Bioterrorism and Health System Preparedness, Issue Brief No. 9
The Agency for Healthcare Research and Quality is holding a series of Web-assisted conferences on bioterrorism and the health care system.
Archived online, each audioconference on bioterrorism is also distilled into
an issue brief. This brief examines some of the ways resources might be deployed in response to a mass casualty event.
Select for the downloadable print version (PDF file 165 KB). PDF Help.
Contents
Introduction
Role of HHS in Achieving Surge Capacity Capabilities
Modeling Surge Capacity
Addressing Pediatric and School-Based Surge Capacity
Addressing Surge Capacity Through Information Exchange
Discussion
For More Information
Introduction
Surge capacity is a health
care system's ability to expand
quickly beyond normal
services to meet an increased
demand for medical care in
the event of bioterrorism or
other large-scale public health
emergencies. |
A bioterrorist attack or other large-scale
public health emergency has the potential
to result in great numbers of human
casualties. Substantial work has been
done throughout the country at local,
regional, State, and Federal levels to
improve health system preparedness, and
many health care organizations and
systems have developed surge capacity
and other medical care preparedness
plans. Planning for a mass casualty event
should also address what can be done
outside the traditional health care system
to cope with a surge in patient flow
before, during, and after an event.
In 2004, the Agency for Healthcare
Research and Quality (AHRQ) expanded
its Bioterrorism Preparedness Research
portfolio to include several projects that
focus on surge capacity issues. Consistent
with that focus, AHRQ sponsored three
Web conferences on surge capacity and
health system preparedness. The first
dealt with education and training; the
second with facilities and equipment; and
the third in the series, held on October 26,
2004, examined some of the ways
resources might be deployed in response
to a mass casualty event. This issue brief
summarizes this third Web conference.
Four panelists made presentations:
- Robert Claypool, M.D., Deputy Chief
Medical Officer, Office of the
Assistant Secretary for Public Health
and Emergency Preparedness, U.S.
Department of Health and Human
Services.
- Nathaniel Hupert, M.D., Assistant
Professor of Public Health and
Medicine, Weill Medical College of
Cornell University.
- Michael Shannon, M.D., Director
of the Center for Biopreparedness
and Chief of the Hospital's
Emergency Medicine Division,
Children's Hospital Boston.
- Gregory Bogdan, Ph.D., Research
Director and Medical Toxicology
Coordinator, Rocky Mountain
Poison and Drug Center, Denver
Health.
Dr. Robert Claypool discussed the role
of the Department of Health and
Human Services (HHS) in developing
surge capacity as part of the National
Response Plan and described an
initiative to develop Public Health
Service Contingency Stations as
alternative care facilities. Dr. Nathaniel
Hupert presented models of surge
capacity and pre-hospital treatment in
a surge event. Dr. Michael Shannon
described the challenges of treating
large numbers of children, and laid out
the main steps in creating a school-based
emergency response plan. Dr.
Gregory Bogdan explained the critical
role that risk and crisis communication
play in reducing the number of people
who might seek health care during a
large-scale public health emergency.
A
short question and answer period
followed the presentations by the
panelists.
Return to Contents
Role of HHS in Achieving Surge Capacity Capabilities
Dr. Robert Claypool began his
presentation by saying that the
definition of surge capacity should
include three factors:
- The volume of cases.
- Time as a function of the volume of cases.
- The complexity of cases involved.
He explained that hospitals
deal with surge fairly regularly, when
they go on bypass or have to send
patients to other places. In dealing
with weapons of mass destruction,
however (Exhibit 1), the health system is at risk of
being severely stressed and so inelastic
that at least two things might happen:
- First, access to care, or having care rendered, might become compromised.
- Second, the quality or the standards of care might have to change to meet the greatest good.
Expanding upon the three-factor
definition of surge, Dr. Claypool
pointed out that surge doesn't
necessarily have to do with large
volumes of patients. A case in point is
the severe acute respiratory syndrome
epidemic that recently occurred in
Canada, where a relatively small
number of patients severely challenged
the health care delivery capability in
Toronto. Burns are another example of
ways in which relatively few cases can
severely stress health system capability.
To meet surge capacity needs, the
Federal Government is currently
developing the National Response
Plan. Within that plan, HHS will lead
the emergency support function that
deals with health and medical issues.
