Mobilizing Mobile Medical Units for Hurricane Relief: The United States Public Health Service and Broward County Health Department Response to Hurricane Wilma, Broward County, Florida

Abstract

To describe the outcomes of a collaborative response of federal, state, county, and local agencies in conducting syndromic surveillance and delivering medical care to persons affected by the storm through the use of mobile medical units.
Nine mobile medical vans were staffed with medical personnel to deliver care in communities affected by the storm. Individual patient encounter information was collected.
A total of 14,033 housing units were approached and checked for occupants. Of residents with whom contact was made, approximately 10 percent required medical assessment in their homes; 3,218 clients were medically evaluated on the mobile medical vans. Sixty-two percent of clients were female. The most common presenting complaints included normal health maintenance (59%), upper respiratory tract illness (10%), and other illness (10%). Injuries occurred in 9 percent. A total of 1,531 doses of medications were dispensed from the mobile medical units during the response.
Mobile medical units provided an efficient means to conduct syndromic surveillance and to reach populations in need of medical care who were unable to access fixed local medical facilities.

Full text

Mobilizing Mobile Medical Units for Hurricane
Relief: The United States Public Health Service and
Broward County Health Department Response to
Hurricane Wilma, Broward County, Florida
Melanie M. Taylor, William S. Stokes, Ronald Bajuscak, Mary Serdula, Karen L. Siegel, Brian Griffin, Jeffrey
Keiser, Lisa Agate, Aaron Kite-Powell, David Roach, Nancy Humbert, Kristin Brusuelas, and Sam S. Shekar
rrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrr
Objectives: To describe the outcomes of a collaborative
response of federal, state, county, and local agencies in
conducting syndromic surveillance and delivering medical care to
persons affected by the storm through the use of mobile medical
units. Methods: Nine mobile medical vans were staffed with
medical personnel to deliver care in communities affected by the
storm. Individual patient encounter information was collected.
Results: A total of 14 033 housing units were approached and
checked for occupants. Of residents with whom contact was
made, approximately 10 percent required medical assessment in
their homes; 3 218 clients were medically evaluated on the
mobile medical vans. Sixty-two percent of clients were female.
The most common presenting complaints included normal health
maintenance (59%), upper respiratory tract illness (10%), and
other illness (10%). Injuries occurred in 9 percent. A total of
1 531 doses of medications were dispensed from the mobile
medical units during the response. Conclusion: Mobile medical
units provided an efficient means to conduct syndromic
surveillance and to reach populations in need of medical care
who were unable to access fixed local medical facilities.
KEY WORDS: disaster relief, hurricane, mobile medical units,
syndromic surveillance
Hurricane Wilma, the 21st named storm of the 2005
Atlantic hurricane season, struck the southwest coast
of Florida on Monday, October 24, as a category 3
storm. Hurricane force winds more than 110 miles per
hour and rain caused widespread damage to homes
J Public Health Management Practice, 2007, 13(5), 447–452
Copyright C
2007 Wolters Kluwer Health |Lippincott Williams & Wilkins
Corresponding Author: LCDR Melanie Taylor, MD, MPH, United States Pub-
lic Health Services, Arizona Department of Health Services, Office of Infectious
Disease Services, 150 N 18th Ave, Suite 140, Phoenix, AZ 85007 (taylorm@
azdhs.gov).
qqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqq
LCDR Melanie M. Taylor, MD, MPH, is a Medical Epidemiologist, United States
Public Health Services Commissioned Corps, Rockville, Maryland, and Division of
STD Prevention, National Center for HIV, Viral Hepatitis, STD, and TB Prevention,
Centers for Disease Control and Prevention, Atlanta, Georgia.
RADM William S. Stokes, DVM, is Deputy Team Commander, Assistant Surgeon
General, United States Public Health Services Commissioned Corps, and Director,
National Toxicology Program Interagency for the Evaluation of Alternative
Toxicological Methods, National Institute of Environmental Health Sciences, National
Institutes of Health, Research Triangle Park, North Carolina.
CAPT Ronald Bajuscak, DMD, MS, is Response Team Commander, National
Consultant for Oral Medicine/Pathology, United States Public Health Services
Commissioned Corps, Rockville, Maryland.
