Abstract
Objectives. We assessed life expectancy increases in the past several decades in South Korea by age and specific causes of death.
Methods. We applied Arriaga's decomposition method to life table data (1970–2005) and mortality statistics (1983–2005) to estimate age- and cause-specific contributions to changes in life expectancy.
Results. Reductions in infant mortality made the largest age-group contribution to the life expectancy increase. Reductions in cardiovascular diseases (particularly stroke and hypertensive diseases) contributed most to longer life expectancy between 1983 and 2005 (30% in males and 28% in females). Lower rates of stomach cancer, liver disease, tuberculosis, and external-cause mortality accounted for 30% of the male and 20% of the female increase in longevity. However, higher mortality from ischemic heart disease, lung and bronchial cancer, colorectal cancer, breast cancer, diabetes, and suicide offset gains by 10% in both genders.
Conclusions. Rapid increases in life expectancy in South Korea were mostly achieved by reductions in infant mortality and in diseases related to infections and blood pressure.
In the past few decades, life expectancy at birth has increased remarkably in South Korea (hereafter Korea). Koreans had an average life expectancy at birth of 52.4 years in 1960, 16 years below the average among member countries of the Organization for Economic Cooperation and Development; it increased to 62.2 years in 1970. In 2005, it reached the average among member countries of 78.5 years (Figure 1), the greatest gain of any country in the coalition and a rise of more than 6 months per year from 1960 to 2005.1 This rapid increase in life expectancy was observed in both men and women, but gender differences in life expectancy varied over time: the gap increased until 1985 and decreased thereafter.
FIGURE 1 Life expectancy at birth among men and women in South Korea compared with Organization for Economic Cooperation and Development (OECD) averages: South Korea, 1970–2005.
The improvement in life expectancy was accompanied by rapid economic growth beginning in the 1960s, with gross national income per capita increasing from less than US $100 in 1960 to US $20 045 in 2007.2 This rapid economic growth and improvement in population health followed the devastation of the national infrastructure during Japanese colonial occupation and World War II (1910 through 1945) and the Korean War (1950 through 1953). Although the steady increase in life expectancy was not affected by 2 recent economic recessions, in the early 1980s and late 1990s,3 the nation's industrialization and the subsequent improvements in living standards, nutrition, and health care have often been cited as major contributors to Koreans' remarkable improvements in health.4
We decomposed Korea's life expectancy gain in the past decades by age and specific causes of death for males and females. Quantification of cause-specific contributions to longer life expectancies in Korea may provide insights into the causes of life expectancy increases in other Asian countries (e.g., China, Japan, and Taiwan), where causes of death and epidemiological patterns of such major risk factors as cigarette smoking and diet changes are similar to those in Korea.5–8
We obtained all available life table (1970 to 2005) and annual mortality data (1983 to 2005) from the Korea National Statistical Office (NSO). We calculated annual life expectancy at birth for males and females between 1970 and 2005 with standard demographic techniques.9 Mortality data included the number of deaths by cause, gender, and 5-year age group (except ages younger than 1 year, 1–4 years, and 80 years or older) for each year from 1983–2005. Causes of death were coded according to the International Classification of Diseases, 10th Revision (ICD-10) for all years.10 The Korean NSO used the conversion table published by the World Health Organization for causes initially coded by ICD-911 criteria.12 We analyzed 14 disease group–specific causes of death (Figure 2). We also analyzed several important single diseases in each disease group.
FIGURE 2 Disease group–specific contributions to life expectancy increase among men and women: South Korea, 1983–2005.
Note. All disease-group codes are from the International Classification of Diseases, 10th Revision (ICD-10). Each bar indicates the net contribution of mortality changes in each cause group to the total increase of life expectancy between 1983 and 2005, in years.
We estimated the contributions of specific ages and causes of death to the change in life expectancy by Arriaga's decomposition method.13,14 This method estimates the contribution of mortality change in each age and cause to the total change in life expectancy at birth for a given period. Because life expectancy at birth is a summary of age-specific mortality rates, the total change in life expectancy at birth between 2 time points is a function of changes in mortality rates in each age for the given period.13 The total change in life expectancy is thus the sum of the number of years added to (positive contribution) or removed from (negative contribution) life expectancy at birth attributable to the decrease or increase, respectively, of mortality at each age for a given period.
