:::
Green Tobacco Sickness in Tobacco Harvesters -- Kentucky, 1992
==============================================================
SOURCE: MMWR 42(13) DATE: Apr 09, 1993
Green tobacco sickness (GTS) is an illness resulting from dermal exposure
to dissolved nicotine from wet tobacco leaves; it is characterized by nausea,
vomiting, weakness, and dizziness and sometimes fluctuations in {*filter*} pressure
or heart rate (1-3). On September 14, 1992, the Occupational Health Nurses in
Agricultural Communities (OHNAC) project of Kentucky * received reports of 27
cases of GTS. The cases occurred among tobacco harvesters who had sought
treatment in several hospital emergency departments in south-central Kentucky
during the preceding 2 weeks. This report summarizes the findings of the
investigation of these cases.
On September 15, OHNAC staff initiated a review of inpatient and
emergency department medical records from May 1 through October 2 at five
hospitals in the Bowling Green and Elizabethtown areas. The review identified
55 persons in whom GTS, nicotine poisoning, or other illnesses compatible with
GTS symptomatology had been diagnosed. On September 25, industrial hygienists
from CDC's National Institute for Occupational Safety and Health (NIOSH)
observed the tobacco-harvesting process. Worker's hands, forearms, thighs, and
backs received the most dermal exposure to wet tobacco. Dew from tobacco
leaves often saturated workers' clothing within minutes of beginning field
work.
To evaluate possible risk factors associated with GTS, NIOSH
investigators and occupational health nurses from the OHNAC project conducted
a case-control study. A case was defined as an emergency department diagnosis
of GTS or nicotine poisoning in a person whose recorded work history included
tobacco harvesting at the time of illness. Forty-nine persons met the case
definition, with episodes occurring from July 25 through September 19, 1992;
two cases were subsequently excluded from analysis because illness onset
coincided with exposure to pesticides (which can induce similar symptoms).
Median age of the 47 case-patients was 29 years (range: 14-54 years); 41 (87%)
were male. Controls were 83 asymptomatic tobacco harvesters referred by case-
patients or local agricultural extension agents. Their median age was 39 years
(range: 16-70 years); 72 (87%) were male.
Twelve (26%) case-patients were hospitalized for 1-2 days; of these, two
(4%) required intensive-care treatment for hypotension and bradycardia. All
case-patients were initially treated in emergency departments with antiemetic
{*filter*}, and 35 (74%) received intravenous fluids.
Forty of 47 case-patients and 83 controls were administered a
questionnaire by telephone. Respondents were asked about the types of jobs
Health InfoCom Network News Page 3
Volume 6, Number 9 April 10, 1993
performed during the tobacco growing season, use of protective clothing,
exposure to wet tobacco leaves, work in wet clothing, work duration, and
personal tobacco use.
Among the 40 case-patients who completed interviews, the median time from
starting work to onset of illness was 10 hours (range: 3-17 hours); most
frequently reported symptoms included weakness (100%), nausea (98%), vomiting
(91%), dizziness (91%), abdominal cramps (70%), headache (60%), and difficulty
breathing (60%). The mean duration of illness was 2.4 days. Thirty-six (90%)
had previous work experience with tobacco. Of these, 14 (39%) had previously
sought medical care for symptoms suggestive of GTS. Seven{*filter*} (85%) of 20
case-patients aged greater than or equal to 30 years attributed their illness
to working in wet tobacco, compared with 12 (60%) case-patients aged less than
30 years.
Age less than 30 years was a risk factor for illness (odds ratio
OR=3.1; 95% confidence interval CI=1.4-7.0). All case-patients and 69
(83%) controls had worked in fields of wet tobacco where their clothes became
wet (OR=infinite; lower confidence limit=1.8). Current use of personal tobacco
products (i.e., cigarettes, snuff, chewing tobacco, pipe, or cigars) appeared
to be weakly protective, but the estimate was not statistically significant
(OR=0.7; 95% CI=0.3-1.5). Sex and work duration (i.e., number of hours per day
or number of days per week) were not associated with illness. The reported use
of protective clothing was similar for case-patients and controls; for case-
patients and controls combined, reported use of protective items worn at least
once during the growing season was 5% for waterproof clothing and 32% for
gloves.
Representative hospital costs were calculated for three levels of care
received by 31 case-patients treated at two participating hospitals. Fees
averaged $250 for outpatient treatment, $566 for hospital admission, and $2041
for intensive-care treatment.