CAPT Mary Serdula, MD, MPH, is a Medical Epidemiologist, United States Public
Health Services Commissioned Corps, Nutrition Branch, Division of Nutrition, Physical
Activity and Obesity, National Center for Chronic Disease Prevention and Health
Promotion, Division of Nutrition and Physical Activity, Centers for Disease Control and
Prevention, Atlanta, Georgia.
CAPT Karen L. Siegel, MA, PT, is Team Leader, United States Public Health Services
Commissioned Corps, and Senior Staff Specialist, National Institutes of Health,
Physical Disabilities Branch (a collaboration between the National Institute of Child
Health and Human Development and the NIH Clinical Center), Bethesda, Maryland.
LCDR Brian Griffin, BSN, RN, is Ready Responder, United States Public Health
Services Commissioned Corps, National Health Service Corps, Health Resources and
Services Administration, Philadelphia, Pennsylvania.
Jeffrey Keiser, BBA, BHS, is Deputy Administrator, Broward County Health
Department, Florida Department of Health, Ft Lauderdale, Florida.
Lisa Agate, MSW, is HIV/AIDS Program Director, Broward County Health
Department, Florida Department of Health, Ft Lauderdale.
Aaron Kite-Powell, MS, is Surveillance Epidemiologist, Bureau of Epidemiology,
Florida, Department of Health, Tallahassee.
David Roach, BA, is Administrator, Broward County Health Department, Florida
Department of Health, Ft Lauderdale.
Nancy Humbert, MSN, ARNP, is Deputy Secretary for Health, Florida Department of
Health, Tallahassee.
Kristin Brusuelas, MPH, is Senior Management Official, Office of the Director,
Centers for Disease Control and Prevention, Atlanta, Georgia.
RADM Sam S. Shekar, MD, MPH, is Assistant Surgeon General, United States Public
Health Services Commissioned Corps, and Director of Clinical Research Grants,
Office of Extramural Research, National Institutes of Health, Bethesda, Maryland.
447

448 ❘Journal of Public Health Management and Practice
and power lines, and temporarily displaced thousands
of residents. Initial estimates indicated that the most
severe damage occurred along the southeastern coast
of Florida.1In Broward County, the hurricane caused
widespread extended power outages affecting approx-
imately 900 000 homes and businesses and lasting over
3 weeks in some neighborhoods.2,3 This effectively shut
down routine services and impeded daily operation
of health and social services for the approximately 1.7
million residents. Hurricane Wilma resulted in an esti-
mated 41 deaths (9 in Broward County)4and was de-
termined to be the third costliest storm in US history,
with damages estimated at $14.4 billion.5
On October 28, 2005, United States Public Health Ser-
vice (USPHS) officers were deployed at the request of
state and local health authorities to the Broward County
Health Department (BCHD) to staff mobile medical
clinics in order to augment local relief efforts and to de-
liver medical services to the areas most heavily affected
by Hurricane Wilma. Nine mobile medical vans were
staffed with USPHS, local health department, and De-
partment of Veterans’ Affairs (DVA) personnel, as well
as Centers for Disease Control and Prevention (CDC)
Public Health Advisors (PHAs) stationed within the
state of Florida. These mobile teams were dispatched
to multistory retirement communities, mobile home
parks, low-income housing facilities, and other areas
that experienced significant structural damage and/or
that requested medical assistance due to the inability of
residents to leave their premises and access local medi-
cal facilities. Because of the widespread power outages
lasting up to 3 weeks in Broward County, many res-
idents with limited mobility living above the ground
level were trapped in their homes with limited supplies
due to inoperable elevators. This was compounded by
the disruption of telephone communications necessi-
tating door-to-door assessment of medical need. Many
of these communities as well as other low-income com-
munities heavily affected by the storm were identified
as priority areas needing medical assistance. The out-
comes of the collaborative efforts of the USPHS, the
CDC, Florida Department of Health, the BCHD, and
the DVA, in delivering medical relief services and con-
ducting syndromic surveillance in response to Hurri-
cane Wilma, are summarized here.
●Methods
Areas of need were identified by preassessment teams
comprising experienced outreach workers, health ed-
ucators, and case managers employed by the BCHD.
