Elderly suicide rates: a replication of cross-national comparisons and association with sex and elderly age-bands using five year suicide data

Abstract: Background: Suicide rates generally increase with age. A recent cross-national study, using one-year cross-sectional data on suicide rates identified regional and cross-national patterns for elderly suicide rates. However, findings from studies using one-year cross-sectional data on suicide rates may be open to bias because of random year on year fluctuations in elderly suicide rates. Methods: One-year average of suicide rates for both sexes in the age-bands 65-74 and 75+ years were calculated from data on suicide rates for five consecutive years ascertained from the World Health Organization. Cross-national variations were examined by segregating different countries into four quartiles of suicide rates. Results: There was wide cross-national variation in elderly suicide rates. Elderly suicide rates were the lowest in Caribbean and Arabic/Islamic countries, and the highest in central and eastern European, countries emerging from the former Soviet Union, some oriental and some west European countries. Conclusions: The regional and cross-national patterns for elderly suicide rates observed in this study were almost identical to a similar earlier study using one-year cross-sectional data on suicide rates. This suggests that the findings of both studies were accurate and robust, and potential explanations for the observed cross-national variations in elderly suicide rates requires further study.


Introduction
n developing and developed countries the proportion of the elderly in the general population is increasing.P 1 P Suicide rates generally increase with age.P 2,3 P Suicides are emotive, distressing to family and professionals, lead to loss of economic productivity (although less likely in older people) and can lead to litigation. Thus, suicides in the elderly are an important public health concern. Comprehensive understanding of the substantial worldwide varia-tion in population patterns of suicide may be critical in the understanding of potential distil risk and protective factors and for developing prevention programmes.P 4 P Cross-national comparisons of elderly suicide rates can allow generation of potentially testable hypotheses about etiological factors.P 5 P Cross-national elderly suicide rates have been compared.P 2,5-8 Another recent cross-national study, using only one-year cross-sectional data on suicide rates, reported that there is wide cross-national variation in elderly suicide rates and elderly suicide rates were the lowest in Caribbean, central American and Arabic countries, and the highest in central and eastern European, some oriental and some west European countries.P 5 P However, the findings of the latter study, using only one-year crosssectional data on suicide rates, may have erroneous because suicide rates can randomly fluctuate year on year.P 9 P Therefore, by using a one-year average of data on suicide rates for five consecutive years, the following facets were examined in a cross-national study using data more recent than those used in the earlier study:P 5 P (i) comparison of elderly suicide rates across different countries; (ii) the comparison of suicide rates between elderly males and females; and, (iii) the comparison of suicide rates between the "old-old" (75+ years) and the "young-old" (65-74 years) age-bands.

Data collection
Data on suicide rates for males and females in the age-bands 65-74 and 75+ years was ascertained manually by the author for ICD 9 code of E54 and ICD-10 code of X60-X84 from the WHO website (3 T U http://www.who.int/whosis/U 3 T U database/mort/table1.cfmU ). For a small number of countries only the raw figures for the number of suicides were available from the WHO website. Suicide rates for these countries were calculated by dividing the number of reported suicides by the population size in the relevant age-band and sex group available on the same website. Data were ascertained for the latest five consecutive years. The one-year average suicide rate was calculated by dividing the sum of suicide rates for the latest five consecutive years by five. The median (range) for the latest year for the suicide rate data was 2005 (1983-2007); the total number of countries with this data was 97.

Data analysis
Cross-national comparisons of elderly suicide rates were examined descriptively by segregating all the countries into the four quartiles of elderly suicide rates to observe any regional trends; these analysis were conducted for males and females in both the age-bands. Wilcoxon's matched-pair signed-rank test was used to compare suicide rates between elderly males and females and between the age-bands 65-74 years and 75+ years across the different countries. SPSS version 16 was used for data analysis.

Results
Data on elderly suicide rates were available for 97 countries from the WHO website (names of these countries are listed in Tables 2-5). The median (range) suicide rate (per 100,000 population in the relevant group) for males aged 65-74 years, males aged 75+ years, females aged 65-74 years and females aged 75+ years was 19.6 (0-93.98), 29.3 (0-161.42), 4.6 (0-32.54) and 5.06 (0-71.36) respectively. Table 1 illustrates the cut-off suicide rates for each of the four quartiles for males and females and for agebands 65-74 years and 75+ years. Tables 2-5 illustrate the different countries in the four quartiles for males and females by age-bands 65-74 years and 75+ years.