Reported By: B Boylan, MS, Lincoln Trail District Health Dept, Elizabethtown;
V Brandt, Barren River District Health Dept, Bowling Green; J Muehlbauer,
Buffalo Trace District Health Dept, Maysville; M Auslander, DVM, C Spurlock,
PhD, Injury Epidemiology Section; R Finger, MD, State Epidemiologist, Kentucky
Dept for Health Svcs. Hazard Evaluations and Technical Assistance Br, and
Surveillance Br, Div of Surveillance, Hazard Evaluations, and Field Studies,
National Institute for Occupational Safety and Health, CDC.
Editorial Note: Before 1992, no cases of GTS had been reported to Kentucky
public health agencies. Increased surveillance of adverse health events in
persons working in agriculture and increased awareness of the condition may
explain the reports in Kentucky during this harvest season (i.e., late
summer). Before the NIOSH investigation was initiated, OHNAC occupational
health nurses had supplied emergency department physicians with literature
about GTS. In addition, rainfall during the 1992 season was
Health InfoCom Network News Page 4
Volume 6, Number 9 April 10, 1993
uncharacteristically heavy, potentially increasing exposure to wet tobacco and
incidence of GTS.
The lower risk for GTS among older workers may result from work practices
developed over time that reduce contact with wet tobacco. In addition, workers
likely to develop symptoms of GTS may leave this work force at a young age.
One potential limitation to these findings is that the age distribution of
controls may not reflect the local population of tobacco workers.
Personal use of tobacco products may be weakly protective, probably
because of development of tolerance to the effects of nicotine among regular
tobacco users. Tolerance may not be protective if dermal absorption
substantially exceeds the user's customary nicotine intake (4), which may have
occurred in this outbreak because of heavier than usual rains.
Approximately 60,000 persons harvest tobacco annually in Kentucky at
least part-time (5). The estimated crude 2-month incidence rate of hospital-
treated GTS among tobacco workers in the five-county study area was 10 per
1000 workers.** Statewide extrapolation of this incidence rate suggests as
many as 600 persons in Kentucky could have sought emergency department care
for the condition. However, this figure may underestimate the true incidence
of GTS because many affected persons may not seek hospital treatment (2).
Use of protective clothing (e.g., water-resistant clothing and {*filter*}
gloves) reduces the amount of nicotine absorbed by workers in contact with
green tobacco (6,7). Tobacco farm owners should inform their employees of the
hazards associated with harvesting wet tobacco and the importance of safe work
practices in preventing GTS; discuss routes of exposure and symptoms
associated with the disease; advise workers to change into clean, dry clothing
and boots during the work day if these become wet; and allow flexible work
hours to avoid work during or immediately after a rainfall. Health-care
providers in areas where tobacco is harvested should consider GTS in workers
who present with symptoms similar to those reported here.
To determine whether GTS regularly occurs or whether this outbreak was
due to an unusually wet growing season, the OHNAC project of Kentucky will
continue active surveillance for GTS in local hospitals and clinics during
tobacco growing seasons. The Kentucky Department for Health Services will
disseminate information on GTS to health-care professionals and institutions
statewide. Workers will be informed about the condition and preventive
measures through the Cooperative Extension Service and through press releases
to community newspapers.
References
1. Gehlbach SH, Williams WA, Perry LD, Woodall JS. Green tobacco sickness: an
illness of tobacco harvesters. JAMA 1974;229:1880-3.
2. Ghosh SK, Parikh JR, Gokani VN, Kashyap SK, Chatterjee SK. Studies on
occupational health problems during agricultural operation of Indian tobacco
Health InfoCom Network News Page 5
Volume 6, Number 9 April 10, 1993
workers. J Occup Health 1979;21:45-7.
3. Gehlbach SH, Perry LD, Williams WA, et al. Nicotine absorption by workers
harvesting green tobacco. Lancet 1975;1:478-80.
4. Goodman AG, Rall TW, Nies AS, Taylor P, eds. Goodman and Gilman's -- the
pharmacological basis of therapeutics. New York: Pergamon Press, 1975:548.
5. United States Department of Agriculture/Kentucky Department of
Agriculture. Kentucky agricultural statistics, 1991-1992. Frankfort, Kentucky:
Kentucky Department of Agriculture, 1992.
6. Gehlbach SH, Williams WA, Freeman JI. Protective clothing as a means of
reducing nicotine absorption in tobacco harvesters. Arch Environ Health
1979;34:111-4.
