Association of Trajectories of Higher-Level Functional Capacity with Mortality and Medical and Long-Term Care Costs Among Community-Dwelling Older Japanese

Association of Trajectories of Higher-Level Functional Capacity with Mortality and Medical and... Abstract Background Higher-level functional capacity is crucial component for independent living in later life. We used repeated-measures analysis to identify aging trajectories in higher-level functional capacity. We then determined whether these trajectories were associated with all-cause mortality and examined differences in medical and long-term care costs between trajectories among community-dwelling older Japanese. Methods 2,675 adults aged 65–90 years participated in annual geriatric health assessments and biennial health monitoring surveys during the period from October 2001 through August 2011. The average number of follow-up assessments was 4.0, and the total number of observations was 10,609. Higher-level functional capacity, which correspond to the fourth and fifth sublevels of Lawton’s hierarchical model, was assessed with the Tokyo Metropolitan Institute of Gerontology-Index of Competence (TMIG-IC). Results We identified four distinct trajectory patterns (high-stable, late-onset decreasing, early-onset decreasing, and low-decreasing) on the TMIG-IC through age 65–90 years. As compared with the high-stable trajectory group, participants in the late-onset decreasing, early-onset decreasing, and low-decreasing TMIG-IC trajectory groups had adjusted hazard ratios for mortality of 1.22 (95% confidence interval: 1.01–1.47), 1.90 (1.53–2.36), and 2.87 (2.14–3.84), respectively. Participants with high-stable and late-onset decreasing higher-level functional capacity trajectories had lower mean monthly medical costs and long-term care costs. In contrast, mean total costs were higher for those with low-decreasing trajectories, after excluding the large increase in such costs at the end of life. Conclusions People with a low-decreasing aging trajectory in higher-level functional capacity had higher risks of death and had high monthly total costs. Higher level functional capacity, Trajectories, Mortality, Medical costs, Long-term care costs The World Health Organization Scientific Group on the Epidemiology of Aging proposed that autonomy or independence in functioning be used as a health index for the elderly (1,2). Functional capacity has been recognized as a hierarchical framework. Lawton (3) defined and systemized seven intercorrelated “sublevels” of competence, namely (in ascending order of complexity), life maintenance, functional health, perception and cognition, physical self-maintenance, instrumental self-maintenance, effectance, and social role. Recent studies of functional capacity (4,5) and disability (6,7) in Western countries analyzed activities of daily living (ADL) and instrumental ADL (IADL), which correspond to the fourth and fifth sublevels of Lawton’s hierarchical model. However, because higher-level functional capacity predicts lower-level function (8), maintenance of such capacity is a crucial component for independent living in later life. The higher sublevels of Lawton’s hierarchical model thus need to be assessed when examining the functional capacity of elders. The Tokyo Metropolitan Institute of Gerontology–Index of Competence (TMIG-IC) (9) was developed and validated to measure the last three sublevels (instrumental self-maintenance, effectance, and social role) of Lawton’s model and has been used to assess higher-level functional capacity in previous studies (8,10–15). An analysis of a nationally representative sample of Japanese (16) found that most elders living in the community had good higher-level functional capacity and that this capacity tended to decline with age. However, to our knowledge no study has used multiple-repeated-measures data from a large-scale longitudinal study to investigate aging trajectories in higher-level functional capacity and identify the population with a lower aging trajectory pattern. Takata et al. (17) reported a longitudinal association between initial higher-level functional capacity assessed by TMIG-IC and all-cause, cardiovascular, and pneumonia deaths among community-dwelling adults aged 80 years. Lubitz et al. (18) used ADL and IADL to examine functional state at age 70 years and linked cumulative health care expenditures by using the Medicare Current Beneficiary Survey from age 70 years until death. Their findings suggest that elders in the lower aging trajectory of higher-level functional capacity have a higher all-cause and cause-specific mortality risk and greater medical and/or long-term care costs than do those with normal aging-related declines. However, existing data are insufficient to confirm this hypothesis. Japan quickly became a world leader in several health metrics, including longevity; however, the country faces challenges related to social health insurance and Long-Term Care Insurance (LTCI). Use of trajectories of higher-level functional capacity might help identify mortality risk and socioeconomic effects and yield new insights that advance health promotion and preventive care initiatives. This prospective study of community-dwelling older adults used repeated-measures data on TMIG-IC from an 11-year longitudinal study launched in Kusatsu Town, a rural community in Japan. We examined local registries to ascertain deaths from any cause and linked these data with Japanese national vital statistics. Then, we obtained data on medical expenses and care expenditures, which encompassed almost all medical provider fees and all care provider fees for the elderly. This study had three objectives: to identify aging trajectories in higher-level functional capacity of community-dwelling older Japanese, to determine whether these trajectories were associated with all-cause and cause-specific mortality, and to examine differences in medical and long-term care costs between aging trajectories of higher-level functional capacity. Methods Participants In collaboration with the local government of Kusatsu Town, Gunma Prefecture, Japan, we launched a longitudinal study of aging and health in 2001. In addition to an annual preventive health check-up for residents aged 40 years or older, participants aged 70 years or older (aged 65 years or older after 2006) underwent a geriatric assessment from 2002 through 2011. Moreover, all residents aged 70 years or older (aged 65 years or older after 2007) underwent biennial health monitoring surveys from 2001 through 2011 (response rate: minimum 91.0%, maximum 98.2%). All annual assessments and biennial health surveys were performed in the same manner. The details of the study design have been previously reported (11–15,19). All participants undergoing geriatric assessment provided written informed consent under conditions approved by the Ethics Committee at Tokyo Metropolitan Institute of Gerontology. The data source for the present study was 2,675 adults aged 65–90 years who underwent surveys conducted during the period from October 2001 through August 2011. We mainly used data from health monitoring surveys in 2001, 2003, 2005, 2007, 2009, and 2011 and used complementary data from biennial geriatric assessments during 2002 through 2010. To be eligible for the study, individuals had to complete TMIG-IC assessments. The average number of follow-up assessments was 4.0, and the total number of observations was 10,609 during the follow-up period. Measurement of TMIG-IC The TMIG-IC (9,20) is designed to measure higher-level functional capacity in community-dwelling older residents. It includes a multidimensional 13-item index of competence (see the Supplementary Appendix). The score ranges from 0 to 13, and lower scores indicate lower functional capacity. The TMIG-IC covers three sublevels of Lawton’s model: instrumental self-maintenance, effectance, and social role. A previous study confirmed that the TMIG-IC has high reliability—as indicated by Cronbach’s alpha, test–retest reliability, and the correlation between the second-order factor score and total score—and high construct, discriminant, and predictive validity (9). Mortality We examined local registries to ascertain deaths from any cause and linked these data with Japanese national vital statistics during the period through December 2015. The underlying cause of death was coded by using the International Classification of Diseases, Tenth Edition (ICD-10). The relevant ICD-10 codes were I00 to I99 for cardiovascular disease (CVD) and C00 to C97 for cancer. Medical and Long-Term Care Costs In Japan, all citizens have access to medical care and long-term care coverage under a universal health insurance system. The official medical insurance system comprises two categories. One is for employees and their dependents, and the other is the National Health Insurance (NHI) and health insurance for older people, which includes farmers, fisherman, and the self-employed, as well as retirees and pensioners, as beneficiaries. Citizens of Japan are automatically enrolled in the health insurance for older people program on their 75th birthday. The NHI and health insurance for older people cover almost all medical treatment and medical provider fees. Payments from insured persons to medical providers are made on a fee-for-service basis, in which the price of each service is determined by a uniform national fee schedule (21–23). The Japanese LTCI system was established to support the need for long-term care services, community-based services, and in-facility services (24). All primary insured persons aged 65 years or older are candidates for care, and secondary insured persons aged 40–64 years with any of 15 specific diseases can also utilize care services. Using data from the NHI, health insurance for older people, and LTCI beneficiaries in Kusatsu Town, we calculated monthly medical costs, monthly long-term care costs, and the sum of these costs as total costs for each participant for the period 1 year before the follow-up survey from 2001 through 2011 (25,26). Previous studies reported that hospital inpatient care and the costs of inpatient and long-term care increased at the end of life (27,28). To exclude dramatic increases in medical and care expenditure, monthly medical and long-term care costs were calculated for the period 1 year before the follow-up survey. However, for participants who died within 1 year of survey completion, costs were calculated for the 1-year interval from 2 years to 1 year before the date of the participant’s death. Costs are expressed in US dollars (1 US dollar = 112 Japanese yen on February 8, 2017). Statistical Analyses First, we identified TMIG-IC trajectory groups for the study period (from October 2001 through August 2011). We used a latent class group-based trajectory modeling approach implemented in the SAS macro PROC TRAJ (29), which assumes that a study population comprises a mixture of finite latent groups within which people follow an approximately homogeneous TMIG-IC trajectory. The assignment probability and parameters for a polynomial trajectory model for each latent group were simultaneously estimated via maximum likelihood for the mixture distribution (14). The number of latent groups and functional form of trajectory models (linear, quadratic, or cubic) were determined by comparing the Bayesian information criterion. After fitting the model, posterior probabilities for each group membership assignment were calculated for each individual, and participants were assigned to the groups with the highest posterior probabilities. Posterior probabilities were also used to assess model adequacy (ie, correctness of classification), by examining whether the posterior probability for an assigned group was sufficiently high as compared with those of the remaining groups (30). We express TMIG-IC trajectories in percentages, which represent the proportions of those assigned to each group, according to the highest posterior probability, rather than the default figure outputs in PROC TRAJ. Second, we examined associations of TMIG-IC trajectory group with mortality by using Cox proportional hazards models that controlled for potential confounders such as baseline sex, age, frequency of going outdoors (31), and self-rated health (32). The underlying time-scale in Cox proportional hazards models starts from first participation (during the period from October 2001 through August 2011) until death or the end of follow-up (December 2015), whichever came first. Because CVD, cancer, and other causes of death can be regarded as competing events, we used the competing-risk subdistribution regression of Fine and Gray (33) to determine hazards of cause-specific death. Finally, trajectory group-specific age trends in medical and long-term care costs were modeled by a generalized estimating equation with negative binomial or Poisson for sex and a linear term for age. Statistical analyses were conducted with SPSS (version 18.0; SPSS, Inc., Chicago, IL) and SAS (version 9.4; SAS Institute, Inc., Cary, NC), and a p value of less than .05 was considered to indicate statistical significance. Results Data from the baseline survey showed that the mean (SD) age of participants was 72.0 (6.2) years; 56.8% were