The total number of years a particular age group contributes to the change in life expectancy for a given period can also be decomposed by causes of death.14 As in the decomposition by age, reductions in mortality from a specific cause in a given age will make positive contributions to the total change in life expectancy at birth, and causes with increased mortality will make negative contributions. Thus the sum of the cause-specific contributions is equal to the total contribution for each age, and the total change in life expectancy is the sum of the age- and cause-specific components.
Figure 3 shows age-specific contributions in years to life expectancy increases among males and females. All age groups contributed positively to life expectancy increases between 1970 and 2005.
FIGURE 3 Age group–specific contributions to life expectancy increase among men and women: South Korea, 1970–2005.
Of the total longevity increase of 16.4 years in males and 16.3 years in females, changes in infant mortality made the largest single contribution in both genders, accounting for 2.7 years (17%) in male and 3 years (18%) in female life expectancy. Decreases in mortality among middle-aged and older adults also made large contributions, accounting for 5.3 years (32%) of the increase among men aged 50 to 69 years and 5.5 years (33%) of the increase among women aged 50 and older. Age-group effects were greater in men than in women aged 40 to 69 years, but greater in women than in men aged 70 years or older.
When we estimated age-specific contributions for multiple shorter periods, the contribution by infants was greater in earlier years and decreased substantially in recent years. For example, contributions by infants accounted for 24% of the male life expectancy increase between 1970 and 1983, 17% between 1983 and 1994, and 5% between 1994 and 2005. The corresponding figures for females were 20%, 22%, and 6%. On the other hand, contributions by older age groups (50 years or older) were greater in recent years. The contribution of men aged 50 years or older increased from 28% in 1970–1983 to 53% in 1994–2005. For women, the corresponding percentages were 19% and 75%.
Figure 2 presents contributions to the total life expectancy increase (11.9 years for males and 10.4 years for females) between 1983 and 2005 by disease group. For males, 65% of the total increase resulted from reductions in cardiovascular disease (CVD; 3.6 years; 30%), digestive system diseases (1.4 years; 12%), infectious and parasitic diseases (0.9 years; 8%), respiratory system diseases (0.8 years; 7%), and external causes (0.9 years; 8%). Similarly, 52% of the total female life expectancy increase was from CVD (2.9 years; 28%), respiratory system diseases (0.9 years; 9%), digestive system diseases (0.8 years; 8%), and infectious and parasitic diseases (0.7 years; 7%).
Ill-defined causes (ICD-10 R00–R99) contributed 30% (3.6 years) of male and 44% (4.6 years) of female total increases in life expectancy in 1983–2005. The elderly accounted for the majority of ill-defined causes of mortality: 68% and 75% of ill-defined causes were assigned to men and women, respectively, who were aged 60 years or older.
Between 1983 and 1994, ill-defined causes of mortality contributed 50% to increased longevity for men and 63% for women. This decreased to 5% for men and 12% for women in 1994–2005 (Table A, available as a supplement to the online version of this article at http://www.ajph.org). This substantial decrease in deaths with ill-defined causes stems from a steady increase in death certification by physicians and a corresponding improvement in the quality of mortality data in Korea.15 Despite the much greater contribution of ill-defined causes in earlier years, the general patterns of cause-specific contributions to life expectancy increase remained similar over time.
Contributions by subgroups of CVD across age groups are shown in Figure 4. The 3.6-year increase in male life expectancy between 1983 and 2005 attributable to CVD largely reflected lowered mortality from hypertensive disease (ICD-10 I10–I13; 1.4 years; 38%), stroke (ICD-10 I60–I69; 1.2 years; 33%), and other heart diseases (1.2 years; 33%). Most of this positive contribution occurred in men aged 45 to 74 years. However, increases in mortality from ischemic heart disease (IHD; ICD-10 I20–I25) lowered life expectancy (−0.3 years; −8%) between 1983 and 2005. In women, contributions of specific CVD subgroups had similar patterns. Overall, a 3-year increase in life expectancy was attributable to CVD subgroups in the study period: hypertensive disease (1.0 year; 33%), stroke (1.0 year; 33%), and other heart disease (1.2 years; 40%) caused fewer deaths, and IHD (−0.3 years; −10%) caused more.