7. Ghosh SK, Gokani VN, Doctor PB, Parikh JR, Kashyap SK. Intervention
studies against "green symptoms" among Indian tobacco harvesters. Arch Environ
Health 1991;46:316-7.
* OHNAC is a national surveillance program conducted by CDC's National
Institute for Occupational Safety and Health (NIOSH) that has placed public
health nurses in rural communities and hospitals in 10 states (California,
Georgia, Iowa, Kentucky, Maine, Minnesota, New York, North Carolina, North
Dakota, and Ohio) to conduct surveillance of agriculture-related illnesses and
injuries that occur among farmers and their family members. These surveillance
data are used to reduce the risk for occupational illness and injury in
agricultural populations.
** The denominator for this rate is based on an estimate of 78.8 person-hours
worked per acre during tobacco harvest, the number of acres planted with
tobacco, and an estimate of 256 harvest-hours worked annually per worker (the
median value reported in the Kentucky GTS case-control study). These figures
generated an estimate of 4730 tobacco-harvest workers in the five affected
counties, of whom 47 sought medical treatment at local hospitals.
Health InfoCom Network News Page 6
Volume 6, Number 9 April 10, 1993
Emergency Mosquito Control Associated with Hurricane Andrew --
Florida and Louisiana, 1992
==============================================================
SOURCE: MMWR 42(13) DATE: Apr 09, 1993
Hurricane Andrew crossed south Florida on August 24, 1992, entered the
Gulf of Mexico, and struck the Louisiana coast on August 26. In Florida, an
estimated 25,000 housing units were destroyed and 37,000 severely damaged in a
200,000-acre area in the southern portion of Dade County; in Louisiana, an
estimated 25,000 housing units were destroyed or severely damaged by the
storm, primarily in the coastal sections of the 36-parish disaster area.
Initial assessment of the disaster areas indicated a need for vector
surveillance and control (1). This report summarizes actions to assess and
alleviate mosquito-related problems in Florida and Louisiana.
Persons residing in the affected areas or returning after the initial
evacuation were exposed to high densities of mosquitoes (e.g., because of
damage to door and window screens and lack of electricity to run air
conditioners). In addition to being a nuisance that hampered recovery efforts
(e.g., repair and reconstruction crews were unable to work during early
morning and late afternoon/early evening hours), this exposure increased the
potential for mosquito-transmitted diseases among recovery workers and
displaced residents, and secondary bacterial infections of mosquito bites
among children were reported in both states. Florida
Dade County Mosquito Control monitored morning and evening mosquito
landing rates at 27 sites beginning September 1. On September 2, carbon
dioxide-baited encephalitis vector survey (EVS) traps were placed at eight
locations and monitored daily by the U.S. Navy Disease Vector Ecology and
Control Center. Aedes taeniorhynchus and Culex nigripalpus were the
pre{*filter*} mosquito species in the area. Daytime landing rates of nuisance
mosquitoes in early September exceeded 20 per minute in sampling sites near
coastal sections of the disaster area. During the 50 days following the
hurricane, 659,458 acres in Dade County were treated by aerial application of
mosquito-control insecticides, resulting in substantially reduced landing rate
counts. For example, after one application to 99,000 acres on September 11,
landing rates in the area were reduced from an average of 14.3 mosquitoes per
minute to 0.4 mosquitoes per minute, and EVS trap collections decreased from
an average of 550 Ae. taeniorhynchus per night to 20 per night.
Mosquito-based surveillance for St. Louis encephalitis (SLE) was
conducted from September 8 through October 15; 28,369 specimens (primarily Cx.
nigripalpus) in 402 pools were tested by antigen capture enzyme-linked
immunosorbent assay (ELISA). No SLE viral antigen was detected.
The presence of competent mosquito vectors (Ae. aegypti and Anopheles
quadrimaculatus) and of recent immigrants from the Caribbean Islands and Latin
America raised the possibility of dengue and malaria transmission in Florida.
Because mosquito-based surveillance for SLE is unable to detect these
Health InfoCom Network News Page 7
Volume 6, Number 9 April 10, 1993
diseases, fliers with information on identification and reporting of dengue
and malaria were distributed to health-care workers in the area. No dengue or
malaria cases were reported to the Florida Department of Health and
Rehabilitative Services.