women, 83.7% were able to go out by themselves, and 9.3% were independent in their home or neighborhood but were unable to go far by themselves. Self-rated health was very good or good for 78.7%. Mean (SD) TMIG-IC score was 11.1 (2.8). We identified four TMIG-IC trajectory patterns: 36.3% of participants were in the high-stable trajectory, 40.1% in the late-onset decreasing trajectory, 17.4% in the early-onset decreasing trajectory, and 6.1% in the low-decreasing trajectory (Figure 1; Table 1). The TMIG-IC score of the high-stable trajectory group was close to the maximum until they reached an age of approximately 85 years. Mean TMIG-IC scores in the late-onset decreasing and early-onset decreasing trajectory groups were 11.5 at age 65. The TMIG-IC score in the late-onset decreasing trajectory group remained relatively constant until around age 80 years but gradually decreased in the early-onset decreasing trajectory group after age 65 years. Mean TMIG-IC score in the low-decreasing trajectory group was approximately half that of the high-stable trajectory group at age 65 and showed a linear decline until around age 85. The average posterior probability of allocating each participant into the four groups was 0.73–0.87, indicating a good fit of the model of group trajectories to individual trajectories. Figure 1. View largeDownload slide TMIG-IC trajectories: a group-based semiparametric mixture model. The solid lines are estimated values. Figure 1. View largeDownload slide TMIG-IC trajectories: a group-based semiparametric mixture model. The solid lines are estimated values. Table 1. Estimated TMIG-IC Values for the Four Trajectories, by Age TMIG-IC Trajectory  Age, Years  65  70  75  80  85  90  High-stable group (n = 972; 36.3%)  12.8 (12.7–13.0)  12.9 (12.8–13.0)  12.9 (12.8–12.9)  12.7 (12.7–12.8)  12.3 (12.2–12.3)  11.0 (10.6–11.5)  Late-onset decreasing group (n = 1,074; 40.1%)  11.5 (11.2–11.9)  12.0 (11.8–12.1)  11.9 (11.8–12.0)  11.3 (11.0–11.5)  9.7 (9.4–10.0)  7.0 (6.6–7.5)  Early-onset decreasing group (n = 465; 17.4%)  11.5 (10.8–12.1)  10.8 (10.4–11.2)  9.6 (9.2–10.0)  7.7 (7.3–8.0)  5.2 (4.8–5.5)  2.3 (1.6–3.0)  Low-decreasing group (n = 164; 6.1%)  7.7 (5.6–9.8)  6.1 (5.3–6.8)  4.3 (3.8–4.9)  2.6 (1.9–3.4)  1.2 (0.2–2.2)  0.4 (0–0.9)  TMIG-IC Trajectory  Age, Years  65  70  75  80  85  90  High-stable group (n = 972; 36.3%)  12.8 (12.7–13.0)  12.9 (12.8–13.0)  12.9 (12.8–12.9)  12.7 (12.7–12.8)  12.3 (12.2–12.3)  11.0 (10.6–11.5)  Late-onset decreasing group (n = 1,074; 40.1%)  11.5 (11.2–11.9)  12.0 (11.8–12.1)  11.9 (11.8–12.0)  11.3 (11.0–11.5)  9.7 (9.4–10.0)  7.0 (6.6–7.5)  Early-onset decreasing group (n = 465; 17.4%)  11.5 (10.8–12.1)  10.8 (10.4–11.2)  9.6 (9.2–10.0)  7.7 (7.3–8.0)  5.2 (4.8–5.5)  2.3 (1.6–3.0)  Low-decreasing group (n = 164; 6.1%)  7.7 (5.6–9.8)  6.1 (5.3–6.8)  4.3 (3.8–4.9)  2.6 (1.9–3.4)  1.2 (0.2–2.2)  0.4 (0–0.9)  Note: Values are averages (95% confidence interval). View Large Table 1. Estimated TMIG-IC Values for the Four Trajectories, by Age TMIG-IC Trajectory  Age, Years  65  70  75  80  85  90  High-stable group (n = 972; 36.3%)  12.8 (12.7–13.0)  12.9 (12.8–13.0)  12.9 (12.8–12.9)  12.7 (12.7–12.8)  12.3 (12.2–12.3)  11.0 (10.6–11.5)  Late-onset decreasing group (n = 1,074; 40.1%)  11.5 (11.2–11.9)  12.0 (11.8–12.1)  11.9 (11.8–12.0)  11.3 (11.0–11.5)  9.7 (9.4–10.0)  7.0 (6.6–7.5)  Early-onset decreasing group (n = 465; 17.4%)  11.5 (10.8–12.1)  10.8 (10.4–11.2)  9.6 (9.2–10.0)  7.7 (7.3–8.0)  5.2 (4.8–5.5)  2.3 (1.6–3.0)  Low-decreasing group (n = 164; 6.1%)  7.7 (5.6–9.8)  6.1 (5.3–6.8)  4.3 (3.8–4.9)  2.6 (1.9–3.4)  1.2 (0.2–2.2)  0.4 (0–0.9)  TMIG-IC Trajectory  Age, Years  65  70  75  80  85  90  High-stable group (n = 972; 36.3%)  12.8 (12.7–13.0)  12.9 (12.8–13.0)  12.9 (12.8–12.9)  12.7 (12.7–12.8)  12.3 (12.2–12.3)  11.0 (10.6–11.5)  Late-onset decreasing group (n = 1,074; 40.1%)  11.5 (11.2–11.9)  12.0 (11.8–12.1)  11.9 (11.8–12.0)  11.3 (11.0–11.5)  9.7 (9.4–10.0)  7.0 (6.6–7.5)  Early-onset decreasing group (n = 465; 17.4%)  11.5 (10.8–12.1)  10.8 (10.4–11.2)  9.6 (9.2–10.0)  7.7 (7.3–8.0)  5.2 (4.8–5.5)  2.3 (1.6–3.0)  Low-decreasing group (n = 164; 6.1%)  7.7 (5.6–9.8)  6.1 (5.3–6.8)  4.3 (3.8–4.9)  2.6 (1.9–3.4)  1.2 (0.2–2.2)  0.4 (0–0.9)  Note: Values are averages (95% confidence interval). View Large Local registries showed 747 (27.9%) incident deaths among the 2,675 participants. The median duration of follow-up for incident death was 2,915 days. Among the 747 incident deaths, 738 (98.8%) were linked with Japanese national vital statistics. As compared with the high-stable TMIG-IC trajectory group during follow-up, the late-onset decreasing, early-onset decreasing, and low-decreasing trajectory groups had hazard ratios (95% confidence interval) of 1.29 (1.07–1.56), 2.33 (1.90–2.84), and 4.67 (3.65–5.99), respectively, for all-cause mortality (Table 2). Even after adjustment for several demographic and health characteristics, the independent association with TMIG-IC trajectory group remained significant: the late-onset decreasing, early-onset decreasing, and low-decreasing trajectory groups had hazard ratios of 1.22 (1.01–1.47), 1.90 (1.53–2.36), and 2.87 (2.14–3.84), respectively, for all-cause mortality. The analysis of cause-specific death, CVD mortality, and other mortality showed significant associations between the four TMIG-IC trajectory patterns. Table 2. Independent Associations of TMIG-IC Trajectory with All-Cause and Cause-Specific Mortality in Community-Dwelling Japanese Aged ≥65 Years TMIG-IC Trajectory  Incident All-Cause Deaths  All-Cause Mortality  CVD Mortality  Cancer Mortality  Other Mortality  Crude HR (95%CI)  Adjusted HR (95% CI)  Crude HR (95% CI)  Adjusted HR (95% CI)  Crude HR (95% CI)  Adjusted HR (95% CI)  Crude HR (95% CI)  Adjusted HR (95% CI)  High-stable group† (n = 972; 36.3%)  191 (19.7%)  1  1  1  1  1  1  1  1  Late-onset decreasing group (n = 1,074; 40.1%)  267 (24.9%)  1.29 (1.07–1.56)**  1.22 (1.01–1.47)*  1.54 (1.09–2.17)*  1.40 (1.01–2.01)*  0.96 (0.68–1.37)  0.93 (0.65–1.33)  1.30 (0.98–1.73)  1.10 (0.82–1.47)  Early-onset decreasing group (n = 465; 17.4%)  195 (41.9%)  2.33 (1.90–2.84)**  1.90 (1.53–2.36)**  2.84 (1.99–4.06)**  2.41 (1.65–3.53)**  1.29 (0.86–1.94)  1.20 (0.79–1.84)  2.13 (1.56–2.90)**  1.52 (1.08–2.15)*  Low-decreasing group (n = 164; 6.1%)  94 (57.3%)  4.67 (3.65–5.99)**  2.87 (2.14–3.84)**  2.65 (1.60–4.40)**  1.97 (1.06–3.66)*  1.98 (1.17–3.36)*  1.67 (0.92–3.05)  4.96 (3.45–7.11)**  3.54 (2.16–5.80)**  TMIG-IC Trajectory  Incident All-Cause Deaths  All-Cause Mortality  CVD Mortality  Cancer Mortality  Other Mortality  Crude HR (95%CI)  Adjusted HR (95% CI)  Crude HR (95% CI)  Adjusted HR (95% CI)  Crude HR (95% CI)  Adjusted HR (95% CI)  Crude HR (95% CI)  Adjusted HR (95% CI)  High-stable group† (n = 972; 36.3%)  191 (19.7%)  1  1  1  1  1  1  1  1  Late-onset decreasing group (n = 1,074; 40.1%)  267 (24.9%)  1.29 (1.07–1.56)**  1.22 (1.01–1.47)*  1.54 (1.09–2.17)*  1.40 (1.01–2.01)*  0.96 (0.68–1.37)  0.93 (0.65–1.33)  1.30 (0.98–1.73)  1.10 (0.82–1.47)  Early-onset decreasing group (n = 465; 17.4%)  195 (41.9%)  2.33 (1.90–2.84)**  1.90 (1.53–2.36)**  2.84 (1.99–4.06)**  2.41 (1.65–3.53)**  1.29 (0.86–1.94)  1.20 (0.79–1.84)  2.13 (1.56–2.90)**  1.52 (1.08–2.15)*  Low-decreasing group (n = 164; 6.1%)  94 (57.3%)  4.67 (3.65–5.99)**  2.87 (2.14–3.84)**  2.65 (1.60–4.40)**  1.97 (1.06–3.66)*  1.98 (1.17–3.36)*  1.67 (0.92–3.05)  4.96 (3.45–7.11)**  3.54 (2.16–5.80)**  Notes: CVD, cardiovascular disease; CI, confidence interval; HR, hazard ratio; TMIG-IC, Tokyo Metropolitan Institute of Gerontology Index of Competence. †Reference group. Cox hazards regression models were run separately. Adjusted for sex, age, frequency of going outdoors, and self-rated health. *p < .05, **p < .01. View Large Table 2. Independent Associations of TMIG-IC Trajectory with All-Cause and Cause-Specific Mortality in Community-Dwelling Japanese Aged ≥65 Years TMIG-IC Trajectory  Incident All-Cause Deaths  All-Cause Mortality  CVD Mortality  Cancer Mortality  Other Mortality  Crude HR (95%CI)  Adjusted HR (95% CI)  Crude HR (95% CI)  Adjusted HR (95% CI)  Crude HR (95% CI)  Adjusted HR (95% CI)  Crude HR (95% CI)  Adjusted HR (95% CI)  High-stable group† (n = 972; 36.3%)  191 (19.7%)  1  1  1  1  1  1  1  1  Late-onset decreasing group (n = 1,074; 40.1%)  267 (24.9%)  1.29 (1.07–1.56)**  1.22 (1.01–1.47)*  1.54 (1.09–2.17)*  1.40 (1.01–2.01)*  0.96 (0.68–1.37)  0.93 (0.65–1.33)  1.30 (0.98–1.73)  1.10 (0.82–1.47)  Early-onset decreasing group (n = 465; 17.4%)  195 (41.9%)  2.33 (1.90–2.84)**  1.90 (1.53–2.36)**  2.84 (1.99–4.06)**  2.41 (1.65–3.53)**  1.29 (0.86–1.94)  1.20 (0.79–1.84)  2.13 (1.56–2.90)**  1.52 (1.08–2.15)*  Low-decreasing group (n = 164; 6.1%)  94 (57.3%)  4.67 (3.65–5.99)**  2.87 (2.14–3.84)**  2.65 (1.60–4.40)**  1.97 (1.06–3.66)*  1.98 (1.17–3.36)*  1.67 (0.92–3.05)  4.96 (3.45–7.11)**  3.54 (2.16–5.80)**  TMIG-IC Trajectory  Incident All-Cause Deaths  All-Cause Mortality  CVD Mortality  Cancer Mortality  Other Mortality  Crude HR (95%CI)  Adjusted HR (95% CI)  Crude HR (95% CI)  Adjusted HR (95% CI)  Crude HR (95% CI)  Adjusted HR (95% CI)  Crude HR (95% CI)  Adjusted HR (95% CI)  High-stable group† (n = 972; 36.3%)  191 (19.7%)  1  1  1  1  1  1  1  1  Late-onset decreasing group (n = 1,074; 40.1%)  267 (24.9%)  1.29 (1.07–1.56)**  1.22 (1.01–1.47)*  1.54 (1.09–2.17)*  1.40 (1.01–2.01)*  0.96 (0.68–1.37)  0.93 (0.65–1.33)  1.30 (0.98–1.73)  1.10 (0.82–1.47)  Early-onset decreasing group (n = 465; 17.4%)  195 (41.9%)  2.33 (1.90–2.84)**  1.90 (1.53–2.36)**  2.84 (1.99–4.06)**  2.41 (1.65–3.53)**  1.29 (0.86–1.94)  1.20 (0.79–1.84)  2.13 (1.56–2.90)**  1.52 (1.08–2.15)*  Low-decreasing group (n = 164; 6.1%)  94 (57.3%)  4.67 (3.65–5.99)**  2.87 (2.14–3.84)**  2.65 (1.60–4.40)**  1.97 (1.06–3.66)*  1.98 (1.17–3.36)*  1.67 (0.92–3.05)  4.96 (3.45–7.11)**  3.54 (2.16–5.80)**  Notes: CVD, cardiovascular disease; CI, confidence interval; HR, hazard ratio; TMIG-IC, Tokyo Metropolitan Institute of Gerontology Index of Competence. †Reference group. Cox hazards regression models were run separately. Adjusted for sex, age, frequency of going outdoors, and self-rated health. *p < .05, **p < .01. View Large Among 2,675 participants, 1,882 (70.4%) incurred monthly medical costs and 379 (14.2%) incurred monthly long-term care costs during the 1-year period before the follow-up survey. The generalized-estimating-equation models showed significant TMIG-IC trajectory-specific age trends in monthly medical and long-term care costs (Table 3; Figures 2 and 3). Mean monthly medical costs in the low-decreasing TMIG-IC trajectory group were estimated at $506.6 and were three times those of the high-stable group at age 65 but declined after age 65 years. The other three TMIG-IC trajectory groups exhibited parallel increases in monthly medical costs after age 65 years. Mean monthly long-term care costs in the low-decreasing TMIG-IC trajectory group were estimated at $275.5 and were obviously higher than in the other three groups at age 65. The low-decreasing TMIG-IC trajectory group had consistently higher monthly long-term care costs than did the remaining groups from age 65 to 90 years. Interestingly, monthly long-term care costs were similar for the early-onset decreasing TMIG-IC trajectory group and the high-stable and late-onset decreasing trajectory groups at age 65. Later in life, however, the early-onset decreasing trajectory group exhibited large increases, while costs in the high-stable and late-onset decreasing trajectory groups remained low. The mean sum of these costs for the low-decreasing trajectory group was estimated at $683.0, which was approximately four times the values for the high-stable ($145.3), late-onset decreasing ($184.2), and early-onset decreasing ($178.9) group at age 65 (Figure 4). The sum of medical and long-term care costs increased after age 65 years in the four TMIG-IC trajectory groups. In particular, the early-onset decreasing trajectory group exhibited a dramatic increase after age 75 years, and the estimated costs exceeded those of the low-decreasing trajectory group until age 87 years. Mean total costs later in life were lower for participants with high-stable and late-onset decreasing TMIG-IC trajectories. Table 3. Estimated Monthly Medical and Care Costs for the Four TMIG-IC Trajectories, by Age. TMIG-IC Trajectory  Age, Years  65  70  75  80  85  90  Medical Cost (US Dollars)  High-stable group (n = 972; 36.3%)  154.3 (124.2–191.6)  195.7 (167.5–228.5)  248.2 (220.4–279.4)  314.7 (276.5–358.3)  399.2 (333.5–477.7)  506.2 (395.9–647.3)  Late-onset decreasing group (n = 1,074; 40.1%)  203.9 (163.8–253.7)  249.8 (212.8–293.3)  306.2 (270.5–346.6)  375.3 (330.1–426.7)  460.0 (387.7–545.7)  563.7 (447.3–710.4)  Early-onset decreasing group (n = 465; 17.4%)  201.4 (135.5–299.3)  249.2 (187.8–330.5)  308.3 (252.1–377.0)  381.4 (312.9–465.0)  472.0 (358.4–621.6)  584.0 (396.2–860.7)  Low-decreasing group (n = 164; 6.1%)  506.6 (258.1–994.3)  459.5 (302.2–698.7)  416.9 (299.3–580.7)  378.2 (227.7–628.2)  343.1 (156.2–753.4)  311.3 (104.4–928.0)  Care cost (US dollars)  High-stable group (n = 972; 36.3%)  1.3 (0.4–3.7)  3.3 (1.4–8.0)  8.7 (4.3–17.6)  22.5 (12.8–39.3)  58.3 (37.4–90.9)  