The capacity to provide health care
represents the most resource-intensive
component of the plan, but the need for resources comes at a time when the
economic realities of the health care
system in the United States are
reflected in the closing of hospitals
and the reduction of bed availability.
Within HHS, the Health Resources
and Services Administration, the
Centers for Disease Control and Prevention, and AHRQ are all looking
at different ways, through grant and
cooperative agreement programs, to
increase bed capacity in our country.
In addition, a collaborative effort
between HHS, the Department of
Homeland Security, the Department of
Defense, and the Department of
Veterans Affairs is developing a
program known as the Public Health
Service Contingency Stations
(PHS-CS). This program borrows a
concept from the Civil Defense System
of the 1950s and 1960s known as
Packaged Disaster Hospitals. These
were deployable hospitals that could be
constituted to support care in the event
of a nuclear attack. At the peak of that
program, 2,000 Packaged Disaster
Hospitals were stationed across the
country, with each providing 200 to 250
beds.
The PHS-CS will have the capability of
a general medical/surgical ward. They
will not have an operating room, an
Intensive Care Unit, or an Emergency
Room. In the initial stage, they will not
have a triage area. Each will be
packaged in 50-bed increments, and
could be deployed in 50-bed increments
up to 250 beds for each station. Each
unit will also contain medical supplies
and a pharmaceutical package. Four
contingency stations have been
purchased in 2004. They are designed to
be used in a shelter of opportunity; i.e.,
they would go into a facility such as a
high school gymnasium or an armory.
In the future, the PHS-CS program will
include modules for burn/surgical
treatment, acute treatment, isolation,
triage, and decontamination. The top
priorities are the capability to treat
burns and the development of a
deployable isolation module. The burn
module is especially important for
bioterrorism preparedness because
vesicants that might be used in a
chemical attack produce lesions that
mimic burns. The deployable isolation
module will be a self-contained, free-standing
unit; that is, it will not require
retrofitting of an existing facility.
Facilities, of course, will require health
care personnel to staff them. HHS is
exploring ways to train health care
professionals to perform multiple
functions. A cardio-pulmonary thoracic
technician, for example, might be cross-trained
to support dialysis care. A current
AHRQ project is to develop, implement,
and test a model to cross-train health care
personnel to provide ventilator support.
The project will then determine the
applicability of the cross-training model
for other required functions.
Return to Contents
Modeling Surge Capacity
Dr. Nathaniel Hupert explained that the
value of modeling surge capacity is to help us question our common
assumptions about what it would take
to respond to a large mass casualty
incident. Furthermore, modeling helps
us analyze the variables involved (time,
numbers of patients, and so on) in
quantitative terms so that concrete
numbers can be generated to use in
planning.
Under a project funded by AHRQ, Dr.
Hupert and his colleagues have
analyzed the determinants of surge
capacity as illustrated in Exhibit 2. The
first is surge arrivals at the hospital,
which is a function of:
- The number of people who have been exposed or involved in an event.
- The possible treatment of patients outside the hospital, or pre-hospital management.
The number of surge
arrivals will be greater or fewer,
depending on how many patients can be
accommodated by pre-hospital
management. In the event of a
biological attack or an outbreak of an
infectious disease, for example, prehospital
management would include the
dispensing of antibiotics or vaccines to
as many people as possible without
admitting them to the hospital.
Once patients arrive at the hospital or
network of health care facilities (such
as clinics, rehabilitation facilities, or
long-term care facilities), the second
determinant of surge capacity is the
availability of hospital resources. The
number of patients who can be treated
during a surge event depends on such
factors as the number of staff, medical
supplies, and the circulation of beds.
The circulation of beds, in turn,
depends to some extent on the third
determinant, which is surge discharge.
The concept of surge discharge is based
on Dr. Hupert's work with Sam Benson,
AEMT-P, of the New York City Office
of Emergency Management. The
concept is that some patients who were
already in the hospital at the time of the
event may be transferred to other
facilities, thus making space available
for surge arrivals. Some percentage of
arriving patients also can be treated and
discharged. Examples of places patients
can go if they are ready for discharge
from the hospital include:
- Home, perhaps with skilled nursing care available at home.
- An out-of-region facility.
- A special, skilled nursing facility.
Dr. Hupert cited the example of his own hospital's
experience on September 11 and 12,
2001, when New York Presbyterian
Hospital was able to discharge
approximately 20 percent of its
patients to be able to treat arriving
casualties.
