrowdedConditions Vary Among Hospitals and Com-

munities. Report No. GAO-03-460. Wash-

ington, DC: U.S. GAO.

U.S. House of Representatives, Committee on

Government Reform, Special Investiga-

tions Division. 2001. National Preparedness:

Ambulance Diversions Impede Access to

Emergency Rooms. Report prepared for Rep.

Henry A. Waxman. October 16, p. i. [On-line report; retrieved 3/11/04.] www.house

.gov/reform/min/pdfs/pdf_com/pdf

_terrorism_diversions_rep.pdf.

leadarticle


20/58


21/58

F

R

O

N

T

I

E

R

S

s uz an ne s . h or to n 17

Increasing Capacity While Improving

the Bottom Line

S U Z A N N E S . H O R T O N , R . N .

Summary Baptist Memorial HospitalMemphis had a capacity prob-

lem, or so we thought. After examining a situation that we considered

virtually unfixable, we implemented high-leverage process changes re-

sulting in significant improvements in patient flow. We learned that we

did, after all, have the capacity we needed without adding any beds or hir-

ing additional staff.

leadarticle

suzanne s. horton, r.n., is director of nursing administration at

baptist memorial hospital in memphis, tennessee.


22/58


BACKGROUND AND

ENVIRONMENT

Baptist Memorial HospitalMemphis

(BMH-M), a 736-bed tertiary care hos-

pital, has routinely exceeded 90 per-cent occupancy. Process mapping for

patient flow and bed turnaround has

been in continual flux, and the hospi-

tal, the busiest in Tennessee, has pur-

sued flow initiatives for more than

five years. Early process-improvement

efforts included centralization of bed

assignments and the addition of a

bed-tracking system to provide bed in-

formation in real time. These efforts

resulted in improvements in turn-around time (TAT) for bed assign-

ments; however, demand continued to

exceed capacity. The emergency de-

partment (ED) implemented best

practices for bedside registration,

triage criteria, initiation of protocols

at triage, and the opening of a fast-

track area. However, the results were

disappointing. In addition, the

tremendous amount of time and en-

ergy expended without visible resultsbecame a drain on the entire team.

Our ED was always in crisis; the ED-

centered process approach was inef-

fective.

In March 2002, Memphis hospi-

tals were diverting ambulances an av-

erage of 70 percent of the time. Am-

bulance services were exhausted daily,

and ambulance crews waited at the

local EDs for as long as 90 minutes.

The crisis prompted both negativepress and frequent regulatory inspec-

tions.

Memphiss hospital leadership and

the Emergency Medical Service Coun-

cil joined together to test eliminating

diversion. Administrative and clinical

personnel were able to redirect their

energies from diversion decision-mak-

ing issues to patient care excellence.

As a result, patients went to their pre-

ferred hospital, ambulance serviceswere no longer exhausted, and the

community TAT for ambulances de-

creased to 60 minutes overall (6 min-

utes at BMH-M).

Upon elimination of diversion, the

ambulance crisis was solved, but the

larger crisis related to patient flow had

now shifted to the ED. We had a deci-

sion to make as to how we would en-

sure care to all patients seeking it in

our ED. From that effort, the BMHmission of preaching, teaching, and

healing that was established in 1912

came alive in 2002.

HOW WE MOVED FORWARD

Administration, nurses, and physi-

cians adopted a zero-tolerance philos-

ophy for barriers to flow and vowed to

find a cure. BMH-M became involved

with the Institute for Healthcare Im-

provement (IHI) in fall 2002, joiningthe following four IHI IMPACT col-

laborative domains:1

1. Improving critical care

2. Achieving workforce excellence

3. Improving flow through acute care

settings

4. Improving patient safety

Strategies and Their Implementation

Our first initiative. The express admis-sion unit (EAU) (Advisory Board

2001) opened in fall 2001, with no

additional labor being used to staff it.

The EAU is a 21-bed dedicated area

that processes direct and emergency

department admissions, relieving

By eliminatingdiversions,

administrative

and clinical

personnel were

able to redirect

their energies

from diversion

decision-making

issues to patient

care excellence.


