Beruflich Dokumente
Kultur Dokumente
................................................................................................
Lines: 0 to 36
E X E C U T I V E S U M M A R Y
———
Managers within the U.S. healthcare system are becoming more aware of the im- * 41.60103pt PgVar
pact of variation in demand on healthcare processes. The UK National Health ———
Service provides a prime example of a system that has experienced the conse- Normal Page
quences when the issue is not dealt with satisfactorily, having suffered from ex- * PgEnds: Eject
cessive queues for a prolonged period. These delays are mostly caused by a lack
of attention to variation and inappropriate responses to the queues, rather than a [309], (1)
capacity shortage.
A number of collaborative programs recently have come to grips with many of
the causes of the queues in both elective care and emergency care. Although there
are still areas that need large-scale improvement, good progress has been made,
especially within emergency care. The authors of this article have acted as technical
advisors to a number of these improvement programs and have been able to doc-
ument many of the practices that have helped to reduce or eliminate unnecessary
queues and delays across the 200 sites in England that have 24-hour emergency
care facilities. Local program managers at these sites continuously reported progress
for a period of 18 months. A number of important lessons for both the design and
control of healthcare processes have emerged from the collaborative work. These
lessons focus on understanding and measurement of demand, capacity planning,
reduction of introduced variation, segmentation and streaming of work, process
design, capacity yield management, and measurement of variation.
For more information on the concepts in this article, please contact Mr.
Walley at paul.walley@wbs.ac.uk. Part of this work was funded by a research
grant to Paul Walley at the University of Warwick.
309
Photocopying and distributing this PDF of the Journal of Healthcare
Management
BookComp/ is prohibited
Health Administration Press/ Journalwithout the permission
of Healthcare Management of 309
/ Vol. 51, No. 5/ Page Health
1st proof
Administration Press, Chicago, Illinois. For permission, please fax
your request to (312) 424-0014 or e-mail hap1@ache.org.
J o urna l o f H e a lt hca re Ma nagement 51:5 September /October 2006
310
Photocopying and distributing this PDF of the Journal of Healthcare
Management
BookComp/ is prohibited
Health Administration Press/ Journalwithout the permission
of Healthcare Management of 310Health
/ Vol. 51, No. 5/ Page 1st proof
Administration Press, Chicago, Illinois. For permission, please fax
your request to (312) 424-0014 or e-mail hap1@ache.org.
Managing Var iation in Demand
311
Photocopying and distributing this PDF of the Journal of Healthcare
Management
BookComp/ is prohibited
Health Administration Press/ Journalwithout the permission
of Healthcare Management of 311Health
/ Vol. 51, No. 5/ Page 1st proof
Administration Press, Chicago, Illinois. For permission, please fax
your request to (312) 424-0014 or e-mail hap1@ache.org.
J o urna l o f H e a lt hca re Ma nagement 51:5 September /October 2006
which lasted 18 months at each site. helped reduce the impact of variation
The final program closed at the end of on the system’s performance.
2004. The national program grouped
30 communities across England, and Really Understand Demand
they were given financial support and Before the collaborative program
advice on how to improve emergency started, demand was believed to be
care flows. All 200 sites with 24-hour random and therefore could not
emergency care facilities participated in be forecast. We have found that
the ESC program. not only is it possible to forecast
Local program managers submitted demand, it is also essential to make
monthly reports to the national appropriate plans for emergency
program team. These reports detailed demand. We found that time of day [312], (4)
the activities of each improvement and day of week drive emergency
team, and the data were rapidly room attendance to a great extent,
Lines: 60 to 78
synthesized to establish which types with weather and a few special causes
———
of improvements had beneficial (e.g., public holidays) also affecting
0.0pt PgVar
effects. At the end of the first wave arrival patterns. When clinicians and ———
of implementation, the program was service managers were given the data in Normal Page
formally reviewed using site visits and usable form, they were better able to * PgEnds: Eject
structured interviews with clinical and understand the concept of variation.
program personnel. In total, 80 people Once repeating seasonal demand
[312], (4)
were interviewed across 33 sites. patterns were recognized and the
Pareto principle was applied (i.e.,
80 percent of the consequences comes
LESSONS FROM THE from 20 percent of the causes), systems
EMERGENCY SERVICES could be designed to meet demand.
C O L L A B O R AT I V E There was also a small cultural shift,
The UK government has established a as staff moved away from feeling panic
series of performance targets for the during busier-than-average days and
healthcare system, rewarding hospitals began recognizing and accepting that
for performance above certain levels. some days are busy due to variation.
