| REPORT
ON DEATHS IN THE CONSTRUCTION INDUSTRY |
|
|
| INTRODUCTION |
A
report, Inequalities and Health in Ireland published in 2001,
confirms the broad international experience of significantly
higher mortality rates for people from lower socio-economic
groups for all causes and for ischaemic heart disease, cancer
and injuries and poisoning. Construction workers in the
main fall into the lower socio-economic groups – craftsmen
are in Social Class 4, semi-skilled manual workers in Social
Class 5 and unskilled manual workers in Social Class 6. |
 |
Dr
Harold Brenner MB, FFPHMI. |
|
Studies
in several countries of deaths of construction workers have
emonstrated higher mortality rates compared with men in their
general populations. In the UK there were significantly increased
risks for deaths from all cancers, lung cancer and accidents
and poisoning. Similar increased risks were found in the USA.
In Japan a recent study revealed significant excess mortalities
for accidents, and also for specific types of cancer among
different trades or jobs within the industry. |
The
Construction Workers Health Trust (CWHT), having recently
researched the health status of construction workers based
on sickness absence and early retirement from the industry
on health grounds, decided to examine the deaths of construction
workers who died in the years 1995 to 2000. The objectives
were: |
|
To
study the mortality of construction workers and ex-construction
workers who had
contributed to the Construction Federation Operative
Pension Scheme (CFOPS) |
|
To
establish the causes of death and to compare the principle
causes with those in the general Irish male population. |
|
| METHOD |
Death
certificates of members of CFOPS who died in the years 1995
to 2000 and whose kin had applied for any type of death benefit
from the Scheme were examined. In most cases copies of the
death certificates had been filed for the purpose of substantiating
the claim for benefit, either on microfiche or more recently
on a computer software programme. Where the copies of certificates
were not filed, microfiche copies of the original certificates
were examined in the General Register Office by permission
of the Superintendent Registrar. The age and cause of death
were extracted from the certificates; the trade or occupation
within the industry of the deceased was established either
from CFOPS records or from the death certificate. The causes
of death were grouped using the ICD- 9 classification. These
data were entered into an Epi Info database for later analysis. |
| |
|
| RESULTS |
Available
records of deaths associated with benefit claims during the
six-year period under study numbered 1,654. |
Types
of benefit |
Types
of benefit associated with the deaths, mortality benefit (MB),
widows pension (WP), widows pension commuted (WPC), funeral
expenses (FE), refund of contributions (RD) and discretionary
payment (DP), are shown in the following table: |
|
| Trade
or occupation |
| The
trade or occupation of the study population for those dying
under the age of 65 and for those aged 65 and over is shown
in the following table: |
|
| Note:
Without relevant baseline data, the above table is for descriptive
purposes only and cannot be used for comparisons of deaths
between trades and occupations. |
| Age
at death |
| Age
at death was established in all cases; the average age of
death in the study population was 70 years (S.D. 13.4) compared
to 72 years (S.D. 14.7) in 1998 for Irish males aged 20 years
and over in the general population. The two-year difference
between these means is statistically significant (95% C.I.
1.02 to 2.96). The average age of death for those aged 65
and over was 76 years although in the late 1980s life expectancy
for males aged 65 in the general population was 13 years. |
| Assuming
a similar age distribution in the two groups, the following
bargraph illustrates the difference in percentage deaths in
10-year age groups between the study population and the Irish
male population (1998) aged 20 years and above. |
|
| Length
of service |
The
length of service in the industry of members who died under
the age of 65 years whose kin were eligible for mortality
benefits was estimated by examination of a 50% sample of their
pension records. The mean length was 12.5 years (median 10
years); the length varied from 1 to 46 years. |
| |
|
| Cause
of death |
| Cause
of death was established in 1634 cases, 1551 from CFOPS records
and 83 cases from the General Register Office records. No
death certificate could be traced in 20 cases. |
| The
causes of death were grouped according to the ICD-9 classification
and are shown in the
following table. |
|
| As
shown in the four following tables, the three principal causes
of death, circulatory disease malignant cancer and respiratory
disease, as well as deaths from injury and poisoning, were sub-grouped
into the commoner causes, for those under the age of 65 and
for those aged 65 and over. |
|
|
|
|
| |
|
| For
lack of appropriate denominators it was not possible to compute
standardised death rates or standard mortality ratios in order
to compare deaths in the study population with those in the
general population. However it is possible to compare mortality
from particular causes among a study population with that
of the general population using proportional mortality ratios
(PMRs), and this method is widely used in mortality studies. |
For
this report proportional mortality ratios were calculated
for malignant cancer, circulatory disease, injury and poisoning
and respiratory disease as well as for selected causes within
these major groupings; they are shown in the following two
tables. The ratios were obtained both for deaths in the age
group 20 to 64 and for all deaths in the study population.
Deaths in the appropriate age groups of the 1998 Irish male
population were used as standards. |
| A
PMR over or under 100 indicates a higher or lower than expected
mortality from a particular cause. The confidence intervals
(C.I.s) for the PMRs have been computed at the 95% level.
If the C.I. does not contain 100, the PMR is statistically
significant. |
| |
|
| DISCUSSION |
This
study is subject to certain limitations and as such the findings
must be interpreted with care. It is recognised that death
certificates may contain considerable inaccuracies in certifying
causes of death. The study is limited to deaths for which
death benefits are claimed, however there is no reason to
think there is any difference in age or causes of deaths among
eligible construction workers whose kin do not claim such
benefits; nor would there be any reason to believe that the
findings would be different for construction workers who are
not members of a pension scheme. In the absence of reliable
denominators, proportional mortality ratios were used for
comparison between the sample and a standard population (the
standard used was the 1998 Irish male population aged 20 years
and above); this method is widely used in studies of occupational
mortality, although PMRs need to be interpreted with care
when mortality is low in a particular group. Length of service
in the industry is difficult to ascertain accurately since
individuals may not necessarily be engaged in construction
work during periods between employments. |
The
mean age at death was 70 years (S.D. 13.4) compared to 72
years (S.D. 14.6) in the general Irish male population aged
20 years and over. This difference is statistically significant.
The bargraph illustrating the percentage deaths occurring
in 10-year age groups clearly shows an excess of deaths in
construction workers between the ages of 40 and 69 over the
general population. This would support a generally held belief
that “construction workers die younger.” For those
deaths occurring under the age of 65, the mean age of death
was 52.5 years, similar to the U.K study where the mean age
at death was 53.1 years. |
| For
descriptive purposes only, the trade or occupation of the
men in the study was established. The categories of general
operative, carpenter, bricklayer and painter accounted for
63% of the total deaths (in the U.K. study this figure was
70%). No attempt was made to compare deaths between categories. |
| |
|
| Length
of service in the industry prior to death was estimated from
the records relating to mortality benefits of workers under
the age of 65. The median length of service was 10 years (mean
12.5 years) compared to the median length of reckonable service
in the U.K. study of 5 years. |
The
study clearly suggests that construction workers experience
an increased mortality from malignant cancer, particularly
from lung and stomach cancer, and from injury and poisoning,
particularly from falls in the under 65 year age group. This
is consistent with studies of construction worker mortality
in other countries. Occupational hazards encountered at work
as well as certain behavioural and socio-economic factors
almost certainly contribute to this excess in mortality in
comparison to the general population. An example would be
seen in cancers of the oropharynx and oesophagus which are
known to be cause-related to the multiplicative effect of
smoking and alcohol intake; in the present study there was
a high PMR (132) for these cancers which, although not statistically
significant, is an important finding. A total of ten deaths
from mesothelioma was recorded in the present study; during
the same six-year period as the study there were over 60 deaths
in the standard population, which would include the ten deaths
in the study. It should be noted that mesothelioma occurs
mainly as a result of exposure to asbestos not only in the
construction industry – it is known that vehicle body
builders, boiler operators, chemical workers and upholsterers
are job groups with significantly high PMRs for mesothelioma.
As would be expected there was a highly significant PMR (244
CI 169 – 319) for mesothelioma in the U.K. study, and
a mortality study of electrical workers in the construction
industry in the U.S.A. showed a PMR of 356 for mesothelioma. |
It
would appear that work in the construction industry does not
contribute to excess deaths from ischaemic heart disease and
this reflects the findings in other similar studies. This
may appear surprising given the reported level of smoking
and degree of obesity among construction workers. However
in this study, there was a significant PMR (130) for deaths
from cerebrovascular disease in all the study population.
There was also a low PMR for respiratory disease especially
in those under the age of 65, again reflecting the experience
in other studies. |
Overall
the mean length of absence was 23.5 days; the median length
was 17 days. The annual trend in mean length of absence is
shown in Figure 4. The mean for 1996 is from data of the first
eight months of the year. Interestingly, there is a strong,
but spurious, inverse correlation (Pearson r2 = .81) between
the mean length of absence and the size of the population. |
| Although
there is concern over anecdotal reports of suicide in the
industry, the study shows significantly less suicides proportional
to the general population. Possible part explanations are
that unemployment is a contributory factor to suicide in the
general population, but in the period under study there was
a high demand for workers in the construction industry, and
that as a whole construction workers derive satisfaction and
pride from the results of their work. |
| Although
the present study sample is small compared to similar studies
in the U.K. (15,000 death certificates examined in one study),
in Japan (17,000 in one study) and the U.S.A. 31,000 in one
study), the sample is sufficiently large that the findings can
be extrapolated to all construction workers in Ireland, estimated
to be in the region of 150,000, with a high degree of certainty. |
| |
|
| ACKNOWLEDGEMENTS |
The
study was commissioned by the Trustees of the Construction
Workers Health Trust and was conducted by Dr. Harold Brenner,
Medical Director of CWHT. He acknowledges with thanks the
permission of the Board of the Construction Federation Operatives
Pension and Sick Pay Scheme and its Administrator, Mr. Pat
Ferguson, for access to its records, and to the staff of CFOPS
for facilitating the extraction of data. Thanks are due to
Mr.Tom Joyce of the General Register Office who readily facilitated
a search for a number of death certificates. Helpful advice
was given by Mr.William Ahern, Chief Executive of CWHT, and
the assistance of Ms Móirlin McGuire, Executive Assistant
of CWHT, ensured that the research proceeded smoothly. |