151.2 (101.8–224.7)  Late-onset decreasing group (n = 1,074; 40.1%)  5.8 (3.2–10.8)  12.3 (7.7–19.8)  26.5 (18.3–38.3)  56.7 (42.3–76.0)  121.6 (93.7–157.9)  260.7 (194.8–349.0)  Early-onset decreasing group (n = 465; 17.4%)  41.5 (25.8–66.8)  76.7 (52.7–111.7)  141.9 (106.8–188.5)  262.3 (211.4–325.5)  485.0 (398.9–589.6)  896.7 (708.2–1135.3)  Low-decreasing group (n = 164; 6.1%)  275.5 (174.3–435.5)  361.3 (256.7–508.6)  473.8 (369.8–607.0)  621.3 (504.7–764.8)  814.8 (635.0–1045.4)  1068.4 (757.5–1507.1)  TMIG-IC Trajectory  Age, Years  65  70  75  80  85  90  Medical Cost (US Dollars)  High-stable group (n = 972; 36.3%)  154.3 (124.2–191.6)  195.7 (167.5–228.5)  248.2 (220.4–279.4)  314.7 (276.5–358.3)  399.2 (333.5–477.7)  506.2 (395.9–647.3)  Late-onset decreasing group (n = 1,074; 40.1%)  203.9 (163.8–253.7)  249.8 (212.8–293.3)  306.2 (270.5–346.6)  375.3 (330.1–426.7)  460.0 (387.7–545.7)  563.7 (447.3–710.4)  Early-onset decreasing group (n = 465; 17.4%)  201.4 (135.5–299.3)  249.2 (187.8–330.5)  308.3 (252.1–377.0)  381.4 (312.9–465.0)  472.0 (358.4–621.6)  584.0 (396.2–860.7)  Low-decreasing group (n = 164; 6.1%)  506.6 (258.1–994.3)  459.5 (302.2–698.7)  416.9 (299.3–580.7)  378.2 (227.7–628.2)  343.1 (156.2–753.4)  311.3 (104.4–928.0)  Care cost (US dollars)  High-stable group (n = 972; 36.3%)  1.3 (0.4–3.7)  3.3 (1.4–8.0)  8.7 (4.3–17.6)  22.5 (12.8–39.3)  58.3 (37.4–90.9)  151.2 (101.8–224.7)  Late-onset decreasing group (n = 1,074; 40.1%)  5.8 (3.2–10.8)  12.3 (7.7–19.8)  26.5 (18.3–38.3)  56.7 (42.3–76.0)  121.6 (93.7–157.9)  260.7 (194.8–349.0)  Early-onset decreasing group (n = 465; 17.4%)  41.5 (25.8–66.8)  76.7 (52.7–111.7)  141.9 (106.8–188.5)  262.3 (211.4–325.5)  485.0 (398.9–589.6)  896.7 (708.2–1135.3)  Low-decreasing group (n = 164; 6.1%)  275.5 (174.3–435.5)  361.3 (256.7–508.6)  473.8 (369.8–607.0)  621.3 (504.7–764.8)  814.8 (635.0–1045.4)  1068.4 (757.5–1507.1)  Note: Values are averages (95% confidence interval). View Large Table 3. Estimated Monthly Medical and Care Costs for the Four TMIG-IC Trajectories, by Age. TMIG-IC Trajectory  Age, Years  65  70  75  80  85  90  Medical Cost (US Dollars)  High-stable group (n = 972; 36.3%)  154.3 (124.2–191.6)  195.7 (167.5–228.5)  248.2 (220.4–279.4)  314.7 (276.5–358.3)  399.2 (333.5–477.7)  506.2 (395.9–647.3)  Late-onset decreasing group (n = 1,074; 40.1%)  203.9 (163.8–253.7)  249.8 (212.8–293.3)  306.2 (270.5–346.6)  375.3 (330.1–426.7)  460.0 (387.7–545.7)  563.7 (447.3–710.4)  Early-onset decreasing group (n = 465; 17.4%)  201.4 (135.5–299.3)  249.2 (187.8–330.5)  308.3 (252.1–377.0)  381.4 (312.9–465.0)  472.0 (358.4–621.6)  584.0 (396.2–860.7)  Low-decreasing group (n = 164; 6.1%)  506.6 (258.1–994.3)  459.5 (302.2–698.7)  416.9 (299.3–580.7)  378.2 (227.7–628.2)  343.1 (156.2–753.4)  311.3 (104.4–928.0)  Care cost (US dollars)  High-stable group (n = 972; 36.3%)  1.3 (0.4–3.7)  3.3 (1.4–8.0)  8.7 (4.3–17.6)  22.5 (12.8–39.3)  58.3 (37.4–90.9)  151.2 (101.8–224.7)  Late-onset decreasing group (n = 1,074; 40.1%)  5.8 (3.2–10.8)  12.3 (7.7–19.8)  26.5 (18.3–38.3)  56.7 (42.3–76.0)  121.6 (93.7–157.9)  260.7 (194.8–349.0)  Early-onset decreasing group (n = 465; 17.4%)  41.5 (25.8–66.8)  76.7 (52.7–111.7)  141.9 (106.8–188.5)  262.3 (211.4–325.5)  485.0 (398.9–589.6)  896.7 (708.2–1135.3)  Low-decreasing group (n = 164; 6.1%)  275.5 (174.3–435.5)  361.3 (256.7–508.6)  473.8 (369.8–607.0)  621.3 (504.7–764.8)  814.8 (635.0–1045.4)  1068.4 (757.5–1507.1)  TMIG-IC Trajectory  Age, Years  65  70  75  80  85  90  Medical Cost (US Dollars)  High-stable group (n = 972; 36.3%)  154.3 (124.2–191.6)  195.7 (167.5–228.5)  248.2 (220.4–279.4)  314.7 (276.5–358.3)  399.2 (333.5–477.7)  506.2 (395.9–647.3)  Late-onset decreasing group (n = 1,074; 40.1%)  203.9 (163.8–253.7)  249.8 (212.8–293.3)  306.2 (270.5–346.6)  375.3 (330.1–426.7)  460.0 (387.7–545.7)  563.7 (447.3–710.4)  Early-onset decreasing group (n = 465; 17.4%)  201.4 (135.5–299.3)  249.2 (187.8–330.5)  308.3 (252.1–377.0)  381.4 (312.9–465.0)  472.0 (358.4–621.6)  584.0 (396.2–860.7)  Low-decreasing group (n = 164; 6.1%)  506.6 (258.1–994.3)  459.5 (302.2–698.7)  416.9 (299.3–580.7)  378.2 (227.7–628.2)  343.1 (156.2–753.4)  311.3 (104.4–928.0)  Care cost (US dollars)  High-stable group (n = 972; 36.3%)  1.3 (0.4–3.7)  3.3 (1.4–8.0)  8.7 (4.3–17.6)  22.5 (12.8–39.3)  58.3 (37.4–90.9)  151.2 (101.8–224.7)  Late-onset decreasing group (n = 1,074; 40.1%)  5.8 (3.2–10.8)  12.3 (7.7–19.8)  26.5 (18.3–38.3)  56.7 (42.3–76.0)  121.6 (93.7–157.9)  260.7 (194.8–349.0)  Early-onset decreasing group (n = 465; 17.4%)  41.5 (25.8–66.8)  76.7 (52.7–111.7)  141.9 (106.8–188.5)  262.3 (211.4–325.5)  485.0 (398.9–589.6)  896.7 (708.2–1135.3)  Low-decreasing group (n = 164; 6.1%)  275.5 (174.3–435.5)  361.3 (256.7–508.6)  473.8 (369.8–607.0)  621.3 (504.7–764.8)  814.8 (635.0–1045.4)  1068.4 (757.5–1507.1)  Note: Values are averages (95% confidence interval). View Large Figure 2. View largeDownload slide TMIG-IC trajectory-specific age trends in monthly medical costs: a generalized estimating equation. The solid lines are estimated values. Figure 2. View largeDownload slide TMIG-IC trajectory-specific age trends in monthly medical costs: a generalized estimating equation. The solid lines are estimated values. Figure 3. View largeDownload slide TMIG-IC trajectory-specific age trends in monthly long-term care costs: a generalized estimating equation. The solid lines are estimated values. Figure 3. View largeDownload slide TMIG-IC trajectory-specific age trends in monthly long-term care costs: a generalized estimating equation. The solid lines are estimated values. Figure 4. View largeDownload slide TMIG-IC trajectory-specific age trends in monthly total costs (medical and long-term care costs): a generalized estimating equation. The solid lines are estimated values. Figure 4. View largeDownload slide TMIG-IC trajectory-specific age trends in monthly total costs (medical and long-term care costs): a generalized estimating equation. The solid lines are estimated values. Discussion This prospective study is the first to show aging trajectories in higher-level functional capacity of community-dwelling older Japanese. The trajectory of higher-level functional capacity was an independent predictor of all-cause, cardiovascular, and non-cancer mortality, and elders with a low-decreasing aging trajectory had higher monthly medical and long-term care costs in later life. Previous studies (4–7) used ADL and IADL to assess functional capacity. These measures correspond to the fourth and fifth sublevels of Lawton’s hierarchical model. TMIG-IC assessment of higher-level functional capacity includes the last three sublevels (instrumental self-maintenance, effectance, and social role) of Lawton’s model. The importance of promoting higher-level functional capacity, especially social participation, is a key proposal for “Active Aging” in the World Health Organization’s policy framework (34). We identified four distinct TMIG-IC trajectory patterns (high-stable, late-onset decreasing, early-onset decreasing, and low-decreasing) among community-dwelling older Japanese aged 65–90 years, and this is the first study to show aging trajectories in higher-level functional capacity in a community-based study. A TMIG-IC score of 10 points or higher indicates normal higher-level function (35). Approximately 80% of Japanese elderly adults (the high-stable and late-onset decreasing trajectory groups) maintained higher-level functional capacity in later life, and 36% of elders were aging successfully. Overall, 17% of elders (the early-onset decreasing trajectory group) exhibited quadratic declines, and 6% of elders (the low-decreasing trajectory group) exhibited impaired higher-level functional capacity and rapid declines after age 65 years. We observed that, as compared with the high-stable TMIG-IC trajectory group, participants in the late-onset decreasing, early-onset decreasing, and low-decreasing trajectory groups had HRs of 1.29, 2.33, and 4.67, respectively, for all-cause mortality. Takata et al. reported that initial TMIG-IC score was associated with all-cause mortality in 697 Japanese aged 80 years (17). Previous studies showed that ADL and IADL were predictors of mortality among community-dwelling older adults in a Western population (36,37). The present prospective study using repeated-measures analysis extends the findings of earlier studies and highlights the importance of intervention for improvements in higher-level functional capacity, even among older adults with low higher-level functional capacity. In an analysis of cause-specific mortality, participants in the lower trajectory groups had high HRs for CVD mortality and non-cancer mortality. Murakami et al. (35) reported that impaired higher-level capacity, as indicated by TMIG-IC at baseline, was associated with incident stroke among community-dwelling older adults with independent basic ADL. Takata et al. (17) reported that higher-level functional capacity, as assessed by TMIG-IC, was associated with cardiovascular and pneumonia mortality but not with cancer mortality. Our results accord with these earlier findings and are the first evidence that potential higher-level functional capacity trajectories are independent predictors of CVD mortality and non-cancer mortality. Our study is the first to find that elders with high-stable and late-onset decreasing higher-level functional capacity trajectories had lower mean monthly medical costs and long-term care costs, while those with low-decreasing trajectories had higher mean total costs, after excluding the dramatic increase in such costs at the end of life. Monthly medical costs in the low-decreasing higher-level functional capacity trajectories tended to decrease with advancing age, but monthly long-term care costs tended to increase after age 65 years. To identify the reason for this cost discrepancy in the low-decreasing TMIG-IC trajectory group, we used data from the geriatric assessment to examine baseline demographic and health characteristics of participants in the four trajectory groups. At baseline, adults in the low-decreasing trajectory group were more likely to have a history of cerebrovascular disease and had higher white blood cell counts (Supplementary Table), which suggests that TMIG-IC trajectory patterns were related to the presence of chronic diseases. Spillman et al. (26) reported that the increase in nursing home costs with advancing age was sufficient to offset the moderating effect of declining medical expenditures in an American population during the last 2 years of life. People with impaired higher level functional capacity, such as the present low-decreasing trajectory group, might have had more chronic diseases and greater mortality risk and may therefore have shifted their needs from medical service to care service in later life. Moreover, in the Mayo Clinic Study of Aging, Leibson et al. (38) reported that annual mean medical costs rose gradually for persons with cognitive normal, mild cognitive impairment, newly discovered dementia, and prevalent dementia. The present study showed that adults in the low-decreasing trajectory group had lower MMSE scores, which suggests that lower trajectories of higher-level functional capacity were associated with lower cognitive function and higher monthly medical and long-term care costs. This study has strengths that warrant mention. First, our large sample of community-dwelling adults allowed us to use a group-based semiparametric mixture model, which showed potential trajectory patterns in TMIG-IC. Previous study focused on ADL and the association with mortality in hospitalized patients (39). A few studies examined the association of higher-level functional capacity with mortality among community-dwelling older adults (17,35). However, the present prospective study used repeated-measures analysis of data from community-dwelling older Japanese to investigate the associations of potential trajectory patterns in higher-level functional capacity with mortality. Second, the repeated-measured data in this study were derived from annual geriatric health assessments and biennial health monitoring surveys. Half of the total of 10,609 observations during the follow-up period were from biennial health monitoring surveys, all of which had high response rates. Third, the data for calculating medical and long-term care costs in this study were derived from the Japanese NHI, health insurance for older people, and LTCI beneficiaries. These systems cover nearly all medical provider fees and all care provider fees. Because Japan has a universal health insurance system, we were able to link aging trajectories in higher-level functional capacity, cause of death, and medical and long-term care costs for community-dwelling older Japanese. We then determined whether these trajectories were associated with all-cause and cause-specific mortality and examined differences in medical and long-term care costs between trajectories among community-dwelling older Japanese. Our findings regarding the socioeconomic effects of trajectory patterns in higher-level functional capacity might help advance policies for health promotion and preventive care initiatives in developed countries with