Louisiana
The disaster area in Louisiana comprised 36 parishes. Nine were
considered candidates for a large build-up of nuisance mosquitoes or for
transmission of mosquitoborne pathogens because early surveillance indicated
that large numbers of mosquitoes were appearing in urban and suburban areas
with large human populations. Emergency mosquito surveillance programs were
established in eight of the 36 parishes, and existing surveillance programs in
three parishes were augmented.
Densities of nuisance mosquitoes were estimated with carbon dioxide-
baited CDC light traps and landing/biting surveillance conducted by the parish
mosquito-control districts. The number of telephone complaints received by
health departments or vector-control agencies were also used to assess
mosquito biting activity. The most common nuisance species collected were
Psorophora columbiae, Cx. salinarius, Psorophora ferox, and Ae. sollicitans.
During the 36 days following the hurricane, approximately 788,000 acres were
treated by aerial and ground application of mosquito-control insecticides by
the cooperating parish mosquito-control programs and private mosquito-control
contractors. Pretreatment and posttreatment surveillance indicated immediate
but short-term reductions in nuisance mosquitoes.
Before and after the storm, the Louisiana Department of Health and
Hospitals and the Louisiana Mosquito Control Association reported no eastern
equine encephalitis (EEE) or SLE virus activity in their wild bird sampling
program, indicating that transmission of arboviral disease was unlikely
following the hurricane. Carbon dioxide-baited CDC light trap collections
after the storm were processed for virus detection by the state public health
laboratory; 2738 mosquitoes (131 pools) of known or suspected vector species
(Culiseta melanura, Coquillettidia perturbans, Cx. pipiens quinquefasciatus,
and Cx. salinarius) were tested for the presence of EEE and SLE viruses. No
arbovirus activity was detected.
Reported by: M Nelms, Dade County Mosquito Control, Miami; US Navy Disease
Vector Ecology and Control Center, Naval Air Station, Jacksonville; WR Opp,
Florida Dept of Agriculture and Consumer Svcs; Florida Dept of Health and
Rehabilitative Svcs. MM Yates, East Baton Rouge Mosquito Abatement and Rodent
Control; GM Stokes, Mosquito Control Contractors, Inc, Iberia Parish, New
Iberia; BB Broussard, Vermilion Parish Mosquito Control, Abbeville; CL Meek,
Louisiana State Univ and Louisiana Mosquito Control Association, Baton Rouge;
MJ Keppinger, Louisiana Dept of Agriculture and Forestry; C Anderson, HB
Bradford, B Savoie, L McFarland, DrPH, State Epidemiologist, Office of Public
Health InfoCom Network News Page 8
Volume 6, Number 9 April 10, 1993
Health, Louisiana Dept of Health and Hospitals. Federal Emergency Management
Agency, regions 4 and 6. US Air Force 910th Airlift Group, Youngstown Air
Reserve Station, Vienna, Ohio. Medical Entomology/Ecology Br, Div of Vector-
Borne Infectious Diseases, National Center for Infectious Diseases, CDC.
Editorial Note: In Florida, Dade County nuisance mosquito-population densities
after the storm were at approximately normal levels for that time of year, and
mosquito species collected were routinely found in large numbers in the area
(Dade County Mosquito Control Program, unpublished surveillance data, 1992).
Increased human exposure to mosquitoes in the Florida disaster area occurred
primarily because of the extensive damage to housing, and mosquito densities
that were tolerable before the storm were unacceptable when human exposure
increased. In the affected Louisiana parishes, storm-associated rainfall
substantially increased nuisance mosquito populations, and displaced persons
were exposed to higher than usual mosquito densities. In East Baton Rouge
Parish, landing rate indices were 212 times higher than the September averages
for 1983-1991, and light trap indices following the storm were 2.1 times
higher than normal in Iberia Parish for 1980-1991.
Federal assistance for emergency vector surveillance and control is
available when a disaster is declared and when one or more of the following
conditions are met* (2): 1) transmission of human or animal disease is in
progress or is deemed imminent, 2) reconstruction efforts are substantially
hampered by large populations of nuisance species, 3) normal functioning of
communities in the disaster area is substantially disrupted, or 4) the large
nuisance populations place additional stress on the human population.
Mosquito-transmitted pathogens were not detected in either disaster area, and
emergency mosquito control was primarily intended to provide relief from high
mosquito densities that hampered recovery efforts. Surveillance after control
measures were implemented indicated that mosquito populations had decreased
markedly.
