1 Division of Epidemiology and Biostatistics, European Institute of Oncology, Milan, Italy; 2 Department of Medical Oncology, St James Hospital, Dublin, Republic of Ireland; 3 Department of Oncology, The Queens University of Belfast, Belfast City Hospital, Belfast, UK
Received 5 December 2001; revised 19 April 2002; accepted 22 July 2002
![]() |
Abstract |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Investigation of long time series of cancer data can still be very useful in helping to identify Cancer Control priorities and achievements. Since the partition of Ireland into the independent Republic of Ireland and Northern Ireland, which remained part of the United Kingdom, cancer mortality data have been published in an essentially similar format in both countries. The information presented here will contribute to providing a basis for the collaborative Cancer Research programme initiated recently.
Patients and methods:
Cancer mortality data have been assembled and analysed separately for the Republic of Ireland and Northern Ireland: the data have then been combined to present mortality rates for the whole of Ireland, covering the period from 1926 to 1995. Several rubrics had to be aggregated to provide data continuously over the time span (e.g. colon and rectum and cervix and body of the uterus). When data were only available in 10-year classes of age, the EM algorithm was employed to obtain 5-year age-specific rates. All rates presented are age-standardised, employing the World Standard Population.
Results:
In women, the death rate from all neoplasms combined increased very slightly from 117 per 100 000 in 19461950 to 120 per 100 000 in 19911995. In men, the death rate increased from 127 per 100 000 to 172 per 100 000 over the same time period. The overall cancer death rate in Ireland is currently similar to the European average in men, although in women it is among the top fifth of national cancer mortality rates in European countries. While cancer is a major cause of death in Ireland, there is no evidence of an evolving epidemic building up: the death rates from most forms of cancer are declining towards the end of the time period considered.
Conclusions:
As demonstrated by falling death rates from Hodgkins disease and testicular cancer, major treatment advances appear to have been incorporated effectively into clinical practice in Ireland. Progress is apparent in tobacco control and further initiatives in this area must be undertaken since tobacco appears to be the only major new carcinogen introduced recently into the Irish environment during the period covered by this study. Effective population-based screening programmes for cervix and breast cancer and, more controversially, consideration of a National Prostate Cancer Screening programme, offer scope for further improvement in mortality. Examination of this long time series of mortality data from Ireland provides information about the evolving cancer pattern and provides the necessary background to evaluate the impact of the cross-border cancer research activities now being launched.
Key words: cancer, control, epidemiology, Ireland
![]() |
Introduction |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Michael Alderson and Robert Case put together Serial Mortality Tables separately for the Republic of Ireland [1] and Northern Ireland [2] using mortality and population data available since partition in the mid-1920s. There are some inconsistencies in how the data have been collected and presented, particularly with the use of different age classes at different time periods. However, there are statistical tools which are useful in removing some of these obstacles. In addition to this historical series, mortality data are available until the mid-1990s from the World Health Organisation (WHO) Mortality Database.
Cancer mortality data have been assembled and analysed separately for the Republic of Ireland and Northern Ireland: the data have then been combined to present mortality rates for the whole of Ireland. The United States National Cancer Institute has recently embarked on an international partnership with the developing cancer programmes in the Republic of Ireland and in Northern Ireland in an attempt to further improve the quality and range of cancer services available for patients. The All IrelandNCI Cancer Consortium held its first conference in Belfast in October 1999 when the aims of the agreement were presented [3]. The information presented here will contribute to providing a basis for prioritising and evaluating the collaborative cancer research programme initiated as an outcome of the peace process.
![]() |
Patients and methods |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
There are some inconsistencies in how the data are presented which need to be allowed for prior to calculating the death rates. In the tables of the Republic of Ireland, data are presented for the age group 7685 for the periods 19261930, 19311935, 19361940, 19411945 and 19461950. For consistency with later data, data are required for the age groups 7680 and 8185. Disaggregation has been achieved using a non-parametric cubic spline [4] applied to the cumulative population.
In the mortality data for both the Republic of Ireland and Northern Ireland for the period from 1926 to 1950, there are many grouped age classes. In order to disaggregate these classes, the EM algorithm [5, 6] was employed, based on an age-period Poisson regression model. For example, data for 19261950 may be available in age classes 4150, while data from 1950 onwards are available in 5-year age groups. Also, combined data for age groups 7685 are presented in the Republic of Ireland from 1926 to 1950. This disaggregation was only necessary for the analysis of age-specific rates and not for the calculation of the age-standardised rates.
