Page 87 - IJPS-3-1

Lagergen M, Kurube N and Saito Y

Zunzunegui, 1999). A similar result was found by Holstein et al., who reported strong
mortality selection, but also that a notable minority improved in functional ability
over time (Holstein et al., 2007). Change in functional status over two years was also
calculated by Crimmins and Saito, who found that improvement and decline in status
were subject to different covariates (Crimmins and Saito, 1993). Calculations of life
expectancy by ADL status have further been made by Zeng, Gu and Land (2004) using
an extended multi-state life table method. They found that the disabled life expectancy
was significantly underestimated if information concerning the changes in disability
status before death were excluded.
All these studies (with the exception of Branch and Ku (1989), and ErnsthBravell et
al. (2008)) deal with either functional dependency and mortality or level of LTC. The
purpose of this study was to calculate probability distributions for dependency and
level of long-term care need combined, starting from varied initial conditions of these
terms at 78 years of age. The results are then compared between Japan and Sweden.
The studied time period in both cases is 15 years, from 78 to 93 years of age, and
results are shown by 3-year time-steps.

2 Material and methods
Longitudinal data on health and LTC level for Japan and Sweden were obtained from
the Nihon University Japanese Longitudinal Study of Aging (NUJLSOA) and the
Swedish National Study on Aging and Care of the population in the Stockholm area of
Kungsholmen (SNAC-K).
For NUJLSOA, data were gathered on several measures of ill health—including each
individual’s ability to perform activities of daily living (ADL), such as taking a bath
or shower, dressing, eating, standing up from a bed or chair, going to the bathroom,
and using the toilet (Katz et al., 1963). Corresponding data regarding instrumental
activities of daily living (IADL) were also recorded. These activities include preparing
meals, purchasing household items or medication, doing light household work, and
taking a bus or train. For each of these activities, it was noted whether the individual
reported the activity as difficult or not. If any difficulty was reported, individuals were
asked if they found the activity somewhat difficult, very difficult, or so difficult they
were unable to do it at all. In the present study, individuals are classified as being
dependent for an activity if they answered “very difficult” or “unable.” Persons were
classified as IADL-dependent if they were dependent in at least one IADL but no ADL,
and ADL-dependent if dependent in at least one ADL .
LTC was classified as no LTC, home-related LTC, or institutional care. Home-related
LTC included home nursing, home rehabilitation, home bathing services, day services,
overnight services, and other services such as welfare equipment rental or purchase
or home improvement services. Facility services, such as welfare facility, insurance
facility, or medical treatment nursing facility, were classified as institutions.
From the NUJLSOA data, a dataset was prepared using the wave 3 (2003) and
wave 4 surveys (2006), and from these people aged 78, 81, 84, 87, 90 and 93 years
at the wave 3 survey. However, to increase statistical power, these age groups were
augmented with ages 1 year below and 1 year above—i.e. 77, 78, and 79; 80, 81
and 82; 83, 84, and 85; 86, 87 and 88; 89, 90 and 91; and 92, 93, and 94. Persons in
institutions were not included in the first 1999-wave of the study, but persons who
transferred to institutions in consecutive waves were followed. Therefore we could
identify those who were institutionalized at wave 3 and 4. The dataset that was used
includes 1,666 persons in wave 3 and 1,246 persons in wave 4. For each gender and
age group in both waves, there were three levels of dependency and three levels of
LTC, i.e. 3 x 3 = 9 states in all.
The Swedish National Study on Aging and Care (SNAC) was initiated by the
Swedish government and involves four areas in Sweden, one of which is the
Kungsholmen area of Stockholm. Each area is studied in two parts: a population part
aimed at monitoring health and disability, and a care-system part that records acute
and long-term care for all inhabitants aged 65 years and older. The Swedish dataset

International Journal of Population Studies 2017, Volume 3, Issue 1 81