Dr. Hupert described a second model
he and his colleagues have developed
for pre-hospital management in
response to a biological attack or an
outbreak of an infectious disease. A
simple, conceptual model is presented
in Exhibit 3.
The model shows that
degree of protection is a function of
how long it takes the community to
respond to the attack or outbreak (delay
in reaction) and the time required to
administer medication to the affected
population (time needed to protect
community). Another way to think
about this is the time that elapses before
the first pill is dispensed, which is the
delay, and the time that passes until the
last pill is dispensed, which is the time
needed for the entire campaign. The
quicker the community can respond,
and the quicker the community can
administer medication, the lower the
potential patient load on community
hospitals will be.
Dr. Hupert and his colleagues have
translated this conceptual model into a
spreadsheet version that calculates
outcomes for pre-hospital anthrax
prophylaxis. The formulas in the
spreadsheet start with consensus
estimates of how inhalational anthrax
might act if there were an outdoor
release of anthrax spores, and
incorporate a set of assumptions about
detection, effectiveness of treatment,
and patient compliance with
medication. Planners can enter
numbers into the spreadsheet to
generate the percentage of exposed
individuals who do not get sick
because they receive antibiotics in
time.
The percentage is highest if the
community responds quickly and is
able to dispense antibiotics to the
affected population in a matter of a
few days. That percentage declines
over time, of course, and those people
who subsequently get sick create an
increase in the numbers of patients
who must be admitted to the hospital.
Thus, the model illustrates the
importance of pre-hospital treatment
in reducing hospital surge.
Applied differently, the model can also
predict when surge is most likely to
occur. The modeling shows that
planners can take one exposure
scenario, play it out in a number of
different ways, and wind up with very
different daily and total casualty loads,
depending upon how the community
has been organized and how
successful the community is in
responding to that event.
The components a community must
have in place to prepare for mass
prophylaxis are explained in detail in
Dr. Hupert's Community-Based Mass
Prophylaxis: A Planning Guide for
Public Health Preparedness. The
Planning Guide was also developed
under an AHRQ grant and is available on the AHRQ Web site (http://www.ahrq.gov/research/cbmprophyl/cbmpro.htm).
A companion piece on the
AHRQ Web site, the Bioterrorism and
Epidemic Outbreak Response Model
(BERM), is an interactive database that
makes it possible for planners to
calculate the actual numbers of
facilities and staff they will need in
their communities to prepare for mass
prophylaxis (http://www.ahrq.gov/research/biomodel.htm).
Dr. Hupert
and his colleagues are developing a
Web-based version of the spreadsheet
that calculates outcomes for pre-hospital
anthrax prophylaxis, the
Regional Hospital Caseload
Calculator, which will also be made
available on the Bioterrorism Preparedness
section of the AHRQ Web site
(http://www.ahrq.gov/prep/).
Return to Contents
Addressing Pediatric and School-Based Surge Capacity
Dr. Shannon and the Center for
Biopreparedness at Children's Hospital
Boston have developed several
pediatric-focused tools under AHRQ
contracts and are developing several
protocols for mass casualty events
involving children. Before discussing
those protocols he described some of
the key issues in planning for surge
capacity involving children.
Dr. Shannon said, in creating surge
capacity plans for children, it is
important to keep in mind and consider
every potential type of disaster, terrorist
or otherwise; this means that planning
should be based on the all-hazards
approach. In addition, surge capacity
planning should include schools, with
two contingencies in mind: one is the
school as a specific target of a terrorist
event or as the scene of a disaster of
some type. The second is a community
disaster that occurs while school is in
session.
Key Challenges in
Consequence Management
Every aspect of consequence
management will be challenged in an
incident that involves large numbers of
children:
- All of the first responders—emergency medical services staff,
firefighters, and police officers—are going to have a difficult time
when they are taking care of many
injured or traumatized children.
- If an event involves an infectious
agent or patients who are
contaminated with a chemical
agent, triage and decontamination
must take place outside the hospital
to protect the hospital campus from
infection or contamination. This
will require establishing control of
all points of ingress and egress.
- Decontamination teams will be
challenged by having to take care of
frightened children or children who
are very small, particularly while
the teams are wearing cumbersome
personal protective equipment.
