Cover Image
close this bookDisaster Reports Number 5: Hurricane Gilbert in Jamaica, September, 1988 (Pan American Health Organization (PAHO) / OrganizaciĆ³n Panamericana de la Salud (OPS), 1988, 42 p.)
View the document(introduction...)
View the documentPrologue
View the documentBackground
View the documentThe hurricane and its effects
View the documentThe surveillance system
View the documentRelief shelters
View the documentPreparedness and response
View the documentLessons learned
View the documentAppendix 1
View the documentAppendix 2
View the documentReferences

The surveillance system

Background Information

The Office of Disaster Preparedness and Emergency Relief Coordination has the overall responsibility of co-ordinating the relief effort and response. They are assisted in this task by the National Disaster Committee comprised of representatives from the security forces, government ministries, voluntary agencies and the ODP. The Prime Minister is the Chairman of the Committee. It was convened for the first time since the issuing of the Hurricane Alert in June 1989.

Within the Ministry of Health there is a Health Action Committee for Disaster Preparedness involving Medical Officers of Health in charge of parishes and hospitals. Hospital management teams were trained in disaster preparedness activities which included developing and writing hospital plans. In addition, disaster plans were developed for parish health teams and, in a few parishes, the process extended to the district/community level [Bullock DuCasse, 1989].

These disaster plans were never implemented. "The impact of the hurricane .... overwhelmed many health workers as they sought to cope with damage to and loss of Ministry of Health facilities, equipment and records as well as personal shelter and belongings" [Bullock- Du Casse, 1989]. Health workers had suffered heavy losses as a result of the storm and were expected to assist in the task of restoring order in the public sector. Yet they enjoyed no advantage in terms of personal relief [PAHO/WHO, 1989]. A Command Post was established at the Ministry of Health with the announcement of the Hurricane Watch on September 10 and efforts made to contact Medical Officers of Health to activate disaster plans. Because of inadequate communication links, these efforts were not all successful. The Ministry issued a press release on Sunday, September 11 urging all facilities to finalize patient discharge. The Command Post was manned until 11.30 a.m. on September 12 when health workers left to look after their homes and belongings.

Because of the lack of road-worthy vehicles and damage to radio equipment, there was no contact between the Ministry of Health and the parishes until September 14. Neither was the Ministry able to assist in damage assessment or the distribution of relief supplies. In fact, the emergency found the island's health service so short of medical supplies that the Salvation Army made supplies available to medical personnel. While in some cases the shortage was absolute, in others it stemmed from the poor distribution system [PAHO/WHO, 1989]. The few health centers that were not damaged were besieged as relief centers, so priority was given to the restoration of services in the hospitals to enable them to offer emergency care.

Emergency Surveillance for Infectious Diseases

One week after the impact, the Caribbean Food and Nutrition Institute was requested by the Ministry of Health to co-ordinate teams for a health and nutrition surveillance of the island. The instructions were to visit parish Medical Officers of Health and obtain information on health and nutrition and to make a rapid assessment of public health concerns. Under normal conditions epidemiological data from sentinel stations, which may be either hospitals or health centers, are supplied to the Ministry of Health once per month. This system was expanded to allow for the inclusion of disaster-related data as well as data from three hospitals that are not normally sentinel stations. Forms A (disease surveillance) and B (public health) were distributed during the visits. (See Appendix 1 and 2).

Four teams visited l2 of the parishes during the week of September 19, while ten parishes were visited by three teams in the week of September 26. Each parish was visited at least once. Nine were visited twice. The teams were composed of physicians nurses and other members of the staff of the Epidemiological Unit of the Ministry of Health. A report was made to the Ministry at the end of the first week and a surveillance report was subsequently published [Patterson, 1988].

This system operated for ten days, from September 19 through 29. Data were collected from what was, in effect, a non-random sample of sentinel stations and relief centers and this must be borne in mind when considering the number of reported cases. The surveillance report that was produced [Patterson, 1988] recorded the occurrence of specified diseases in "selected sentinel stations, three other hospitals and a few reporting shelters." A breakdown of the numbers of cases by the type of facility was not given. Neither were the number of sentinel stations, shelters and their locations known. In any case, as Patterson recorded, sentinel stations established in normal times are not appropriate in times of disaster. Key hospitals such the University Hospital of the West Indies are not designated sentinel stations although the U.H.W.I. was among the locations treating a large number of injured and ill persons.