These teams conducted assessments of communities af-
fected by the storm. Environmental health workers and
community nurses reported additional areas of medical
need to the BCHD. Priority for mobile van deployment
was given to areas without electricity, areas with homes
blocked by fallen debris, and communities where citi-
zens with known medical needs resided such as devel-
opments housing elderly persons. In addition, teams
consisting of staff from the CDC, the North Carolina
Department of Health, and the Florida Department of
Health conducted a rapid needs assessment that in-
cluded 210 interviews within 42 individual county cen-
sus blocks that represented the approximately 375 580
households in northeastern Broward County on the ba-
sis of 2000 US Census estimates. The assessments indi-
cated that 5 percent to 15 percent of the households
had members with one or more medical needs. From
this information, public health department officials and
the county Emergency Operations Center prioritized
additional areas of potential medical need and iden-
tified locations for mobile medical van deployment.
These locations consisted of retirement communities,
mobile home parks, low-income housing facilities/
apartments, churches, and other areas heavily affected
by the storm. Advanced notice of mobile van deploy-
ment to local communities was performed by distribu-
tion of flyers in the communities, contact with commu-
nity leaders or association presidents, and notification
of local law enforcement officials 1 to 2 days prior to
the planned arrival of the medical unit.
Nine mobile medical vans provided by local hospital
and clinic organizations and the DVA were staffed with
USPHS officers, BCHD staff, CDC PHAs, and DVA per-
sonnel. Each mobile van unit was typically staffed with
9 to 10 personnel that included a team leader, one med-
ical doctor, one mid-level provider (nurse practitioner
or physician’s assistant), one pharmacist, two nurses
and a medical assistant, one to two staff members from
the Broward County Health Department, a CDC PHA,
a county social worker, and one or two members of
the Florida Department of Elder and Veterans Affairs.
Broward County provided nurses that assisted in pa-
tient care needs and in the advance site inspections to
identify high-need areas for the mobile units. Services
provided on the mobile medical vans included health
maintenance assessments, primary medical care, moni-
toring oxygen and respiratory tract therapy, dispensing
medications, and arranging for continued services. The
BCHD coordinated the procurement of the mobile med-
ical vans, medical supplies, pharmaceuticals, staffing,
and transportation logistics. Advance agreements for
use of mobile medical units were not obtained prior to
the storm.
Data collection and analyses
Individual patient encounter information including
gender, age, race, ethnicity, zip code, and presenting

Mobilizing Mobile Medical Units for Hurricane Relief ❘449
medical complaint was collected using a standardized
surveillance form developed by the Florida Depart-
ment of Health for use in disaster response. A conve-
nience sampling method was used. Persons who sought
care on the mobile vans or were approached in their
homes were included in the analysis of reasons for visit.
Separate data analyses were not conducted on the ba-
sis of client self-presentation to the mobile vans versus
door-to-door visits by mobile unit staff. Client identi-
fiers were not collected. Data were compiled and en-
tered daily into an Excel database. Data analyses were
performed using Excel (Version 2000, Microsoft, Red-
mond, Washington).
Definitions of reason for visit to the mobile medi-
cal unit were defined on the standardized surveillance
form and included 15 syndrome categories: animal bite,
asthma, carbon monoxide poisoning, dermatologic dis-
order (rashes, sores insect bite, sunburn), febrile ill-
ness, gastrointestinal illness, heat-related injury, other
injury, lower respiratory tract illness (cough, shortness
of breath, pneumonia), mental health (stress, anxiety),
musculoskeletal system trauma (fracture, sprain, dis-
location), minor wound/injury (abrasion, laceration,
puncture, foreign body removal), normal health main-
tenance (blood pressure and blood glucose monitoring,
prescription refills), other (not fitting into other cate-
gories), poisoning (pesticides, gas fumes, ingestions),
and upper respiratory tract illness (cold, sore throat).
For analysis purposes, some definitions were combined
into larger categories including respiratory complaints
(asthma, upper and lower respiratory tract complaints),
and injuries (animal bite, heat-related injury, muscu-
loskeletal trauma, minor wound).