23/58

F

R

O

N

T

I

E

R

S


unit nurses of work associated with

admission assessments and initial

admission orders while removing re-

sponsibility for particularly time-

intensive activity from busy unitnurses.

At the start of the initiative, the

unit was open Monday through Friday

from 7 a.m. to 11 p.m. All initial pa-

perwork and diagnostics were com-

pleted in the EAU with a projected

TAT of 60 to 70 minutes. The EAU

provided some relief, but we realized

in the winter crunch 2002 that 16-

hour access to it was an impediment

to overall flow. During the wintercrunch, which saw high volume and

higher acuity, we noticed that to close

the EAU by 11 p.m., no beds could go

to the ED patients for approximately

four to five hours to ensure that all 21

EAU patients had rooms. This stand-

still occurred in the busiest part of the

day for the ED. When we began oper-

ating the EAU 24 hours a day, the flow

became smooth.

Since expanding the hours, theEAUs volume has increased to 35 to

50 patients per day. An unexpected

benefit was that physicians no longer

sent patients to get their workup in

the ED. Patients initial workup oc-

curred expeditiously in the EAU,

which resulted in a 50 percent reduc-

tion in ED holding hours for admitted

patients. The EDs winter crunch of

2002 dissipated, and unit staff nurses

turned their energies from admissionduties to patient care with fewer inter-

ruptions.

The emergency department. As a

member of IHIs IMPACT collabora-

tive on flow, our first area of focus,

after establishing the EAU, was the

ED. We committed to the following

collaborative measures:

Patients were to be placed from the

ED to an inpatient bed within onehour of the decision to admit.

Patients were to be moved from the

post-anesthesia care unit (PACU)

to an inpatient bed within one

hour from the time the patient is

deemed ready to move.

Patients were to be placed from the

intensive care unit (ICU) to an in-

patient bed within four hours from

the time the patient is deemed

ready to move. Patients were to be physically

transferred from the inpatient facil-

ity to a long-term-care facility

within 24 hours after the patient is

deemed ready to transfer.

After the first IHI learning ses-

sion, we began rapid process improve-

ment, a technique of testing ideas for

change on a small scale, altering

processes to improve them, and ex-panding or spreading the processes to

other areas when they are successful.

Little did we know that it would revo-

lutionize and energize every aspect of

everything we did. Our first test of

change involved faxing reports from

the ED to the receiving unit, which

eliminated time spent holding and re-

turning calls. We began with one unit

and eventually spread this process to

the entire facility within threemonths. We then worked through a

package of ED high-leverage changes

(changes that give the organization

the most return in terms of outcome)

provided by IHI faculty. The ED staff,

nursing leadership, and ED medical

leadarticle


24/58


director met weekly to initiate

changes and modify them as needed.

Changes included the following:

Enhancing triage staff (added para-medic and triage nurse during

peak time)

Enhancing triage criteria

Beginning lab and x-ray diagnostic

procedures at triage per protocol

when the ED was at capacity

Taking patients directly to a room

when one is available with bedside

registration

Matching both physician and nurs-

ing labor to peak times Providing e-mail access to staff to

improve communication

One of the highest-leverage

changes was segmenting the urgent

care population within the ED and de-

veloping a fast-track area, essentially

providing a minor medical unit within

the ED. Patients with nonemergent

needs typically wait the longest and re-

quire the least amount of time fortreatment, leading to great dissatisfac-

tion. It is important to note that fast

tracking patients was not new to this

team, but a commitmentto this group

of patients was. Previously, this area

was only open seemingly when

staffing allowed, and the stars in the

heavens were aligned just right. The

ED medical director and the nursing

management team agreed that a suc-

cessful fast-track process was criticalto improving performance. Time was

invested daily to coach and mentor

staff and provide positive feedback to

those who made it happen. Tough

decisions had to be made when staff

were not successful.

Changes that enhanced this

process were as follows:

An area adjacent to, but not in, the

main ED was dedicated to fasttrack.

A nurse practitioner or physician

assistant, nurse, and emergency

medical technician were dedicated

to fast track.

Criteria for fast tracking were re-

vised and included in the triage cri-

teria.

Departments (e.g., x-ray and lab)

partnered for optimal TAT.

Data micromanagement and prob-lem solving occurred on a daily

basis.

Access case management. Access

case management is a process change

we began in the ED, but our plan is to

spread to all access (entry) points of

the hospital. The process uses case

managers to ensure that the patients

are admitted to the right level of care,

evidence-based protocols are initiatedby diagnosis, and appropriate hospital

and social needs are identified and

acted on at the entry point. At BMH-

M, a case manager and social worker

were assigned to the ED to ensure our

patients received timely and appropri-

ate services, and evidence-based pro-

tocols were initiated with the help of

the case manager in the ED for treat-

ing congestive heart failure and pneu-

monia. This team assisted in access-ing community resources to those in

need. It also mobilized internal re-

sources on complex cases that re-

quired a multidisciplinary approach

to discharge planning. Having the

team in the ED connected them to the

Huddle meetingsimprove bed-flow

planning and

provide more

information to all

stakeholders.


25/58

F

R

O

N

T

I

E

R

S


family at the soonest possible point to

participate in identifying any dis-

charge concerns.

The critical care unit. Parallel to the

flow changes were changes initiatedin critical care. In January 2003, an

intensivist program was implemented

on nights and weekends in our 38

general critical care units. Multidisci-

plinary rounds were held every night

by core staff as well as three days a

week with an expanded team. These

rounds stressed compliance with evi-

dence-based practices such as reduc-

tion in ventilator-acquired pneumonia

and other common complications incritical care. Such practices included

elevating the head of the bed at 30 de-

grees, interrupting continuous seda-

tion at least once every 24 hours, and

preventing deep-venous thrombosis

and upper-gastrointestinal bleeding.

In addition, because the intensivists

were now present, patients who would

normally wait until the next day for

extubation were weaned from the ven-

tilator during the night.Huddle meetings. Huddle meetings

were initiated to improve bed-flow

planning and provide more informa-

tion to all stakeholders. In these meet-

ings, the house supervisor, bed-

assignment nurse, housekeeping su-

pervisor, and key charge nurses came

together three times a day to plan for

surgeries, patients coming from outly-

ing facilities, other admissions, and

transfers to higher and lower levels ofcare. This process has facilitated an

enduring team effort to provide pa-

tient-centered bed flow.

The discharge process. Orchestration

of discharges is another breakthrough

practice. A patient-centered approach

to care, orchestrated discharge is a col-

laboration with the physician that an-

ticipates when the patient will be dis-

charged. Once the physician agrees

that the patient is ready for planneddischarge, an appointment time is

agreed on with the patient and family

for the discharge date (usually the

next day, but it can be the same day).

All disciplines, including the physi-

cian and physician consultants, are

alerted to the discharge appointment

time. All actions required for dis-

charge are completed prior to the ap-

pointed time to allow the patient to

leave within 30 minutes of that time.This process allows the patient, fam-

ily, and staff to be much more proac-

tive in the discharge process. Even the

ambulance service, our new partner in

the no-diversion world, appreciates

the importance of an orchestrated dis-

charge and assists us in meeting the

goal. In terms of capacity, the process

allows the centralized bed-assignment

area to plan on beds that will be va-

cated at specific times during the dayto slot admissions, transfers, and sur-

gical patients who are in the queue.

RESULTS

After one year, the following outcomes

were achieved in overall IHI measures

and in the ED, the PACU, and the

ICU.

IHI Measures

The following results were achievedafter implementing several IHI mea-

sures:

1. Patients placed from the ED to an

inpatient bed within one hour of

the decision to admit increased

leadarticle


26/58


from a baseline of 52 percent to 73.1

percent.

2. Patients placed from the PACU to an

inpatient bed within one hour from

the time they are deemed ready tomove increased from a baseline of

76 percent to 78.3 percent.

3. Patients placed from the ICU to an

inpatient bed within four hours

from the time they are deemed

ready to move increased from a

baseline of 45 percent to 62.7 per-

cent.

4. Patients transferred from the inpa-

tient facility to a long-term-care fa-

cility within 24 hours after they aredeemed ready to transfer increased

from a baseline of 68 percent to

80.1 percent.

Additional ED Results

The emergency department saw the

following results after implementing

the new policies:

TAT for overall ED was reduced by

9 percent, while ED volume in-creased by 6 percent.

Fast-track volume increased from

12 percent to 20 percent of the

adult volume in the ED.

TAT for the fast-track area de-

creased from 2.21 hours to approxi-

mately 1.5 hours.

Patients who left without being

seen by a physician decreased from

5 percent to 1.4 percent.

Holding hours for admitted pa-tients were reduced by 50 percent.

Patient satisfaction improved from

the 10th percentile to the 85th per-

centile, according to a survey con-

ducted by the Gallup Organization

(2003).

Employee satisfaction improved 30

percent in the highest 2 points (6

and 7) of a 7-point scale.

TAT for ambulances averaged 7

minutes, as compared to the com-munitys average of 60 minutes.

PACU Results

Holding hours were reduced by 36

percent.

ICU Results

The results achieved in the ICU are as

follows:

Length of stay was reduced by 2daysthe equivalent of building 12

ICU bedswith no construction or

additional staff.

Mortality rate decreased 40 per-

cent.

Volume increased 20 percent from

fiscal year 20022003.

Ventilator-acquired pneumonia de-

creased 80 percent.

Readmission rate did not change.

WHAT NEXT?

Future plans include expanding the

access case management initiative to

all points of entry, implementing mul-

tidisciplinary rounds on all units, and

providing 80 percent of those patients

with an orchestrated discharge with

their appointment time met at least

80 percent of the time. In addition,

we plan to smooth surgical flow by

capping elective cases to decreasedaily variability.

LESSONS LEARNED

Improving flow is not about building

more beds or hiring more staff. It is

about developing and improving

Improving flow isnot about

building more

beds or hiring

more staff. It is

about developing

and improving

processes to

decrease

variability and

smooth flow.


27/58

F

R

O

N

T

I

E

R

S


processes to decrease variability and

smooth flow. For BMH-M, it required

a culture change of leadership and

staff that moved us from Were doing

the best we can to Sorry, that is notgood enough. Our mantra became,

What is the barrier, and how do we

eliminate it? Overcrowding of the ED

is not all about the ED but is a

symptom of a broken system. It is

linked to the admission, discharge,

and transition processes throughout

the organization. The added payoff is

that improved processes result in im-

proved quality of care as well as pa-

tient and employee satisfaction.

NOTE

1. Please see Note 1 of Patient Flow in

Hospitals: Understanding and Con-

trolling It Better, by Carol Haraden

and Roger Resar, in this issue of Fron-tiers for a discussion of the IHI IM-

PACT collaborative project.

REFERENCE

Advisory Board Company. 2001. Heart of the

Enterprise: Optimizing Nursing Productiv-

ity in an Era of Deepening Shortage.

Washington, DC: Advisory Board Com-

pany.

Gallup Organization. 2003. Emergency Pa-

tient Loyalty Results. Washington, DC:

Gallup Organization.

leadarticle


28/58


29/58

F

R

O

N

T

I

E

R

S

diana henderson, christy dempsey, and debra appleby 25

A Case Study of Successful Patient

Flow Methods: St. Johns Hospital

D I A N A H E N D E R S O N , C H R I S T Y

D E M P S E Y , A N D D E B R A A P P L E B Y

Summary Participating in the Institute for Healthcare Improvement

(IHI) IMPACT collaborative has given St. Johns Hospital the opportu-

nity to improve patient flow and the delivery of patient care. This part-

nership has allowed us to experience a wealth of information shared by

a collaborative network of hospitals. IHI has introduced rapid-cycle im-

provement methodologies, variability-reduction strategies, and strate-

gies to aid in planning for the expected as methods that have enhanced

our already established performance-improvement program. St Johns

has achieved breakthrough improvement with patient flow.

leadarticle

diana henderson is executive director for quality; christy

dempsey is director of perioperative services; and debra appleby

is supervisor of quality resources at st. johns hospital in

springfield, missouri.