Although Cx. nigripalpus, the vector of SLE virus, is present in Dade
County, SLE virus activity in the area is historically low (Florida Department
of Health and Rehabilitative Services, unpublished surveillance data). SLE
virus activity throughout Florida and Louisiana was low before and after the
hurricane, and the potential for SLE virus transmission in the area was low.
SLE virus surveillance was initiated because of the increased exposure of the
displaced residents and recovery workers to mosquito bites.
In 1989, emergency arbovirus surveillance following Hurricane Hugo was
based on virus isolation in cell culture, and all mosquitoes collected were
identified and tested. Turnaround time was 2-3 weeks using this protocol.
Following Hurricane Andrew, surveillance programs in both disaster areas
tested only known vector species using antigen capture ELISA techniques; with
this protocol, results were available in 3-6 days. This substantial
improvement in turnaround time should enable timely detection and response to
a mosquitoborne disease in emergency situations (CDC, unpublished data, 1990).
Health InfoCom Network News Page 9
Volume 6, Number 9 April 10, 1993
References
1. CDC. Rapid health needs assessment following Hurricane Andrew --Florida
and Louisiana, 1992. MMWR 1992;41:685-8.
2. CDC. Centers for Disease Control emergency response plan, 1990. Atlanta:
US Department of Health and Human Services, Public Health Service, 1990.
* Federal Emergency Response Plan (Public Law 93-288, as amended April 1992).
Health InfoCom Network News Page 10
Volume 6, Number 9 April 10, 1993
Injuries and Illnesses Related to Hurricane Andrew --
Louisiana, 1992
=====================================================
SOURCE: MMWR 42(13) DATE: Apr 09, 1993
On August 26, 1992, Hurricane Andrew struck Louisiana. On August 24, in
anticipation of hurricane-related injuries and illnesses, the Office of Public
Health (OPH), Louisiana Department of Health and Hospitals, in cooperation
with hospital emergency room (ER) and public utility personnel and coroners,
established an active emergency surveillance system in 19 parishes to monitor
these events. This report summarizes the findings from this emergency
surveillance system.
A hurricane-related fatal or nonfatal injury/illness was defined as one
that occurred from 12 noon August 24 through 12 midnight September 21 that
resulted from the preparation for, impact of, or clean-up after the hurricane
and required treatment in a hospital ER or caused death. The OPH developed a
questionnaire to collect data on demographic variables (i.e., age, sex,
marital status, and parish); nature of injury/ illness (i.e., cut, fall,
electrocution, or rash); body part affected; location, etiology, and time of
injury/illness; and reporting institution. To facilitate reporting of these
hurricane-related events, the OPH made periodic telephone calls to ER
personnel and coroners who had administered the questionnaire to or for
persons with injuries/ illnesses that met the case definition.
Twenty-one (50%) of 42 hospital ERs, five (26%) of 19 coroners' offices,
and one of two public utilities participated in the emergency surveillance
system and reported a total of 462 hurricane-related events. Of 406 events
with a reported date of occurrence, 15 (4%) occurred before landfall; 70
(17%), during the hurricane; and 321 (79%), after the hurricane (Figure 1,
page 249). Of 310 events with a reported place of occurrence, 244 (79%)
occurred outside, and most (237 69% of 343) occurred in or around the home.
Of the 462 hurricane-related events, 445 (96%) had nonfatal outcomes
(Table 1, page 249). Of the 17 (4%) fatal outcomes, eight occurred before the
hurricane made landfall: six were due to drowning; one, to an impact injury
sustained in a motor-vehicle crash during the evacuation; and one, to a crush
injury sustained during a tornado that preceded the hurricane (Figure 1). Of
the 445 nonfatal events, 383 (86%) were injuries, and 62 (14%) were illnesses;
319 (72%) occurred among males. The most common nonfatal injury was a
cut/laceration/puncture wound (184 41% of 445), followed by a strain/sprain
(49 11%) (Table 1). The most common body parts reported affected by a
nonfatal hurricane-related injury/illness were the upper extremities,
including the fingers, hands, and arms (157 38% of 411), followed by the
lower extremities, including the toes, feet, and legs, (89 22%).
Three parishes -- St. Mary's, St. John's, and Iberia -- had hurricane-
related injury/illness rates higher than 200 per 100,000 population (Figure 2,
page 250); two parishes -- Iberville and Assumption -- had rates of 50-200 per
Health InfoCom Network News Page 11
Volume 6, Number 9 April 10, 1993
100,000 population. All other affected parishes had rates less than 50 per
100,000 population.