Diagnostic procedures such as sputum or nasopha-
ryngeal bacterial and viral cultures were not performed
on the mobile vans. Thus, tracking of diagnostic trends
was not available for purposes of epidemic detection.
Patients requiring more intensive care weretransported
by ambulance to functioning emergency care centers.
Clinical criteria for emergency referral were clinician
dependent and not defined by the mobile unit response.
●Results
During the period, October 28 to November 9, 2005,
mobile medical van units visited 51 housing develop-
ments. A total of 140 33 housing units were approached
and checked for occupants. Of these, approximately
40 percent (n=5 646) were occupied. Of residents
with whom contact was made, an estimated 10 percent
needed medical assessment in their homes. Most resi-
dents needing medical assessment in their homes were
not able to ambulate to mobile medical units due to
physical limitations. Health issues encountered among
these clients were similar, per staff reports, to those of
clients presenting as “walk-ins” to the mobile medical
units.
A total of 3 218 clients self-presented for medical
evaluation on the mobile medical vans or were med-
ically evaluated in their homes. The majority of these
clients (approximately 85%) presented as walk-ins to
the mobile medical units. Sixty-two percent of clients
were female (n=2 009); 67 percent (n=2 153) of
clients were White and 19 percent were Black (n=901).
Twenty-one percent (n=673) of clients reported being
of Hispanic ethnicity (Table 1).
The most common presenting syndromes included
normal health maintenance (59%, n=1 898), upper res-
piratory tract illness (10%, n=319), and other illness
(10%, n=310). The most frequent reasons for catego-
rization as normal health maintenance (N=1 898) in-
cluded blood pressure check (64%, n=1 215), medica-
tion refill (17%, n=325), supportive reassurance (6%,
n=118), and glucose check (4%, n=79). There were no
presentations of suspected carbon monoxide poisoning
(see Table 1). Current diagnosis of hypertension was re-
ported by 26 percent (n=836), diabetes by 10 percent
(n=328), and chronic obstructive pulmonary disease
by 3 percent (n=92) of clients. (see Table 1). The most
frequently dispensed prescription and nonprescription
medication types included pulmonary (15%, n=231),
pain relief (15%, n=229), antibiotics (13%, n=192),
and hypertension (12%, n=167). Forty-two tetanus
vaccines were administered (Table 2).
Evaluation of the reason for visit to the mobile med-
ical vans by date revealed an increasing proportion of
acute care needs emerging during the period of the re-
sponse (Figure 1). The proportion of visits for respira-
tory tract complaints, injuries, and mental health issues
increased gradually, with peaks in these complaints 4
to 6 days after the initiation of the medical response.
The most common reasons for emergency transport in-
cluded severe hypertension and hyperglycemia. The es-
timated number of persons transported for emergency
care for the above conditions during the response was
10.
●Discussion
Hurricane Wilma followed hurricanes Katrina and Rita
in the unforgettable storm season of 2005. Less attention
was given to this storm despite it being one of the most
costly storms in history. In part, this may be owing to
the minimal mortality caused by the storm; only an esti-
mated 41 fatalities were reported in comparison to the 1
200 plus deaths attributed to Hurricane Katrina.4Rapid
response by local and state governments to the needs of
citizens affected by the storm is likely to have prevented

450 ❘Journal of Public Health Management and Practice
TABLE 1 ●Demographics of clients presenting for medical
care in mobile van units or in homes visited by United
States Public Health Service personnel during the
Hurricane Wilma response (N=3 218)
qqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqq
Variable
Age
0–18 334 (10)
19–30 201 (6)
31–50 773 (24)
51–65 766 (24)
>65 1 109 (35)
Unknown 35 (1.1)
Gender
Male 1 204 (37)
Female 2 009 (62)
Unknown 5 (0.2)
Race
White 2 153 (67)
Black 901 (28)
Other/unknown 164 (5)
Ethnicity
Hispanic 673 (21)
Non-Hispanic 2 446 (76)
History of chronic disease
Hypertension 836 (26)
Diabetes 328 (10)
Chronic obstructive pulmonary disease 92 (3)
Clinical syndrome category
Animal bite 4 (0.2)
Asthma 76 (2)
Carbon monoxide poisoning 0 (0)
Dermatologic disorder 112 (3)
Febrile illness 0 (0)
Gastrointestinal tract illness 80 (2)
Heat related 0 (0)
Injury, other 95 (3)
Lower respiratory tract illness 29 (9)
Mental health 122 (4)
Musculoskeletal system trauma 124 (4)
Minor wound injury 62 (2)
Normal health maintenance 1 898 (59)
Other 310 (10)
Poisoning, other 0 (0)
Upper respiratory tract illness 319 (10)
Values given are number (percentage).