growing social security cost burdens. This study has some limitations. First, a previous study found that elders in a Japanese sample were more likely to be competent than those in a US sample (20). Trajectory patterns of higher-level functional capacity might differ between Japanese and Western populations; thus, future studies should examine the aging patterns of higher-level functional capacity among Western populations. Moreover, the data source for the present study was limited to Kusatsu Town, Japan. Data from other study sites or nationally representative data for Japan are needed in order to confirm the present findings. Second, use of death certificates to classify major causes of death may lead to misclassification. However, this limitation is not unique to the present study, and use of death certificates for this purpose was reported to be accurate in Japan (40). Third, the present study examined baseline demographic and health characteristics of participants in four trajectory groups; however, these data were limited to half the total observations. Future studies should attempt to identify risk markers for the lower aging trajectory in higher-level functional capacity and apply those findings to health promotion and development of programs that reduce social costs. In conclusion, this prospective study found that community-dwelling older adults exhibit four major trajectories in TMIG-IC, a measure of higher-level functional capacity. Approximately 80% of elders (the high-stable and late-onset decreasing trajectory groups) maintained higher-level functional capacity in later life. In contrast, 6% of elders (the low-decreasing trajectory group) exhibited rapid declines after age 65 years and had greater risks for cardiovascular mortality, non-cancer mortality, and all-cause mortality. Furthermore, the low-decreasing trajectory group had higher monthly medical and long-term care costs. These findings indicate that interventions that improve higher-level functional capacity may improve health and longevity, and lessen the socioeconomic burden, among elders. Supplementary Material Supplementary data is available at The Journals of Gerontology, Series A: Biological Sciences and Medical Sciences online. Funding This study was supported by TMIG, JST/RISTEX, Grants-In-Aid for Scientific Research (B) JP20390190, (B) JP21390212, (B) JP24390173, and (B) JP26310111, a Grant-In-Aid for Research Activity Start-up JP24890302, and a Grant-In-Aid for Young Scientists (B) JP15K16539 from the Ministry of Education, Culture, Sports, Science and Technology, Japan. Conflict of Interest The authors have no potential conflicts of interest related to this research. Acknowledgements We thank the staff members of the Public Health Center and town office in Kusatsu and our colleagues on the Research Team for Social Participation and Community Health, Tokyo Metropolitan Institute of Gerontology (TMIG). References 1. The Uses of Epidemiology in the Study of the Elderly. Report of a WHO scientific group on the epidemiology of aging. World Health Organ Tech Rep Ser . 1984; 706: 1– 84. PubMed  2. Ishizaki T, Kai I, Kobayashi Y, Imanaka Y. Functional transitions and active life expectancy for older Japanese living in a community. Arch Gerontol Geriatr . 2002; 35: 107– 120. Google Scholar CrossRef Search ADS PubMed  3. Lawton MP. Assessing the competence of older people. In: Kent DP, Kastenbaum R, Sherwood S, eds. Research Planning and Action for the Elderly: The Power and Potential of Social Science . New York: Human Sciences Press; 1972: 122– 143. 4. Vermeulen J, Neyens JC, van Rossum E, Spreeuwenberg MD, de Witte LP. Predicting ADL disability in community-dwelling elderly people using physical frailty indicators: a systematic review. BMC Geriatr . 2011; 11: 33. doi: 10.1186/1471-2318-11-33 Google Scholar CrossRef Search ADS PubMed  5. Jefferson AL, Paul RH, Ozonoff A, Cohen RA. Evaluating elements of executive functioning as predictors of instrumental activities of daily living (IADLs). Arch Clin Neuropsychol . 2006; 21: 311– 320. doi: 10.1016/j.acn.2006.03.007 Google Scholar CrossRef Search ADS PubMed  6. Buurman BM, Han L, Murphy TEet al.   Trajectories of disability among older persons before and after a hospitalization leading to a skilled nursing facility admission. J Am Med Dir Assoc . 2016; 17: 225– 231. doi: 10.1016/j.jamda.2015.10.010 Google Scholar CrossRef Search ADS PubMed  7. Ferrante LE, Pisani MA, Murphy TE, Gahbauer EA, Leo-Summers LS, Gill TM. Functional trajectories among older persons before and after critical illness. JAMA Intern Med . 2015; 175: 523– 529. doi: 10.1001/jamainternmed.2014.7889 Google Scholar CrossRef Search ADS PubMed  8. Fujiwara Y, Shinkai S, Kumagai Set al.   Longitudinal changes in higher-level functional capacity of an older population living in a Japanese urban community. Arch Gerontol Geriatr . 2003; 36: 141– 153. Google Scholar CrossRef Search ADS PubMed  9. Koyano W, Shibata H, Nakazato K, Haga H, Suyama Y. Measurement of competence: reliability and validity of the TMIG index of competence. Arch Gerontol Geriatr . 1991; 13: 103– 116. Google Scholar CrossRef Search ADS PubMed  10. Fujiwara Y, Yoshida H, Amano Het al.   Predictors of improvement or decline in instrumental activities of daily living among community-dwelling older Japanese. Gerontology . 2008; 54: 373– 380. doi: 10.1159/000151221 Google Scholar CrossRef Search ADS PubMed  11. Taniguchi Y, Yoshida H, Fujiwara Y, Motohashi Y, Shinkai S. A prospective study of gait performance and subsequent cognitive decline in a general population of older Japanese. J Gerontol A Biol Sci Med Sci . 2012; 67: 796– 803. doi: 10.1093/gerona/glr243 Google Scholar CrossRef Search ADS PubMed  12. Taniguchi Y, Shinkai S, Nishi Met al.   Nutritional biomarkers and subsequent cognitive decline among community-dwelling older Japanese: a prospective study. J Gerontol A Biol Sci Med Sci . 2014; 69: 1276– 1283. doi: 10.1093/gerona/glt286 Google Scholar CrossRef Search ADS PubMed  13. Taniguchi Y, Fujiwara Y, Nofuji Yet al.   Prospective study of arterial stiffness and subsequent cognitive decline among community-dwelling older Japanese. J Epidemiol . 2015; 25: 592– 599. doi: 10.2188/jea.JE20140250 Google Scholar CrossRef Search ADS PubMed  14. Taniguchi Y, Fujiwara Y, Murayama Het al.   Prospective study of trajectories of physical performance and mortality among community-dwelling older Japanese. J Gerontol A Biol Sci Med Sci . 2016; 71: 1492– 1499. doi: 10.1093/gerona/glw029 Google Scholar CrossRef Search ADS PubMed  15. Taniguchi Y, Kitamura A, Seino Set al.   Gait performance trajectories and incident disabling dementia among community-dwelling older Japanese. J Am Med Dir Assoc . 2017; 18: 192.e13– 192.e20. doi: 10.1016/j.jamda.2016.10.015 Google Scholar CrossRef Search ADS   16. Koyano W, Hashimoto M, Fukawa T, Shibata H, Gunji A. Functional capacity of the elderly: measurement by the TMIG Index of Competence [in Japanese]. Nihon Koshu Eisei Zasshi. Japanese J Public Health . 1993; 40: 468– 474. 17. Takata Y, Ansai T, Akifusa S, Soh I, Sonoki K, Takehara T. High-level functional capacity and 4-year mortality in an 80-year-old population. Gerontology . 2007; 53: 46– 51. doi: 10.1159/000095934 Google Scholar CrossRef Search ADS PubMed  18. Lubitz J, Cai L, Kramarow E, Lentzner H. Health, life expectancy, and health care spending among the elderly. N Engl J Med . 2003; 349: 1048– 1055. doi: 10.1056/NEJMsa020614 Google Scholar CrossRef Search ADS PubMed  19. Shinkai S, Yoshida H, Taniguchi Yet al.   Public health approach to preventing frailty in the community and its effect on healthy aging in Japan. Geriatr Gerontol Int . 2016; 16 ( Suppl 1): 87– 97. doi: 10.1111/ggi.12726 Google Scholar CrossRef Search ADS PubMed  20. Shibata H, Sugisawa H, Watanabe S. Functional capacity in elderly Japanese living in the community. Geriatr Gerontol Int . 2001; 1: 8– 13. doi:10.1046/j.1444-1586.2001.00004.x Google Scholar CrossRef Search ADS   21. Nakamura K, Okamura T, Kanda H, Hayakawa T, Okayama A, Ueshima H; Health Promotion Research Committee of the Shiga National Health Insurance Organizations. Medical costs of obese Japanese: a 10-year follow-up study of National Health Insurance in Shiga, Japan. Eur J Public Health . 2007; 17: 424– 429. doi: 10.1093/eurpub/ckl271 Google Scholar CrossRef Search ADS PubMed  22. Tsuji I, Takahashi K, Nishino Yet al.   Impact of walking upon medical care expenditure in Japan: the Ohsaki Cohort Study. Int J Epidemiol . 2003; 32: 809– 814. Google Scholar CrossRef Search ADS PubMed  23. Nakamura K, Okamura T, Hayakawa T, Kanda H, Okayama A, Ueshima H; Health Promotion Research Committee of the Shiga National Health Insurance Organizations. Medical expenditures of men with hypertension and/or a smoking habit: a 10-year follow-up study of National Health Insurance in Shiga, Japan. Hypertens Res . 2010; 33: 802– 807. doi: 10.1038/hr.2010.81 Google Scholar CrossRef Search ADS PubMed  24. Ministry of Health, Labour and Welfare. The comprehensive reform of social security and tax. http://www.mhlw.go.jp/english/social_security/kaikaku.html. Accessed March, 2018. 25. Hashimoto H, Horiguchi H, Matsuda S. Micro data analysis of medical and long-term care utilization among the elderly in Japan. Int J Environ Res Public Health . 2010; 7: 3022– 3037. doi: 10.3390/ijerph7083022 Google Scholar CrossRef Search ADS PubMed  26. Spillman BC, Lubitz J. The effect of longevity on spending for acute and long-term care. N Engl J Med . 2000; 342: 1409– 1415. doi: 10.1056/NEJM200005113421906 Google Scholar CrossRef Search ADS PubMed  27. Ishizaki T, Shimmei M, Fukuda Het al.   Cumulative number of hospital bed days among older adults in the last year of life: a retrospective cohort study. Geriatr Gerontol Int . 2017; 17: 737– 743. doi: 10.1111/ggi.12777 Google Scholar CrossRef Search ADS PubMed  28. Tanuseputro P, Wodchis WP, Fowler Ret al.   The health care cost of dying: a population-based retrospective cohort study of the last year of life in Ontario, Canada. PLoS One . 2015; 10: e0121759. doi: 10.1371/journal.pone.0121759 Google Scholar CrossRef Search ADS PubMed  29. Jones B, Nagin D, Roeder K. A SAS procedure based on mixture modelling for estimating developmental trajectories. Social Methods Res . 2001; 29: 374– 393. Google Scholar CrossRef Search ADS   30. Hines KE, Middendorf TR, Aldrich RW. Determination of parameter identifiability in nonlinear biophysical models: a Bayesian approach. J Gen Physiol . 2014; 143: 401– 416. doi: 10.1085/jgp.201311116 Google Scholar CrossRef Search ADS PubMed  31. Shinkai S, Fujita K, Fujiwara Yet al.   Prognosis of different types of homeboundness among community-living older adults: two-year prospective study [in Japanese]. Nihon Koshu Eisei Zasshi. Japanese J Public Health . 2005; 52: 627– 638. 32. Mossey JM, Shapiro E. Self-rated health: a predictor of mortality among the elderly. Am J Public Health . 1982; 72: 800– 808. Google Scholar CrossRef Search ADS PubMed  33. Fine JP, Gray RJ. A proportional hazards model for the subdistribution of a competing risk. J Am Stat Assoc . 1999; 94: 496– 509. doi: 10.1080/01621459.1999.10474144 Google Scholar CrossRef Search ADS   34. World Health Organization. Active ageing: a policy framework. http://apps.who.int/iris/bitstream/10665/67215/1/WHO_NMH_NPH_02.8.pdf. Accessed March, 2018. 35. Murakami K, Tsubota-Utsugi M, Satoh Met al.   Impaired higher-level functional capacity as a predictor of stroke in community-dwelling older adults: the Ohasama study. Stroke . 2016; 47: 323– 328. doi: 10.1161/STROKEAHA.115.011131 Google Scholar CrossRef Search ADS PubMed  36. Lo AX, Donnelly JP, McGwin GJr, Bittner V, Ahmed A, Brown CJ. Impact of gait speed and instrumental activities of daily living on all-cause mortality in adults ≥65 years with heart failure. Am J Cardiol . 2015; 115: 797– 801. doi: 10.1016/j.amjcard.2014.12.044 Google Scholar CrossRef Search ADS PubMed  37. Scott WK, Macera CA, Cornman CB, Sharpe PA. Functional health status as a predictor of mortality in men and women over 65. J Clin Epidemiol . 1997; 50: 291– 296. Google Scholar CrossRef Search ADS PubMed  38. Leibson CL, Long KH, Ransom JEet al.   Direct medical costs and source of cost differences across the spectrum of cognitive decline: a population-based study. Alzheimers Dement . 2015; 11: 917– 932. doi: 10.1016/j.jalz.2015.01.007 Google Scholar CrossRef Search ADS PubMed  39. Ponzetto M, Maero B, Maina Pet al.   Risk factors for early and late mortality in hospitalized older patients: the continuing importance of functional status. J Gerontol A Biol Sci Med Sci . 2003; 58: 1049– 1054. Google Scholar CrossRef Search ADS PubMed  40. Hasuo Y, Ueda K, Kiyohara Yet al.   Accuracy of diagnosis on death certificates for underlying causes of death in a long-term autopsy-based population study in Hisayama, Japan; with special reference to cardiovascular diseases. J Clin Epidemiol . 1989; 42: 577– 584. Google Scholar CrossRef Search ADS PubMed  © The Author(s) 2018. Published by Oxford University Press on behalf of The Gerontological Society of America. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png The Journals of Gerontology Series A: Biomedical Sciences and Medical Sciences Oxford University Press

Loading next page...
