METHODS

Sampling

A representative sample of residents born in 1903, 1908, 1913, or 1918 was selected from resident registration records in two cities in the Tokyo metropolitan area—one in the east part and the other in the west part—in 1989. The following sampling rates were applied: 1/2 for people born in 1913 or 1918, and 1/1 for people born in 1903 or 1908. A total of 7362 people were selected and asked to participate in a mail follow up survey of their individual functional status and health. A total of 5924 people consented for this mail questionnaire follow up survey. In 1992, a subsequent questionnaire survey was conducted of those who agreed to participate. We asked about their residence’s environmental conditions, and about their functional status, lifestyles, living arrangements, and socioeconomic status. A total of 3144 people responded to the survey by consenting to use their data with their resident registry records in later years. This population was defined as the 1992 cohort population to be followed up in 1997.

Baseline assessment

A baseline assessment of the 1992 cohort, which consisted of 1339 men and 1805 women, was completed during April to June 1992. At the 1992 survey, nine items relating to residential environment were asked and all the responses were recorded. Percentage of responders to the individual questionnaire items were from 94.2% to 96.9%. The questionnaire were: space near the residence for taking a stroll, a park, and tree lined streets near the residence, noise from automobiles and factories near the residence, the level of crime in the community, hours of sunlight at the residence, existence of a garden at the residence, whether the residence faced a road with a regular bus service, active communication among neighbouring residents, and preference to continue to live in the current community. The amount of monthly living expenses was asked and its answer was used as an indicator of the socioeconomic status of participants. Marital status was asked and its answer was used as an indicator of participants’ living arrangements.

Follow up

Survival of the respondents as of 31 May 1997 were followed up by our obtaining official residence records from their local governments, based on each participant’s informed consent thereto.

Analysis

The association between baseline residential-environment characteristics and the five year survival of the subject senior citizens was analysed by use of the χ², and Kruskal-Wallis tests.

Principal component analysis with Varimax rotation was performed to extract residential-environmental factors from the variables of environmental characteristics that showed significant univariate statistical associations with the five year survival of male and/or female subjects, as calculated in the earlier tests.

Associations between each of the baseline residential-environmental factors and five year survival were examined by multiple logistic regression analysis. Independent influences of residential-environment conditions, represented by factor scores obtained by the earlier analysis, were calculated after excluding the effects of age, sex, marital status, and socioeconomic status. To exclude the effects of lower functional status at the baseline in individual regression models, we examined logistic regression coefficients for the subjects of any levels of functional status, and those for subjects who did not need help to get out of bed at the baseline time.

RESULTS

The survival of 2211 and the death of 897 of the 3144 persons in the 1992 cohort population were confirmed five years later, in 1997. The follow up rate was 98.9%. The five year survival percentage was higher among women (75.2%) than among men (65.9%). The five year survival percentage decreased as the baseline age of the subjects increased, being 85.7% among those 73 years of age in the baseline year and 43.7% among those 88 years of age in the baseline year.

Table 1 shows five year survival percentages of older people in relation to the characteristics of their residential environment, sex, and age. For the male and female subjects in total, the five year survival percentages were greater for subjects having either a space near their residence for taking strolls (p<0.01), or parks and tree lined streets near their residence (p<0.05). These results were observed for each of the age categories. The five year survival percentage of senior citizens was greater among those who preferred to continue to live in their current community than among those who did not express such a preference (p< 0.01); this was observed for each of the age categories. The five year survival percentage of senior citizens increased in accordance with the number of hours of sunlight at their residence (p<0.01); this was also observed for the age categories of 73 and 78 in the baseline year.

Statistically greater survival rates were observed among male subjects who were not subject to noise from automobiles and factories near their residence (p<0.01) or who enjoyed sunlight at their residence (p<0.05), but such statistically greater survival rates were not observed among female subjects under the same conditions.

Among female subjects, five year survival percentages were greater for those who claimed to have active communication with their neighbours (p<0.05) and among those who preferred to continue to live in their current community (p<0.01), but such greater survival percentages were not observed among comparable male subjects.

Table 2 shows the factor loading of the obtained two factors from the principal component analysis with Varimax rotation of six living environment variables. The obtained two factors explained 50.0% of the variances of the six variables entered. The first factor represented “walkable green streets and spaces near the residence,” and the other represented “a positive attitude towards one’s own community.” The average of the standardised factor score of “walkable green streets and spaces near the residence” among men (−0.0471) was lower than that among women (0.0375) (p<0.05). The average of the standardised factor score of “a positive attitude to one’s own community” among men (−0.0707) was lower than that among women (0.0562) (p<0.01).

Table 3 summarises the results of multiple logistic regression analysis. It presents odds ratios with 95% confidence intervals and their significance level for the residential-environment factor of walkable green streets and spaces, in order to predict the five year survival of older people. It shows the odds ratios after adjustment for factor score of another residential-environment factor, age, sex, marital status, and living expenses for all of the people who were followed up and for people who did not need help to get out of bed at baseline. A calculation involving only the male subjects showed that the factor of walkable green streets and spaces near the residence was a significant predictor of survival in five years (p<0.05), after excluding the effects of age, marital status, attitude towards one’s own community, and living expenses.

The following five living environment variables were found to have a significantly independently impact on five year survival of older people: younger age, being a female in sex, living with one’s spouse, higher living expenses, and positive attitude towards one’s own community individually independently significantly contributed to predicting the five year survival of older people. These associations were found in the both models with and without subjects who did not need help to get out of bed at the baseline time.

DISCUSSION

The quality of the physical environments near a residence—observed as the space for taking a stroll, tree lined streets, the number of hours of sunlight at the residence, and less noise from automobiles and factories—showed a positive association with the longevity of senior citizens. The factor of walkable green streets and spaces near the residence significantly and positively influenced the five year survival of senior citizens independent of a person’s age, sex, marital status, attitude toward their own community, and socioeconomic status. This association between walkable green spaces and longevity remained after excluding subjects of very low baseline functional status. These results clearly show that living in an area with walkable green spaces positively influenced on the longevity of senior citizens.

Five year survival of the subjects was confirmed by examining resident registration records, which are one of the reliable sources of official information regarding the life and death of people in Japan.¹⁸ The follow up percentage of 98.9% after the five year period was considered high enough to reflect the entire population. It is considered that the average five year survival percentage of the baseline group, 71.1%, as compared with the follow up percentage, was an appropriate proportion for conducting statistical analysis throughout this study.

It is probable that the availability of space for taking a stroll positively contributed to longevity. Booth et al¹⁹ demonstrated that the availability of space for walking increased the amount of a senior citizen’s activity. It has also been reported that more active older people gain a higher functional status and live longer.²⁰ In addition, the positive influence of walking to reduce risk for coronary heart disease among women was recently indicated.²¹ The availability of space near your own residence for taking a stroll is believed to increase the chances of walking outside of the residence, which helps to maintain a high physical functional status.

The results with the factor of walkable green streets and spaces that are independent of personal characteristics suggested that the value of parks and tree lined streets near residences is particularly high in densely populated urban areas for the senior citizens living there. Tanaka et al²² have demonstrated, by using aggregated data, a positive correlation between woodland/farmland area and lower mortality rates for residents of cities having a density of more than 4000 inhabitants per km², but not among cities having a sparse population density. An apparent association between senior citizens’ longevity and walkable green spaces near their residence, which was observed in the present cohort study, was consistent with our previous ecological study. The present cohort study was carried out in highly densely populated urban areas; the average population density of our study areas was 13 050 inhabitants per km². The per capita park space in Tokyo was 5.18 m² in 1997, and the “greenery rate” as a percentage of the sum of the areas of forests, grass area, agricultural land, garden greens, parks, trees along streets, riverside and pondside space was 29% in the centre area of Tokyo.²³ If you consider that most of the Tokyo metropolitan area is a highly densely populated urban area, the results obtained in this study can be regarded as important evidence on which urban planning in Tokyo should be based.

Urbanisation has accelerated the spread of rapid motorisation and the development of road networks in densely populated urban areas.²⁴ Congested traffic on narrow roads, highly concentrated development, and little open space for relaxing places are some examples of characteristics about which we are concerned.^25,26 Although all of the measurements of atmospheric sulphur oxides, 85.2% of measurements of atmospheric nitrogen oxides, and 78.6% of measurements of suspended particulate matter from general monitoring stations have met the environmental standard in the urban centre of Tokyo,²⁷ the long term influence of slightly unpreferable air quality on senior citizens is a matter of concern.²⁸ In such a congested environment, existence of greenery along the streets and small parks within residential zones is considered to improve air quality and provide favourable conditions for the living of senior citizens.²³ A number of studies have shown that active living helps to maintain and improve the functional status of older people.^29–31 The current situations in urban environments might discourage the elderly from moving around easily outside of their residence. It is generally acknowledged that green spaces provide settings for healthy exercise and that they induce citizens to walk.³² It was reported that community groups are walking along newly developed walking trails with greenery by a river in the study area.³³ It was considered that living in an area that provides good access to walkable green spaces facilitates activities outside a person’s residence. Studies have shown that walkable green spaces in residential areas contribute to the health of senior citizens, in part by facilitating their activity outside of their residence.

It has been pointed out that the ability to live in a better residential environment is associated with higher socioeconomic status.³⁴ It is also known that there is an association between higher income and a lower risk of mortality.^35,36 This study showed a significant association between socioeconomic status and longevity. However, it is noteworthy that we found a clear and significant association between walkable green spaces and longevity, even after excluding the influence of socioeconomic conditions on longevity. Furthermore, after excluding the influence of a person’s baseline functional status, there still was a significant association between walkable green spaces and longevity.

Our findings suggested that if favourably walkable green streets and spaces were provided, the health of senior citizens would be promoted further, regardless of their socioeconomic status. These results could be generalised in a densely populated urban centre where senior citizens live in medium rise to high rise apartment housing complexes without their own gardens or in detached houses with limited areas for your own garden. The obtained evidence that walkable green streets and spaces promote the health of senior citizens was considered to be useful in supporting a decisive shift in the longstanding urban planning policy debate in favour of the proponents of public policies that put priority on health. This evidence also strongly supports the preference of people for comfortable living environments. Vision sharing and consensus building in the community for city planning would be more constructive if supported not only by intangible preferences but also by substantial evidence showing health outcomes.

Master plans for urban development should pay more attention to maintaining and increasing greenery filled public areas that are easy to walk in and are within easy walking distance of every household. Closer and more effective collaboration among health, construction, civil engineering, planning, and other relevant sectors including government agencies, which has been minimal until now, should be encouraged more, on the basis of the present evidence.

In conclusion, based on a five year follow up cohort study of 3144 older people in Tokyo, a densely populated megacity, the longevity of senior citizens is positively influenced by their living in areas with walkable green spaces, independent of a person’s age, sex, marital status, baseline functional status, and socioeconomic status. Greenery filled public areas that are near to residences and easy to walk in should be further emphasised in the development and redevelopment of densely populated areas in a megacity through intersectoral collaboration.