some morbidity and psychological stress. Mobile med-
ical units staffed by local health department person-
nel, DVA personnel, CDC PHAs, and USPHS officers
played an integral role in finding and providing care to
persons with medical conditions who were unable to
access medical care locally caused by or exacerbated as
a result of conditions produced by the storm. Interven-
TABLE 2 ●Medication types dispensed from mobile
medical units during the Hurricane Wilma response
qqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqq
Medication type
Diabetes 74 (5)
Hypertension 167 (11)
Cardiac 97 (6)
Antibiotic 192 (13)
Dermatologic 63 (4)
Wound care 108 (7)
Eye 30 (2)
Pulmonary 231 (15)
Pain 229 (15)
Immunization 42 (3)
Gastrointestinal 110 (7)
Other 188 (12)
Values given are number (percentage).
tion likely minimized the number of patients whose
condition may have otherwise deteriorated to a life-
threatening or fatal situation. Collaboration between
federal, state, and local agencies facilitated the delivery
of care to thousands of persons affected by Hurricane
Wilma.
The active illness and injury surveillance imple-
mented on the mobile medical units was similar to
that used in responses to hurricanes Katrina and Rita.6
These surveillance activities served multiple functions
during this response that included prompt evaluation
of trends in diagnoses, assessing unmet needs related
to medication refills, and informing mobile van de-
ployment or redeployment. It has been demonstrated
that less than 10 percent of persons arriving at a hous-
ing shelter following a disaster are in need of medi-
cal care.7This was also true for our convenience sam-
ple; most of the health issues encountered were not re-
lated to Wilma and although acute care needs were ad-
dressed in some patients, far more presented for normal
health maintenance needs (blood pressure checks, glu-
cose checks, medication refills) that might have been
prepared for prehurricane relief. A large percentage of
patients had chronic diseases. Following disasters, it is
these groups of persons that become unable to manage
their chronic conditions due to disruptions in the med-
ical care system.8For example, many health-related fa-
cilities such as clinics and pharmacies remained closed
for 1 to 2 weeks after Hurricane Wilma, and home
healthcare services were also disrupted. Indeed, sev-
eral cases of severe hypertension and hyperglycemia
were identified on the mobile units following Wilma.
Telephone service was also interrupted for extended
periods, which made it difficult or impossible for many
individuals to request needed medical or social as-
sistance. Social services personnel working with the
medical units were able to link persons to temporary

Mobilizing Mobile Medical Units for Hurricane Relief ❘451
FIGURE 1 ●Reasons for Mobile Medical Unit Visit by Day of Response
qqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqq
housing, elderly care services, and mental health ser-
vices. Mobile unit staff participating in the Hurricane
Wilma response were able to reach a large number of
persons with acute and chronic medical, mental health,
and social services needs as they arose in the days af-
ter the event. Without these interventions, it is likely
that many of the chronic medical conditions would
have progressed to more severe conditions requiring
emergency department visits or even hospitalization,
resulting in further stress on limited medical resources
already overwhelmed by the disaster.
Priority was given to developments housing elderly
persons owing to the concern regarding limited mobil-
ity due to the loss of electricity resulting in inoperable
elevators, limited lighting, and the inability to recharge
power-assisted wheelchairs and other devices. Door-
to-door visits conducted by mobile van staff concur-
rent with the provision of medical care on the mobile
vans identified persons in medical need who were not
otherwise able to present for care. Other health needs
such as the need for food, water, medication refills,
and assistance with ambulating were identified and
fulfilled through these visits as well. The needs of the
elderly during this and another disasters differ from
those of younger populations and require specialized
attention and response that may include door-to-door
visits.