Separately for each site and gender, the first estimate of the number of cases in each 5-year age class was obtained by dividing the total number of deaths in the aggregated class by the number of 5-year age groups within it. Usually there were only two which meant assigning half the deaths to each age class. An age-period model was then fitted using all the data in the 5-year groups. This gives the predicted number of deaths in each age group (eij). The number of deaths in the 5-year age groups of the aggregated age classes is then re-estimated in proportion to the predicted number of deaths. If there are two 5-year age groups aggregated with predicted values e1 and e2 and total deaths (d) in the aggregated group, then e1/(e1 + e2)*d and e2/(e1 + e2)*d are used as the new estimates of the numbers of deaths. This process is then repeated until the estimated deaths converge.
Due to changes in coding practice and publishing strategy, the data for oral cavity and larynx cancer are unavailable for the period from 1951 to 1955. In this event, we have taken the average of the preceding and following rates for the periods for men and women in this time period to get an estimate of the rate.
In order to minimise the effect of differences in age composition in different periods and areas, the rates are age-adjusted to the World Standard Population [7]. The relative risk of a particular cancer in Northern Ireland compared with the Republic of Ireland was calculated by the ratio of the two age-adjusted rates, for men and women separately. The 95% confidence interval for the ratios was calculated using the variance of the ratio [8] and a normal approximation [8].
![]() |
Results |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
|
|
|
|
|
Cancer of the oesophagus
The coding rubric for cancer of the oesophagus remained unchanged and unique during the period studied. In men and women, the mortality rate has increased by around 50%. However, there are some important differences in the pattern observed. Since 1965, the death rate in women has remained fairly constant while there has been a consistent increase in the death rate observed in men, rising from 4.2 per 100 000 in 19561960 to 7.5 per 100 000 in 19911995 (Figure 1).
Cancer of the stomach
The coding and publication of data on cancer of the stomach remained unchanged throughout the period of observation. In both gender groups, mortality reached a peak in 19411945 and has subsequently experienced a consistent decline. In men the mortality rate peaked at 30 per 100 000 and has declined to 11 per 100 000 in the most recent period. In women, the corresponding decline was from 22 to 5 per 100 000 (Figure 2).
|
Cancer of the liver and gallbladder
Although of questionable epidemiological value, these two sites have been combined in the published data for the earlier decades. The combined site has been presented here for completeness (Figure 2). There has been a substantial and consistent decline in mortality rates during the period covered. In men, the death rate fell from 8.1 in 19261930 to 1.3 in 19911995. In women the corresponding declines were from 9.4 to 1.3 per 100 000 (Figure 2).
Cancer of the pancreas
This rubric was preserved throughout the period studied. Mortality rates in men were consistently higher than in women, although they paralleled one another throughout the study period. Death rates increased consistently until the period 19761980 and have declined marginally in both gender groups subsequently (Figure 2).
Cancer of the trachea, bronchus and lung and pleura
This rubric remained consistent throughout the period studied although the identity of cancer of the pleura has been obscured by the all-embracing rubric: for simplicity it shall be referred to as lung cancer. Lung cancer increased in men from an annual death rate of 0.8 per 100 000 in 19261930 to reach a peak of 51.6 per 100 000 in 19811986 (Figure 3). The rate has subsequently fallen to 46.3 per 100 000 in the final (19911995) time period (Figure 3). There is a clear cohort effect underlying this pattern of lung cancer change (Figure 4). In women, the death rate at the beginning of the observation period (0.84 per 100 000) was very similar to that in men. The death rates in women have risen and the peak death rate reached (in 19851990 and 19911995) was 18 per 100 000 per annum (Figure 3).
|
|
|
|
Cancer of the prostate and testes
These rubrics remained identifiable throughout the period covered. Mortality from cancer of the testes rose to attain a peak in the period 19661970 (of 0.82 per 100 000), remained constant over the next decade (at 0.79 per 100 000) and subsequently fell to 0.35 per 100 000 in the final time period (19911995) (Figure 7). Cancer of the prostate rose throughout the three-quarters of a century covered by this study, starting at a rate of 3 per 100 000 and rising to a death rate of 17.5 per 100 000 in 19911995 (Figure 7).
|
|
Cancer of the thyroid
This rubric remained unchanged and unique throughout the period of the study. Survival from thyroid cancer is very good which makes interpretation of temporal trends in mortality more difficult. Thyroid cancer mortality rate was higher in women than in men throughout the study period. In women, the mortality rate peaked in the period 19561960 (at 1.1 per 100 000 per annum) and has subsequently declined consistently (to reach 0.5 per 100 000 in 19911995) (Figure 8). In men, the death rate has remained unchanged since the start of the 1960s (Figure 8).