One of these is a four-hour wait target
in emergency departments. Emergency Maintain the Right Capacity Plans
departments are expected to assess, Problems occur when there is insuf-
stabilize, and admit to a bed or to ficient capacity to meet demand. It is
discharge 98 percent of patients within remarkable how few systems formally
four hours. Before the ESC started, tried to match capacity to predicted
on average, only about 65 percent of demand.
waiting patients got through the system
in four hours. The mean is now near Example: The daily emergency demand
96 percent across the 200 sites. The profile. At the first site we investigated,
following recommendations show how weekday demand followed a consistent
a number of simple principles have profile through a 24-hour period. There
312
Photocopying and distributing this PDF of the Journal of Healthcare
Management
BookComp/ is prohibited
Health Administration Press/ Journalwithout the permission
of Healthcare Management of 312
/ Vol. 51, No. 5/ Page Health
1st proof
Administration Press, Chicago, Illinois. For permission, please fax
your request to (312) 424-0014 or e-mail hap1@ache.org.
Managing Var iation in Demand
•••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••
FIGURE 1
Hourly Patient Arrival Pattern in a Typical UK Emergency Room
[313], (5)
% of Total per Hour
3
Lines: 78 to 103
———
2
0.0pt PgVar
———
1 Normal Page
* PgEnds: Eject
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
[313], (5)
Time of Day
•••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••
was a morning rush of patients with the emergency room, where the peak
minor injuries and a slight blip in in capacity occurs four hours after the
demand after schools closed and on peak in demand. We should not be
into the evenings. Patients with major surprised that queues occur. However,
illnesses, usually requiring admission, staff did not always realize that the
tended to arrive more regularly during capacity plan was inappropriate. In
daylight hours, partly determined by some cases, additional staff members
ambulance availability. Figure 1 shows were deliberately introduced when the
the demand by hour. queues were at their worst—to manage
The capacity strategy for the the queues, rather than to prevent the
queues from happening.
emergency department misread
demand considerably because there is a Understand How Your System
tradition within the NHS of assigning Introduces Variation
staff on early and late shifts to cope It is normally assumed that elective
with peaks and troughs in workload demand has less variation because it
on hospital wards. Using a ward-based can be reduced by careful scheduling.
shift pattern is quite inappropriate in However, the ESC program discovered
313
Photocopying and distributing this PDF of the Journal of Healthcare
Management
BookComp/ is prohibited
Health Administration Press/ Journalwithout the permission
of Healthcare Management of 313
/ Vol. 51, No. 5/ Page Health
1st proof
Administration Press, Chicago, Illinois. For permission, please fax
your request to (312) 424-0014 or e-mail hap1@ache.org.
J o urna l o f H e a lt hca re Ma nagement 51:5 September /October 2006
•••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••
FIGURE 2
Example of Demand Variation Between Emergency Care and Elective Care
35
30
25
Admissions (Surgery)
20
15
[314], (6)
10
[314], (6)
•••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••
that much in-hospital variation in of delays, or not ill but have not left
demand is actually introduced by the the hospital for some reason.
planning processes. We found some surprises when we
investigated the source of variation in
Example: Variation in emergency and elective demand. Figure 2 shows a comparison
admissions. In most UK hospitals between emergency demand and
some wards take both emergency elective demand variation. In most
patients and elective patients. Lack of
circumstances, elective demand
availability of beds is often a major
variation is greater than emergency
cause of delay for patients who need
demand variation because of the
emergency admission. Most hospitals
way in which operating theaters are
claim to work at average bed-occupancy
scheduled. Hospital planning systems
levels of more than 90 percent,
meaning that congestion and queuing create their own problems by making
are significant problems. Audits suggest elective schedules busiest on the days
that typically 60 percent to 70 percent when emergency demand is also
of the patients occupying those beds expected to be high.
are receiving active treatment and the When we compare admissions and
rest are either waiting for initial visits departures, we see that not only are
by doctors, stuck in the system because they partly out of sync, but that there is
314
Photocopying and distributing this PDF of the Journal of Healthcare
Management
BookComp/ is prohibited
Health Administration Press/ Journalwithout the permission
of Healthcare Management of 314
/ Vol. 51, No. 5/ Page Health
1st proof
Administration Press, Chicago, Illinois. For permission, please fax
your request to (312) 424-0014 or e-mail hap1@ache.org.
Managing Var iation in Demand
•••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••
FIGURE 3
Example of Demand Amplification in Healthcare Processes
200
180
160
140
120
Patients
100
[315], (7)
80
60
Lines: 134 to 155
40
———
20 0.0pt PgVar
0 ———
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 Normal Page
Time (Days) PgEnds: TEX
Admissions Discharges
[315], (7)
•••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••
315
Photocopying and distributing this PDF of the Journal of Healthcare
Management
BookComp/ is prohibited
Health Administration Press/ Journalwithout the permission
of Healthcare Management of 315
/ Vol. 51, No. 5/ Page Health
1st proof
Administration Press, Chicago, Illinois. For permission, please fax
your request to (312) 424-0014 or e-mail hap1@ache.org.
