|Assessing Needs in the Health Sector after Floods and Hurricanes (Pan American Health Organization (PAHO) / Organización Panamericana de la Salud (OPS), 1987, 90 p.)|
Disasters may have an indirect negative influence on health because of the way they alter the environment. The frequency of diseases will not always increase, but flood disasters do present the potential for outbreaks of communicable diseases. According to Western (1982), the risk factors involved in communicable disease outbreaks after disasters are:
1. Alterations in pre-existing morbidity: It is highly unlikely that the disaster will spark outbreaks of diseases that are not endemic to the region, unless such diseases are introduced by emergency relief personnel. The problem lies in the fact that reliable figures on disease frequencies in a given region are generally nonexistent.
2. Ecological changes resulting from the disaster: Droughts, floods, and hurricanes, in particular, will have a noticeable effect on the region's ecology by exerting an impact on the habitats of vectors and on water sources, thereby increasing the risk of vector-transmitted diseases.
3. Population displacements: If affected populations move far, they may come upon diseases against which they are not immunized or may carry a disease or its vector to an area previously free of it. Furthermore, if they move to populated areas, they may overburden the services and facilities of the community sheltering them.
4. Changes in population density: Overcrowding, especially in evacuation camps or centers, will contribute to the possibility of disease outbreak created either by direct or airborne contact.
5. Disorganized public services.
6. Interruption of basic public health services: Water treatment plants may be destroyed or severely damaged and programs, such as malaria control and follow-up of tuberculosis cases, may be interrupted.
Although they tend to produce few injuries among survivors, heavy floods, and particularly slow ones over a long period of time, may easily bring about a combination of all of the risk factors cited by Western.
Changes in the frequency of certain diseases must be detected so that the necessary steps may be taken to forestall or lessen the impact of epidemic outbreaks. The assessment of the morbidity and mortality situation should not become a lengthy research project, but should focus on eight key questions, the answers to which provide the basis for a sound public health response:
1. How has the flooding affected disease frequency in general?
2. What risk factors have appeared as a result of the floods (even though no increase in disease has yet been reported)?
3. What measures are needed to control these risks?
4. What specific diseases have been reported on the increase?
5. What should be done to cope with that increase?
6. What pharmaceuticals, supplies, equipment, and manpower are needed? What is available locally?
7. What are the projected cost, timing, and duration of need?
8. How will the imported items identified in 6 be transported and distributed?
The assessment teams should consult the epidemiologist, area health director, and other local health workers to determine what diseases are common in the area, particularly the communicable diseases whose frequency may be affected by the risk factors cited above.
If there is no information available on endemicity or if the data available are questionable, certain reasonable hypotheses must be formulated. The team should draw up a tentative list of diseases to be investigated based on experience with past floods:
· Acute gastrointestinal diseases ("dehydration" is frequently recorded as a separate entity, especially when a program of oral rehydration exists)
· Acute respiratory diseases
· Tuberculosis (recorded as a separate entity from "respiratory diseases"; a special register usually exists for such cases)
· Yellow fever
· Equine encephalitis
· "Childhood" diseases, usually preventable by immunization
- Whooping cough
(Tetanus, which is included in immunization programs, is not a disease whose frequency increases during floods)
· Typhoid fever
· Skin lesions
· Certain conditions typical of the region or country (e.g., Bolivian hemorrhagic fever)
Diseases that have never been reported in the region, or that have been completely eradicated, should be eliminated from the list.
Ideally, any changes in morbidity and mortality found by the team should be given as incidence rates, expressed as the number of new disease cases per 1,000 or 10,000 persons at risk. In most instances, however, these rates cannot be computed reliably because baseline data are lacking and population figures that should serve as the denominator are outdated, unreliable, or unknown for the geographic area surveyed.
It is equally difficult to estimate the number of people covered by the health center or post, except in small, isolated communities. Moreover, when two or more health units are closely clustered, it is not uncommon for the same patient to attend several in search of free care and drugs.