All neoplasms
All neoplasms include cancers of all forms and other tumours whether of benign or unspecified behaviour. The data concerning such deaths in the Republic of Ireland are not available before 1927 and in the period 19401948. The number of deaths in 19271930 and 19361939 has been multiplied by 5/4 in order to approximate the true number of deaths in the periods 19261930 and 19361940, respectively. Moreover, the number of deaths in 19491950 is multiplied by 5/2 in order to approximate the true number of deaths in the period 19461950. The data concerning the deaths in Northern Ireland are not available before 1944. The number of deaths in 19441945 has been multiplied by 5/2 in order to approximate the true number of deaths in the period 19411945. In women, the death rate from all neoplasms combined increased very slightly from 117 per 100 000 in 19461950 to 120 per 100 000 in 19911995. In men, the death rate increased from 127 per 100 000 to 172 per 100 000 over the same time period.
![]() |
Discussion |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Cancer mortality data exist for Ireland since partition took place in the early 1920s. Of course, cancer incidence data would have been more helpful, but they simply do not exist for the entire country historically and have only become available in more recent years [20, 21]. However, cancer mortality data can be very useful, although issues such as the introduction of successful treatment and abrupt changes in coding practices must be kept in mind when considering such information.
The separate data for the Republic of Ireland and Northern Ireland have been aggregated having been shown not to be significantly heterogeneous when sites and gender groups were considered across the two countries. The aggregations necessary to have comparable and continuous data over the 70 year period has, of course, led to some loss of information. It would have been useful to have separate data on cancer of the pleura, given the major shipbuilding industry which existed in Northern Ireland for many decades. There are known difficulties in coding practices which limit the utility of individual death rates from cancer of the colon and rectum [22] and cancer of the cervix and body of uterus [23]. Thus, although amalgamation of these rubrics means that no data are discarded, it still represents a loss of useful information.
The overall cancer mortality rates in Ireland are slightly below the European average in men and above the average in women [24]. The overall cancer death rate is dominated by lung cancer and it is reassuring to see a fall in mortality clearly established in men (Figure 3). This decline has occurred later than among men in the UK [25]. There is a suggestion of stabilisation in the death rate from lung cancer among Irish women in the last time period which, if it continues, would be a source of satisfaction. Tobacco control activities in Ireland are hopefully having the desired effect of reducing tobacco-related cancers but there are no grounds for complacency. It is encouraging to note that the majority of smokers wish to quit smoking and many wish to reduce their smoking activity [26].
The aetiology of cancers of the oral cavity, oesophagus and larynx have strong similarities: independent effects of cigarette smoking, alcohol consumption and the joint (multiplicative) effects of both risk factors influence the risk of all three forms of cancer [27]. However, the pattern of temporal change in death rates from all three cancers in Ireland is quite different. The most recent increases seen in oral cavity and oesophagus cancers in men, which have been taking place for a longer time for the latter, probably reflect an increasing alcohol consumption: it is strange, however, that no change is apparent in cancer of the larynx over the same period (Figure 1). It could be that the effect of alcohol on larynx cancer is less than that for oral cavity and oesophageal cancers. Some of the leading regional rates in oesophageal cancer in women in Europe have been apparent in regions of Ireland [22]. The diverging trends seen in oesophageal cancer in men (increasing noticeably) and women (decreasing slowly) in Ireland is noteworthy and deserves some further attention. For similar reasons, so too does colorectal cancer.
Long-time decreases in stomach cancer in both gender groups is a common finding in Ireland as well as most other western European countries (Figure 2), as are the recent declines in intestinal cancer. The declines in death rates from cancer of the liver and gallbladder are quite noticeable and difficult to explain simply (although the incorrect attribution of metastatic liver cancers as primary tumours in the past may be an important contributing factor). There has been no major change in coding practice which could explain this nor any major change in risk factors which could explain this change: hepatitis in Ireland over the time period is not a major factor nor is changing alcohol consumption. The increase in death rate from pancreas cancer in the early part of the century is a notable feature of the current data set as is the recent decrease in death rates seen in both gender groups.
Some of the highest regional rates in the European Community for breast cancer have been found in regions of both Northern Ireland and the Republic of Ireland [22]. The overall breast cancer mortality rate in Ireland is among the highest in Europe [24]. Although the death rate has increased from 1926 until 1975, it appears to have stabilised subsequently (Figure 5), although it is not yet decreasing. The impact of a national population-based screening programme in Northern Ireland in 1988 has had an impact on reducing death rates in middle-aged women who have been eligible for screening (P. Boyle, S. Golia, C. Robertson, N. OHiggins, unpublished data). In contrast, there has been no such change in the Republic of Ireland in this latter time period (19911995) when there has been no organised national screening programme. This supports the efficacy of organised mammographic screening programmes.