 
/lp/ou_press/association-of-trajectories-of-higher-level-functional-capacity-with-sU8LbZqcF1
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of The Gerontological Society of America. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.
ISSN
1079-5006
eISSN
1758-535X
D.O.I.
10.1093/gerona/gly024
Publisher site
See Article on Publisher Site

Abstract

Abstract Background Higher-level functional capacity is crucial component for independent living in later life. We used repeated-measures analysis to identify aging trajectories in higher-level functional capacity. We then determined whether these trajectories were associated with all-cause mortality and examined differences in medical and long-term care costs between trajectories among community-dwelling older Japanese. Methods 2,675 adults aged 65–90 years participated in annual geriatric health assessments and biennial health monitoring surveys during the period from October 2001 through August 2011. The average number of follow-up assessments was 4.0, and the total number of observations was 10,609. Higher-level functional capacity, which correspond to the fourth and fifth sublevels of Lawton’s hierarchical model, was assessed with the Tokyo Metropolitan Institute of Gerontology-Index of Competence (TMIG-IC). Results We identified four distinct trajectory patterns (high-stable, late-onset decreasing, early-onset decreasing, and low-decreasing) on the TMIG-IC through age 65–90 years. As compared with the high-stable trajectory group, participants in the late-onset decreasing, early-onset decreasing, and low-decreasing TMIG-IC trajectory groups had adjusted hazard ratios for mortality of 1.22 (95% confidence interval: 1.01–1.47), 1.90 (1.53–2.36), and 2.87 (2.14–3.84), respectively. Participants with high-stable and late-onset decreasing higher-level functional capacity trajectories had lower mean monthly medical costs and long-term care costs. In contrast, mean total costs were higher for those with low-decreasing trajectories, after excluding the large increase in such costs at the end of life. Conclusions People with a low-decreasing aging trajectory in higher-level functional capacity had higher risks of death and had high monthly total costs. Higher level functional capacity, Trajectories, Mortality, Medical costs, Long-term care costs The World Health Organization Scientific Group on the Epidemiology of Aging proposed that autonomy or independence in functioning be used as a health index for the elderly (1,2). Functional capacity has been recognized as a hierarchical framework. Lawton (3) defined and systemized seven intercorrelated “sublevels” of competence, namely (in ascending order of complexity), life maintenance, functional health, perception and cognition, physical self-maintenance, instrumental self-maintenance, effectance, and social role. Recent studies of functional capacity (4,5) and disability (6,7) in Western countries analyzed activities of daily living (ADL) and instrumental ADL (IADL), which correspond to the fourth and fifth sublevels of Lawton’s hierarchical model. However, because higher-level functional capacity predicts lower-level function (8), maintenance of such capacity is a crucial component for independent living in later life. The higher sublevels of Lawton’s hierarchical model thus need to be assessed when examining the functional capacity of elders. The Tokyo Metropolitan Institute of Gerontology–Index of Competence (TMIG-IC) (9) was developed and validated to measure the last three sublevels (instrumental self-maintenance, effectance, and social role) of Lawton’s model and has been used to assess higher-level functional capacity in previous studies (8,10–15). An analysis of a nationally representative sample of Japanese (16) found that most elders living in the community had good higher-level functional capacity and that this capacity tended to decline with age. However, to our knowledge no study has used multiple-repeated-measures data from a large-scale longitudinal study to investigate aging trajectories in higher-level functional capacity and identify the population with a lower aging trajectory pattern. Takata et al. (17) reported a longitudinal association between initial higher-level functional capacity assessed by TMIG-IC and all-cause, cardiovascular, and pneumonia deaths among community-dwelling adults aged 80 years. Lubitz et al. (18) used ADL and IADL to examine functional state at age 70 years and linked cumulative health care expenditures by using the Medicare Current Beneficiary Survey from age 70 years until death. Their findings suggest that elders in the lower aging trajectory of higher-level functional capacity have a higher all-cause and cause-specific mortality risk and greater medical and/or long-term care costs than do those with normal aging-related declines. However, existing data are insufficient to confirm this hypothesis. Japan quickly became a world leader in several health metrics, including longevity; however, the country faces challenges related to social health insurance and Long-Term Care Insurance (LTCI). Use of trajectories of higher-level functional capacity might help identify mortality risk and socioeconomic effects and yield new insights that advance health promotion and preventive care initiatives. This prospective study of community-dwelling older adults used repeated-measures data on TMIG-IC from an 11-year longitudinal study launched in Kusatsu Town, a rural community in Japan. We examined local registries to ascertain deaths from any cause and linked these data with Japanese national vital statistics. Then, we obtained data on medical expenses and care expenditures, which encompassed almost all medical provider fees and all care provider fees for the elderly. This study had three objectives: to identify aging trajectories in higher-level functional capacity of community-dwelling older Japanese, to determine whether these trajectories were associated with all-cause and cause-specific mortality, and to examine differences in medical and long-term care costs between aging trajectories of higher-level functional capacity. Methods Participants In collaboration with the local government of Kusatsu Town, Gunma Prefecture, Japan, we launched a longitudinal study of aging and health in 2001. In addition to an annual preventive health check-up for residents aged 40 years or older, participants aged 70 years or older (aged 65 years or older after 2006) underwent a geriatric assessment from 2002 through 2011. Moreover, all residents aged 70 years or older (aged 65 years or older after 2007) underwent biennial health monitoring surveys from 2001 through 2011 (response rate: minimum 91.0%, maximum 98.2%). All annual assessments and biennial health surveys were performed in the same manner. The details of the study design have been previously reported (11–15,19). All participants undergoing geriatric assessment provided written informed consent under conditions approved by the Ethics Committee at Tokyo Metropolitan Institute of Gerontology. The data source for the present study was 2,675 adults aged 65–90 years who underwent surveys conducted during the period from October 2001 through August 2011. We mainly used data from health monitoring surveys in 2001, 2003, 2005, 2007, 2009, and 2011 and used complementary data from biennial geriatric assessments during 2002 through 2010. To be eligible for the study, individuals had to complete TMIG-IC assessments. The average number of follow-up assessments was 4.0, and the total number of observations was 10,609 during the follow-up period. Measurement of TMIG-IC The TMIG-IC (9,20) is designed to measure higher-level functional capacity in community-dwelling older residents. It includes a multidimensional 13-item index of competence (see the Supplementary Appendix). The score ranges from 0 to 13, and lower scores indicate lower functional capacity. The TMIG-IC covers three sublevels of Lawton’s model: instrumental self-maintenance, effectance, and social role. A previous study confirmed that the TMIG-IC has high reliability—as indicated by Cronbach’s alpha, test–retest reliability, and the correlation between the second-order factor score and total score—and high construct, discriminant, and predictive validity (9). Mortality We examined local registries to ascertain deaths from any cause and linked these data with Japanese national vital statistics during the period through December 2015. The underlying cause of death was coded by using the International Classification of Diseases, Tenth Edition (ICD-10). The relevant ICD-10 codes were I00 to I99 for cardiovascular disease (CVD) and C00 to C97 for cancer. Medical and Long-Term Care Costs In Japan, all citizens have access to medical care and long-term care coverage under a universal health insurance system. The official medical insurance system comprises two categories. One is for employees and their dependents, and the other is the National Health Insurance (NHI) and health insurance for older people, which includes farmers, fisherman, and the self-employed, as well as retirees and pensioners, as beneficiaries. Citizens of Japan are automatically enrolled in the health insurance for older people program on their 75th birthday. The NHI and health insurance for older people cover almost all medical treatment and medical provider fees. Payments from insured persons to medical providers are made on a fee-for-service basis, in which the price of each service is determined by a uniform national fee schedule (21–23). The Japanese LTCI system was established to support the need for long-term care services, community-based services, and in-facility services (24). All primary insured persons aged 65 years or older are candidates for care, and secondary insured persons aged 40–64 years with any of 15 specific diseases can also utilize care services. Using data from the NHI, health insurance for older people, and LTCI beneficiaries in Kusatsu Town, we calculated monthly medical costs, monthly long-term care costs, and the sum of these costs as total costs for each participant for the period 1 year before the follow-up survey from 2001 through 2011 (25,26). Previous studies reported that hospital inpatient care and the costs of inpatient and long-term care increased at the end of life (27,28). To exclude dramatic increases in medical and care expenditure, monthly medical and long-term care costs were calculated for the period 1 year before the follow-up survey. However, for participants who died within 1 year of survey completion, costs were calculated for the 1-year interval from 2 years to 1 year before the date of the participant’s death. Costs are expressed in US dollars (1 US dollar = 112 Japanese yen on February 8, 2017). Statistical Analyses First, we identified TMIG-IC trajectory groups for the study period (from October 2001 through August 2011). We used a latent class group-based trajectory modeling approach implemented in the SAS macro PROC TRAJ (29), which assumes that a study population comprises a mixture of finite latent groups within which people follow an approximately homogeneous TMIG-IC trajectory. The assignment probability and parameters for a polynomial trajectory model for each latent group were simultaneously estimated via maximum likelihood for the mixture distribution (14). The number of latent groups and functional form of trajectory models (linear, quadratic, or cubic) were determined by comparing the Bayesian information criterion. After fitting the model, posterior probabilities for each group membership assignment were calculated for each individual, and participants were assigned to the groups with the highest posterior probabilities. Posterior probabilities were also used to assess model adequacy (ie, correctness of classification), by examining whether the posterior probability for an assigned group was sufficiently high as compared with those of the remaining groups (30). We express TMIG-IC trajectories in percentages, which represent the proportions of those assigned to each group, according to the highest posterior probability, rather than the default figure outputs in PROC TRAJ. Second, we examined associations of TMIG-IC trajectory group with mortality by using Cox proportional hazards models that controlled for potential confounders such as baseline sex, age, frequency of going outdoors (31), and self-rated health (32). The underlying time-scale in Cox proportional hazards models starts from first participation (during the period from October 2001 through August 2011) until death or the end of follow-up (December 2015), whichever came first. Because CVD, cancer, and other causes of death can be regarded as competing events, we used the competing-risk subdistribution regression of Fine and Gray (33) to determine hazards of cause-specific death. Finally, trajectory group-specific age trends in medical and long-term care costs were modeled by a generalized estimating equation with negative binomial or Poisson for sex and a linear term for age. Statistical analyses were conducted with SPSS (version 18.0; SPSS, Inc., Chicago, IL) and SAS (version 9.4; SAS Institute, Inc., Cary, NC), and a p value of less than .05 was considered to indicate statistical significance. Results Data from the baseline survey showed that the mean (SD) age of participants was 72.0 (6.2) years; 56.8% were women, 83.7% were able to go out by themselves, and 9.3% were independent in their home or neighborhood but were unable to go far by themselves. Self-rated health was very good or good for 78.7%. Mean (SD) TMIG-IC score was 11.1 (2.8). We identified four TMIG-IC trajectory patterns: 36.3% of participants were in the high-stable trajectory, 40.1% in the late-onset decreasing trajectory, 17.4% in the early-onset decreasing trajectory, and 6.1% in the low-decreasing trajectory (Figure 1; Table 1). The TMIG-IC score of the high-stable trajectory group was close to the maximum until they reached an age of approximately 85 years. Mean TMIG-IC scores in the late-onset decreasing and early-onset decreasing trajectory groups were 11.5 at age 65. The TMIG-IC score in the late-onset decreasing trajectory group remained relatively constant until around age 80 years but gradually decreased in the early-onset decreasing trajectory group after age 65 years. Mean TMIG-IC score in the low-decreasing trajectory group was approximately half that of the high-stable trajectory group at age 65 and showed a linear decline until around age 85. The average posterior probability of allocating each participant into the four groups was 0.73–0.87, indicating a good fit of the model of group trajectories to individual trajectories. Figure 1. View largeDownload slide TMIG-IC trajectories: a group-based semiparametric mixture model. The solid lines are estimated values. Figure 1. View largeDownload slide TMIG-IC trajectories: a group-based semiparametric mixture model. The solid lines are estimated values. Table 1. Estimated TMIG-IC Values for the Four Trajectories, by Age TMIG-IC Trajectory  Age, Years  65  70  75  80  85  90  High-stable group (n = 972; 36.3%)  12.8 (12.7–13.0)  12.9 (12.8–13.0)  12.9 (12.8–12.9)  12.7 (12.7–12.8)  12.3 (12.2–12.3)  11.0 (10.6–11.5)  Late-onset decreasing group (n = 1,074; 40.1%)  11.5 (11.2–11.9)  12.0 (11.8–12.1)  11.9 (11.8–12.0)  11.3 (11.0–11.5)  9.7 (9.4–10.0)  7.0 (6.6–7.5)  Early-onset decreasing group (n = 465; 17.4%)  11.5 (10.8–12.1)  10.8 (10.4–11.2)  9.6 (9.2–10.0)  7.7 (7.3–8.0)  5.2 (4.8–5.5)  2.3 (1.6–3.0)  Low-decreasing group (n = 164; 6.1%)  7.7 (5.6–9.8)  6.1 (5.3–6.8)  4.3 (3.8–4.9)  2.6 (1.9–3.4)  1.2 (0.2–2.2)  0.4 (0–0.9)  TMIG-IC Trajectory  Age, Years  65  70  75  80  85  90  High-stable group (n = 972; 36.3%)  12.8 (12.7–13.0)  12.9 (12.8–13.0)  12.9 (12.8–12.9)  12.7 (12.7–12.8)  12.3 (12.2–12.3)  11.0 (10.6–11.5)  Late-onset decreasing group (n = 1,074; 40.1%)  11.5 (11.2–11.9)  12.0 (11.8–12.1)  11.9 (11.8–12.0)  11.3 (11.0–11.5)  9.7 (9.4–10.0)  7.0 (6.6–7.5)  Early-onset decreasing group (n = 465; 17.4%)  11.5 (10.8–12.1)  10.8 (10.4–11.2)  9.6 (9.2–10.0)  7.7 (7.3–8.0)  5.2 (4.8–5.5)  2.3 (1.6–3.0)  Low-decreasing group (n = 164; 6.1%)  7.7 (5.6–9.8)  6.1 (5.3–6.8)  4.3 (3.8–4.9)  2.6 (1.9–3.4)  1.2 (0.2–2.2)  0.4 (0–0.9)  Note: Values are averages (95% confidence interval). View Large Table 1. Estimated TMIG-IC Values for the Four Trajectories, by Age TMIG-IC Trajectory  Age, Years  65  70  75  80  85  90  High-stable group (n = 972; 36.3%)  12.8 (12.7–13.0)  12.9 (12.8–13.0)  12.9 (12.8–12.9)  12.7 (12.7–12.8)  12.3 (12.2–12.3)  11.0 (10.6–11.5)  Late-onset decreasing group (n = 1,074; 40.1%)  11.5 (11.2–11.9)  12.0 (11.8–12.1)  11.9 (11.8–12.0)  11.3 (11.0–11.5)  9.7 (9.4–10.0)  7.0 (6.6–7.5)  Early-onset decreasing group (n = 465; 17.4%)  11.5 (10.8–12.1)  10.8 (10.4–11.2)  9.6 (9.2–10.0)  7.7 (7.3–8.0)  5.2 (4.8–5.5)  2.3 (1.6–3.0)  Low-decreasing group (n = 164; 6.1%)  7.7 (5.6–9.8)  6.1 (5.3–6.8)  4.3 (3.8–4.9)  2.6 (1.9–3.4)  1.2 (0.2–2.2)  0.4 (0–0.9)  TMIG-IC Trajectory  Age, Years  65  70  75  80  85  90  High-stable group (n = 972; 36.3%)  12.8 (12.7–13.0)  12.9 (12.8–13.0)  12.9 (12.8–12.9)  12.7 (12.7–12.8)  12.3 (12.2–12.3)  11.0 (10.6–11.5)  Late-onset decreasing group (n = 1,074; 40.1%)  11.5 (11.2–11.9)  12.0 (11.8–12.1)  11.9 (11.8–12.0)  11.3 (11.0–11.5)  9.7 (9.4–10.0)  7.0 (6.6–7.5)  Early-onset decreasing group (n = 465; 17.4%)  11.5 (10.8–12.1)  10.8 (10.4–11.2)  9.6 (9.2–10.0)  7.7 (7.3–8.0)  5.2 (4.8–5.5)  2.3 (1.6–3.0)  Low-decreasing group (n = 164; 6.1%)  7.7 (5.6–9.8)  6.1 (5.3–6.8)  4.3 (3.8–4.9)  2.6 (1.9–3.4)  1.2 (0.2–2.2)  0.4 (0–0.9)  Note: Values are averages (95% confidence interval). View Large Local registries showed 747 (27.9%) incident deaths among the 2,675 participants. The median duration of follow-up for incident death was 2,915 days. Among the 747 incident deaths, 738 (98.8%) were linked with Japanese national vital statistics. As compared with the high-stable TMIG-IC trajectory group during follow-up, the late-onset decreasing, early-onset decreasing, and low-decreasing trajectory groups had hazard ratios (95% confidence interval) of 1.29 (1.07–1.56), 2.33 (1.90–2.84), and 4.67 (3.65–5.99), respectively, for all-cause mortality (Table 2). Even after adjustment for several demographic and health characteristics, the independent association with TMIG-IC trajectory group remained significant: the late-onset decreasing, early-onset decreasing, and low-decreasing trajectory groups had hazard ratios of 1.22 (1.01–1.47), 1.90 (1.53–2.36), and 2.87 (2.14–3.84), respectively, for all-cause mortality. The analysis of cause-specific death, CVD mortality, and other mortality showed significant associations between the four TMIG-IC trajectory patterns. Table 2. Independent Associations of TMIG-IC Trajectory with All-Cause and Cause-Specific Mortality in Community-Dwelling Japanese Aged ≥65 Years TMIG-IC Trajectory  Incident All-Cause Deaths  All-Cause Mortality  CVD Mortality  Cancer Mortality  Other Mortality  Crude HR (95%CI)  Adjusted HR (95% CI)  Crude HR (95% CI)  Adjusted HR (95% CI)  Crude HR (95% CI)  Adjusted HR (95% CI)  Crude HR (95% CI)  Adjusted HR (95% CI)  High-stable group† (n = 972; 36.3%)  191 (19.7%)  1  1  1  1  1  1  1  1  Late-onset decreasing group (n = 1,074; 40.1%)  267 (24.9%)  1.29 (1.07–1.56)**  1.22 (1.01–1.47)*  1.54 (1.09–2.17)*  1.40 (1.01–2.01)*  0.96 (0.68–1.37)  0.93 (0.65–1.33)  1.30 (0.98–1.73)  1.10 (0.82–1.47)  Early-onset decreasing group (n = 465; 17.4%)  195 (41.9%)  2.33 (1.90–2.84)**  1.90 (1.53–2.36)**  2.84 (1.99–4.06)**  2.41 (1.65–3.53)**  1.29 (0.86–1.94)  1.20 (0.79–1.84)  2.13 (1.56–2.90)**  1.52 (1.08–2.15)*  Low-decreasing group (n = 164; 6.1%)  94 (57.3%)  4.67 (3.65–5.99)**  2.87 (2.14–3.84)**  2.65 (1.60–4.40)**  1.97 (1.06–3.66)*  1.98 (1.17–3.36)*  1.67 (0.92–3.05)  4.96 (3.45–7.11)**  3.54 (2.16–5.80)**  TMIG-IC Trajectory  Incident All-Cause Deaths  All-Cause Mortality  CVD Mortality  Cancer Mortality  Other Mortality  Crude HR (95%CI)  Adjusted HR (95% CI)  Crude HR (95% CI)  Adjusted HR (95% CI)  Crude HR (95% CI)  Adjusted HR (95% CI)  Crude HR (95% CI)  Adjusted HR (95% CI)  High-stable group† (n = 972; 36.3%)  191 (19.7%)  1  1  1  1  1  1  1  1  Late-onset decreasing group (n = 1,074; 40.1%)  267 (24.9%)  1.29 (1.07–1.56)**  1.22 (1.01–1.47)*  1.54 (1.09–2.17)*  1.40 (1.01–2.01)*  0.96 (0.68–1.37)  0.93 (0.65–1.33)  1.30 (0.98–1.73)  1.10 (0.82–1.47)  Early-onset decreasing group (n = 465; 17.4%)  195 (41.9%)  2.33 (1.90–2.84)**  1.90 (1.53–2.36)**  2.84 (1.99–4.06)**  2.41 (1.65–3.53)**  1.29 (0.86–1.94)  1.20 (0.79–1.84)  2.13 (1.56–2.90)**  1.52 (1.08–2.15)*  Low-decreasing group (n = 164; 6.1%)  94 (57.3%)  4.67 (3.65–5.99)**  2.87 (2.14–3.84)**  2.65 (1.60–4.40)**  1.97 (1.06–3.66)*  1.98 (1.17–3.36)*  1.67 (0.92–3.05)  4.96 (3.45–7.11)**  3.54 (2.16–5.80)**  Notes: CVD, cardiovascular disease; CI, confidence interval; HR, hazard ratio; TMIG-IC, Tokyo Metropolitan Institute of Gerontology Index of Competence. †Reference group. Cox hazards regression models were run separately. Adjusted for sex, age, frequency of going outdoors, and self-rated health. *p < .05, **p < .01. View Large Table 2. Independent Associations of TMIG-IC Trajectory with All-Cause and Cause-Specific Mortality in Community-Dwelling Japanese Aged ≥65 Years TMIG-IC Trajectory  Incident All-Cause Deaths  All-Cause Mortality  CVD Mortality  Cancer Mortality  Other Mortality  Crude HR (95%CI)  Adjusted HR (95% CI)  Crude HR (95% CI)  Adjusted HR (95% CI)  Crude HR (95% CI)  Adjusted HR (95% CI)  Crude HR (95% CI)  Adjusted HR (95% CI)  High-stable group† (n = 972; 36.3%)  191 (19.7%)  1  1  1  1  1  1  1  1  Late-onset decreasing group (n = 1,074; 40.1%)  267 (24.9%)  1.29 (1.07–1.56)**  1.22 (1.01–1.47)*  1.54 (1.09–2.17)*  1.40 (1.01–2.01)*  0.96 (0.68–1.37)  0.93 (0.65–1.33)  1.30 (0.98–1.73)  1.10 (0.82–1.47)  Early-onset decreasing group (n = 465; 17.4%)  195 (41.9%)  2.33 (1.90–2.84)**  1.90 (1.53–2.36)**  2.84 (1.99–4.06)**  2.41 (1.65–3.53)**  1.29 (0.86–1.94)  1.20 (0.79–1.84)  2.13 (1.56–2.90)**  1.52 (1.08–2.15)*  Low-decreasing group (n = 164; 6.1%)  94 (57.3%)  4.67 (3.65–5.99)**  2.87 (2.14–3.84)**  2.65 (1.60–4.40)**  1.97 (1.06–3.66)*  1.98 (1.17–3.36)*  1.67 (0.92–3.05)  4.96 (3.45–7.11)**  3.54 (2.16–5.80)**  TMIG-IC Trajectory  Incident All-Cause Deaths  All-Cause Mortality  CVD Mortality  Cancer Mortality  Other Mortality  Crude HR (95%CI)  Adjusted HR (95% CI)  Crude HR (95% CI)  Adjusted HR (95% CI)  Crude HR (95% CI)  Adjusted HR (95% CI)  Crude HR (95% CI)  Adjusted HR (95% CI)  High-stable group† (n = 972; 36.3%)  191 (19.7%)  1  1  1  1  1  1  1  1  Late-onset decreasing group (n = 1,074; 40.1%)  267 (24.9%)  1.29 (1.07–1.56)**  1.22 (1.01–1.47)*  1.54 (1.09–2.17)*  1.40 (1.01–2.01)*  0.96 (0.68–1.37)  0.93 (0.65–1.33)  1.30 (0.98–1.73)  1.10 (0.82–1.47)  Early-onset decreasing group (n = 465; 17.4%)  195 (41.9%)  2.33 (1.90–2.84)**  1.90 (1.53–2.36)**  2.84 (1.99–4.06)**  2.41 (1.65–3.53)**  1.29 (0.86–1.94)  1.20 (0.79–1.84)  2.13 (1.56–2.90)**  1.52 (1.08–2.15)*  Low-decreasing group (n = 164; 6.1%)  94 (57.3%)  4.67 (3.65–5.99)**  2.87 (2.14–3.84)**  2.65 (1.60–4.40)**  1.97 (1.06–3.66)*  1.98 (1.17–3.36)*  1.67 (0.92–3.05)  4.96 (3.45–7.11)**  3.54 (2.16–5.80)**  Notes: CVD, cardiovascular disease; CI, confidence interval; HR, hazard ratio; TMIG-IC, Tokyo Metropolitan Institute of Gerontology Index of Competence. †Reference group. Cox hazards regression models were run separately. Adjusted for sex, age, frequency of going outdoors, and self-rated health. *p < .05, **p < .01. View Large Among 2,675 participants, 1,882 (70.4%) incurred monthly medical costs and 379 (14.2%) incurred monthly long-term care costs during the 1-year period before the follow-up survey. The generalized-estimating-equation models showed significant TMIG-IC trajectory-specific age trends in monthly medical and long-term care costs (Table 3; Figures 2 and 3). Mean monthly medical costs in the low-decreasing TMIG-IC trajectory group were estimated at $506.6 and were three times those of the high-stable group at age 65 but declined after age 65 years. The other three TMIG-IC trajectory groups exhibited parallel increases in monthly medical costs after age 65 years. Mean monthly long-term care costs in the low-decreasing TMIG-IC trajectory group were estimated at $275.5 and were obviously higher than in the other three groups at age 65. The low-decreasing TMIG-IC trajectory group had consistently higher monthly long-term care costs than did the remaining groups from age 65 to 90 years. Interestingly, monthly long-term care costs were similar for the early-onset decreasing TMIG-IC trajectory group and the high-stable and late-onset decreasing trajectory groups at age 65. Later in life, however, the early-onset decreasing trajectory group exhibited large increases, while costs in the high-stable and late-onset decreasing trajectory groups remained low. The mean sum of these costs for the low-decreasing trajectory group was estimated at $683.0, which was approximately four times the values for the high-stable ($145.3), late-onset decreasing ($184.2), and early-onset decreasing ($178.9) group at age 65 (Figure 4). The sum of medical and long-term care costs increased after age 65 years in the four TMIG-IC trajectory groups. In particular, the early-onset decreasing trajectory group exhibited a dramatic increase after age 75 years, and the estimated costs exceeded those of the low-decreasing trajectory group until age 87 years. Mean total costs later in life were lower for participants with high-stable and late-onset decreasing TMIG-IC trajectories. Table 3. Estimated Monthly Medical and Care Costs for the Four TMIG-IC Trajectories, by Age. TMIG-IC Trajectory  Age, Years  65  70  75  80  85  90  Medical Cost (US Dollars)  High-stable group (n = 972; 36.3%)  154.3 (124.2–191.6)  195.7 (167.5–228.5)  248.2 (220.4–279.4)  314.7 (276.5–358.3)  399.2 (333.5–477.7)  506.2 (395.9–647.3)  Late-onset decreasing group (n = 1,074; 40.1%)  203.9 (163.8–253.7)  249.8 (212.8–293.3)  306.2 (270.5–346.6)  375.3 (330.1–426.7)  460.0 (387.7–545.7)  563.7 (447.3–710.4)  Early-onset decreasing group (n = 465; 17.4%)  201.4 (135.5–299.3)  249.2 (187.8–330.5)  308.3 (252.1–377.0)  381.4 (312.9–465.0)  472.0 (358.4–621.6)  584.0 (396.2–860.7)  Low-decreasing group (n = 164; 6.1%)  506.6 (258.1–994.3)  459.5 (302.2–698.7)  416.9 (299.3–580.7)  378.2 (227.7–628.2)  343.1 (156.2–753.4)  311.3 (104.4–928.0)  Care cost (US dollars)  High-stable group (n = 972; 36.3%)  1.3 (0.4–3.7)  3.3 (1.4–8.0)  8.7 (4.3–17.6)  22.5 (12.8–39.3)  