In addition, as mentioned earlier, the vast majority of shelters used during the emergency period comprised what was termed "unofficial shelters" without identifiable managers. Few parishes were able to give an accurate number of the shelters that had been established. Moreover, by the week beginning September 19 when the surveillance teams were mobilized, many of the centers had been closed. In two parishes all the centers had been closed by September 26. To these one must add the fact that many rural areas were inaccessible and a large number of persons isolated in these areas would have been unable to seek medical attention outside of the local areas [Patterson, 1988].

Data were collected during the two weeks beginning September 19 and September 26. Different centers, and in some cases, parishes, reported during those two weeks on occurrences commencing September 13. Only the findings of the second week were available for inclusion in the surveillance report. Original data had been given to the Ministry of Health because of an expressed urgency and had been misplaced and unavailable for inclusion. Discrepancies between the surveillance report and the figures subsequently produced by the Ministry of Health [Bullock-DuCasse, 1989] create problems of interpretation. Although the Ministry's figures for the second period are identical to those given in the surveillance report, the two weekly reports are represented as a "trend" over the two-week period [Bullock- DuCasse, 1989]. Patterson stressed that even if the report of the first week should become available, the fact that the figures came from different centers means that the data are not comparable. There are other reasons why the suggestion of a trend should be avoided. The figures presented for the earlier period relate to 12 parishes-those for the second week, ten parishes. The number of days covered by the two reports also differed.

As the surveillance report concluded, "Accurate surveillance data were very difficult, if not impossible, to obtain during the emergency period ...." Unfortunately, the results of the Ministry of Health's hospital survey which, under the circumstances, could have been most useful, and was reported as being "underway" in April 1989, were not available in August.

No significant increase in the rate of communicable diseases was noted during the period [Bullock-DuCasse, 1989]. There were, however, isolated outbreaks of gastroenteritis and an increase in the number of cases of trauma immediately after the hurricane.

Table 1 provides the data collected during the two weeks during which the surveillance system operated.

Table 1. Number of cases reported during the weeks beginning September 19 and September 26.


Sept. 19

Sept. 26

Trauma

551

142

Fever

112

120

Gastroenteritis

60

14

Skin rash

65

17

Fever and cough

175

82

Wheezing, SOB

37

41

Because of the shortcomings of the surveillance system already noted, no meaningful analysis of these figures could be undertaken.

A retrospective analysis of attendance at three hospitals in the K.M.A. during the month of September 1988 was undertaken by the author. For the Bustamante Hospital for Children (BHC ) where disruption was minimal, both Casualty Department and Admissions records were studied. Analysis was restricted to the records of the Casualty Departments of the University Hospital of the West Indies (UHWI) and the Kingston Public Hospital (KPH). However, the PAHO/WHO study looked at hospital admissions in these two as well as in the Cornwall Regional, among Type A hospitals. Their sample also included Type B and C hospitals.

The Bustamante Hospital for Children

Casualty Department Gastroenteritis is, numerically, the single most important complaint seen at the BHC in normal times and this was true of the immediate post-disaster period also. As Fig. 4 shows, the number of cases of gastroenteritis treated at the hospital on September 19, that is, one week after the storm increased sharply. The number treated on that day was 2.5 times the average for the seven days which preceded Hurricane Gilbert. From September 14, when near normal casualty services were restored, through September 30, about 324 cases were seen. As a percentage of total attendance, gastroenteritis increased in the same period (Fig. 5). The PAHO/WHO report also noted increases in Casualty attendance at the Spanish Town Hospital among children and adults. Vigorous treatment, which included administering of intravenous fluids and holding patients for twenty-four hours observation, reduced the number admitted to the wards.


Figure 4. Casualty attendance—Bustamante Hospital for Children, September 5-30,


Figure 5. Casualty attendance---Busbmante Hospital for Children, % of all case seen (3-day period).

The number of cases of fever and skin ailments is shown in Fig. 6. The numbers are given for three-day periods except for the last two days of the month. Three cases of conjunctivitis were treated between September 21 and 28.


Figure 6. Casualty attendance—Bustamante Hospital for Children (3-day period).

Hospital Admissions Fig. 7 shows the number of admissions to the BHC over the 26-day period. The only unusual feature was the sharp drop in admissions on September 12 and 13 when only emergency cases were seen. The discharge rate for gastroenteritis for the month of September was 14.6 percent. This compares with a rate of 15.1 percent for 1986, the most recent figures available [Social and Economic Survey, 1988]. The 56 cases of gastroenteritis admitted during the month compares with a six-month average of 57.6 for the K.M.A. in 1983 [Bailey, 1988].

Casualty Departments, UHWI and KPH

Analysis of the records of the Casualty Department of the UHMI is hampered by the fact that all or a part of the records of four days in September were missing (Fig. 8). However, in both institutions there was a rise in the number of injuries treated immediately after the hurricane.


Figure 7. Admissions to Bustamante Hospital for Children (3-day period).


Figure 8. Casualty attendance-Univesity Hospital of the West Indies (3-day period).

Fig. 9 shows the cases of asthma and trauma as a percentage of all patients treated at the KPH. In addition, at KPH, the number of gun-shot wounds (10) seen on September 12, the day of Hurricane Gilbert, was five times the average treated in the seven days preceding the hurricane. These were mainly wounds inflicted by the security forces on looters.


Figure 9. Casualty attendance—Kingston Public Hospital (3-day period).

There were no significant increases in communicable disease patterns in the post-disaster period. The PAHO/WHO report [1989] sees this as the result of the rapid institution of preventive measures which included the prompt attention to water quality monitoring and waste management, emergency measures for vector control, the focused activity in typhoid endemic areas, the rapid clearance of shelters, intensification of health education activities and the provision of funds to meet emergency needs at parish levels.

Surveillance of Environmental Health and Water Quality

With the assistance of PAHO, a surveillance system was put in place to monitor water quality and the water distribution system as well as sewage treatment plants and environmental conditions at hospitals and shelters. Teams composed of two members each visited water supply and treatment plants where samples for bacteriological analysis were taken and free residual chlorine determinations made. These activities continued until the end of October.

There was also surveillance for vectors, especially the Aedes aegypti mosquito. Routine surveillance continued especially in known typhoid endemic areas.

Except for those systems that were fed by gravity, all water systems were affected by the hurricane. Priority in restoration was given to those systems in which hospitals were located. Within two days, 30 percent of the service in the metropolitan area had been re-established an intermittent basis [PAHO/WHO, 1989]. Within ten days of the hurricane, about 50 percent of the country's normal water supply had been restored. There were delays in rural areas largely because of the shortage of transportation and problems of access to plants in rugged localities [Barrett, 1989b]. Assistance with transportation was given by the Jamaica Defence Force and the USAID. In addition, PAHO and Medecins sans Frontieres provided 22 portable water tanks which helped to free up the hard pressed water trucks. The tanks were placed at central locations in rural communities and were filled by the water trucks. Canada donated two portable treatment plants which were located near Maggoty in St. Elizabeth and in Duckenfield in St. Thomas [Barrett, 1989]. These two areas have, in the past, had a high incidence of typhoid [PAHO/ECD].

Initial results indicated that water was unsafe for human consumption and monitoring began on September 14, with assistance from PAHO [PAHO/WHO,1989]. Tests were performed by four teams, first, in Kingston, and later in the rural areas. In the first round of sampling (September 26-30) a very high percentage of samples were found to have high bacteriological content and a low concentration of chlorine (Table 2). The Nwc was advised to increase and ensure continuous chlorination, to initiate sanitary surveys at points of high coliforms concentration and to recommend that the public boil water and use chlorine products.

Purification tablets were made available to members of the public by the Ministry of Health and the Jamaica Red Cross since the boiling of water was impossible in many cases. Considerable effort had to be exerted to counter alternative and ineffective suggestions in the news media for purifying water.

In the second round of sampling, eight of the fifteen hospitals sampled had unsafe bacteriological water and nine low chlorine content [PAHO/ECD, Oct. 3- 6).This represented an improvement in the bacteriological quality of the water (Table 2). However, at three hospitals the quality had deteriorated while at two others the quality, though improved, was still unsafe. Purification tablets, chlorine comparators and manual filter pumps were given to the ECD for distribution to the most needy hospitals. The third round of sampling found improved conditions and tests in November and December indicated that the water quality was back to normal [PAHO/WHO, 1989].

Table 2. Variation of Bacteriological Quality (bq) (% Pos.) and Low Chlorine Concentration (Ic) (% Low).

PARISH

SEPT.26-30

OCT.3-6

OC1.11-21


(bq)

(Ic)

(bq)

(Ic)

(bq)

(Ic)

Clarendon

-

-

-

-

50

38

Hanover

100

0

57

83

57

67

Kingston/ St. Andrew

50

50

40

48

27

16

Manchester

50

50

33

33

14

0

Portland

88

100

53

56

9

36

St. Ann

69

69

100

100

25

14

St. Catherine

30

70

50

67

43

29

St. Elizabeth

90

40

49

67

0

8

St. James

70

50

14

71

88

88

St. Mary

67

78

60

68

67

33

St. Thomas

82

45

50

50

40

50

Trelawny

90

80

29

67

67

67

Westmoreland

56

78

60

60

60

60

Source: PAHO/ECD, Oct. 11-21.

With regards to excrete disposal and sanitation, many latrines were destroyed and sewage treatment plants made inoperative by a lack of electricity. There was, therefore, an increase of pollutants reaching water bodies.

Of the 42 treatment plants visited by PAHo/EcD, only fifteen were fully and two partially operating. Fourteen samples were analyzed for Biochemical Oxygen Demand (BOD) and Suspended Solids (Ss). None of the discharges met the recommended discharge limits for BOD and only two met the discharge limits for SS. It was recommended that monitoring should continue until the end of 1988, expanded to cover all parishes and monthly reports on the characteristics of the effluent made. Maximum discharge limits were set at:

BOD

20 mg/1

SS

30 mg/1

Total Coliform

200/100 ml

Free Chlorine Residual

0.5 mg/1

Source: PAHO/ECD.

The surveillance report found that few areas had addressed the problem of garbage disposal and that the numbers of vectors, especially flies and mosquitoes, were increasing [Patterson, 1988]. Some parishes had started oiling and fogging.

PAHO/WHO [1989] recommended that an environmental health activities plan should be elaborated for public health inspectors based on water systems, sewage plants and dump sites. Survey and sampling forms should be created and the importance of reporting on these forms emphasized. They also recommended that equipment for monitoring, such as pH meters and chlorine comparators be acquired as soon as possible. They saw a need for the approval and gazetting of regulations for drinking water, soil and air pollution. Finally, they emphasized the need for co-operation and co-ordination of activities between ECD and the PHC Unit and for a disaster plan involving public health inspectors.

Routine Surveillance

Routine surveillance based on operative sentinel stations and reports from physicians on the occurrence of a number of infectious diseases continued. Endemic areas for typhoid and leptospirosis were closely monitored and a close check was also kept on the incidence of malnutrition in children.

Gastroenteritis: The reported incidence of gastroenteritis increased sharply in October. In fact, the figure for October was almost 26 percent that for the same month in 1987. The Surveillance Report [Patterson, 1988] maintains that there were almost as many cases among adults as among children. PAHO/WHO [1989], in a study of selected conditions in seven parishes, also reported an increase in admissions to hospitals in the period October 1988 to March 1989 (Table 3) over the period October 1987 to March 1988, despite a decline in annual totals between 1987 and 1988. This increase took place in all the institutions studied except the University and Noel Holmes hospitals where there has been a consistent decline, and in the Spanish Town Hospital for reasons previously mentioned [PAHO/WHO, 1989]. The parishes covered by the survey were Kingston/St. Andrew, St. Thomas, St. Mary, St. James, Hanover, Manchester and St. Catherine.

Table 3. Admissions from a sample of hospitals, 1987-1988.