We will therefore refer to "change in frequency of diseases," a loose indication of a trend that can be obtained from the proportion of observed or attended symptoms or symptom complexes among the total observed or attended population (see Chapter 10, Surveillance Systems). The same rough measurement applies in the rapid assessment of changes in mortality.
Changes in Frequency
The effect of time must be kept in mind. If the evaluation is made during the early stages of the flood, it is unlikely that any significant changes will be noted in disease frequency. The flood's influence will vary according to the disease and how long the flood lasts. Increases in malaria, for example, are not detected until several months after the waters have begun to recede, leaving an ideal habitat for vector proliferation.
When a highly organized system of epidemiologic surveillance exists in the affected area, changes in frequency or even incidence rates may be easily observed. Comparisons can be made with patterns of previous years to verify that changes detected are not seasonal. In addition, any disease not initially considered that has significant increases can be added to the list.
Unfortunately, a responsive system of epidemiologic surveillance will be rare. Whatever exists will probably be inadequate, and very often with an information lag of several months or years, depending on how far away one is from the source of such data.
Sources of Information
The team will often have to seek needed information at the very site of the disaster, probably before any statistical process has been initiated.
A. Health centers and posts:
Whether the health facilities are manned by doctors, nurses, or auxiliaries, there is always a "daily report" or attendance log in which some data about the patients seen during the day (such as name, age, diagnosis, and treatment) are entered. In moat countries this daily report is the basis for a center's monthly report and includes the number of cases treated by diagnosis and often by age groups. These monthly reports then go to a higher administrative level, where they are processed together with those of other centers. From there they continue to another level, and eventually provide the basic data for the health ministry's reports on the country's health status. Usually a monthly report is also prepared on the trends of certain communicable diseases; it, too, is based on daily case records.
The monthly reports can be used to calculate the number of disease cases on the list to be investigated among the total number of patients bested and/or the number of inhabitants covered by the center, if known. It should be kept in mind that the catchment area figure may include those people covered by a smaller unit.
The morbidity rates, or the percentage of cases treated (when rates cannot be obtained) at the flood's onset, should be compared with the previous year's figures for the same period. If the health care conditions are very dissimilar (e.g., when a doctor is now available where there was none before), the team will have to compare the new figures with those for the months prior to the flood.
1) Department of statistics:
In moat cases, statistics departments are behind in preparation of their reports sad the data available are not current enough to be useful in an emergency. The assessment team will probably have to get information from the outpatient and emergency departments and from inpatient wards.
2) Outpatient and emergency departments:
The procedure will be the same as for a health center.
3) Hospital wards:
Depending on the hospital's level of sophistication, there may be an infectious diseases ward, a "diarrhea" and "rehydration" ward, a malnutrition ward, and so forth, from which one can secure information. Generally, there will be at least a medical and a pediatric ward where data can be collected (the gynecology-obstetrics and surgery wards' information would probably not be pertinent).
In most hospitals, each ward keeps its own statistical record apart from that required by the department of statistics. There will probably be an "admissions book" or a "patients' register," in which each person admitted to the ward is recorded, usually by name, age, and initial diagnosis. If such a book exists, it would be the fastest way to compare the number of cases of selected diseases with those for the preflood periods. When calculating the proportion of diseases, the team should realize that the total number of patients seen by the facility may have increased or decreased due to the flood.
The laboratory will have records of presumed diagnoses and test results.
5) Malaria eradication program:
The surveillance system for malaria is usually better than that for other communicable diseases. Nonetheless, there may be a delay of a year or more before the figures from outlying areas reach the central level (e.g., provinces, departments) and are processed. Once again, data should come from as close to the community in question as possible. At the level of health area/hospital area, province, or department, there may well be a malaria eradication service that collects field data and processes them to a higher level. Even so, considerable time may elapse from the moment the information originates at the most primary health level to that moment when the service collects it. The health centers, however, usually have a separate malaria report that registers presumptive or suspected cases for which blood samples have been taken. The center usually keeps the stub of the entry or form that accompanies the sample, on which the patient's basic data are noted.