- The challenge of decontaminating
children will be compounded by
inclement weather, and especially
cold weather.
- If the event requires mass
distribution of antibiotics, antidotes,
or vaccines, health care workers
should expect the triage and
screening of a child to take twice as
long as it would for an adult,
because children will not be able to
provide details of their medical
history.
- General Emergency Departments—those that take care of mixed
populations, children as well as
adults—are going to be challenged
by having to treat very large
numbers of children with limited
amounts of pediatric supplies.
Hospitals in general will be
similarly challenged.
- There will be enormous needs in
terms of mental health response and
recovery, not only for children but
also for parents and health care
workers.
Special Needs of Children
Because children are different from
adults, they need to be treated
differently in a mass casualty event.
- In
terms of a biological event, children
have an immature immune system,
which means that they are less able to
resist some types of infections, thus
leading to greater morbidity and
mortality in children.
- In a chemical
event, children are a challenge for
anyone to assess, manage, and guide
through decontamination.
- If there were
a radiological event, children are much
more susceptible to the consequences,
particularly the development of cancer.
- If there were a blast injury of some
type or any blunt trauma, treating large
numbers of children in what
Emergency Medicine calls the "golden
hour of trauma" would be difficult.
- In an event involving burns, children have
a greater likelihood of life-threatening
fluid loss and susceptibility to infection.
The School Crisis Response
Manual: Guiding Principles
Dr. Shannon and his colleagues are
currently developing a protocol for
collaboration between hospitals and
schools to prepare and train for
emergency response. Visits to school
districts around the country have
convinced him that guidelines are
desperately needed to prepare for a
school's being the target of a terrorist
event.
The main steps in creating a school-based
emergency response plan are:
- Get a sense of the Principal's level
of understanding and his or her
communication with local public
health authorities.
- Work with school nurses to ensure
there is a plan for taking the
children's medications with them
during evacuation and relocation.
- Include a sheltering and/or a lock-down
plan.
- Determine the best sites for
sheltering or lock-down, based on
an understanding of the unique
architecture of the school, the
location of the heating systems,
routes of ingress and egress, and so
on.
- Include after-school programs in the
planning and identify the person
who would be in charge.
- Integrate with existing plans; for
example, with the evacuation and
relocation plan for a fire drill.
- Take inclement or stormy weather
into account.
- Conduct drills, including tabletop
exercises with the key leadership of
the school.
Protocols for Disaster Medical
Assistance Teams and
Emergency Departments
Dr. Shannon and his colleagues are also
developing protocols for Disaster
Medical Assistance Teams (DMAT)
and Emergency Departments in
response to a mass casualty incident
involving children. DMATs are part of
the National Disaster Medical System
and have been in existence for some
time. They are designed to respond to a
disaster of any type.
Of approximately
30 existing DMATs, however, only two
are dedicated pediatric teams. More are
needed; at the very least, the system
should consider having these teams
regionalized so that the site of the
emergency can be reached more
quickly if needed. In addition, existing
DMAT teams should receive additional
training in taking care of children.
The challenges Emergency
Departments will face include:
- Having adequate number of
pediatric supplies.
- Having in place pediatric
decontamination equipment and
protocols.
- Having developed an effective
response plan to manage large
numbers of children.
- Having staff skilled at assessing
young, non-verbal children.
- Taking care of children while
wearing personal protective
equipment.
- Creating systems for identifying,
tracking, and reuniting children
with their families.
- Having identified the rational use
of alternate sites for taking care of
children, such as schools or
neighborhood health centers.
- Allocating resources to treat the
"second wave" of patients.
Dr. Shannon said schools and
neighborhood health centers can be
used to accommodate the "worried
well" and to treat the "walking
wounded" so that the Emergency
Department can be reserved for those
who really need immediate help. As
was learned from the Sarin incident in
Tokyo in 1995, the first wave of
patients arriving at the hospital are able
to get there on their own and are
relatively healthy. The more seriously
injured and ill have to be treated at the
scene and transported to the hospital;
they are the "second wave."
The AHRQ products being developed
by Children's Hospital Boston are in
various stages of completion. Their
availability will be announced on the
AHRQ Web site.
Return to Contents
Addressing Surge Capacity Through Information Exchange
Dr. Bogdan explained that one of the
key components of surge response is
risk and crisis communication: getting
out the appropriate messages to the
public so they know what is going on
and what is available to help them. This
would reduce the number of people
presenting to health care facilities who
may not need to be seen—the "worried
well"—which would allow hospital
resources to be used most effectively to
treat people who need these resources
the most.
