The health benefits of physical activity have been well documented (10,35). There is an abundance of physical activity surveillance literature published focused on US populations that indicates relatively few people practice a physically active lifestyle (1,30). Large proportions of the population in Japan and in many other countries in the world are also insufficiently active (14,36). Thus, physical activity surveillance and promotion continue to be important public health priorities around the world.
Surveillance of population physical activity provides the basic information for the planning, implementation, and evaluation of public health practice. To date, most physical activity surveillance activities have been conducted using self-report instrumentation. However, the potential for information bias from self-reported physical activity is a well-known limitation (28). For example, the proportion of people meeting recommended levels of physical activity is discrepant between self-report and accelerometry (8,30). Indeed, recent progress in physical activity assessment technologies now permits surveillance using objective methods, including accelerometers and pedometers. For example, the US National Health and Nutrition Examination Survey (NHANES) used an accelerometer to objectively monitor physical activity in 2003-2004 (30) and again in 2005-2006 (34). Pedometers have also been used in surveillance (2,3,13).
The National Health and Nutrition Survey of Japan (NHNS-J) has been monitoring pedometer-determined physical activity among Japanese since 1989 (14). However, the results of this survey are not well recognized because they have not been published in the English language. Also, time trends for step-determined physical activity have not been examined across survey years. Although the data from this survey are not available on an open-access basis, it is possible to synthesize the published cross-sectional summarized data from each year adjusted for age and to describe time trends for pedometer-determined physical activity among the Japanese population.
In the present study, the 1995-2007 levels of step-determined physical activity are reported, and age-adjusted time trends among Japanese were examined using data extracted from the series of summary reports available from the NHNS-J (14-22).
The NHNS-J has been conducted annually since 1945 by the Ministry of Health, Labour and Welfare (MHLW) in Japan. It dictates collection of fundamental information for health promotion mainly from the viewpoint of lifestyle among Japanese. The survey consists of three parts: 1) physical condition, 2) nutritional aspects, and 3) lifestyle (14). The physical condition aspect of the survey includes anthropometry, blood pressure, blood sampling, and pedometer-determined physical activity (since 1989) among other parameters. In this study, we focused solely on the results of the pedometer survey.
The results of NHNS-J are summarized by MHLW and released in a brief website report in the year after data collection. A more detailed final report is also released a few years after the survey. Because raw data from this survey are not openly accessible, we used the final reports and summarized data for secondary analyses herein (14-22). As noted above, the NHNS-J first implemented pedometer-based surveillance in 1989. However, in the early stage of this pedometer surveillance strategy, participant inclusion criteria and the method of data summarization in MHLW reports were not consistent. Thus, this study examined the most consistently collected and presented data available annually from 1995 to the newly released 2007 survey. Pedometer data extracted from the 1995-2007 NHNS-J included mean steps per day and the number of participants in MHLW-defined steps per day categories (0-1999, 2000-3999, 4000-5999, 6000-7999, 8000-9999, 10,000+ steps per day) by gender (men, women) and age categories (20-29, 30-39, 40-49, 50-59, 60-69, 70+ yr old) (14-22).
Participants in NHNS-J.
The annual administration of the NHNS-J begins with a selection of 300 census units. These 300 were randomly chosen from a total of 2000 census units previously selected from the whole of Japan (about 1,960,000 census units) using a stratified random sampling method as part of the Comprehensive Survey of Living Conditions of the People on Health and Welfare. Each census unit included approximately 20 households. All households and residents aged 1 yr or older in these 300 units were asked to participate in the NHNS-J. Only individuals age 15 yr or older were invited to participate in the pedometer component of the survey. This analysis focused on step-determined physical activity of adults, which is defined as 20 yr or older in Japan. The NHNS-J sampling strategy was consistent from 1995 to 2007. For example, in the NHNS-J of 2007, the 300 census units selected included approximately 6000 households, composed of approximately 18,000 family members age 1 yr or older. All of these individuals were asked to join the survey. As a result, 8885 residents participated in the nutrition component of the NHNS-J, and 7131 participated in the pedometer component. Among these 7131, 6768 were 20 yr old or older. Numbers of participants in each year organized by gender and age are indicated in Table 1.