Reported by: K Kelso, S Wilson, L McFarland, DrPH, State Epidemiologist,
Office of Public Health, Louisiana Dept of Health and Hospitals. Disaster
Assessment and Epidemiology Section, Health Studies Br, Div of Environmental
Hazards and Health Effects, National Center for Environmental Health; Div of
Field Epidemiology, Epidemiology Program Office, CDC.
Editorial Note: Emergency surveillance systems can facilitate public health
decision- making during natural disasters and have an impact on policies for
future disasters. For example, in this report, after Hurricane Andrew made
landfall, Louisiana public health officials monitored for outbreaks of
diarrheal illness to identify and repair damaged waste-disposal systems and
determine allocation of potable water. In addition, previous surveillance
during other hurricanes affected the public health response to Hurricane
Andrew. Louisiana public health officials were aware that hurricanes trigger
secondary effects (such as tornadoes and flash floods) that, together with
storm surges, can cause fatalities (e.g., drownings), even before making
landfall, and that most injuries/illnesses related to hurricanes occur during
the postimpact (i.e., clean-up) phase (1,2). Using this information, officials
alerted Louisiana residents through radio announcements before and after
Hurricane Andrew made landfall to the dangers that would be present during the
preimpact, impact, and postimpact phases (e.g., drownings, crush injuries, and
electrocutions, respectively).
Information on natural disaster-related morbidity and mortality is
available from many sources, including medical examiners' and coroners'
reports, death certificates, the American Red Cross, meteorologic services,
police and fire departments, and emergency medical services (3-5). However,
these sources use different methods and criteria for case selection (e.g.,
each uses a different definition of disaster-related injury), and no one
source collects complete information on deaths and injuries. Similarly, no
universally accepted definition exists of a disaster-related death. For
example, following Hurricane Hugo in 1989, two coroners in South Carolina
reported "heart attacks" that occurred during the hurricane as caused by
hurricane-induced stress, but coroners and medical examiners in other regions
of the state did not consider any heart attacks hurricane-related, regardless
of when they occurred, and did not report them as such (6). The lack of
standardized definitions for disaster-related death and injury presents
difficulties in enumerating related deaths and injuries following a natural
disaster. Furthermore, comparison of death and injury data from different
sources is problematic.
This report demonstrates the feasibility of collecting emergency
surveillance data that can be used to prevent injury and death related to a
natural disaster. Better epidemiologic knowledge of the types of injury and
Health InfoCom Network News Page 12
Volume 6, Number 9 April 10, 1993
illness and causes of death related to hurricanes is essential for the
planning and provision of public health responses (e.g., distribution of
relief supplies, equipment, and personnel) during such disasters (7). To
assist efficient data collection and to facilitate decisions made by emergency
personnel following disasters, CDC, in collaboration with state health
departments, has developed disaster-related injury/illness surveillance
questionnaires that can be quickly modified for specific situations. In
addition, to enable comparisons of disaster-related injury/illness data from
different sources, CDC is standardizing surveillance variables and methods of
data collection. Development of robust methods for collecting and analyzing
these questionnaires should assist public health professionals in guiding
their emergency responses during future disasters.
References
1. Seaman J. Epidemiology of natural disasters. Contributions of Epidemiology
and Biostatistics 1984;5:1-177.
2. CDC. Update: work related electrocutions associated with Hurricane Hugo --
Puerto Rico. MMWR 1989;38:718-20,725.
3. CDC. Hurricanes and hospital emergency-room visits -- Mississippi, Rhode
Island, Connecticut (Hurricanes Elena and Gloria). MMWR 1986;34:765-70.
4. CDC. Preliminary report: medical examiner reports of deaths associated
with Hurricane Andrew -- Florida, August 1992. MMWR 1992;41:641-4.
5. Patrick P, Brenner SA, Noji EK, Lee J. The American Red Cross-Centers for
Disease Control natural disaster morbidity and mortality surveillance system
Letter. Am J Public Health 1992;82:1690.
6. Philen R, Combs DL, Miller L, et al. Hurricane Hugo, 1989. Disasters
1992;15:177-9.
7. Noji EK. Disaster epidemiology: challenges for public health action. J
Public Health Policy 1992:13:332-40.
--------- end of part 1 ------------
---
HICN609 Medical News Part 1/4