The challenges for communicating
with the public during an emergency are
illustrated by the findings of the
Redefining Readiness Project from the
Center for Advancement of
Collaborative Strategies and Health
(http://www.cacsh.org). According to
this study, 60 percent of the public said
they would not heed official
instructions to get vaccinated during a
smallpox outbreak, and 40 percent of
the public would not heed official
instructions to shelter in place during a
dirty bomb incident. The conclusion to
be drawn is that people won't
necessarily accept and follow
instructions, but want to decide for
themselves what is best for their
families. Call centers provide a
mechanism for the public to ask their
questions and express their concerns to
enable them to make their own
decisions.
The public needs to know general event
information—who, what, where, when,
and why—and they need to trust the
source of the information. They want to
find out how they can protect
themselves and their families. They
want to know what State and local
health departments recommend and, if
they have further needs, they want to be
able to refer to the most appropriate
agencies to take care of those needs.
They also want help with making
decisions. They want to know, based on
their own circumstances, their own
problems, and their own underlying
medical conditions, what they should
do next to take care of themselves and
their families.
The "natural fits" to provide needed
information are the Poison Control
Centers that exist across the country;
nurse advice lines, which many HMO
agencies utilize; drug information
centers; and public health agencies.
Exhibit 4 illustrates how a medical call
center works. The general public and
health care providers can contact these
centers in a variety of ways, but the
most common mode of communication
is the telephone. Some callers will
describe symptoms that require a visit to
a health care facility for further
evaluation, but many questions or
concerns can be answered effectively
over the telephone.
In a surge event, the
medical call center fulfills its purpose by
reducing the number of patients who seek care at a hospital or other health
care facility, allowing those agencies to
focus on helping those with the greatest
need.
The Colorado Health Emergency Line
for the Public, or CO-HELP, is a
medical call center that has been in
operation since January 2003. This
service was created to provide a
standardized, prepared response to
health events in Colorado and to
provide consistent and accurate
information to callers. It was also
designed to collect and maintain
structured data that could help
characterize events for future responses.
The service was funded by the State
Health Department to provide the
capacity and the capability to adapt to
emerging health threats.
CO-HELP refers users to the right
agencies for further assistance. These
include the State's poison control
center, nurse advice lines, health care
providers, and guidelines and protocols
available from local and State health
departments. CO-HELP's users include
the general public, health care
providers, hospitals, clinics, health
agencies, and schools. It was accessed,
for example, by tourists who were
planning to visit Colorado in 2003,
when West Nile virus and influenza
were concerns.
Dr. Bogdan and his colleagues
developed the Health Emergency
Assistance Line and Triage Hub
(HEALTH) Model under a contract with
AHRQ. The report describes how the
Denver Health Medical Information
Centers determined the requirements,
specifications, and resources needed to
develop a public health emergency
contact center that is highly integrated
with public health agencies. The report
is available on the Bioterrorism
Preparedness section of the AHRQ Web
site, as is the HEALTH Contact
Center Assessment Tool
(http://www.ahrq.gov/research/health/). The Assessment Tool was
developed to assist other agencies in
developing the capabilities and
functions of the HEALTH Model.
"The greatest benefit of using
a medical call center is to be
able to provide the public with
a mechanism for one-on-one
information exchange." |
Return to Contents
Discussion
The question and answer period
included discussion of the PHS-CS,
including time needed to deploy them
and how they will be staffed. Staffing
of Medical Call Centers was also
discussed.
Public Health Service
Contingency Stations
Dr. Claypool explained that the time
required to deploy the PHS-CS is
expected to evolve as the Stations
themselves are developed in stages and
as they are procured in greater
numbers. The initial set of four Stations
that were purchased in 2004 were tested
in late 2004 and early 2005. The
component parts of these Stations are
palletized and are designed to be set up
in a facility such as a high school
gymnasium. Dr. Claypool estimated
that, initially, response to a medical
event might take 24 to 48 hours.
The goal of the PHS-CS program over
the next few years is to place the
Stations in ready positions all around
the country, in each of the Federal
Emergency Management Agency's
Health Emergency Service Regions.