FIGURE 4 Age group–specific and disease group–specific contributions to life expectancy change caused by cardiovascular disease (CVD) among (a) men and (b) women: South Korea, 1983–2005.
Note. All disease-group codes are from the International Classification of Diseases, 10th Revision (ICD-10).
Although cancer (all types) contributed only 0.3 years (3%) and 0.2 years (2%) total to increased life expectancy in males and females, respectively (Figure 2), we found large variations across types of cancer. As shown in Figure 5, in males, reductions in stomach cancer (ICD-10 C16; 0.6 years; 5% of the total increase in life expectancy) and liver cancer (ICD-10 C22; 0.2 years; 1.6%) contributed most. These positive contributions were offset by rises in colorectal cancer (ICD-10 C18–C21; −0.1 years; −1%) and lung and bronchial cancer (ICD-10 C33–C34; −0.2 years; −2%).
FIGURE 5 Age group–specific and disease group–specific contributions to life expectancy change caused by cancer among (a) men and (b) women: South Korea, 1983–2005.
Note. All disease-group codes are from the International Classification of Diseases, 10th Revision (ICD-10).
We observed similar patterns among females. Mortality decreases in stomach cancer made the largest positive contribution (0.5 years; 5%) followed by cervical and uterine cancer (ICD-10 C53–C55; 0.1 years; 1%) and liver cancer (ICD-10 C22; 0.1 years; 1%). However, mortality increased in colorectal cancer (ICD-10 C18–C21; −0.1 years; −1%) and lung and bronchial cancer (ICD-10 C33–C34; −0.1 years; −1%). Breast cancer (ICD-10 C50) mortality also decreased longevity (−0.06 years; −0.6%).
Fewer deaths from respiratory system diseases accounted for 7% of the male and 9% of the female life expectancy increase (Figure 2), with pneumonia (ICD-10 J12–J18) the most important single cause in both males (0.51 years; 64%) and females (0.56 years; 60%). Infants and children younger than 5 years were the most affected by lower pneumonia mortality rates (73% of the longevity increase in males and 79% in females occurred in these age groups). Fewer deaths from tuberculosis (ICD-10 A15–A19) contributed to the greatest increase in longevity of the infectious and parasitic diseases, accounting for 72% (0.6 years) in males and 57% (0.4 years) in females. Among diseases of the digestive system, mortality decreases in liver disease (ICD-10 K70–K76) made the largest contribution in both males (64%) and females (44%).
Changes in mortality from endocrine, nutritional, and metabolic diseases; perinatal conditions; and mental and behavioral disorders (males only) negatively affected life expectancy (Figure 2). Diabetes (ICD-10 E10–E14) accounted for −0.15 years (−1%) in males and −0.24 years (−2%) in females.
Mortality changes attributable to external causes also had different patterns across specific causes (Figure A, available as a supplement to the online version of this article at http://www.ajph.org). The 0.9-year increase in male life expectancy from combined external causes was largely explained by a decrease in drowning (ICD-10 W65–W74; 0.26 years) and poisoning (ICD-10 X40–X49; 0.28 years). However, mortality from suicide (ICD-10 X60–X84) decreased longevity between 1983 and 2005 (−0.2 years; 22% of the total negative contribution from external causes of death) among men, mostly among men aged 30 years or older.
The effect of transport injuries (ICD-10 V01–V99) on longevity changed over the study period. They decreased life expectancy between 1983 and 1994 (in males, −0.66 years; −11%; in females, −0.32 years; −6%) but increased life expectancy in 1994–2005 (in males, 0.85 years; 14%; in females, 0.39 years; 8%; Table A). Mortality from this type of injury also varied with age: persons younger than 45 years (except adolescents aged 15–19 years) accounted for a 0.13-year increase in life expectancy, offset by increased mortality among persons aged 45 years or older (−0.08 years). This age-specific pattern did not vary over time (results not shown). The overall contribution from transport injuries was an increase of 0.5 years in life expectancy across all ages. Among females, mortality changes from specific external causes were similar to, but smaller than, those of males.
