Gains and Losses in Life Expectancy with Care Needs from the Elimination of Diseases in Japan

Background: Evaluating the impact of diseases on expected years of life with care needs is important in older populations. We examined gains or losses in expected years with care needs at age 65 years in Japan. Methods: We used Japanese national health statistics data, including care needs, based on long-term care insurance. The number of expected years with and without care needs at age 65 years, after the elimination of ve selected diseases in 2010, 2013, and 2016 were calculated using a proposed method. Results: Expected years without care needs at age 65 years increased from 17.15 years in 2010 to 17.82 years in 2016 for men and from 20.24 to 20.74 years for women; expected years with care needs were between 1.70 and 1.73 years for men and between 3.59 and 3.66 years for women. Elimination of malignant neoplasms and heart diseases led to gains in expected years with and without care needs. Elimination of cerebrovascular diseases, arthropathies, and dementia showed gains in expected years without care needs and losses in expected years with care needs. Gains and losses from elimination of dementia expanded each year but those from elimination of other diseases compressed. Conclusions: Our results suggest that the prevention of dementia, cerebrovascular diseases, and arthropathies lead to more expected years without care needs and fewer expected years with care needs. The impact of dementia might increase and that of other diseases might decrease each year.


Background
Life expectancy with and without disability or care needs is important for the assessment of health in older populations. 1 Disability-free life expectancy has been investigated in several countries. [2][3][4][5] In Japan, expected years with and without care needs have been calculated using long-term care insurance data, and their trends between 2005 and 2009 have previously been examined. 6 Reducing the expected number of years with care needs, together with increasing the expected years without care needs, is a primary goal in older populations. 7,8 In planning how to achieve these goals, evaluating the impact of disease on disability-free life expectancy is essential. As an indicator of disease impact, gains in disability-free life expectancy owing to the elimination of diseases has been reported in several countries, including Japan. 1,9−11 However the impact of diseases on expected years with and without disability or care needs differ between older and younger populations and change over time. 1 In the present study, gains or losses in expected years with and without care needs, at age 65 years, after the elimination of selected diseases, were calculated using a previously proposed method and using data from longterm care insurance in Japan. We then examined their trends during 2010-2016.

Long-term care insurance in Japan
The long-term care insurance system in Japan has been described elsewhere. 12 This system, organized by the Ministry of Health, Labour and Welfare, was started on April 1, 2000. Every Japanese person aged 65 years or more is eligible. The level of care needs is assessed based on each individual's physical and mental status. The care level determines the number of services covered by insurance.

Data
We used population, death, and life-table data of Japan in 2010, 2013, and 2016. 13,14 The number of individuals with care needs was obtained from the Survey of Long-Term Care Bene t Expenditures at the end of September in 2010, 2013, and 2016, which was based on all claims for long-term care bene t expenditures in Japan. 15 The proportion of primary causative diseases leading to care needs among persons with care needs living at home was obtained from a long-term care questionnaire of the Comprehensive Survey of Living Conditions in 2010, 2013, and 2016, which covered about 8,000 individuals with care needs in randomly selected households nationwide. 16 An additional pdf le shows the question about causes of care needs [see Additional le 1]. Data from the surveys were used with permission from the Ministry of Internal Affairs and Communications and the Ministry of Health, Labour and Welfare of Japan.
Because the data we used were national statistical data that was published or used with permission and these statistical data do not include personally identi able information, no ethical review was needed.