This is one of few reports to describe the use of
mobile medical units to provide medical care and to
gather medical surveillance information during hur-
ricane relief efforts in the United States. Provision of
medical care and medical surveillance using a mobile
medical response effort should be considered for nat-
ural or manmade disasters where there is limited abil-
ity of individuals to access centralized facilities due to
physical disabilities, fuel shortages, limitations in pub-
lic transportation, or medical quarantine. These results
are based on a convenience sample of persons who pre-
sented to or were contacted by the staff of the mobile
medical units, and thus the findings may not represent
other populations affected by this hurricane. Despite
the sampling method used, the race/ethnicity distribu-
tion of this sample mirrors that of the US Census Bureau
2004 Community Survey for Broward County.8This
disaster response relied upon the functioning infras-
tructure of the local health department and emergency
medical services. These activities may be difficult to co-
ordinate in situations where a breakdown in the public
health infrastructure occurs, such as was demonstrated
after Hurricane Katrina.8,10
The Hurricane Wilma response demonstrated the
critical role of the state and local health departments
in organizing the disaster response effort. Equipped
with knowledge of health, socioeconomic, and social
demographics of their communities, the BCHD was
able to coordinate with the county Emergency Op-
erations Center to identify the highest priority areas
for mobile medical unit deployments. The BCHD ex-
pended considerable effort in trying to locate, in this
emergency situation, the proper persons to arrange for-
mal agreements for use of the mobile medical units.
Ideally, advance agreements should be made with or-
ganizations that operate and maintain mobile medical
units for their use during disaster response. This type of

452 ❘Journal of Public Health Management and Practice
coordination will prove critical in future disaster events
that may threaten the health of communities such as
pandemic influenza, bioterrorist events, or additional
natural disasters. The use of properly staffed mobile
medical van units may facilitate dispensing methods,
such mass vaccination, and antibiotic stockpile distri-
bution. Local, state, and federal public health agencies
should consider the use of mobile medical clinics in
their disaster preparedness planning and emergency
response.
REFERENCES
1. National Oceanic and Atmospheric Administration’s Na-
tional Weather Service. http://www.srh.noaa.gov/mfl/
events/?id=Wilma. Published December 1, 2005. Accessed
February 28, 2006.
2. Broward County Board of County Commissioners. Com-
munity update, November 2005. Broward County con-
tinues to recover from Hurricane Wilma. http://www.
broward.org/enews/communityupdate/november05.htm.
Published November 2005. Accessed March 15, 2006.
3. Office of Electricity Delivery and Energy Reliability (OE).
U.S. Department of Energy. Gulf Coast Hurricanes Situa-
tion Report #34, November 2, 2005. http://www.oe.netl.
doe.gov/docs/gulf coast/gulf 110205-1500.pdf. Published
November 2, 2005. Accessed March 15, 2006.
4. Ragan P, Schulte J, Nelson S. 2004–2005 Florida hurricane
mortality [abstract]. Presented at: Council of State and Ter-
ritorial Epidemiologists 2006 Annual Conference; June 4–8,
2006; Anaheim, California.
5. National Oceanic and Atmospheric Administration’s Na-
tional Weather Service. http://www.srh.noaa.gov/data/
NHC/TWSAT. Published October 30, 2005. Accessed Febru-
ary 28, 2006.
6. Centers for Disease Control and Prevention. Injury and ill-
ness surveillance in hospitals and acute-care facilities after
hurricanes Katrina and Rita—New Orleans area, Louisiana,
September 25–October 15, 2005. MMWR. 2006;55:35–38.
7. Mattox K. Hurricanes Katrina and Rita: role of individuals
and collaborative networks in mobilizing/coordinating soci-
etal and professional resources for major disasters. Crit Care.
2006;10:205–210.
8. Greenough PG, Kirsch TD. Public health response-assessing
needs. N Engl J Med. 2005;353:1544–1546.
9. United States Census Bureau. American Community Survey.
Broward County, Florida. http://factfinder.census.gov. Pub-
lished 2005. Accessed May 19, 2006.
10. Centers for Disease Control and Prevention. Surveillance for
Illness and Injury After Hurricane Katrina—New Orleans,
Louisiana, MMWR. September 8–25, 2005. 2005;54:1018–
1021.