Our finding that reductions in infant mortality represented the largest single age-group contribution to increases in life expectancy in Korea is consistent with results from other countries that experienced rapid improvement in life expectancy.16,17 Infant mortality from pneumonia, infectious and parasitic diseases, congenital malformations, and nervous system diseases dramatically decreased in 1983–2005. Steady declines in fertility since the 1970s18 and comprehensive universal health insurance, established by 1989 (universal coverage increased from 15% in 197716), may have contributed to the remarkable reduction in infant mortality.20
We found that the largest contribution to the increase in life expectancy was from CVD, accounting for approximately 30% of the total improvement for both genders. These findings are similar to those from other countries, including Taiwan,7 Italy,21 Germany,22 the United States,23 and Japan.24 Our detailed analysis showed that 72% of the CVD contribution for males and 68% for females was attributable to mortality reduction from hypertensive disease and stroke. Age-specific contributions showed similar patterns for hypertensive disease and stroke, suggesting common risk factors for both causes.
Approximately 70% of deaths from hypertensive disease (ICD-10 I10–I13) were attributable to essential hypertension (ICD-10 I10), according to our analysis of detailed cause-of-death data. Essential hypertension itself cannot directly lead to death absent complications such as stroke or IHD, and this classification may reflect premorbid detection of hypertension, which could be an indicator of severity. Previous radiological studies showed that hemorrhagic stroke is more prevalent than ischemic stroke in Korea.25 Hypertension is more strongly related to hemorrhagic stroke than to ischemic stroke in Korean data,26 and ischemic stroke shares common risk factors and temporal trends with IHD.26,27 Prevalence of hypertension has decreased among Korean adults, especially in men,28 and hypertension treatment rates have increased.29 Improvements in stroke and hypertension treatments and expanded access to such medical care provided by national health insurance presumably contributed to the mortality reductions from stroke and hypertensive disease.
Early life circumstances have been found to be of etiological significance for both hypertension and hemorrhagic stroke.30,31 Song et al.32 and Song and Sung33 showed that adult height, a marker of early life exposures, increased with birth year among Korean men and women and that it was strongly inversely associated with stroke. In men, this association was greater for hemorrhagic than ischemic stroke.32 Early life circumstances are also important for stomach cancer and liver diseases because they can originate in early childhood Helicobacter pylori infections34 and mother-to-infant transmission of hepatitis B,35 respectively. Declining levels of H. pylori infection36 and viral hepatitis B infection37 may have played a major role in mortality reduction from stomach cancer, liver cancer, and liver disease (mainly liver cirrhosis in Korea). Evidence has also been found of birth cohort effects in stomach cancer mortality38 and H. pylori infection rate36; mortality rate decreases are greatest among individuals born since the 1920s. This parallels the large positive contributions observed in our study and suggests that deaths from these diseases will continue to decline.
Institutionalization of public hygiene in Korea in the late 19th and early 20th centuries39 may have led to reductions in infectious and parasitic diseases. Successive improvements in clean water supply after the Korean War (1950–1953)40 contributed to the reduction of waterborne communicable diseases in the short term and of chronic diseases with infectious origins in the long term. Saemaul Undong (the New Village Movement), a government-led initiative, was launched in 1970 to modernize rural areas by building infrastructure and improving housing conditions. Korea is now the most urbanized country in Asia, with 81% of the total population living in cities in 2005, an increase from 21% in 1950 and 41% in 1970.41 This rapid urbanization may have facilitated better access to health care facilities as well as reduced exposure to waterborne infections.