The strong decline in cancer of the uterus is probably a reflection of cervical cancer screening although further decline could be possible due to the introduction of an organised programme throughout the island. The increasing death rate from ovarian cancer is consistent and substantial and could usefully be investigated.
The increase in prostate cancer (Figure 7), greater than five-fold over the 70 years covered, reflects the increasing importance of this disease in men. Prostate cancer is the third commonest form of cancer death in Irish men at the present time after lung cancer and intestinal cancer. Unless some successful action is initiated soon, this situation will change for the worse, especially given the striking changes in lung cancer. The nature of this action is not known at present. In particular, prostate cancer screening remains controversial [28] although there have been some strong suggestions of its efficacy in populations with high-quality treatment facilities available [29].
The decline in the death rate from testicular cancer (Figure 7) and Hodgkins disease (Figure 8) both occur at similar points when new, effective therapy has been introduced: similar changes have been seen for both testicular cancer [3033] and Hodgkins disease [34, 35] in other communities. It is obvious that the advances in therapy which have been effective against these cancers have been introduced successfully into clinical practice in Ireland to the benefit of patients with these diseases. This is also true to some extent for leukaemia (Figure 8), where a decline has also taken place.
Generally, examination of this long time series of mortality data from Ireland provides reassuring information. This analysis demonstrates that there is no overall cancer epidemic building up; that the death rates from most forms of cancer considered are declining; that major treatment advances have been incorporated effectively into clinical practice; that progress is apparent in tobacco control and that there is no evidence from these data of a major new carcinogen being introduced into the Irish environment (apart from tobacco). Currently, the overall cancer death rate in Irish men is in line with other European countries, although the high mortality rate for cancer in women, among the top fifth of national cancer mortality rates in European countries, highlights an important priority.
However, it has recently been shown, in an evaluation of the Europe Against Cancer programme, that between 1985 and 2000, Ireland was tenth out of 15 member states in terms of reductions in cancer mortality (P. Boyle, A. dOnofrio, P. Maisonneuve et al., unpublished data). There is clearly scope for improving this situation, notably with the introduction of an effective population-based screening programme for cervix and breast cancer, continuing successful tobacco control initiatives and, more controversially, with consideration of a National Prostate Cancer Screening programme. There could be some major local factor contributing to the divergent patterns seen in cancers of the oral cavity, larynx and oesophagus and, together with the aetiology of breast and ovarian cancer, it may be worthwhile pursuing. These are areas worthy of further investigation and research.
These data provide a baseline for further cancer control and treatment activities in Ireland.
![]() |
Acknowledgements |
---|
![]() |
Footnotes |
---|
![]() |
References |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
2. Case RAM, Coghill C, Davies JM et al. Serial Mortality Tables. Neoplastic Diseases, Volume 3. Northern Ireland, 19221970. Deaths and Death Rates by Sex, Age, Site and Calendar Period. London, UK: Institute of Cancer Research 1976.
3. Johnston PG, Daly PA, Liu E. The NCI All Ireland Cancer Conference. Oncologist 1999; 4: 275277.
4. Green PJ, Silverman BW. Non-parametric Regression and Generalized Linear Models. London, UK: Chapman and Hall 1994.
5. Dempster A, Laird NM, Rubin DB. Maximum likelihood from incomplete data via the EM algorithm. J R Stat Soc (B) 1977; 39: 138.[ISI]
6. McLachlan GJ, Krishnam T. The EM algorithm and extensions. New York, NY: John Wiley 1977.
7. Boyle P, Parkin DM. Statistical Methods for Registries. In Jensen OM, Parkin DM, McLennan R et al. (eds): Cancer Registration: Principles and Methods. IARC Scientific Publication No. 95. Lyon, France: IARC 1991; 126158.
8. Stuart A, Ord JK. Kendalls Advanced Theory of Statistics. Volume 1: Distribution Theory, 6th edition. London, UK: E. Arnold 1994.
9. La Vecchia C, Lucchini F, Negri E et al. Trends of cancer mortality in Europe, 19551989. I. Digestive sites. Eur J Cancer 1992; 28: 132235.[CrossRef][Medline]
10. Levi F, Lucchini F, Negri E et al. Cancer mortality in Europe, 19901994, and an overview of trends from 1955 to 1994. Eur J Cancer 1999; 35: 14771516.[CrossRef][ISI][Medline]
11. McCartan BE, Crowley M. Oral cancer in Ireland, 19841988. Eur J Cancer B Oral Oncol 1993; 29B: 127130.[CrossRef]
12. Buttimer J, McInerney D. Development and utilization of mammographic services in Ireland 19851991. J Med Sci 1995; 164: 122124.