Disease status
We selected ve diseases: malignant neoplasms (International Classi cation of Diseases, Tenth Revision (ICD-10): C00-C97), heart diseases (ICD-10: I01-I02.0, I05-I09, I20-I25, I27, I30-I52), cerebrovascular diseases (ICD-10: I60-I69), dementia (ICD-10: F00-F03 and G30), and arthropathies (ICD-10: M00-M25). 13 The primary causative disease for individuals with care needs living at home was evaluated using responses to the question, "What is the primary causative disease that has led to your care needs?" 16 The response to the question was one of 12 diseases and injuries, including the ve abovementioned diseases, "other disease or injury", and "unknown." A response to any of the ve selected diseases was classi ed as having care needs primarily caused by these diseases.
Calculation of gains or losses in expected years with and without care needs after the elimination of diseases We calculated the expected years of life with and without care needs at age 65 years in 2010, 2013, and 2016, using the method in a previous study. 6 The care needs for individuals aged 65 years and older was evaluated using the care need levels certi ed by the long-term care insurance of Japan. 12 We classi ed level 2, 3, 4 and 5 as having care needs and the rest as having no care needs. We then calculated the sex-and age-speci c prevalence of care needs, according to the following age groups: 65-69, 70-74, 75-79, 80-84, 85-89, 90-94, and 95 years or older.
We calculated the expected years with and without care needs after elimination of each of the above ve diseases, using the proposed method. 17 According to a previous study, the method is described as follows. 18 The life table eliminating deaths caused by diseases was constructed using data of the number of deaths and life tables with no disease elimination. 14 The probability of survival at age group x with a disease eliminated (p x e ) was given by the probability with no disease elimination (p x ), the number of deaths (D x ) from all diseases and injuries, and the number of deaths from the disease (D x e ), as follows: where ln is a natural logarithm function and the age groups are as given above. Using the Chiang life-table method, the number of survivors (l x e ), the stationary population (L x e ), and the expected years of life (e x e ) after eliminating a disease was calculated from the values of p x e . 19 We assumed that the proportion of individuals with care needs primarily caused by disease among those with care needs living at home (obtained from the Comprehensive Survey of Living Conditions) was equal to that among all individuals with care needs, including residents in health care and welfare facilities for older people requiring long-term care, by sex, age and care needs level group. Under this assumption, the number of individuals with care needs primarily caused by disease was given to be the proportion among those with care needs living at home multiplied by the number of all individuals with care needs. We calculated the sex-and agespeci c prevalence of care needs after eliminating each disease. The numerator of prevalence was individuals with care needs, excluding those with care needs primarily caused by disease. The denominator was all individuals, excluding those with care needs primarily caused by disease.
Using the Sullivan method, we divided the expected years of life in age group x after eliminating disease according to those with and without care needs, as follows: e x e = Σπ y e L y e /l x e + Σ (1 − π y e ) L y e /l x e where Σ represents the sum from age group x to the oldest age group in the age group y, and π y e is the agespeci c prevalence of care needs after eliminating disease. 1,18,20 Results Table 1 shows the prevalence of patients with care needs and death rates in the older population in 2010, 2013, and 2016. The prevalence and death rate increased with age in both men and women in all three years. The sexand age-speci c prevalence and death rates decreased each subsequent year, except in the age group 95 years or more.  Table 2 shows the proportion of patients with care needs and the proportion of deaths owing to the selected diseases in the older population in 2010, 2013, and 2016. The proportion of deaths from malignant neoplasms, heart diseases, and cerebrovascular diseases was relatively high and decreased each year. The proportion of patients with care needs owing to cerebrovascular diseases and the proportion of female patients with care needs owing to arthropathies were relatively high and decreased each year. The proportion of patients with care needs owing to dementia was relatively high and increased with subsequent years.  Table 3 shows

Discussion
We found that the elimination of dementia, cerebrovascular diseases, and arthropathies led to some gains in expected years without care needs and some losses in the expected years with care needs at age 65 years, and that the elimination of malignant neoplasms and heart diseases led to gains in both expected years without and with care needs among older populations in Japan. Although the elimination of these diseases is hypothetical and not realistic, our ndings demonstrate the burden of these diseases on expected years with and without care needs. 1,18 Our ndings imply that the prevention of dementia, cerebrovascular diseases, and arthropathies is important in planning measures for reducing the number of expected years with care needs among older people. 7 We revealed that the gains in expected years without care needs and losses in expected years with care needs from the elimination of dementia expanded with each subsequent year during 2010-2016 in Japan, and those from the elimination of cerebrovascular diseases and arthropathies compressed with each year. These results suggest that the impact of dementia increases and the impact of other diseases decreases over time, which can serve as a target of measures aiming to reduce the expected years with care needs.
The results observed in the present study were based on the prevalence of care needs, as well as death rates. 1,18 As shown in Table 2, dementia, cerebrovascular diseases, and arthropathies led to a high proportion of patients with care needs and malignant neoplasms and heart diseases led to a lower proportion. The proportion of patients with care needs in dementia increased with each year and that in cerebrovascular diseases and arthropathies decreased. It has been reported that dementia, cerebrovascular diseases, and arthropathies are major causes of care needs and the impact of dementia increases each year in Japan. 16,21−23 The present study provided no information regarding the reasons for these trends.
Several previous studies have reported the effects of elimination of several diseases on the expected number of years with disability among older populations. 1,8,9 In Australia in 1993, the elimination of mental disorders and musculoskeletal disorders led to some losses in the expected years with disability at age 65 years, and the elimination of malignant neoplasms and cardiovascular disease in men led to some gains. 1 In the Netherlands during 1982-1991, elimination of arthritis/back complaints led to losses in the expected years with disability at age 65 years, and the elimination of cancer and heart disease led to some gains. 8 In Brazil, the elimination of malignant neoplasms, heart diseases, and cerebrovascular diseases showed gains in the expected years with disability at age 60 years among men and losses in women in 2000 and led to losses in both sexes in 2010. 9,24 Those results are uncommon. One reason for those ndings could be that the extent of death and disability from disease varies widely by population and year.
Gains in the expected years with and without activity limitation at birth from the elimination of several diseases in 2007 in Japan have been calculated. 18 That study reported that the elimination of dementia and cerebrovascular diseases led to gains in expected years without activity limitation and losses in expected years with activity limitation. Although activity limitation in that study involved much milder disability than the care needs investigated in the present study, the results were similar to ours. 6,16,18 That previous study also reported that the elimination of shoulder lesions and lower back pain led to large losses in expected years with activity limitation. We did not target shoulder lesions and low back pain because these diseases are not primary causes of either death or nursing care needs. 16 There are some limitations in the present study. We calculated the expected years with and without care needs by the elimination of diseases, using a previously proposed method. 17,18,20 Although it is assumed that the agespeci c prevalence of care needs in the stationary population is equivalent to that observed in the real population, the method has been applied in many studies. 2-6,9−11 We used long-term care insurance data of Japan and classi ed level 2 of care or greater as having care needs. 12 The data source and classi cation were the same as in a previous study. 6 The coverage of care needs in the insurance system would be su ciently high and stable. 6,12 We selected the ve diseases, including malignant neoplasms, heart diseases, and cerebrovascular diseases, which are leading causes of death in Japan. 13 Dementia, cerebrovascular diseases, and arthropathies are primary causes of nursing care needs in Japan. 16 However, targeting other diseases would also be important.
The underlying cause of death was used in the present study. If deaths indirectly caused by a disease were not considered, the total effects of the disease on death rates would be underestimated. 13,25 The underestimation of the malignant neoplasms, heart diseases, and cerebrovascular diseases selected in the present study would be relatively small, unlike that of hypertension and diabetes. 26 A primary causative disease leading to care needs was used. 16 Similar to use of the underlying cause of death, this would result in underestimating the total effects of diseases on the prevalence of care needs. Nevertheless, the problem of using an underlying cause of death and a primary causative disease of care needs is common in many studies. 1,[9][10][11]18 We assumed that the proportion of individuals with care needs primarily caused by disease among those with care needs living at home was equal to the proportion among all individuals with care needs, including residents in health care and welfare facilities for older people requiring long-term care, by sex, age and care needs level group. In the present study, we used data on primary causative disease leading to care needs among individuals with care needs living at home obtained from the Comprehensive Survey of Living Conditions. 16 Data for residents in health care and welfare facilities for older people requiring long-term care were unavailable. Many people with care needs were living at home, and about 23% were resident in these facilities during 2016 in Japan. 15,27 It is important to examine the above assumption.

Conclusions
In conclusion, our results suggest that the prevention of dementia, cerebrovascular diseases, and arthropathies lead to more expected years without care needs and fewer expected years with care needs. The impact of dementia might increase and that of other diseases might decrease each year.
Abbreviations ICD-10: International Classi cation of Diseases, Tenth Revision Declarations