58.3 (37.4–90.9)  151.2 (101.8–224.7)  Late-onset decreasing group (n = 1,074; 40.1%)  5.8 (3.2–10.8)  12.3 (7.7–19.8)  26.5 (18.3–38.3)  56.7 (42.3–76.0)  121.6 (93.7–157.9)  260.7 (194.8–349.0)  Early-onset decreasing group (n = 465; 17.4%)  41.5 (25.8–66.8)  76.7 (52.7–111.7)  141.9 (106.8–188.5)  262.3 (211.4–325.5)  485.0 (398.9–589.6)  896.7 (708.2–1135.3)  Low-decreasing group (n = 164; 6.1%)  275.5 (174.3–435.5)  361.3 (256.7–508.6)  473.8 (369.8–607.0)  621.3 (504.7–764.8)  814.8 (635.0–1045.4)  1068.4 (757.5–1507.1)  TMIG-IC Trajectory  Age, Years  65  70  75  80  85  90  Medical Cost (US Dollars)  High-stable group (n = 972; 36.3%)  154.3 (124.2–191.6)  195.7 (167.5–228.5)  248.2 (220.4–279.4)  314.7 (276.5–358.3)  399.2 (333.5–477.7)  506.2 (395.9–647.3)  Late-onset decreasing group (n = 1,074; 40.1%)  203.9 (163.8–253.7)  249.8 (212.8–293.3)  306.2 (270.5–346.6)  375.3 (330.1–426.7)  460.0 (387.7–545.7)  563.7 (447.3–710.4)  Early-onset decreasing group (n = 465; 17.4%)  201.4 (135.5–299.3)  249.2 (187.8–330.5)  308.3 (252.1–377.0)  381.4 (312.9–465.0)  472.0 (358.4–621.6)  584.0 (396.2–860.7)  Low-decreasing group (n = 164; 6.1%)  506.6 (258.1–994.3)  459.5 (302.2–698.7)  416.9 (299.3–580.7)  378.2 (227.7–628.2)  343.1 (156.2–753.4)  311.3 (104.4–928.0)  Care cost (US dollars)  High-stable group (n = 972; 36.3%)  1.3 (0.4–3.7)  3.3 (1.4–8.0)  8.7 (4.3–17.6)  22.5 (12.8–39.3)  58.3 (37.4–90.9)  151.2 (101.8–224.7)  Late-onset decreasing group (n = 1,074; 40.1%)  5.8 (3.2–10.8)  12.3 (7.7–19.8)  26.5 (18.3–38.3)  56.7 (42.3–76.0)  121.6 (93.7–157.9)  260.7 (194.8–349.0)  Early-onset decreasing group (n = 465; 17.4%)  41.5 (25.8–66.8)  76.7 (52.7–111.7)  141.9 (106.8–188.5)  262.3 (211.4–325.5)  485.0 (398.9–589.6)  896.7 (708.2–1135.3)  Low-decreasing group (n = 164; 6.1%)  275.5 (174.3–435.5)  361.3 (256.7–508.6)  473.8 (369.8–607.0)  621.3 (504.7–764.8)  814.8 (635.0–1045.4)  1068.4 (757.5–1507.1)  Note: Values are averages (95% confidence interval). View Large Table 3. Estimated Monthly Medical and Care Costs for the Four TMIG-IC Trajectories, by Age. TMIG-IC Trajectory  Age, Years  65  70  75  80  85  90  Medical Cost (US Dollars)  High-stable group (n = 972; 36.3%)  154.3 (124.2–191.6)  195.7 (167.5–228.5)  248.2 (220.4–279.4)  314.7 (276.5–358.3)  399.2 (333.5–477.7)  506.2 (395.9–647.3)  Late-onset decreasing group (n = 1,074; 40.1%)  203.9 (163.8–253.7)  249.8 (212.8–293.3)  306.2 (270.5–346.6)  375.3 (330.1–426.7)  460.0 (387.7–545.7)  563.7 (447.3–710.4)  Early-onset decreasing group (n = 465; 17.4%)  201.4 (135.5–299.3)  249.2 (187.8–330.5)  308.3 (252.1–377.0)  381.4 (312.9–465.0)  472.0 (358.4–621.6)  584.0 (396.2–860.7)  Low-decreasing group (n = 164; 6.1%)  506.6 (258.1–994.3)  459.5 (302.2–698.7)  416.9 (299.3–580.7)  378.2 (227.7–628.2)  343.1 (156.2–753.4)  311.3 (104.4–928.0)  Care cost (US dollars)  High-stable group (n = 972; 36.3%)  1.3 (0.4–3.7)  3.3 (1.4–8.0)  8.7 (4.3–17.6)  22.5 (12.8–39.3)  58.3 (37.4–90.9)  151.2 (101.8–224.7)  Late-onset decreasing group (n = 1,074; 40.1%)  5.8 (3.2–10.8)  12.3 (7.7–19.8)  26.5 (18.3–38.3)  56.7 (42.3–76.0)  121.6 (93.7–157.9)  260.7 (194.8–349.0)  Early-onset decreasing group (n = 465; 17.4%)  41.5 (25.8–66.8)  76.7 (52.7–111.7)  141.9 (106.8–188.5)  262.3 (211.4–325.5)  485.0 (398.9–589.6)  896.7 (708.2–1135.3)  Low-decreasing group (n = 164; 6.1%)  275.5 (174.3–435.5)  361.3 (256.7–508.6)  473.8 (369.8–607.0)  621.3 (504.7–764.8)  814.8 (635.0–1045.4)  1068.4 (757.5–1507.1)  TMIG-IC Trajectory  Age, Years  65  70  75  80  85  90  Medical Cost (US Dollars)  High-stable group (n = 972; 36.3%)  154.3 (124.2–191.6)  195.7 (167.5–228.5)  248.2 (220.4–279.4)  314.7 (276.5–358.3)  399.2 (333.5–477.7)  506.2 (395.9–647.3)  Late-onset decreasing group (n = 1,074; 40.1%)  203.9 (163.8–253.7)  249.8 (212.8–293.3)  306.2 (270.5–346.6)  375.3 (330.1–426.7)  460.0 (387.7–545.7)  563.7 (447.3–710.4)  Early-onset decreasing group (n = 465; 17.4%)  201.4 (135.5–299.3)  249.2 (187.8–330.5)  308.3 (252.1–377.0)  381.4 (312.9–465.0)  472.0 (358.4–621.6)  584.0 (396.2–860.7)  Low-decreasing group (n = 164; 6.1%)  506.6 (258.1–994.3)  459.5 (302.2–698.7)  416.9 (299.3–580.7)  378.2 (227.7–628.2)  343.1 (156.2–753.4)  311.3 (104.4–928.0)  Care cost (US dollars)  High-stable group (n = 972; 36.3%)  1.3 (0.4–3.7)  3.3 (1.4–8.0)  8.7 (4.3–17.6)  22.5 (12.8–39.3)  58.3 (37.4–90.9)  151.2 (101.8–224.7)  Late-onset decreasing group (n = 1,074; 40.1%)  5.8 (3.2–10.8)  12.3 (7.7–19.8)  26.5 (18.3–38.3)  56.7 (42.3–76.0)  121.6 (93.7–157.9)  260.7 (194.8–349.0)  Early-onset decreasing group (n = 465; 17.4%)  41.5 (25.8–66.8)  76.7 (52.7–111.7)  141.9 (106.8–188.5)  262.3 (211.4–325.5)  485.0 (398.9–589.6)  896.7 (708.2–1135.3)  Low-decreasing group (n = 164; 6.1%)  275.5 (174.3–435.5)  361.3 (256.7–508.6)  473.8 (369.8–607.0)  621.3 (504.7–764.8)  814.8 (635.0–1045.4)  1068.4 (757.5–1507.1)  Note: Values are averages (95% confidence interval). View Large Figure 2. View largeDownload slide TMIG-IC trajectory-specific age trends in monthly medical costs: a generalized estimating equation. The solid lines are estimated values. Figure 2. View largeDownload slide TMIG-IC trajectory-specific age trends in monthly medical costs: a generalized estimating equation. The solid lines are estimated values. Figure 3. View largeDownload slide TMIG-IC trajectory-specific age trends in monthly long-term care costs: a generalized estimating equation. The solid lines are estimated values. Figure 3. View largeDownload slide TMIG-IC trajectory-specific age trends in monthly long-term care costs: a generalized estimating equation. The solid lines are estimated values. Figure 4. View largeDownload slide TMIG-IC trajectory-specific age trends in monthly total costs (medical and long-term care costs): a generalized estimating equation. The solid lines are estimated values. Figure 4. View largeDownload slide TMIG-IC trajectory-specific age trends in monthly total costs (medical and long-term care costs): a generalized estimating equation. The solid lines are estimated values. Discussion This prospective study is the first to show aging trajectories in higher-level functional capacity of community-dwelling older Japanese. The trajectory of higher-level functional capacity was an independent predictor of all-cause, cardiovascular, and non-cancer mortality, and elders with a low-decreasing aging trajectory had higher monthly medical and long-term care costs in later life. Previous studies (4–7) used ADL and IADL to assess functional capacity. These measures correspond to the fourth and fifth sublevels of Lawton’s hierarchical model. TMIG-IC assessment of higher-level functional capacity includes the last three sublevels (instrumental self-maintenance, effectance, and social role) of Lawton’s model. The importance of promoting higher-level functional capacity, especially social participation, is a key proposal for “Active Aging” in the World Health Organization’s policy framework (34). We identified four distinct TMIG-IC trajectory patterns (high-stable, late-onset decreasing, early-onset decreasing, and low-decreasing) among community-dwelling older Japanese aged 65–90 years, and this is the first study to show aging trajectories in higher-level functional capacity in a community-based study. A TMIG-IC score of 10 points or higher indicates normal higher-level function (35). Approximately 80% of Japanese elderly adults (the high-stable and late-onset decreasing trajectory groups) maintained higher-level functional capacity in later life, and 36% of elders were aging successfully. Overall, 17% of elders (the early-onset decreasing trajectory group) exhibited quadratic declines, and 6% of elders (the low-decreasing trajectory group) exhibited impaired higher-level functional capacity and rapid declines after age 65 years. We observed that, as compared with the high-stable TMIG-IC trajectory group, participants in the late-onset decreasing, early-onset decreasing, and low-decreasing trajectory groups had HRs of 1.29, 2.33, and 4.67, respectively, for all-cause mortality. Takata et al. reported that initial TMIG-IC score was associated with all-cause mortality in 697 Japanese aged 80 years (17). Previous studies showed that ADL and IADL were predictors of mortality among community-dwelling older adults in a Western population (36,37). The present prospective study using repeated-measures analysis extends the findings of earlier studies and highlights the importance of intervention for improvements in higher-level functional capacity, even among older adults with low higher-level functional capacity. In an analysis of cause-specific mortality, participants in the lower trajectory groups had high HRs for CVD mortality and non-cancer mortality. Murakami et al. (35) reported that impaired higher-level capacity, as indicated by TMIG-IC at baseline, was associated with incident stroke among community-dwelling older adults with independent basic ADL. Takata et al. (17) reported that higher-level functional capacity, as assessed by TMIG-IC, was associated with cardiovascular and pneumonia mortality but not with cancer mortality. Our results accord with these earlier findings and are the first evidence that potential higher-level functional capacity trajectories are independent predictors of CVD mortality and non-cancer mortality. Our study is the first to find that elders with high-stable and late-onset decreasing higher-level functional capacity trajectories had lower mean monthly medical costs and long-term care costs, while those with low-decreasing trajectories had higher mean total costs, after excluding the dramatic increase in such costs at the end of life. Monthly medical costs in the low-decreasing higher-level functional capacity trajectories tended to decrease with advancing age, but monthly long-term care costs tended to increase after age 65 years. To identify the reason for this cost discrepancy in the low-decreasing TMIG-IC trajectory group, we used data from the geriatric assessment to examine baseline demographic and health characteristics of participants in the four trajectory groups. At baseline, adults in the low-decreasing trajectory group were more likely to have a history of cerebrovascular disease and had higher white blood cell counts (Supplementary Table), which suggests that TMIG-IC trajectory patterns were related to the presence of chronic diseases. Spillman et al. (26) reported that the increase in nursing home costs with advancing age was sufficient to offset the moderating effect of declining medical expenditures in an American population during the last 2 years of life. People with impaired higher level functional capacity, such as the present low-decreasing trajectory group, might have had more chronic diseases and greater mortality risk and may therefore have shifted their needs from medical service to care service in later life. Moreover, in the Mayo Clinic Study of Aging, Leibson et al. (38) reported that annual mean medical costs rose gradually for persons with cognitive normal, mild cognitive impairment, newly discovered dementia, and prevalent dementia. The present study showed that adults in the low-decreasing trajectory group had lower MMSE scores, which suggests that lower trajectories of higher-level functional capacity were associated with lower cognitive function and higher monthly medical and long-term care costs. This study has strengths that warrant mention. First, our large sample of community-dwelling adults allowed us to use a group-based semiparametric mixture model, which showed potential trajectory patterns in TMIG-IC. Previous study focused on ADL and the association with mortality in hospitalized patients (39). A few studies examined the association of higher-level functional capacity with mortality among community-dwelling older adults (17,35). However, the present prospective study used repeated-measures analysis of data from community-dwelling older Japanese to investigate the associations of potential trajectory patterns in higher-level functional capacity with mortality. Second, the repeated-measured data in this study were derived from annual geriatric health assessments and biennial health monitoring surveys. Half of the total of 10,609 observations during the follow-up period were from biennial health monitoring surveys, all of which had high response rates. Third, the data for calculating medical and long-term care costs in this study were derived from the Japanese NHI, health insurance for older people, and LTCI beneficiaries. These systems cover nearly all medical provider fees and all care provider fees. Because Japan has a universal health insurance system, we were able to link aging trajectories in higher-level functional capacity, cause of death, and medical and long-term care costs for community-dwelling older Japanese. We then determined whether these trajectories were associated with all-cause and cause-specific mortality and examined differences in medical and long-term care costs between trajectories among community-dwelling older Japanese. Our findings regarding the socioeconomic effects of trajectory patterns in higher-level functional capacity might help advance policies for health promotion and preventive care initiatives in developed countries with growing social security