Pedometer-determined physical activity.
The same pedometer, AS-200 (Yamasa Co., Ltd., Tokyo, Japan), was used for the NHNS-J from 1995 to 2007. Yamasa is the Japanese generic name for Yamax, a commonly used research-grade pedometer. The survey was conducted on a single individually specified day between Monday and Saturday in November every year. The specific dates for survey administration and pedometer monitoring were dependent on census units and on participants. Specifically, the survey office of each census unit set the survey period (e.g., a weeklong period in the month of November). Participants then selected a single "typical" day during that period to monitor their physical activity with a pedometer. Craig et al. (4) have shown that a single day of pedometer data can be used for population surveillance purposes. Participants were asked to wear the device on their waist from the time they got up in the morning until the time they went to bed at night, removing the device only to engage in water-based activities. Participants recorded their steps per day on a survey log and returned it on a subsequently scheduled physical examination day.
Mean ± SD steps per day and the proportion of participants classified in the MHLW step-defined activity categories were described by gender and age groups for the 2007 data. Regarding the time trend analysis between 1995 and 2007, 1) the mean steps per day, 2) the proportion taking ≥10,000 steps per day, and 3) the proportion taking <4000 steps per day by gender and age groups were presented. Because the age distributions of survey samples shifted over time to represent an aging population, these analyses were adjusted by age (i.e., setting the age distribution of the 1995 survey as standard).
The survey was conducted on the basis of the Health Promotion Law of Japan. The Ministry of Internal Affairs and Communications of Japan reviewed and approved the survey protocols, and informed consent was obtained from participants. In this analysis, the data came from summarized reports that have been already published and do not include personal information.
In the NHNS-J 2007 (Table 2), mean ± SD steps per day were 7321 ± 4588 among men and 6267 ± 3827 among women. The proportion taking ≥10,000 steps per day was 23.3% among men and 16.0% among women, whereas the proportion taking <4000 steps per day was 24.8% among men and 30.6% among women. Men averaged more steps per day than women in all age groups. The gender differences were the largest in the 20- to 29-yr-old age group (1717 steps per day) and the smallest in the 60- to 69-yr-old age group (603 steps per day). Age-related declines of mean steps per day were observed in men. In women, participants age 40-49 yr old averaged the highest number of steps per day. In both genders, large declines in mean steps per day were observed among participants age 70 yr or older. Specifically, compared with 20-29 yr olds, 70-yr-old participants accumulated 3614 steps per day fewer in men and 3036 steps per day fewer in women.
The time trends for age-adjusted mean steps per day, age-adjusted proportion taking ≥10,000 steps per day, and age-adjusted proportion taking <4000 steps per day are presented in Table 3. For these indicators, 1998-2000 were the most active years in terms of pedometer-determined physical activity for both men and women. Age-adjusted mean steps per day declined from peak values in 1998-2000 to 2007 by 529 steps per day among men and by 857 steps per day among women. The age-adjusted proportion of active persons (≥10,000 steps per day) also declined among both men and women during this same time frame (−5.1% and −5.0%, respectively). In contrast, the age-adjusted proportion taking <4000 steps per day increased by 4.8% among men and by 8.2% among women.
Figure 1 shows the time trends for age-adjusted mean steps per day. A decline in steps per day was observed across the annual administration of the survey (Fig. 1A). Age-adjusted mean steps per day increased from 1995 to 2000 among men and from 1995 to 1998 among women but have been steadily decreasing since that time. Analyses by age groups within each gender showed steady or slight declines in all age groups recently (Fig. 1B, C).
The proportion taking ≥10,000 steps per day demonstrates the same trend as mean steps per day; that is, it has decreased in recent years from peak values in 1998-2000 (Fig. 2A-C).
The proportion taking <4000 steps per day generally showed a reciprocal change compared with the proportion taking ≥10,000 steps per day (Fig. 3A-C). Analyzed by gender and by age, the increase in the proportion taking <4000 steps per day was most pronounced in women in recent years, especially in the oldest age group (≥70 yr old) (Fig. 3B, C).