13. Gavin AT, OReilly D. Cancer in Northern Ireland by 2002. Ulster Med J 1996; 65: 106112.[ISI][Medline]
14. Kunst AE, Groenhof F, Mackenbach JP, Health EW. Occupational class and cause specific mortality in middle aged men in 11 European countries: comparison of population based studies. EU Working Group on Socioeconomic Inequalities in Health. BMJ 1998; 316: 16361642.
15. Codd MB, Buttimer J, Comber H et al. Mortality from breast cancer in Ireland prior to the introduction of population-based mammographic screening. Ir J Med Sci 1999; 168: 8792.[ISI][Medline]
16. Dolan SJ, Russell CF. Medullary thyroid carcinoma in Northern Ireland, 19671997. Ann R Coll Surg Engl 2000; 82: 156161.[ISI][Medline]
17. Nasca PC, Greenwald P, Burnett WS et al. Cancer among the foreign-born in New York State. Cancer 1981; 48: 23232328.[ISI][Medline]
18. Harding S. The incidence of cancers among second-generation Irish living in England and Wales. Br J Cancer 1998; 78: 958961.[ISI][Medline]
19. Abbotts J, Williams R, Smith GD. Mortality in men of Irish heritage in West Scotland. Public Health 1998; 112: 229232.[ISI][Medline]
20. National Cancer Registry Ireland. Cancer in Ireland, 1996. Incidence and Mortality. Cork, Republic of Ireland: National Cancer Registry Board 1999.
21. Northern Ireland Cancer Registry. Cancer Incidence Data 19931996 and Mortality Data 199399. Belfast, Ireland: Northern Ireland Cancer Registry 2000.
22. Smans M, Boyle P, Muir CS (eds). Cancer Mortality Atlas of EEC. IARC Scientific Publication No. 107. Lyon, France: IARC 1993.
23. Cuzick J, Boyle P. Trends in cervix cancer mortality. Cancer Surv 1988; 7: 417439.[ISI][Medline]
24. Levi F, Lucchini F, Boyle P et al. Cancer incidence and mortality in Europe, 19881992. J Epi Bio 1998; 3: 295373.
25. Peto R, Darby S, Deo H et al. Smoking, smoking cessation, and lung cancer in the UK since 1950: combination of national statistics with two casecontrol studies. BMJ 2000; 321: 323329.
26. Boyle P, Gandini S, Robertson C et al. and the International Smokers Survey Group. Characteristics of smokers attitudes towards stopping: survey of 10 295 smokers in representative samples from 17 European countries. Eur J Publ Health 2000; 10: 514.
27. Boyle P, La Vecchia C, Maisonneuve P et al. Cancer epidemiology and prevention. In Peckham M, Pinedo H, Veronesi U (eds): Oxford Textbook of Oncology. Oxford, UK: Oxford University Press 1995; 199273.
28. Boyle P. Prostate specific antigen (PSA) testing as screening for prostate cancer. The current controversy. Ann Oncol 1998; 9: 12631264.[CrossRef][ISI][Medline]
29. Bartsch G, Horninger W, Klocker H et al. on behalf of the Tyrol Prostate Cancer Screening Group. Decrease in prostate cancer mortality following introduction of prostate specific antigen (PSA) screening in the Federal State of Tyrol, Austria. Urology 2001; 58: 417424.[CrossRef][ISI][Medline]
30. Boyle P, Kaye SB, Robertson AG. Testicular cancer in Scotland. Eur J Cancer Clin Oncol 1987; 23: 827830.[ISI][Medline]
31. Boyle P, Maisonneuve P, Kaye SB. Testicular cancer in central Europe. Lancet 1990; 335: 1033.
32. Plesko I, Ondrus D, Boyle P. Testicular cancer incidence and mortality in Slovakia, 19681990. Lancet 1996; 347: 900901.[ISI][Medline]
33. Becker N, Boyle P. Decline in mortality from testicular cancer in West Germany after re-unification. Lancet 1997; 350: 744.[ISI][Medline]
34. Boyle P, Soukop M, Scully C et al. Improving prognosis of Hodgkins disease in Scotland. Eur J Cancer Clin Oncol 1988; 24: 229234.[CrossRef][ISI][Medline]
35. La Vecchia C, Levi F, Lucchini F et al. Hodgkins disease mortality in Europe. Br J Cancer 1991; 64: 723734.[ISI][Medline]