Deaths from external causes had more immediate associations with industrialization, but the contributions to mortality reduction varied across specific causes and years. Mortality from transport injuries increased until the mid-1990s, possibly because of rapid industrialization. A previous study showed that gasoline prices and transport volume associated with an economic recession as well as a rigorous nationwide campaign explained the decrease in transport accident mortality after the late 1990s.42 Steady upward trends in suicide, with a rapid increase after the economic crisis in the late 1990s,42 may be the principal cost of social disruption associated with industrialization. In Korea, the most important agents associated with deaths by poisoning were herbicides and pesticides, which were mostly available to agricultural workers.43 The reduction in poisoning might be the consequence of agricultural and fishery workforce reduction, which plummeted from 25% of the total workforce in 1985 to 7.5% in 2005 because of urbanization and associated industrial restructuring.44
The substantial lifestyle changes brought about by rapid urbanization and industrialization could also help explain increased mortality from IHD; lung, colorectal, and breast cancer; and diabetes. These causes jointly reduced longevity by 0.8 years (−6%) in males and 1.6 (−15%) years in females. Although IHD mortality has markedly increased in Korea during the past 2 decades,45 the absolute mortality rate is still lower than in Western countries,46 which may be partly explained by continued adherence to the traditional Korean diet, which is high in vegetables and low in fat.47
Smoking rates rose before World War II in the United States48 and the United Kingdom,49 much earlier than in Asian countries, including Korea, whose populations have a relatively smaller risk of mortality from smoking-related lung cancer.50–53 Increases in consumption and mass production of cigarettes occurred after the 1970s,54 and the age of smoking initiation is in the 20s among men and 30s or 40s among women in Korea,55 much older than in Western countries. The later advent of widespread cigarette smoking and the older age of initiation seem to provide a plausible explanation for the lower lung cancer risk in Asian countries.56 Despite the rapid decrease in smoking among Korean men over the past decade, prevalence is still high (approximately 50%).57 In addition, smoking rates are stable or increasing among young Korean women. Previous studies found increasing trends in fat intake,47 obesity and diabetes,58 lung cancer,59 and breast cancer60 among recent Korean birth cohorts. Therefore, preventing deaths from smoking- and diet-related causes will be 1 of the major challenges for public health efforts to further extend life expectancy in Korea.
In Korea, it is mandatory to report all deaths to the NSO within 1 month. Death registration reached 100% in the mid-1980s, increasing from 93% for men and 70% for women in 1971–1975.61 However, questions have been raised about infant mortality data because neonatal deaths are less likely to be reported to the NSO when a baby dies before the birth has been reported by a family member.62 The NSO therefore estimates life table data from adjusted infant mortality rates supplemented by periodic infant mortality surveys carried out by the government and supplemented by national health insurance and crematorium data.63 These efforts to enhance the completeness of infant mortality statistics led to a sharp increase in recorded perinatal infant mortality (ICD-10 P00–P96) in the late 1990s, which we observed as a negative contribution to life expectancy. The cause-specific contributions we found for other ages were valid because they were derived from complete data. We also used the adjusted infant mortality rates to estimate age-specific contributions.
Death certification by physicians has steadily increased in Korea; more than 90% of deaths among persons younger than 60 years were certified by physicians in 2002.15 However, the rate of death certification by physicians decreased with age among persons of both genders who were aged 60 years or older, which might explain why senility (ICD-10 R54) accounted for most of the ill-defined causes of death, particularly in the 1980s. This might be responsible for the substantial contribution of ill-defined causes (ICD-10 R00–R99) to the life expectancy increase we found. The existence of vague causes such as senility in the 1980s could not explain the contributions to increased life expectancy from reduced mortality rates from several major causes that we observed, including from stroke and stomach cancer, but may have obscured some of the contribution of such diseases to the life expectancy increase. In addition, results for the major contributing causes to the increase in life expectancy were similar when we used better-quality data from the late 1990s and early years of the 21st century (e.g., 1997–2005).
Infant mortality decline was a major driver of longer life expectancy between 1970 and 2005. Decreases in CVD (stroke and hypertensive diseases in particular), liver disease, stomach cancer, tuberculosis, pneumonia, and external-cause mortality explained 60% of male life expectancy increase and 48% of female life expectancy increase between 1983 and 2005. In addition, increases in longevity stemmed from the same major causes for both genders.
Koreans exhibited patterns of health transition observed approximately 1 century earlier in Western countries (albeit more rapidly in Korea). For example, the pattern of cause-specific contributions to increases in life expectancy and their dependence on early life influences are reminiscent of mortality declines in the late 19th and early 20th centuries in Britain.64
Mortality from IHD, colorectal cancer, lung cancer, and diabetes is on the rise in Korea. The harmful health effects of climbing smoking rates after World War II and recent Westernization of diets, especially among younger Koreans, will likely push mortality from these diseases higher. Future public health efforts should aim to improve preventive measures and treatment for these diseases.
Acknowledgments
This study was supported in part by an Interdisciplinary Capacity Enhancement grant to Seungmi Yang from the Canadian Institute of Health Research (grant HOA-80072).
Human Participant Protection
No protocol approval was necessary because only publicly available secondary data were used for this study.