30/58


BACKGROUND

St. Johns Hospital is a tertiary hospi-

tal located in Springfield, Missouri,

serving 32 counties that cover 22,000

square miles in southwest Missouriand northwest Arkansas. The hospital

has more than 30,000 admissions a

year and an average length of stay

(LOS) of 4.53 days for all patients.

St. Johns senior leadership formed

an alliance with the Institute for

Healthcare Improvement (IHI) in May

2002 to assist our organization in

achieving a superior level of perfor-

mance. IHI measures of improvement

include improved health status, betterclinical outcomes, lower cost, greater

access, greater ease of use, and im-

proved satisfaction for individuals and

their communities.

St. Johns leadership team, includ-

ing representatives from St. Johns

Hospital senior leaders and medical

management services, selected im-

proving patient flow through the acute

care setting from the five domains of-

fered by IHI. Perioperative serviceswas identified as the initial focus of

the project. St. Johns perioperative

services consists of 26 operating

rooms in the hospital and 6 ambula-

tory surgery center rooms. An average

of 25,000 cases are performed annu-

ally in these rooms. All specialties are

represented, with the exception of

organ transplantation. The rationale

for selecting perioperative services was

that it is a high-volume, high-risk,multifaceted area and displays strong

leadership support. Issues with on-

time starts, turnover times, and pa-

tient flow through the various areas of

perioperative services are significant in

terms of patient, staff, and physician

satisfaction; cost and revenue; and

quality of care.

Variability in caseloads, patient

acuity, and specialty needs has a direct

impact on not only perioperative ser-vices but also the hospital as a whole.

This variability leads to the down-

stream effect of hospital bed capacity

constraints, LOS issues, and intensive

care unit (ICU) bed availability as well

as the upstream effect of excessive

emergency department (ED) waiting

times. Therefore, to accelerate im-

provements in perioperative services,

teams involving the ED (upstream ef-

fect) and the surgical ICU (down-stream effect) were commissioned

concurrently.

St. Johns overall aim for this pro-

ject is to ensure that patients receive

timely access to appropriate care and

move safely and efficiently through

the system without unnecessary and

unproductive delays. As discussed

below, each team has established

goals that contribute to achieving the

overall aim.

METHODOLOGY

Project teams were formed, and a

physician champion and team leader

were identified for each team. St.

Johns leadership team collaborated

with the physician champion and

team leader to determine what disci-

plines or stakeholders should be rep-

resented on the team. Team members

were selected from each area affectedby the improvement process. A walk-

through of each teams area was con-

ducted to identify opportunities for

improvement. Frontline staff not in-

volved in the teams also played a cru-

cial role in identifying opportunities

The PDSAimprovement

model helps to

build the

organizations

capacity for

change by

focusing on

foundational

issues.


31/58

F

R

O

N

T

I

E

R

S


for improvement by participating in

individual department surveys. Per-

formance-improvement tools were

used to identify delays or bottlenecks

in the process. Decisions were madescientifically, based on data rather

than hunches. Once data were ana-

lyzed, changes were developed and

tested.

St. Johns uses the Plan-Do-Study-

Act improvement model devised by

Langley, Nolan, and Nolan (Berwick

1996). The Plan-Do-Study-Act

(PDSA) cycle describes inductive

learningthe growth of knowledge

through making changes and the re-flecting on the consequences of those

changes. Nolans model intends that

the enterprise of testing change in in-

formative cycles should be part of nor-

mal daily activity throughout an orga-

nization (Berwick 1996, 620).

The PDSA improvement model

helps to build the organizations ca-

pacity for change by focusing on foun-

dational issues such as the following:

Creating a change culture

Developing an infrastructure to

support improvement

Building the business case for

quality

Implementing strategies for

rapidly spreading innovation

St. Johns uses the rapid-cycle im-

provement methodology by including

the key stakeholders on each team,conducting tests of change (trying a

change on a small scalee.g., one pa-

tient, one doctor, one room) to see

what works and what does not, under-

standing that a failed test is an oppor-

tunity to learn, accelerating successes

by holding weekly team meetings to

evaluate effectiveness and to plan next

steps to spread improvements, remov-

ing barriers, and communicating

strategies and outcomes throughoutthe organization.

