number of observations in the geriatric assessment is over 10,000. All participants in
health checkups provided written informed consent under conditions approved by
the Ethics Committee at Tokyo Metropolitan Institute of Gerontology.
The statistical method used in the study series was repeated-measures analysis,
such as the group-based semiparametric mixture model analyzed with the SAS
macro PROC TRAJ (Jones, Nagin, & Roeder, 2001). This approach uses a multinomial modeling strategy to identify relatively homogeneous clusters of developmental trajectories within a sample population; that is, the modeling strategy allows
for the emergence of more than two trajectories. Trajectory parameters are derived
by latent class analysis using maximum likelihood estimation. The best-fitting model
was determined by comparing the Bayesian Information Criterion values. To assess
model fit, posterior probabilities of group membership assignments were calculated
for each individual. A high posterior probability of belonging to a single group
represents a good fit (Hines, Middendorf, & Aldrich, 2014). In addition, Cox
proportional hazards models were used to examine associations of trajectories with
adverse health outcomes, such as mortality and incident dementia, while controlling
for potential confounders.
The next section will discuss the five health indexes (physical performance, gait
function, nutritional biomarkers, cognitive function, and functional capacity) for
aging trajectories and trajectories over time among community-dwelling older
Japanese.
2 Trajectories of Aging in Physical Performance (Taniguchi
et al., 2016)
Physical performance measures for older adults are good predictors of adverse health
outcomes, such as disability, hospitalization, cognitive decline and mortality
(Taniguchi et al., 2012; Shinkai et al., 2000; Volpato et al., 2011; Rantanen et al.,
2003; Cooper et al., 2010). However, no study had used multiple repeated-measures
data from a longitudinal study to investigate age-related trends and trajectories of
aging in physical performance. This study had two objectives: to identify trajectories
of aging and to determine whether potential physical performance trajectories were
associated with all-cause mortality after adjustment for important confounders
among community-dwelling older Japanese.
The data source for the present study was 1449 adults aged 65 years or older
during the period from 2002 through 2010 in the Kusatsu Longitudinal Study. In a
sample of the 1449 adults with valid and complete baseline data, 1048 had participated in at least one follow-up assessment by 2011. The average number of followup assessments among the 1048 adults was 4.5, and the total number of observations
was 4747. Analysis of local registries showed 89 (8.5%) deaths from any cause
among the 1048 participants through 2011. The physical performance measures
evaluated were handgrip strength (kg) of the dominant hand, as measured with a
A New Informatics: Trajectory Patterns in Health Indexes for Japanese Elders. . .
61
health checkups provided written informed consent under conditions approved by
the Ethics Committee at Tokyo Metropolitan Institute of Gerontology.
The statistical method used in the study series was repeated-measures analysis,
such as the group-based semiparametric mixture model analyzed with the SAS
macro PROC TRAJ (Jones, Nagin, & Roeder, 2001). This approach uses a multinomial modeling strategy to identify relatively homogeneous clusters of developmental trajectories within a sample population; that is, the modeling strategy allows
for the emergence of more than two trajectories. Trajectory parameters are derived
by latent class analysis using maximum likelihood estimation. The best-fitting model
was determined by comparing the Bayesian Information Criterion values. To assess
model fit, posterior probabilities of group membership assignments were calculated
for each individual. A high posterior probability of belonging to a single group
represents a good fit (Hines, Middendorf, & Aldrich, 2014). In addition, Cox
proportional hazards models were used to examine associations of trajectories with
adverse health outcomes, such as mortality and incident dementia, while controlling
for potential confounders.
The next section will discuss the five health indexes (physical performance, gait
function, nutritional biomarkers, cognitive function, and functional capacity) for
aging trajectories and trajectories over time among community-dwelling older
Japanese.
2 Trajectories of Aging in Physical Performance (Taniguchi
et al., 2016)
Physical performance measures for older adults are good predictors of adverse health
outcomes, such as disability, hospitalization, cognitive decline and mortality
(Taniguchi et al., 2012; Shinkai et al., 2000; Volpato et al., 2011; Rantanen et al.,
2003; Cooper et al., 2010). However, no study had used multiple repeated-measures
data from a longitudinal study to investigate age-related trends and trajectories of
aging in physical performance. This study had two objectives: to identify trajectories
of aging and to determine whether potential physical performance trajectories were
associated with all-cause mortality after adjustment for important confounders
among community-dwelling older Japanese.
The data source for the present study was 1449 adults aged 65 years or older
during the period from 2002 through 2010 in the Kusatsu Longitudinal Study. In a
sample of the 1449 adults with valid and complete baseline data, 1048 had participated in at least one follow-up assessment by 2011. The average number of followup assessments among the 1048 adults was 4.5, and the total number of observations
was 4747. Analysis of local registries showed 89 (8.5%) deaths from any cause
among the 1048 participants through 2011. The physical performance measures
evaluated were handgrip strength (kg) of the dominant hand, as measured with a
A New Informatics: Trajectory Patterns in Health Indexes for Japanese Elders. . .
61