6) If there are no records available:
It is possible that no information system will exist in the flooded area or that, as a result of the flood, all records are destroyed. As a last resort, the team can use the patients awaiting treatment at the health center or at the hospital's outpatient department as a source of information. They should be asked to identify their main symptoms, which allows for a tentative (although rough) diagnosis and also helps determine whether one of the diseases to be investigated is involved. The team can also get a general idea from the health center staff about the number of cases handled compared with the normal situation. Frequently the nursing auxiliary, the health promoter-in short, the primary health worker who has close day-to-day contact with the community-is quite sensitive to the changes taking place in his or her community.
Print and other communications media will often report outbreaks. Although these reports are often unfounded rumors, they should also be checked. If the reports are valid, decision makers can take appropriate action; if no outbreak exists, authorities may decide to publicly dispel the rumors.
C. Evacuation centers:
A detailed review of evacuation centers appears in Chapter 8. When there is a disease registration system and a fixed, nondynamic population, any changes in incidence can be identified during the life-span of the camp. Since a more realistic scenario would include a dynamic population, the best that can be obtained may be the proportion of attendances for specific symptom complexes or readily diagnosed diseases.
Possible Causes of Error
There are several factors that could effect an apparent change in the frequency of the diseases under investigation and lead to error:
A. Different diagnostic criteria, or a different type of health care source; for example, comparing the data reported by a doctor with those reported by a sanitarian.
B. An increase in health care coverage, caused, for instance, by an increase in the availability of health care or a greater number of persons seeking care.
C. A decrease in available health care delivery, such as from the isolation caused by the flood; this could make certain health care centers inaccessible to the population and result in a drop in the number of cases reported by those centers.
D. Improved capacity for diagnosis, for example, unusual access to laboratory facilities. It is not uncommon for a request for external emergency aid to result in the provision of laboratory equipment- sometimes relatively simple, like microscopes or slides-that would make possible, for instance, a greater number of blood smear tests for malaria parasites. This in turn would give a false impression that the incidence of malaria had risen.
E. A greater "alertness" toward certain diseases, particularly when the diagnosis is only clinical: the health worker is "seeking" given diseases. During floods a large number of cases of typhoid fever are often diagnosed clinically without laboratory confirmation. Most are not actually typhoid cases but, faced with that suspicion, the examiner feels more secure in diagnosing a disease to which the population (and, at times, health officials) is sensitive.
F. Seasonal variations; certain diseases endemic to the region regularly show a "peak" in frequency during given months. The flood could coincide with one of those "peaks" and, unless a comparison can be made with previous years, the increase in cases could be attributed erroneously to the flood.
G. In the case of malaria there are two special problems. The frequency of this disease has persistently risen over the last decade throughout most of Latin America. Several factors are responsible for this, above all the economic problems that have led to a reduction in national eradication programs. After a flood it is very difficult to determine whether the increase in frequency stems directly from the disaster or is part of the generally observed rise (research addressing this problem is currently under way in Ecuador). Moreover, variations in the availability of slides, transportation, techniques, and so forth, may bring about marked variations in the number of samples taken and processed.
Number of Centers
The number of centers to be investigated actually depends upon the number of officers making the assessment and the amount of time they have to do so.
If the area in question is very homogeneous, a "representative" center provides a quick idea of the situation, although the information could be biased as a result of numerous factors (e.g., accessibility, dedication of the personnel). In most cases, information would be more accurate if it came from three health centers in geographically distinct areas, although if differences in frequency were found, they might not be statistically significant. These tests would have to be carefully designed according to each circumstance (i.e., random selection, taking into account characteristics of the population, its size, etc.)-often an impractical luxury. Ideally, data should be from both the affected area and one that is not involved. In practice - particularly, when the geographic area is large-it is almost impossible to find a "control" center and a "case" center similar enough to provide reliable information.
Changes in causes of mortality over time will confirm the flood's effects on morbidity, its severity, and the urgency and extent of steps that should be taken or strengthened to safeguard the population's health.
Different kinds of disasters affect mortality in varying ways. In general, hurricanes and slow floods, such as those brought about by the "El Niño" phenomenon in South America in 1983, cause few direct deaths. However, some of their indirect results, for example, landslides, can be deadly.
The effects on mortality (if any) secondary to increased morbidity will take some time to be felt and will be more marked in the cases of the community's most vulnerable groups, particularly young children. Therefore, the assessment team should concentrate on children under 5 years of age.
Sources of Information
A. Birth and death registration office:
There is usually an office that registers births and deaths at the city or district level. The registration sheets (death certificates) often have a detachable stub that is kept in the local office when the original certificate is sent to a higher administrative level. The minimum information needed is age and cause of death. As with the morbidity figures, the team should compare the number of deaths by cause and age for similar time periods in different years. If this is not possible, the data should be compared with a time period before the flooding started.
The registration office is the first and most important place in which information should be sought; if for some reason it cannot be obtained there, the following places can be tried:
B. Health centers/sanitary posts:
In some countries the local registry offices send a monthly death report to the pertinent health center.
By referring to the "admissions book," "daily report," or "patients' log" in the various wards, one can secure information on the deaths in those wards and their causes.
D. If no records are available:
A general estimate of increased mortality might come from contacts with the local clergy, the sanitarian or nurse, and the doctor (if there is one). In some cases, especially in the more remote rural areas, there may be "clandestine cemeteries" where very young children are buried; their deaths are not registered because of the community's isolation, bureaucratic procedures, or expense.
Possible Causes of Error
A. Insufficient time elapsed for recording.
As was mentioned earlier, if the information is collected shortly after the flood's onset, changes in mortality may not yet be evident.
B. Delayed registration.
There may be a significant delay between the time of death and the date on which the family registers it.
C. Inhibited travel.
The disaster itself may hinder family members from reaching the place of registration.
Most Latin American countries and the Caribbean underregister their deaths-especially those of children. It is quite possible the disaster will augment the proportion of unreported cases in already isolated communities.
· How has the flood modified the frequency of illnesses?
· What specific illnesses have shown an increase in frequency?
· What is the best way to combat the increase?
· What pharmaceuticals, material, equipment, and manpower are needed?
· What is the estimated timing, duration, and cost?
· How will the people and supplies be transported and the material distributed?
· Illnesses common in the area
· Tentative list of illnesses to be investigated
· Changes in frequency: time period to investigate
- From the flood's onset, compared with the same period in the previous year
- Weeks (or months) immediately before and immediately after onset of the flooding
Sources of information
· Epidemiological surveillance system
· Health centers and dispensaries (outposts)
- Monthly attendance record (number of cases classified by
- Monthly record of communicable diseases
- Attendance book
- Department of statistics
- Outpatient department (casualty department)* monthly attendance record (number of cases classified by diagnosis)
* monthly record of communicable diseases
* attendance book
- Hospital wards* registry of patients ("admissions" book)
* nurses' report book
* wards to investigate: medicine, pediatrics, malnutrition, gastroenteritis, rehydration
- Malaria Eradication Program
- Surveillance/report system
- Record of samples (health centers)
· Evacuation centers
· If no records
- Survey in clinic/hospital waiting room
- Survey in hospital wards
- Health staff views
· Sources of error
- Different diagnostic criteria
- Increase in the available services
- Decrease in attendance
- Better diagnostic facilities
- Increased "sensitivity" among health staff
- Seasonal variations
- Malaria* global trend of increase in incidence
* lack of diagnostic equipment, personnel, and transportation
Sources of information
· Registration office
- Copy or stub of death certificate
- Registration book
· Health centers/outposts
- Monthly death record
- Admissions book
· If no records
- Church books
- Interview with clergy
- Interview with doctor/health worker
- "Clandestine cemeteries"