The corresponding goal is to be able to
deploy one or more of the Stations in a
matter of hours. Dr. Claypool gave the
example of having to quarantine the
passengers on a 747 jet at an airport:
Rather than hold the passengers on the
airplane, parked on the tarmac for 24 or
48 or 72 hours, the target would be to
set up a 250-bed Contingency Station
with its quarantine capability in a
hangar at the affected airport in as few
as 4 hours.
The staffing of a Contingency Station is
separate from its physical components.
Dr. Claypool said that the basic unit,
that is, a 50-bed module that has
medical service ward capability, would
require two doctors who would be
supported by physician's assistants,
registered nurses, medical technicians,
and some administrative staff. The total
would be 20 or more qualified staff
members.
In response to a question about the
relationship between DMATs and the
PHS-CS, Dr. Claypool said it is likely
that both would be activated in response
to a major event, and they would be
expected to work in concert and to
coordinate with each other.
Regarding both time to deploy and the
number of staff needed, Dr. Shannon
cited his experience as a member of one
of the Pediatric DMATs. He said that
the actual transport time to get the
Children's Hospital Boston team to the
site of an emergency is "as quick as any
type of plane ride you can imagine."
More time is involved, however, in
mobilizing the team, including
contacting team members, and
mobilizing their equipment. That
mobilization can add many, many hours
to the time for deployment, which is
another argument for regionalization of
Pediatric DMATs. The size of the
Pediatric DMAT varies from 10 to 20
members, depending on the size of the
emergency and whether or not that
Pediatric DMAT team is incorporated
into another DMAT team.
Dr. Claypool pointed out that the
Contingency Stations will have pediatric
capability, including pediatric
formulations of drugs and pediatric beds.
Medical Call Centers
A question submitted by E-mail asked,
"In the event that call centers are
utilized as a tool for mass casualty
events and/or infectious disease
outbreaks, what personnel will be used
to support these call center
infrastructures already in place?"
Assuming the surge of incoming calls
would be more than regular staff could
handle, "Where will the additional
personnel come from? Volunteers?
Public health people already in the
organizations?"
Dr. Bogdan responded that the medical
call centers have the technology
infrastructure to handle large volumes
of calls, "But we always come up short
with personnel." He explained that
additional personnel can be trained to
perform limited functions, especially
those such as setting up prepared
answers to questions and helping the
caller to understand the answers. His
experience in Colorado has included
considering pools of volunteers and
establishing relationships with the
Public Health Departments so they can
supply personnel to help staff the
phones.
Another successful approach
has been to use recorded messages that
callers hear when they first connect
with the call center. Those messages
can be changed to reflect the
information that is most commonly
required at different points in time.
Dr. Bogdan cited his experience with the
influenza outbreak in Colorado in 2003.
Initially, the most common question was
"Where can I get my shot?" Callers who
asked that question could select from a
menu of options that provided Web
addresses containing information on
locations in their neighborhoods where
they could get their shot. The question
most frequently asked then quickly
became "What are the signs and
symptoms of influenza?" The call
center's message was changed to provide
that information with the result that
callers could learn what they wanted to know without necessarily having to be
connected to a live human being.
Return to Contents
For More Information
The Web conference on which this
Issue Brief is based, Addressing Surge
Capacity in a Mass Casualty Event, is
available as a streaming presentation
and as a text transcript on the
Bioterrorism Preparedness section of
AHRQ's Web site
(http://www.ahrq.gov/news/ulp/btsurgemass/).
Several resources on health system
preparedness for bioterrorism have been
developed with funding from AHRQ
and are available on the AHRQ Web
site. The Rocky Mountain Regional
Care Model for Bioterrorist Events
(http://www.ahrq.gov/research/altsites.htm) presents information on alternative
care facilities in response to a surge
event and the personnel resources
needed to staff those facilities.
Return to Contents
This issue brief was prepared for
AHRQ by the National Academy for
State Health Policy under contract
number 290-98-0009. |
AHRQ Publication No. 06-0027
Current as of January 2006
Internet Citation:
Addressing Surge Capacity in a Mass Casualty Event. Bioterrorism and Health System Preparedness, Issue Brief No. 9. AHRQ Publication No. 06-0027, January 2006. Agency for Healthcare Research and Quality, Rockville, MD. http://www.ahrq.gov/news/ulp/btbriefs/btbrief9.htm