cost burdens. This study has some limitations. First, a previous study found that elders in a Japanese sample were more likely to be competent than those in a US sample (20). Trajectory patterns of higher-level functional capacity might differ between Japanese and Western populations; thus, future studies should examine the aging patterns of higher-level functional capacity among Western populations. Moreover, the data source for the present study was limited to Kusatsu Town, Japan. Data from other study sites or nationally representative data for Japan are needed in order to confirm the present findings. Second, use of death certificates to classify major causes of death may lead to misclassification. However, this limitation is not unique to the present study, and use of death certificates for this purpose was reported to be accurate in Japan (40). Third, the present study examined baseline demographic and health characteristics of participants in four trajectory groups; however, these data were limited to half the total observations. Future studies should attempt to identify risk markers for the lower aging trajectory in higher-level functional capacity and apply those findings to health promotion and development of programs that reduce social costs. In conclusion, this prospective study found that community-dwelling older adults exhibit four major trajectories in TMIG-IC, a measure of higher-level functional capacity. Approximately 80% of elders (the high-stable and late-onset decreasing trajectory groups) maintained higher-level functional capacity in later life. In contrast, 6% of elders (the low-decreasing trajectory group) exhibited rapid declines after age 65 years and had greater risks for cardiovascular mortality, non-cancer mortality, and all-cause mortality. Furthermore, the low-decreasing trajectory group had higher monthly medical and long-term care costs. These findings indicate that interventions that improve higher-level functional capacity may improve health and longevity, and lessen the socioeconomic burden, among elders. Supplementary Material Supplementary data is available at The Journals of Gerontology, Series A: Biological Sciences and Medical Sciences online. Funding This study was supported by TMIG, JST/RISTEX, Grants-In-Aid for Scientific Research (B) JP20390190, (B) JP21390212, (B) JP24390173, and (B) JP26310111, a Grant-In-Aid for Research Activity Start-up JP24890302, and a Grant-In-Aid for Young Scientists (B) JP15K16539 from the Ministry of Education, Culture, Sports, Science and Technology, Japan. Conflict of Interest The authors have no potential conflicts of interest related to this research. Acknowledgements We thank the staff members of the Public Health Center and town office in Kusatsu and our colleagues on the Research Team for Social Participation and Community Health, Tokyo Metropolitan Institute of Gerontology (TMIG). References 1. The Uses of Epidemiology in the Study of the Elderly. Report of a WHO scientific group on the epidemiology of aging. World Health Organ Tech Rep Ser . 1984; 706: 1– 84. PubMed  2. Ishizaki T, Kai I, Kobayashi Y, Imanaka Y. Functional transitions and active life expectancy for older Japanese living in a community. Arch Gerontol Geriatr . 2002; 35: 107– 120. Google Scholar CrossRef Search ADS PubMed  3. Lawton MP. Assessing the competence of older people. In: Kent DP, Kastenbaum R, Sherwood S, eds. Research Planning and Action for the Elderly: The Power and Potential of Social Science . New York: Human Sciences Press; 1972: 122– 143. 4. Vermeulen J, Neyens JC, van Rossum E, Spreeuwenberg MD, de Witte LP. Predicting ADL disability in community-dwelling elderly people using physical frailty indicators: a systematic review. BMC Geriatr . 2011; 11: 33. doi: 10.1186/1471-2318-11-33 Google Scholar CrossRef Search ADS PubMed  5. Jefferson AL, Paul RH, Ozonoff A, Cohen RA. Evaluating elements of executive functioning as predictors of instrumental activities of daily living (IADLs). Arch Clin Neuropsychol . 2006; 21: 311– 320. doi: 10.1016/j.acn.2006.03.007 Google Scholar CrossRef Search ADS PubMed  6. Buurman BM, Han L, Murphy TEet al.   Trajectories of disability among older persons before and after a hospitalization leading to a skilled nursing facility admission. J Am Med Dir Assoc . 2016; 17: 225– 231. doi: 10.1016/j.jamda.2015.10.010 Google Scholar CrossRef Search ADS PubMed  7. Ferrante LE, Pisani MA, Murphy TE, Gahbauer EA, Leo-Summers LS, Gill TM. Functional trajectories among older persons before and after critical illness. JAMA Intern Med . 2015; 175: 523– 529. doi: 10.1001/jamainternmed.2014.7889 Google Scholar CrossRef Search ADS PubMed  8. Fujiwara Y, Shinkai S, Kumagai Set al.   Longitudinal changes in higher-level functional capacity of an older population living in a Japanese urban community. Arch Gerontol Geriatr . 2003; 36: 141– 153. Google Scholar CrossRef Search ADS PubMed  9. Koyano W, Shibata H, Nakazato K, Haga H, Suyama Y. Measurement of competence: reliability and validity of the TMIG index of competence. Arch Gerontol Geriatr . 1991; 13: 103– 116. Google Scholar CrossRef Search ADS PubMed  10. Fujiwara Y, Yoshida H, Amano Het al.   Predictors of improvement or decline in instrumental activities of daily living among community-dwelling older Japanese. Gerontology . 2008; 54: 373– 380. doi: 10.1159/000151221 Google Scholar CrossRef Search ADS PubMed  11. Taniguchi Y, Yoshida H, Fujiwara Y, Motohashi Y, Shinkai S. A prospective study of gait performance and subsequent cognitive decline in a general population of older Japanese. J Gerontol A Biol Sci Med Sci . 2012; 67: 796– 803. doi: 10.1093/gerona/glr243 Google Scholar CrossRef Search ADS PubMed  12. Taniguchi Y, Shinkai S, Nishi Met al.   Nutritional biomarkers and subsequent cognitive decline among community-dwelling older Japanese: a prospective study. J Gerontol A Biol Sci Med Sci . 2014; 69: 1276– 1283. doi: 10.1093/gerona/glt286 Google Scholar CrossRef Search ADS PubMed  13. Taniguchi Y, Fujiwara Y, Nofuji Yet al.   Prospective study of arterial stiffness and subsequent cognitive decline among community-dwelling older Japanese. J Epidemiol . 2015; 25: 592– 599. doi: 10.2188/jea.JE20140250 Google Scholar CrossRef Search ADS PubMed  14. Taniguchi Y, Fujiwara Y, Murayama Het al.   Prospective study of trajectories of physical performance and mortality among community-dwelling older Japanese. J Gerontol A Biol Sci Med Sci . 2016; 71: 1492– 1499. doi: 10.1093/gerona/glw029 Google Scholar CrossRef Search ADS PubMed  15. Taniguchi Y, Kitamura A, Seino Set al.   Gait performance trajectories and incident disabling dementia among community-dwelling older Japanese. J Am Med Dir Assoc . 2017; 18: 192.e13– 192.e20. doi: 10.1016/j.jamda.2016.10.015 Google Scholar CrossRef Search ADS   16. Koyano W, Hashimoto M, Fukawa T, Shibata H, Gunji A. Functional capacity of the elderly: measurement by the TMIG Index of Competence [in Japanese]. Nihon Koshu Eisei Zasshi. Japanese J Public Health . 1993; 40: 468– 474. 17. Takata Y, Ansai T, Akifusa S, Soh I, Sonoki K, Takehara T. High-level functional capacity and 4-year mortality in an 80-year-old population. Gerontology . 2007; 53: 46– 51. doi: 10.1159/000095934 Google Scholar CrossRef Search ADS PubMed  18. Lubitz J, Cai L, Kramarow E, Lentzner H. Health, life expectancy, and health care spending among the elderly. N Engl J Med . 2003; 349: 1048– 1055. doi: 10.1056/NEJMsa020614 Google Scholar CrossRef Search ADS PubMed  19. Shinkai S, Yoshida H, Taniguchi Yet al.   Public health approach to preventing frailty in the community and its effect on healthy aging in Japan. Geriatr Gerontol Int . 2016; 16 ( Suppl 1): 87– 97. doi: 10.1111/ggi.12726 Google Scholar CrossRef Search ADS PubMed  20. Shibata H, Sugisawa H, Watanabe S. Functional capacity in elderly Japanese living in the community. Geriatr Gerontol Int . 2001; 1: 8– 13. doi:10.1046/j.1444-1586.2001.00004.x Google Scholar CrossRef Search ADS   21. Nakamura K, Okamura T, Kanda H, Hayakawa T, Okayama A, Ueshima H; Health Promotion Research Committee of the Shiga National Health Insurance Organizations. Medical costs of obese Japanese: a 10-year follow-up study of National Health Insurance in Shiga, Japan. Eur J Public Health . 2007; 17: 424– 429. doi: 10.1093/eurpub/ckl271 Google Scholar CrossRef Search ADS PubMed  22. Tsuji I, Takahashi K, Nishino Yet al.   Impact of walking upon medical care expenditure in Japan: the Ohsaki Cohort Study. Int J Epidemiol . 2003; 32: 809– 814. Google Scholar CrossRef Search ADS PubMed  23. Nakamura K, Okamura T, Hayakawa T, Kanda H, Okayama A, Ueshima H; Health Promotion Research Committee of the Shiga National Health Insurance Organizations. Medical expenditures of men with hypertension and/or a smoking habit: a 10-year follow-up study of National Health Insurance in Shiga, Japan. Hypertens Res . 2010; 33: 802– 807. doi: 10.1038/hr.2010.81 Google Scholar CrossRef Search ADS PubMed  24. Ministry of Health, Labour and Welfare. The comprehensive reform of social security and tax. http://www.mhlw.go.jp/english/social_security/kaikaku.html. Accessed March, 2018. 25. Hashimoto H, Horiguchi H, Matsuda S. Micro data analysis of medical and long-term care utilization among the elderly in Japan. Int J Environ Res Public Health . 2010; 7: 3022– 3037. doi: 10.3390/ijerph7083022 Google Scholar CrossRef Search ADS PubMed  26. Spillman BC, Lubitz J. The effect of longevity on spending for acute and long-term care. N Engl J Med . 2000; 342: 1409– 1415. doi: 10.1056/NEJM200005113421906 Google Scholar CrossRef Search ADS PubMed  27. Ishizaki T, Shimmei M, Fukuda Het al.   Cumulative number of hospital bed days among older adults in the last year of life: a retrospective cohort study. Geriatr Gerontol Int . 2017; 17: 737– 743. doi: 10.1111/ggi.12777 Google Scholar CrossRef Search ADS PubMed  28. Tanuseputro P, Wodchis WP, Fowler Ret al.   The health care cost of dying: a population-based retrospective cohort study of the last year of life in Ontario, Canada. PLoS One . 2015; 10: e0121759. doi: 10.1371/journal.pone.0121759 Google Scholar CrossRef Search ADS PubMed  29. Jones B, Nagin D, Roeder K. A SAS procedure based on mixture modelling for estimating developmental trajectories. Social Methods Res . 2001; 29: 374– 393. Google Scholar CrossRef Search ADS   30. Hines KE, Middendorf TR, Aldrich RW. Determination of parameter identifiability in nonlinear biophysical models: a Bayesian approach. J Gen Physiol . 2014; 143: 401– 416. doi: 10.1085/jgp.201311116 Google Scholar CrossRef Search ADS PubMed  31. Shinkai S, Fujita K, Fujiwara Yet al.   Prognosis of different types of homeboundness among community-living older adults: two-year prospective study [in Japanese]. Nihon Koshu Eisei Zasshi. Japanese J Public Health . 2005; 52: 627– 638. 32. Mossey JM, Shapiro E. Self-rated health: a predictor of mortality among the elderly. Am J Public Health . 1982; 72: 800– 808. Google Scholar CrossRef Search ADS PubMed  33. Fine JP, Gray RJ. A proportional hazards model for the subdistribution of a competing risk. J Am Stat Assoc . 1999; 94: 496– 509. doi: 10.1080/01621459.1999.10474144 Google Scholar CrossRef Search ADS   34. World Health Organization. Active ageing: a policy framework. http://apps.who.int/iris/bitstream/10665/67215/1/WHO_NMH_NPH_02.8.pdf. Accessed March, 2018. 35. Murakami K, Tsubota-Utsugi M, Satoh Met al.   Impaired higher-level functional capacity as a predictor of stroke in community-dwelling older adults: the Ohasama study. Stroke . 2016; 47: 323– 328. doi: 10.1161/STROKEAHA.115.011131 Google Scholar CrossRef Search ADS PubMed  36. Lo AX, Donnelly JP, McGwin GJr, Bittner V, Ahmed A, Brown CJ. Impact of gait speed and instrumental activities of daily living on all-cause mortality in adults ≥65 years with heart failure. Am J Cardiol . 2015; 115: 797– 801. doi: 10.1016/j.amjcard.2014.12.044 Google Scholar CrossRef Search ADS PubMed  37. Scott WK, Macera CA, Cornman CB, Sharpe PA. Functional health status as a predictor of mortality in men and women over 65. J Clin Epidemiol . 1997; 50: 291– 296. Google Scholar CrossRef Search ADS PubMed  38. Leibson CL, Long KH, Ransom JEet al.   Direct medical costs and source of cost differences across the spectrum of cognitive decline: a population-based study. Alzheimers Dement . 2015; 11: 917– 932. doi: 10.1016/j.jalz.2015.01.007 Google Scholar CrossRef Search ADS PubMed  39. Ponzetto M, Maero B, Maina Pet al.   Risk factors for early and late mortality in hospitalized older patients: the continuing importance of functional status. J Gerontol A Biol Sci Med Sci . 2003; 58: 1049– 1054. Google Scholar CrossRef Search ADS PubMed  40. Hasuo Y, Ueda K, Kiyohara Yet al.   Accuracy of diagnosis on death certificates for underlying causes of death in a long-term autopsy-based population study in Hisayama, Japan; with special reference to cardiovascular diseases. J Clin Epidemiol . 1989; 42: 577– 584. Google Scholar CrossRef Search ADS PubMed  © The Author(s) 2018. Published by Oxford University Press on behalf of The Gerontological Society of America. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.

Journal

The Journals of Gerontology Series A: Biomedical Sciences and Medical SciencesOxford University Press

Published: Mar 26, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off