Comparison of elderly suicide rates across different countries
Suicide rates in males aged 65-74 years were the lowest in Caribbean (Anguilla, Antigua, Aruba, Bahamas, Bermuda, British Virgin Islands, Dominica, Jamaica, Montserrat, St Vincent and Grenadines Islands, and Turks and Caicos Islands), central American (Belize, Ecuador and Guatemala) and Arabic and Islamic countries (Bahrain, Kuwait and Brunei) countries. Suicide rates in males aged 65-74 years were the highest in central and east European and countries emerging from the former Soviet Union (Belarus, Bulgaria, Croatia, Estonia, Georgia, Hungary, Kazakhstan, Latvia, Lithuania, Moldavia, Russia, Slovenia and Ukraine), some oriental (South Korea and Japan), and some west European (Austria, Belgium, France, Finland, Luxembourg and Switzerland) countries (Table 2). A similar pattern, with some variations, was also observed for suicide rates in males aged 75+ years, females aged 65-74 years and females aged 75+ years (Tables 3-5).

Comparison of suicide rates in males and females across countries
Using the Wilcoxon's matched-pair signed-rank test, suicide rates across different countries were significantly

Comparison of suicide rates between the two agebands
Using the Wilcoxon's matched-pair signed-rank test, suicide rates across different countries were significantly higher in the age-band 75+ years than in the age-band 65-74 years in males (Z=-5.95, P<0.00001) and females (Z = -4.88, P<0.00001).

Discussion
Some methodological issues need consideration. Data on suicide rates in cross-national studies should be viewed cautiously because: data were not available from all countries;P 10,11 P the validity of the data was unclear;P 6,11 P the legal criteria for the proof of suicide vary between countries and different regions within a country;P 7,11 P some countries have poor death registration facilities;P 7 P and, cultural and religious factors and stigma attached to suicide, particularly in some Arabic and Asian countries, may lead to under-reporting of suicides.P 11,12 P Additionally, some suicides may be misclassified as deaths due to undetermined causes,P 13 P but data on undetermined deaths were not available for all countries. Some of these methodological difficulties were reduced by using a one-year average of five consecutive year data on elderly suicide rates. Moreover, the latest and best available data set from the WHO were used.
The regional and cross-national patterns in elderly suicide rates observed in the current study were similar to those reported previously.P 2,5  Turks & Caicos Venezuela Trinidad Ukraine Turkmenistan regional and cross-national patterns in elderly suicide rates were almost identical to a recent study using similar methodology.P 5 P However, there were four important methodological differences between the current and the latter study. First, one-year average of five consecutive years data on elderly suicide were used in the current study, whereas one-year cross-sectional data on suicide rates were used in the previous study. This would serve to reduce any bias due to year on year random fluctuations in elderly suicide rates. Second, in the current study countries only reporting raw numbers of elderly suicides were also included and suicide rates were specifically calculated for those countries, whereas such countries were excluded in the previous study. Third, consequently, the sample size in the current study was significantly larger. Finally, the current study used much more recent data than the earlier study and thus also provides data on the current status of regional and cross-national patterns of elderly suicide rates. Moreover, the observations that suicide rates are higher in males than females for both the elderly agebands and higher in the 75+ years age-band than 65-74 years age-band in males and females were also consistent with previous studies.P 2,5 The findings of the regional and cross-national patterns, the gender differences and age differences in elderly suicide rates in the current study and the previous study were almost identical despite the methodological improvements in the current study, and this suggests that the findings of both studies were accurate and robust. Therefore, the observed and persistent regional and crossnational patterns in elderly suicide rates requires further consideration because there may be several potential explanations including regional and cross-national variations in: genetic factors predisposing to mental illness and suicidal behaviors; the prevalence of mental illness (especially depression); life expectancy; socioeconomic deprivation and income inequality; social fragmentation; cultural and religious factors; availability of appropriate mental healthcare services; and presence of public health policy initiatives to reduce suicide rates.P 3,5,14 P Some studies addressing these issues have emerged, including socioeconomic status, education, population growth, service provision and funding, elderly dependency ratios, household size and ageing,P 3,14-21 P and others are underway. 5 T Funding:5 T None 5 T Competing interests5 T : None declared 5 T Ethical approval:5 T There is no need for the Ethical approval