Although surveillance activities incorporating objectively monitored physical activity have appeared in the scientific literature recently (2,3,8,13,30,34), the NHNS-J represents the singular and therefore unique source of ongoing pedometer-based surveillance data, dating back to 1989 but consistently administered since 1995. Although the raw data are not publicly available, summary data reported in Japanese-language reports still represent an important source of objectively monitored physical activity trends.
According to the NHNS-J 2007, men took 7321 ± 4588 steps per day and women took 6267 ± 3827 steps per day. Men took more steps per day than women in all age groups. Steps per day were lower with older age groups among men, whereas among women, the 40- to 49-yr-old age group took the highest mean steps per day. The trend for steps per day in recent years indicated declines (−529 age-adjusted mean steps per day among men and −857 age-adjusted mean steps per day among women) from peak values recorded in 1998-2000 to most recently reported values collected in 2007. In addition, the growing segment of those taking <4000 steps per day and the diminishing segment taking ≥10,000 steps per day are a concerning trend over time.
Accelerometer-determined step data (treated to approximate pedometer-determined scaling) collected as part of the 2005-2006 US NHANES indicated that American men took 7431 steps per day and women took 5756 steps per day (34). In comparison, this Japanese sample displayed approximately the same level among men and was more active in women. Bassett et al. (2) reported that men took 5340 pedometer-determined steps per day and women took 4912 steps per day among US adults using a separate nationally representative sample. Compared with the Bassett et al. (2) study, which used a similar type of pedometer, Japanese took more steps per day than both US men and women. The pedometer-based study in Western Australia reported mean values of 10,079 steps per day among men and 9169 steps per day among women (13). These figures were much higher than these NHNS-J Japanese data.
Direct comparison of these different surveys from around the world is hampered by use of different methods, including sampling, data collection, and device used (e.g., accelerometer vs pedometer and different brands of pedometers). The Japanese survey was conducted with a nationally representative adult sample monitored for a single individually selected day between Monday and Saturday in November each year using the Yamasa AS200 pedometer. In contrast, the NHANES (30), which also uses a nationally representative sample, used a 7-d survey (although fewer days are typically accepted for analyses) rolled out during a 2-yr cycle using the ActiGraph AM7164 accelerometer (Fort Walton Beach, FL). The Bassett et al. (2) study recruited participants through an online survey panel for a 2-d survey in June and used the Accusplit AE120 (Livermore, CA), which the authors report has the same measurement mechanism as the Yamax pedometer. The Australian survey (13) was based on a randomly selected sample and a 7-d monitoring period in November to December (obviously in the southern hemisphere so different from November in Japan) and used the Yamax SW700. They required four or more valid days for inclusion in their analyses, a requirement that might selectively exclude the most sedentary individuals (34). Across these surveys, response rates were either not clearly reported or not very high. There is also some evidence that suggests that the pedometer survey respondents generally tend to be more active than nonrespondents (2,11). Despite these concerns and overall differences in survey administration and analyses, it could be said that the Australians walked the most, the Japanese were intermediary, and the US residents walked the least on the basis of these studies.
Three to four days of data collection are often cited as necessary to assess habitual physical activity by accelerometry (31). However, this is based on a requirement to establish a stable estimate of time in moderate-to-vigorous physical activity and not necessarily a volume indicative of physical activity collected and expressed as steps per day. For surveillance to assess population levels of physical activity, a 1-d protocol with a sufficient sample size may be sufficient (4). Regarding the device, Yamasa (same as Yamax) pedometers are well validated (6,29). Although recent reports have raised concern about this instrument's ability to accurately capture steps per day taken by overweight/obese individuals (5), the difference in accelerometer-determined steps per day across body mass index-defined weight status categories displays a similar pattern (33,37). The potential for international comparison of objectively monitored physical activity is apparent from a previous report comparing accelerometer-determined physical activity from the United States and Sweden; however, steps per day were not presented in that report (9). Clearly, more research is needed to standardize population-level surveillance efforts. Until then, researchers are encouraged to clearly report their methods, including documenting response rates, monitoring periods (including seasons), and instrumentation choices.
The 500- to 900-steps-per-day decline documented in the NHNS-J may seem trivial from an individual point of view but is likely relevant in terms of a population-level statistic. