ENVIRONMENTAL RESEARCH52, 34--46 (1990)
Absenteeism and Respiratory Disease among Children and Adults in Helsinki in Relation to Low-Level Air Pollution and Temperature ANTTI PONK,~
Helsinki City Health Department, Viipurinkatu 2, SF 00510 Helsinki, Finland Received September 28, 1989 The weekly changes in ambient sulfur dioxide, nitrogen dioxide, and temperature were compared with the figures for respiratory infection in children and adults and for absenteeism from day-care centers (DCC), schools, and workplaces during a 1-year period in Helsinki. The annual average level of sulfur dioxide was 21 i~g/m3 and of nitrogen dioxide 47 i~g/m3; the average temperature was + 3. I°C. The levels of these pollutants and the temperature were significantly correlated with the number of upper respiratory infections reported from health centers. Low temperature also correlated with increased frequency of acute tonsillitis, of lower respiratory tract infection among DCC children, and of absenteeism from day-care centers, schools and workplaces. Furthermore, a significant association w a s found between levels of sulfur dioxide and absenteeism. After statistical standardization for temperature, no other correlations were observed apart from that between high levels of sulfur dioxide and numbers of upper respiratory tract infections diagnosed at health centers (P = 0.04). When the concentrations of sulfur dioxide were above the mean, the frequency of the upper respiratory tract infections was 15% higher than that during the periods of low concentration. The relative importance of the effects of low-level air pollution and low temperature on health is difficult to assess. © 1990AcademicPress, Inc.
INTRODUCTION
Exposure to high levels of sulfur dioxide ( 8 0 2 ) , nitrogen dioxide (NO2), and particulate matter in the ambient air is known to cause acute and chronic adverse effects on health, mainly on the respiratory tract (World Health Organization, 1977, 1979, 1987; Colley and Brasser, 1980; Lan and Shy, 1981; Love et al., 1981; Mostardi et al., 1981; Schenker et al., 1983, 1986; Arossa et al., 1987; Charpin et al., 1988). The World Health Organization has issued guidelines for exposure limits consistent with the protection of public health (World Health Organization, 1977, 1979, 1987). These are based on data from reliable epidemiological studies and take into account sensitive groups such as children and persons with chronic respiratory disease. However, clear-cut limits are difficult to draw (World Health Organization, 1987). Some studies in 1980s suggest that SO2 and NOE cause respiratory tract infections in children at concentrations lower than the given global guidelines (Lan and Shy, 1981; Love et al., 1981; Mostardi et al., 1981; Paunio et al., 1987; Goren and Hellmann, 1988). These results are inconsistent with some other studies among children or persons with chronic respiratory disease (Schenker et al., 1983; Pershagen et al., 1984; Marty et al., 1985; Schenker et al., 1986; Vedal et al., 1987). In suggesting causal interpretations of associations, however, 34 0013-9351/90 $3.00 Copyright© 1990 by AcademicPress, Inc. All rightsof reproductionin any formreserved.
A B S E N T E E I S M A N D RESPIRATORY DISEASE
35
one should be cautious, because the agents measured may be merely indexes of other pollutants or the health effects may be due to the combined action of various compounds. In addition, the effect of low temperature should be taken into account. Cold weather is known to be associated with an increase in the frequency of respiratory infections (Pershagen et al., 1984; Marty et al., 1985; Vedal et al., 1987). This report presents correlations between (i) ambient levels of SO2, NO2, and temperature and (ii) respiratory diseases and absenteeism among children and adults. MATERIALS AND METHODS
This survey was carried out among children in day-care centers (DCC), children of school age, and adults in Helsinki during an 1-year period (1987). Helsinki, the capital of Finland, has a population of 492,000 inhabitants. Because of Helsinki's coastal position and rather low degree of industrialization, the ambient air is only mildly polluted, although energy is mainly supplied by coal-fired plants and despite exhaust gases from more than 210,000 cars. District municipal heating covers 89% of all households in Helsinki, cooking is done primarily by electricity, and the use of gas burners is negligible. Air Pollution M e a s u r e m e n t s
Air pollution measurements are carried out in Helsinki by district municipal authorities. SO2 is measured at four automatic monitoring stations by means of coulometric instruments, and NO2 at one station with a chemiluminescent apparatus. The arithmetic weekly mean of hourly levels was determined as well as the arithmetic weekly mean of maximum hourly levels for each 24-hr period. Temperature was recorded at one automatic station. From these data, weekly mean temperatures were calculated, and weekly minimum levels recorded. The city of Helsinki covers a rather small area, 185 km 2. The main sources of air pollutants are energy production by three coal-fired and six oil-fired power plants, road traffic, and, to a lesser extent, industrialization. In 1986, the emissions of SO2 in Helsinki were calculated to total 23,800 tons, of which 87% was derived from energy production, 1% from traffic, and 12% from other sources. The proportion of remote transmission has been estimated to be about one-fourth of the SO2 concentration, i.e., 5 p~g/m3. The corresponding emission of N O 2 was 17,600 tons in 1986, and the proportions from energy production and from traffic were 65 and 34%. Because of the relatively small and flat area of Helsinki and the numerous sources of air pollution, the geographical differences in the concentrations of pollutants within the populated areas are rather small; therefore, the mean of SO2 concentrations measured at four monitoring stations is used in the statistical analysis. In a previous study on the greater Helsinki area, measurements of air pollutants in the different areas were found to be closely correlated, despite the greater distances (Pershagen et al., 1984).
36
ANTTI PONKA
Study Groups and Disease Recording Several indicators of illness were used: the respiratory tract infections diagnosed at communal health centers, the respiratory tract infections in DCC children, and the absenteeism due to febrile illnesses among DCC children, schoolchildren, and adults. In Finland, individual communes are responsible for arranging free health care for the whole population. Helsinki has 25 communal health centers, handling 54% of all acute cases of illness requiring treatment. The patients represent all social classes and age groups, although those of working age largely use medical services rendered by the private sector as the law obliges free occupational medical services for all employees. These occupational medical services comprise the major part of the private medical sector in Helsinki. Communal health centers routinely send weekly reports of all cases of respiratory tract infection and other common infections to the Health Department of Helsinki. Thus, these reports give a good impression of the incidence of respiratory tract infections among the inhabitants of the city. The number of cases of upper respiratory tract infection (URI) reported during the 1-year survey was 46,826 and that of tonsillitis 10,780. The diagnoses were made by physicians or by trained nurses. Forty-four percent of 1- to 6-year-old children in Helsinki are cared for in about 300 communal day-care centers. The frequency of common communicable diseases among all these children is followed continuously by day-care personnel and reported to the Health Department monthly. The incidence of diseases was followed on a weekly basis among a subsample of 1532 children (10.9% of all 14,882 DCC children) in 14 day-care centers. These day-care centers were selected to represent various geographical areas in the city. The location of day-care centers, workplaces, power plants, and monitoring stations is presented in the Fig. 1. The number of reported cases of URI in the subsample of DCC children was 3579, of tonsillitis 84, of otitis media 482, and of lower respiratory tract infection (LRI) 131. Absenteeism due to febrile illnesses among these children is also recorded and reported weekly. Similarly, the corresponding absenteeism was recorded by trained nurses among all 55,506 schoolchildren in 210 schools and among 28,481 working adults in 12 workplaces, the latter being 10.3% of the total population of working age in Helsinki.
Statistical Analysis Four methods were used. First, simple product-moment correlations were computed among the weekly levels of ambient pollutants and temperatures, the numbers of reported cases of respiratory disease, and the frequency of absenteeism. The means for air pollutants and temperatures were calculated weekly from Saturday to Friday, and the means for diseases from Monday to Sunday. A time lag of 2 days was taken to correspond to the incubation period of febrile illnesses. The correlations were also calculated by using lags of 0, 1, or 3 days. Second, mean levels of pollutants and temperatures were divided at the mean
37
A B S E N T E E I S M A N D RESPIRATORY DISEASE
/t" 1 "-~...~ /'
/"
\ o
[ \',J'
0
t~ /
1
O
i O so~ ~1=
~"1
i ,
/
"h I
"\-~ o
"~"
.
o
°
o
,.o
_
.bJ'b
_~7
/J
?
e..,,q
i
"~
t
2,.
,/
I
Z'" /
( \
i I
i i i
FIG. 1. A map of the city of Helsinki representing the location of power plants, monitoring stations (SO2 and NO2), workplaces (closed circles) and day-care centers (open circles).
into high- and low-level periods. The incidences of diseases during these two periods were compared with Student's t test. Third, the incidences of diseases were compared after statistical standardization for temperature as described earlier by Pershagen et al. (1984). This approach was used to reduce the influence of low temperature on the incidence of diseases, in an attempt to clarify the effects of SO2 and NOz on the frequency of illness. The standardization for temperature has been done as follows: the weeks have been divided into two groups, those with a temperature above and those with a temperature under the mean temperature of the year. The following difference is regarded as the effect of temperature on each disease variable: the mean value of the variable in the above-mentioned class (above or under the mean) minus the total mean of the variable. This difference has been subtracted from each value of the dependent variable of the present week. The values of the variables for each
38
ANTTI PONKA
week have been used as dependent variables in the calculations. Student's t test has been used in the comparison. Fourth, the effect of various ambient air variables on illnesses were tested by using multiple stepwise regression. RESULTS
Concentrations of Air Pollutants and Sickness The means of air pollutants and temperatures are presented in Table 1 and the frequency of respiratory tract infections and absences due to febrile illnesses is presented in Table 2. The weekly means of SO2 concentration ranged from 9 to 62 ixg/m3, with an average of 21 ixg/m3. The annual mean concentrations of total suspended particles were 48-123 ixg/m3; smoke was not determined. The average weekly mean of NO 2 concentration was 47 p~g/m3. Because NO 2 was measured only at one station, the concentrations were calculated later on the areas of 29 day-care centers, situating all over the Helsinki. The estimates were done by using two different models (Petersen, 1980; Yamartino and Wiegand, 1986) and also taking into account the remote transmission of NO2. The calculated annual means of NO2 concentrations were 16--63 i~g/m3. Thus the results confirm the measurements of the automatic monitoring station, as well as the later measurements. During the follow-up year the 24-hr guidelines given by WHO for combined exposure to SO 2 and particulate matter (World Health Organization, 1987) was not exceeded as regards SO2. The highest l-hr value was 391 p~g/m3. The highest weekly mean concentration of SO2 was 61.5 ~g/m 3. The short-term guideline for NO 2 was exceeded only once, during 1 hr when the value reached 406 ~g/m 3. Thus, taking both short-term and long-term values into consideration, it is justified to classify Helsinki as a low-level pollution area. The correlations between the concentrations of SO2 and NO2 and cold were statistically highly significant (P < 0.0001) when calculated on a weekly or daily TABLE 1 MEAN WEEKLY AIR POLLUTANT CONCENTRATIONS AND TEMPERATURES IN HELSINKI IN 1987
SO 2 (~g/m3) Arithmetic mean
Mean
Median
Standard deviation
Range
21.1
Mean of daily maximums
53.0
17.0 48.0
11.7 20.8
9.0--61.5 25.%130.3
NO 2 (~g/m 3) Arithmetic mean Mean of daily maximums
46.5 84.5
43.8 78.8
12.5 26.3
28.5-81.1 54.6--184.9
Temperature (°C) Arithmetic mean Mean of weekly minimums
+ 3.1 + 0.1
+6.1 +3.4
10.9 12.3
- 2 9 . 0 - + 19.5 -37.0--+ 16.9
39
ABSENTEEISM AND RESPIRATORY DISEASE TABLE 2 REPORTED WEEKLY FREQUENCY OF RESPIRATORYTRACT INFECTIONS AND ABSENTEEISM IN I-IELSINKI IN 1987 Follow-up period (weeks)
Mean (number or percentage)
Standard deviation
Range