OUTCOMES

The key to improving flow lies in re-

ducing process variation that impedes

flow. While some variability is normal,

other variation is not and should be

eliminated (IHI 2003, 1). Litvak sug-

gests that variability could be de-

creased by providing an operating

room dedicated to unscheduled oradd-on cases (McManus et al. 2003).

By doing so, overall efficiency would

be increased and the variability in sur-

gical case flow would actually be re-

duced.

Operating Room Use

The concept of dedicating an operat-

ing room to unscheduled cases,

thereby limiting its use for scheduled

procedures or block time, is not one tobe considered lightly in an era of in-

creasing surgical case volumes, lim-

ited capacity both in terms of physical

space and staff availability, and com-

petition for managed care covered

lives. However, St. Johns, in collabo-

ration with its surgeons, set aside an

operating room as a trial project in

November 2002. Prior to this time,

the operating room had been blocked

for a general/trauma surgeon groupfor elective cases. By agreeing to re-

lease this room from use for their

elective cases, the general/trauma sur-

geons actual elective weekly block

time was reduced. The trauma sur-

geons were agreeable to this test of

leadarticle


32/58


change because potential existed to

improve overall efficiency (Henderson

et al. 2003).1

After approximately three months

of segmenting this room as an add-on room, the data were reassessed.

The following improvements were

achieved (Henderson et al. 2003):

During the hours of 7:30 a.m. to

1:30 p.m. on weekdays, the num-

ber of surgical cases increased by

5.1 percent.

The number of operating rooms

needed for surgical cases at 3 p.m.,

5 p.m., 7 p.m., and 11 p.m. onweekdays decreased by 45 percent.

A 2 percent overall reduction in

overtime was achieved.

The general/trauma surgeon

group involved in the project real-

ized a greater than 4.6 percent in-

crease in revenue.

The nursing floors are able to pre-

dict more accurately their evening

and night-shift staffing require-

ments.

ED Patient Flow

St. Johns is a Level I trauma center

with 64,000+ visits annually. To re-

duce prolonged waiting times in the

ED, we chose to focus on delays in

admitting ED patients. Admission de-

lays tie up ED beds for up to an hour

or more after the ED has completed

its work with the patient and he or

she is simply awaiting a bed. Wefound that a key process step that ac-

counted for much of the delay in-

volved communication between the

ED nurse and the accepting nurse on

the floor. The communication

process between the two departments

sometimes required multiple phone

calls to connect with the staff mem-

ber needing to receive the patient sta-

tus details.

In an effort to expedite this com-munication process, a standardized

fax report form was tested for admit-

ting patients to an intermediate car-

diac floor. The fax report form was de-

veloped collaboratively by the ED and

cardiac staff and evaluated by both

staff each time a fax was sent. The fax

reporting procedure provided great

staff satisfaction, because it elimi-

nated waiting time on the phone for

the ED nurse and the floor nurse ex-perienced less interruption from pa-

tient care activity. Multiple phone calls

from both units were eliminated,

which was a great time saver (Hender-

son et al. 2003).

After the initial trial, the use of this

form was implemented as routine

practice for this floor and on an addi-

tional intermediate cardiac floor, with

similar success. The ED faxed report

is now a standard part of the processthroughout the hospital. In December

2003, the median time from the deci-

sion to admit to physical placement in

an inpatient bed was 61 minutes. This

is a 67 percent decrease, compared to

baseline data. The implementation of

the faxed report clearly has had a posi-

tive impact on this measure (Hender-

son et al. 2003).

Additional successes in the ED in-

clude the following:

ED acute care center (fast track)

hours have been expanded from 12

to 16 hours per day. Visits have in-

creased from an average of 38 to

60 visits per day.

A key step thataccounted for

much of the delay

involved

communication

between the ED

nurse and the

accepting nurse

on the floor.


33/58

F

R

O

N

T

I

E

R

S


ED internal patient-satisfaction

Documents

Frontiers - Capacity Management