TOTAL

TOTAL

%

OCT '.87

OCT ' 88

%

1987

1988

CHANCE

MARCH 88

MARCH 89

CHANCE

819

736

-10

402

442

10

Source: PAHO/WHO, 1989.

In view of the declining trend in admissions for gastroenteritis prior to Hurricane Gilbert, and the damage to hospitals which would have discouraged admissions, the increase in admissions is significant [PAHO/WHO,1989].

Febrile Illness: Figures for febrile illness for 1987 are incomplete and this makes comparisons difficult. However, the figure for October 1988 was, roughly, 5 percent below that for 1987. The figures for early 1989 were also below those for the corresponding months of 1988.

One suspects that the decrease in the post- Gilbert period is not so much a reflection of an overall improvement as the dislocation in the health system and the inadequacy of the reporting system based on sentinel stations in situations such as these. The case of typhoid provides an excellent example of the problem.

Typhoid: Seventeen cases of typhoid were reported between April and December 1988, nine in the period following the storm. Up to September 1989, the Epidemiological Unit of the Ministry of Health had no case on record for 1989. Yet, there was an outbreak in the parish of Westmoreland. It occurred in areas where the disease is endemic [Silva, 1989]. However, the first case, on November 23 was diagnosed as a viral illness [Dalina, 1989]. It was not until January that the disease was identified and by the end of March eleven of the suspected thirty-six cases had been confirmed by laboratory. The outbreak occurred in three subdivisions of the Savanna-la-Mar Health District: Savanna-la-Mar, Smithfield, on the outskirts of the town, and Petersfield, to the northeast. It has been associated with the destruction of pit latrines by the hurricane and open defecation which followed.

In so far as hospitalization was concerned, the number admitted has always been small and this makes the interpretation of the data difficult [PAHO/WHO, 1989]. Thirty cases were admitted between October 1987 and March 1988 in the hospitals visited. Only 12, however, were admitted between October 1988 and March 1989 [PAHO/WHO, 1989]. The emphasis on the rapid improvement in water quality, especially in endemic areas, and public education undoubtedly kept the situation under control.

Immunization: There was a small increase in the number of BCG immunizations done in the post-Gilbert period in most parishes. The increase in this activity in St. Mary, as shown in Table 4, was marked and involved health centre staff and volunteers including school principals who had previously received training in identifying and motivating families who could benefit from this programme [PAHO/WHO, 1989].

Table 4. BCG Immunization by Parish
Oct. 1987 March 1988 and Oct. 1988-March 1989.

PARISH

BCG IMMUNIZATIONS

% CHANGE


(1)

(2)

+/-

Kingston/St. Andrew

5,694

7,311

28

St. Thomas

1,052

924

-12

St. Mary

1,404

2,770

97

St. James

2,420

2,332

- 4

Hanover

682

773

13

Manchester

1,733

1,923

11

St. Catherine

4,115

4,359

6

TOTAL

17,000

20,392

19


(1) Oct. 1987 - March 1988



(2) Oct. 1988 - March 1989


Source:PAHO/WHO, 1989

There was a decline in all other immunization programmes ranging from 10 percent in the case of measles to 21 percent in the case of polio. Here, too, the achievement of St. Mary was outstanding, it being the only parish to show an increase in the number of immunizations done in the post-Gilbert period [Table 5].

Table 5. Changes in the level of Immunization against OPT, polio and measles by selected Parishes Oct. 1987-March 1988 and Oct. 1988 - March 1989.

PARISH

IMMUNIZATIONS ( %+/-)


DPT

POLIO

MEASLES

Kingston/St. Andrew

-19

-26

-15

St. Thomas

-20

-33

-22

St. Mary

51

52

66

St. James

- 40

- 41

-12

Hanover

-34

-35

-23

Manchester

-17

-19

-19

St. Catherine

-15

-22

-13

Source: PAHO/WHO, 1989

The PAHO/WHO report warned that, in view of the decrease in the number of post-disaster immunizations and the fact that the percentage of the 0-11 target population immunized against these diseases by the end of 1988 was well below expected levels (Table 6), a special effort should be targeted on Kingston and St. Andrew, St. Catherine and Hanover. A measles epidemic began in December 1989 and by January 31, 1990 more than 5,000 children had been infected and eight had died. [Epidemiological Unit]. The parishes that were most affected were Kingston and St. Andrew where the epidemic began and St. Catherine. Few cases have been reported from St. Mary.

Table 6. Percentage of 0-11 year-old age group immunized against DPT, polio and measles, 1987 and 1988.

PARISH

DPT

POLIO

MEASLES


1987

1988

1987

1988

1987

1988

Kingston & St. Andrew

85.2

61.2

84.8

62.8

23.3

36.6

St. Thomas

70.0

88.8

73.3

90.8

49.3

68.6

St. Mary

88.6

96.8

90.3

95.7

62.7

63.7

St. James

93.8

104.6

92.8

104.3

59.4

76.3

Hanover

92.2

78.6

90.4

79.8

66.7

65.3

Manchester

77.4

101.2

78.2

91.1

60.3

69.0

St. Catherine

73.4

74.6

74.8

70.7

38.6

53.0

Thus, there was no significant increase in communicable diseases in the immediate post-disaster period. This was due to the rapidity with which the surveillance of the environment was initiated, the attention given to the management of waste and vector control as well as an intensification of health education activities [PAHO/WHO, 1989]. Parishes were provided with the means to meet their emergency needs. An important element also was the fact that the population in relief shelters very quickly returned to their homes.

Malnutrition: Fig. 10, based on figures obtained from the Nutrition Division of the Ministry of Health, shows the trend in moderate and severe malnutrition (Gomez Grades 11 and 111) ¹ It is clear from the figure that the upswing occurred well before the hurricane. The survey of seven parishes [PAHO/WHO, 1989] revealed a total of 5,900 reported cases in 1987 and 7,809 in 1988, an increase of 32 percent (Table 7). These figures refer to mild and moderate malnutrition (Gomez Grades I and 11). The increase in the period October 1988 to March 1989 over October 1987 to 1988 was marked in most of the parishes surveyed, but was particularly so in the parishes of St. Mary (134 percent! and Kingston/St. Andrew (72 percent). The average for the seven parishes was a 51 percent increase for the 1988 -1989 period. This could have resulted from increased vigilance in the wake of Gilbert or from improved record keeping [PAHO/WHO]. However, the level of the increase demands attention, especially since clinic statistics tend to underestimate parish wide prevalence [Rainford, 1987].

1. The Gomez Classification is used to evaluate the nutritional status of children. It compares a child's actual weight, at a specific point of development, with the weight he/she should have, or normal weight, according to his/her chronological age. This determination is based on standardized international values. According to this classification, the severity of malnutrition is classified by grades:

Grade I

- 75-85% of normal weight

Grade II

- 60-75% of normal weight

Grade III

less than 60% of normal weight


Figure 10. Jamaica -children severely malnourished (Gomez Grades II & III).

There was no corresponding increase in hospital admissions, however. The reduction in available beds and the necessity for greater selectivity may could have contributed to this decline.

Table 7. Reported cases of malnutrition in selected Parishes: 1987 and 1988; Oct. 1987 March 1988; Oct. 1988-March 1989.

PARISH

TOTAL

TOTAL

% CHANGE

OCT.87

OCT.88

% CHANGE


1987

1988

+/-

MAR.88

MAR.89

+/-

Kingston/ St. Andrew

2,009

2,290

14

799

1,376

72

St. Thomas

556

621

12

276

422

53

St. Mary

392

670

71

217

508

134

St. James

478

856

79

300

452

51

Hanover

305

633

108

233

279

20

Manchester

752

1,060

41

380

591

56

St. Catherine

1,408

1,679

19

780

886

14

TOTAL

5,900

7,809

32

2,985

4,514

51

Summary

There was an upward trend in the reported cases of gastroenteritis and malnutrition. The size of the increase in the cases of malnutrition was cause for concern and it was strongly recommended [PAHO/WHO, 1989] that further investigation be targeted on the parishes of St. Mary, Kingston/St. Andrew, Manchester, St. Thomas and St. James.

In addition, the decline in the number of immunizations done post Gilbert suggested that special effort be targeted on Kingston/St. Andrew, St. Catherine and Hanover.