Respiratory tract infections diagnosed at health centers URI Tonsillitis
52 52
900.5 207.3
262.6 39.3
376-1514 126-296
Respiratory tract infections in DCC children URI Otitis media Tonsillitis Lower respiratory tract infection
37 37 37 37
96.7 13.0 2.3 3.5
28.2 5.0 1.6 2.9
52-173 5-27 0-7 0-11
Absenteeism DCC children Schoolchildren Adults
38 37 38
9.2 2.8 1.1
2.6 0.8 0.3
4.6-14.7 1.1-4.3 0.7-1.8
basis. This was as expected, because energy production is highest during the cold period of the year and also because the dilution of pollutants due to weather conditions is not as comprehensive as during other seasons. The correlations between the concentrations of SO2 measured on the four monitoring stations were all statistically highly significant, the mean value of the correlation coefficients being 0.693 (P < 0.0001). The average frequency of absences due to febrile illnesses among the DCC children was 9.2%, when the summer months and the Christmas and Easter holidays were excluded. The actual values varied rather widely from 4.6 to 14.7%. The average frequency of absences for the schoolchildren was 2.8% and for the people of working age 1.1%. The numbers reported to the Department of Health by the health centers averaged 901 cases of URI and 207 of tonsillitis weekly. The Correlation between Air Pollutants and Sickness The correlations among SO2, NO2, temperature, and sickness are presented in Table 3. The mean SO2 concentration and mean temperature correlated significantly with the incidences of URI and tonsillitis reported from the health centers. Similar correlations were seen with the frequency of absenteeism among DCC children and adults, and for temperature also with the absenteeism of schoolchildren. The NO 2 concentrations correlated only with the incidences of URI reported from the health centers. Because the distribution of SO2, NO2, and temperature was not normally distributed, the correlations between these and sickness were also calculated using the logarithmic values of the variables. However, this method gave similar results as the use of the non-log-transformed values.
40
ANTTI PONK,~
,
k,
V ,.,~ V ,..~
~
VI
I
0 l-
Z
Z
Z
Z v I v I
~
z
z v ~ v
& © Z
i
0 Z 0
Absenteeism and Respiratory Disease among Children and Adults in Helsinki in Relation to Low-Level Air Pollution and Temperature ANTTI PONK,~
Helsinki City Health Department, Viipurinkatu 2, SF 00510 Helsinki, Finland Received September 28, 1989 The weekly changes in ambient sulfur dioxide, nitrogen dioxide, and temperature were compared with the figures for respiratory infection in children and adults and for absenteeism from day-care centers (DCC), schools, and workplaces during a 1-year period in Helsinki. The annual average level of sulfur dioxide was 21 i~g/m3 and of nitrogen dioxide 47 i~g/m3; the average temperature was + 3. I°C. The levels of these pollutants and the temperature were significantly correlated with the number of upper respiratory infections reported from health centers. Low temperature also correlated with increased frequency of acute tonsillitis, of lower respiratory tract infection among DCC children, and of absenteeism from day-care centers, schools and workplaces. Furthermore, a significant association w a s found between levels of sulfur dioxide and absenteeism. After statistical standardization for temperature, no other correlations were observed apart from that between high levels of sulfur dioxide and numbers of upper respiratory tract infections diagnosed at health centers (P = 0.04). When the concentrations of sulfur dioxide were above the mean, the frequency of the upper respiratory tract infections was 15% higher than that during the periods of low concentration. The relative importance of the effects of low-level air pollution and low temperature on health is difficult to assess. © 1990AcademicPress, Inc.
INTRODUCTION
Exposure to high levels of sulfur dioxide ( 8 0 2 ) , nitrogen dioxide (NO2), and particulate matter in the ambient air is known to cause acute and chronic adverse effects on health, mainly on the respiratory tract (World Health Organization, 1977, 1979, 1987; Colley and Brasser, 1980; Lan and Shy, 1981; Love et al., 1981; Mostardi et al., 1981; Schenker et al., 1983, 1986; Arossa et al., 1987; Charpin et al., 1988). The World Health Organization has issued guidelines for exposure limits consistent with the protection of public health (World Health Organization, 1977, 1979, 1987). These are based on data from reliable epidemiological studies and take into account sensitive groups such as children and persons with chronic respiratory disease. However, clear-cut limits are difficult to draw (World Health Organization, 1987). Some studies in 1980s suggest that SO2 and NOE cause respiratory tract infections in children at concentrations lower than the given global guidelines (Lan and Shy, 1981; Love et al., 1981; Mostardi et al., 1981; Paunio et al., 1987; Goren and Hellmann, 1988). These results are inconsistent with some other studies among children or persons with chronic respiratory disease (Schenker et al., 1983; Pershagen et al., 1984; Marty et al., 1985; Schenker et al., 1986; Vedal et al., 1987). In suggesting causal interpretations of associations, however, 34 0013-9351/90 $3.00 Copyright© 1990 by AcademicPress, Inc. All rightsof reproductionin any formreserved.
A B S E N T E E I S M A N D RESPIRATORY DISEASE
35
one should be cautious, because the agents measured may be merely indexes of other pollutants or the health effects may be due to the combined action of various compounds. In addition, the effect of low temperature should be taken into account. Cold weather is known to be associated with an increase in the frequency of respiratory infections (Pershagen et al., 1984; Marty et al., 1985; Vedal et al., 1987). This report presents correlations between (i) ambient levels of SO2, NO2, and temperature and (ii) respiratory diseases and absenteeism among children and adults. MATERIALS AND METHODS
This survey was carried out among children in day-care centers (DCC), children of school age, and adults in Helsinki during an 1-year period (1987). Helsinki, the capital of Finland, has a population of 492,000 inhabitants. Because of Helsinki's coastal position and rather low degree of industrialization, the ambient air is only mildly polluted, although energy is mainly supplied by coal-fired plants and despite exhaust gases from more than 210,000 cars. District municipal heating covers 89% of all households in Helsinki, cooking is done primarily by electricity, and the use of gas burners is negligible. Air Pollution M e a s u r e m e n t s
Air pollution measurements are carried out in Helsinki by district municipal authorities. SO2 is measured at four automatic monitoring stations by means of coulometric instruments, and NO2 at one station with a chemiluminescent apparatus. The arithmetic weekly mean of hourly levels was determined as well as the arithmetic weekly mean of maximum hourly levels for each 24-hr period. Temperature was recorded at one automatic station. From these data, weekly mean temperatures were calculated, and weekly minimum levels recorded. The city of Helsinki covers a rather small area, 185 km 2. The main sources of air pollutants are energy production by three coal-fired and six oil-fired power plants, road traffic, and, to a lesser extent, industrialization. In 1986, the emissions of SO2 in Helsinki were calculated to total 23,800 tons, of which 87% was derived from energy production, 1% from traffic, and 12% from other sources. The proportion of remote transmission has been estimated to be about one-fourth of the SO2 concentration, i.e., 5 p~g/m3. The corresponding emission of N O 2 was 17,600 tons in 1986, and the proportions from energy production and from traffic were 65 and 34%. Because of the relatively small and flat area of Helsinki and the numerous sources of air pollution, the geographical differences in the concentrations of pollutants within the populated areas are rather small; therefore, the mean of SO2 concentrations measured at four monitoring stations is used in the statistical analysis. In a previous study on the greater Helsinki area, measurements of air pollutants in the different areas were found to be closely correlated, despite the greater distances (Pershagen et al., 1984).
36
ANTTI PONKA
Study Groups and Disease Recording Several indicators of illness were used: the respiratory tract infections diagnosed at communal health centers, the respiratory tract infections in DCC children, and the absenteeism due to febrile illnesses among DCC children, schoolchildren, and adults. In Finland, individual communes are responsible for arranging free health care for the whole population. Helsinki has 25 communal health centers, handling 54% of all acute cases of illness requiring treatment. The patients represent all social classes and age groups, although those of working age largely use medical services rendered by the private sector as the law obliges free occupational medical services for all employees. These occupational medical services comprise the major part of the private medical sector in Helsinki. Communal health centers routinely send weekly reports of all cases of respiratory tract infection and other common infections to the Health Department of Helsinki. Thus, these reports give a good impression of the incidence of respiratory tract infections among the inhabitants of the city. The number of cases of upper respiratory tract infection (URI) reported during the 1-year survey was 46,826 and that of tonsillitis 10,780. The diagnoses were made by physicians or by trained nurses. Forty-four percent of 1- to 6-year-old children in Helsinki are cared for in about 300 communal day-care centers. The frequency of common communicable diseases among all these children is followed continuously by day-care personnel and reported to the Health Department monthly. The incidence of diseases was followed on a weekly basis among a subsample of 1532 children (10.9% of all 14,882 DCC children) in 14 day-care centers. These day-care centers were selected to represent various geographical areas in the city. The location of day-care centers, workplaces, power plants, and monitoring stations is presented in the Fig. 1. The number of reported cases of URI in the subsample of DCC children was 3579, of tonsillitis 84, of otitis media 482, and of lower respiratory tract infection (LRI) 131. Absenteeism due to febrile illnesses among these children is also recorded and reported weekly. Similarly, the corresponding absenteeism was recorded by trained nurses among all 55,506 schoolchildren in 210 schools and among 28,481 working adults in 12 workplaces, the latter being 10.3% of the total population of working age in Helsinki.
Statistical Analysis Four methods were used. First, simple product-moment correlations were computed among the weekly levels of ambient pollutants and temperatures, the numbers of reported cases of respiratory disease, and the frequency of absenteeism. The means for air pollutants and temperatures were calculated weekly from Saturday to Friday, and the means for diseases from Monday to Sunday. A time lag of 2 days was taken to correspond to the incubation period of febrile illnesses. The correlations were also calculated by using lags of 0, 1, or 3 days. Second, mean levels of pollutants and temperatures were divided at the mean
37
A B S E N T E E I S M A N D RESPIRATORY DISEASE
/t" 1 "-~...~ /'
/"
\ o
[ \',J'
0
t~ /
1
O
i O so~ ~1=
~"1
i ,
/
"h I
"\-~ o
"~"
.
o
°
o
,.o
_
.bJ'b
_~7
/J
?
e..,,q
i
"~
t
2,.
,/
I
Z'" /
( \
i I
i i i
FIG. 1. A map of the city of Helsinki representing the location of power plants, monitoring stations (SO2 and NO2), workplaces (closed circles) and day-care centers (open circles).
into high- and low-level periods. The incidences of diseases during these two periods were compared with Student's t test. Third, the incidences of diseases were compared after statistical standardization for temperature as described earlier by Pershagen et al. (1984). This approach was used to reduce the influence of low temperature on the incidence of diseases, in an attempt to clarify the effects of SO2 and NOz on the frequency of illness. The standardization for temperature has been done as follows: the weeks have been divided into two groups, those with a temperature above and those with a temperature under the mean temperature of the year. The following difference is regarded as the effect of temperature on each disease variable: the mean value of the variable in the above-mentioned class (above or under the mean) minus the total mean of the variable. This difference has been subtracted from each value of the dependent variable of the present week. The values of the variables for each
38
ANTTI PONKA
week have been used as dependent variables in the calculations. Student's t test has been used in the comparison. Fourth, the effect of various ambient air variables on illnesses were tested by using multiple stepwise regression. RESULTS
Concentrations of Air Pollutants and Sickness The means of air pollutants and temperatures are presented in Table 1 and the frequency of respiratory tract infections and absences due to febrile illnesses is presented in Table 2. The weekly means of SO2 concentration ranged from 9 to 62 ixg/m3, with an average of 21 ixg/m3. The annual mean concentrations of total suspended particles were 48-123 ixg/m3; smoke was not determined. The average weekly mean of NO 2 concentration was 47 p~g/m3. Because NO 2 was measured only at one station, the concentrations were calculated later on the areas of 29 day-care centers, situating all over the Helsinki. The estimates were done by using two different models (Petersen, 1980; Yamartino and Wiegand, 1986) and also taking into account the remote transmission of NO2. The calculated annual means of NO2 concentrations were 16--63 i~g/m3. Thus the results confirm the measurements of the automatic monitoring station, as well as the later measurements. During the follow-up year the 24-hr guidelines given by WHO for combined exposure to SO 2 and particulate matter (World Health Organization, 1987) was not exceeded as regards SO2. The highest l-hr value was 391 p~g/m3. The highest weekly mean concentration of SO2 was 61.5 ~g/m 3. The short-term guideline for NO 2 was exceeded only once, during 1 hr when the value reached 406 ~g/m 3. Thus, taking both short-term and long-term values into consideration, it is justified to classify Helsinki as a low-level pollution area. The correlations between the concentrations of SO2 and NO2 and cold were statistically highly significant (P < 0.0001) when calculated on a weekly or daily TABLE 1 MEAN WEEKLY AIR POLLUTANT CONCENTRATIONS AND TEMPERATURES IN HELSINKI IN 1987
SO 2 (~g/m3) Arithmetic mean
Mean
Median
Standard deviation
Range
21.1
Mean of daily maximums
53.0
17.0 48.0
11.7 20.8
9.0--61.5 25.%130.3
NO 2 (~g/m 3) Arithmetic mean Mean of daily maximums
46.5 84.5
43.8 78.8
12.5 26.3
28.5-81.1 54.6--184.9
Temperature (°C) Arithmetic mean Mean of weekly minimums
+ 3.1 + 0.1
+6.1 +3.4
10.9 12.3
- 2 9 . 0 - + 19.5 -37.0--+ 16.9
39
ABSENTEEISM AND RESPIRATORY DISEASE TABLE 2 REPORTED WEEKLY FREQUENCY OF RESPIRATORYTRACT INFECTIONS AND ABSENTEEISM IN I-IELSINKI IN 1987 Follow-up period (weeks)
Mean (number or percentage)
Standard deviation
Range
Respiratory tract infections diagnosed at health centers URI Tonsillitis
52 52
900.5 207.3
262.6 39.3
376-1514 126-296
Respiratory tract infections in DCC children URI Otitis media Tonsillitis Lower respiratory tract infection
37 37 37 37
96.7 13.0 2.3 3.5
28.2 5.0 1.6 2.9
52-173 5-27 0-7 0-11
Absenteeism DCC children Schoolchildren Adults
38 37 38
9.2 2.8 1.1
2.6 0.8 0.3
4.6-14.7 1.1-4.3 0.7-1.8
basis. This was as expected, because energy production is highest during the cold period of the year and also because the dilution of pollutants due to weather conditions is not as comprehensive as during other seasons. The correlations between the concentrations of SO2 measured on the four monitoring stations were all statistically highly significant, the mean value of the correlation coefficients being 0.693 (P < 0.0001). The average frequency of absences due to febrile illnesses among the DCC children was 9.2%, when the summer months and the Christmas and Easter holidays were excluded. The actual values varied rather widely from 4.6 to 14.7%. The average frequency of absences for the schoolchildren was 2.8% and for the people of working age 1.1%. The numbers reported to the Department of Health by the health centers averaged 901 cases of URI and 207 of tonsillitis weekly. The Correlation between Air Pollutants and Sickness The correlations among SO2, NO2, temperature, and sickness are presented in Table 3. The mean SO2 concentration and mean temperature correlated significantly with the incidences of URI and tonsillitis reported from the health centers. Similar correlations were seen with the frequency of absenteeism among DCC children and adults, and for temperature also with the absenteeism of schoolchildren. The NO 2 concentrations correlated only with the incidences of URI reported from the health centers. Because the distribution of SO2, NO2, and temperature was not normally distributed, the correlations between these and sickness were also calculated using the logarithmic values of the variables. However, this method gave similar results as the use of the non-log-transformed values.
40
ANTTI PONK,~
,
k,
V ,.,~ V ,..~
~
VI
I
0 l-
Z
Z
Z
Z v I v I
~
z
z v ~ v
& © Z
i
0 Z 0