J o urna l o f H e a lt hca re Ma nagement 51:5 September /October 2006
316
Photocopying and distributing this PDF of the Journal of Healthcare
Management
BookComp/ is prohibited
Health Administration Press/ Journalwithout the permission
of Healthcare Management of 316
/ Vol. 51, No. 5/ Page Health
1st proof
Administration Press, Chicago, Illinois. For permission, please fax
your request to (312) 424-0014 or e-mail hap1@ache.org.
Managing Var iation in Demand
•••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••
FIGURE 4
The Redesign of Minor Patient Treatment
317
Photocopying and distributing this PDF of the Journal of Healthcare
Management
BookComp/ is prohibited
Health Administration Press/ Journalwithout the permission
of Healthcare Management of 317
/ Vol. 51, No. 5/ Page Health
1st proof
Administration Press, Chicago, Illinois. For permission, please fax
your request to (312) 424-0014 or e-mail hap1@ache.org.
J o urna l o f H e a lt hca re Ma nagement 51:5 September /October 2006
service time variation because of the system failures. Delays can be caused
design configuration. In operations by any of the following:
management terminology, this redesign
has moved the process from a “long 1. Hospital-acquired infections
and thin” structure to a “short and fat” 2. Patients getting bedsores
structure, which is much more flexible.
3. Waits for inpatient diagnostics
Now that triage has been removed,
there is additional treatment capacity 4. Lack of physiotherapist/occupa-
because the highly skilled triage nurse tional therapist assessment
can now be used to treat patients 5. Planning errors
instead of managing patient flow. This 6. Not informing relatives of the
[318], (10)
process innovation has been one of the discharge plan
fastest ever to be adopted in the NHS.
7. Pharmacy closures after 5.00 p.m.,
Minor patient waiting times have fallen preventing patients from picking Lines: 208 to 255
by 70 percent. up their medication before they go ———
home 0.0pt PgVar
———
Look at Capacity Yield Losses Long Page
All of these quality-related PgEnds: TEX
A number of hospitals have now
problems reduce bed capacity, forcing
virtually eliminated bed shortages
the system to operate in a more
and have reduced the number of [318], (10)
congested manner. Our estimates
patients “outlying” on inappropriate
suggested that typically 25 percent of
wards. Ward staff set an expected date
bed capacity was used wastefully.
of discharge as soon as a patient is
admitted, with a project plan for what
Use Process Control Measures to
needs to happen before discharge can
Understand Variation
take place. The tasks to be scheduled
One of the approaches that the NHS
might include, for example, a visit
has imported from the United States
from an occupational therapist to
is the use of statistical process control
determine the patient’s ability to cope
(SPC) to understand variation at a
with his or her disabilities after going
local level. The IHI and authors such as
home. This makes the processes more Carey (2002) and Wheeler (2000) have
deterministic, and controlling variation provided a methodology that is prac-
is a consequence. tical enough to be implemented on a
The discharge planning document large scale. Although use has not spread
unexpectedly became a goldmine as far as we would like, SPC is used to
of information for the ESC. When
discharge dates are extended beyond • monitor demand changes,
the original plan, the document can
• measure performance over time,
be studied to understand the reasons
patients have been delayed. A high • understand the behavior of processes,
proportion of the discharge delays have • develop a body of knowledge about
been caused by quality errors and other systems, and
318
Photocopying and distributing this PDF of the Journal of Healthcare
Management
BookComp/ is prohibited
Health Administration Press/ Journalwithout the permission
of Healthcare Management of 318
/ Vol. 51, No. 5/ Page Health
1st proof
Administration Press, Chicago, Illinois. For permission, please fax
your request to (312) 424-0014 or e-mail hap1@ache.org.
Managing Var iation in Demand
319
Photocopying and distributing this PDF of the Journal of Healthcare
Management
BookComp/ is prohibited
Health Administration Press/ Journalwithout the permission
of Healthcare Management of 319
/ Vol. 51, No. 5/ Page Health
1st proof
Administration Press, Chicago, Illinois. For permission, please fax
your request to (312) 424-0014 or e-mail hap1@ache.org.
J o urna l o f H e a lt hca re Ma nagement 51:5 September /October 2006
NHS Modernisation Agency. 2004. Ten Warburton, R. 2004. “An Analytical Investi-
High Impact Changes for Service Improve- gation of the Bullwhip Effect.” Production
ment and Delivery: A Guide for NHS and Operations Management 13: 150–60.
Leaders. London: Department of Wheeler, D. J. 2000. Understanding Varia-
Health. tion: The Key to Managing Chaos, 2nd ed.
Noon, C. E., C. T. Hankins, and M. J. Cote. Knoxville, TN: SPC Press.
2003. “Understanding the Impact of Yang, S., C. P. McLaughlin, R. W. Vaughan,
Variation in the Delivery of Healthcare and J. J. Aluise. 1992. “Factory Focus in
Services.” Journal of Healthcare Management Hospital-Owned Ambulatory Surgery.”
48 (2): 82–98. International Journal of Service Industry
Management 3 (4): 63–75.
[320], (12)
P R A C T I T I O N E R A P P L I C A T I O N
Lines: 283 to 312
James B. Conway, CHE, senior fellow, Institute for Healthcare Improvement,
———
Cambridge, Massachusetts 0.795pt PgVar
———
Normal Page
A cross the United States, the healthcare system struggles with issues of demand
and capacity and associated waits, delays, and diversions. Hospitals closed to
emergency ambulance patients for parts of days are commonplace, and diversion
PgEnds: TEX
[320], (12)
has been referred to as an epidemic. The Institute of Medicine (2006) recently
reported that 45 percent of emergency departments reported going on diversion
at some point in 2003; overall, it is estimated that 501 ambulances were diverted
that year. Waits in emergency departments of five or more hours are not usual, and
patients routinely leave the emergency service without being seen. The most recent
national data available, collected in 2002 by the Lewin Group and reported by the
American Hospital Association (Lewin Group 2002), show the average time waiting
for transfer from an emergency department to an acute or critical care bed was
about 3.2 hours. In hospitals on diversion more than 20 percent of the time, the
wait was 5.8 hours. More than half of urban hospitals reported being on diversion
some times, and one in eight urban hospitals reported being on diversion 20
percent or more of the time. Many believe that additional resources must be added
to appropriately resolve the issue of capacity, and the need to address these issues
is receiving considerable attention.
John Kelsch, an executive at Xerox and a consultant for the Quality of Health
Care in America project, is reported to have said, “To do things differently, we must
see things differently. When we see things we haven’t noticed before, we can ask
questions we didn’t know to ask before.” This article is an important contribution
and a jarring reminder to us as leaders to look differently at things “we already
know.” The results this study presents from 200 hospitals with emergency depart-
ments are striking and are no doubt humbling for many readers. After 18 months
of intervention, data show that these emergency departments assess, stabilize, and
320
Photocopying and distributing this PDF of the Journal of Healthcare
Management
BookComp/ is prohibited
Health Administration Press/ Journalwithout the permission
of Healthcare Management of 320
/ Vol. 51, No. 5/ Page Health
1st proof
Administration Press, Chicago, Illinois. For permission, please fax
your request to (312) 424-0014 or e-mail hap1@ache.org.
Managing Var iation in Demand
References
Institute for Healthcare Improvement. 2003. “Optimizing Patient Flow: Moving Patients
Smoothly Through Acute Care Settings.” [Online information; retrieved 04/06/06.]
http://www.ihi.org/IHI/Results/Whitepapers.
Institute of Medicine. 2006. Hospital-Based Emergency Care: At the Breaking Point. Washington, DC:
National Academies Press.
Lewin Group. 2002. “Overcrowded Emergency Departments Leading to More Diversions,
Longer Wait Times.” [Online information; retrieved 04/06/06.] http://www.hospitalconnect
.com/aha/press room-info/releasedisplay.jsp?dcrpath=AHA/Press Release/data/AHAPRESS
RELEASES-0708x9&domain=AHA.
Litvak, E., and M. C. Long. 2000. “Cost and Quality Under Managed Care: Irreconcilable Differ-
ences?” American Journal of Managed Care 6: 305–12.
Litvak, E., M. C. Long, A. B. Cooper, M. L. McManus. 2001. “Emergency Department Diversion:
Causes and Solutions.” Academic Emergency Medicine 8 (11): 1108–10.
321
Photocopying and distributing this PDF of the Journal of Healthcare
Management
BookComp/ is prohibited
Health Administration Press/ Journalwithout the permission
of Healthcare Management of 321
/ Vol. 51, No. 5/ Page Health
1st proof
Administration Press, Chicago, Illinois. For permission, please fax
your request to (312) 424-0014 or e-mail hap1@ache.org.
J o urna l o f H e a lt hca re Ma nagement 51:5 September /October 2006
[Last Page]
[322], (14)
[322], (14)
322
Photocopying and distributing this PDF of the Journal of Healthcare
Management
BookComp/ is prohibited
Health Administration Press/ Journalwithout the permission
of Healthcare Management of 322
/ Vol. 51, No. 5/ Page Health
1st proof
Administration Press, Chicago, Illinois. For permission, please fax
your request to (312) 424-0014 or e-mail hap1@ache.org.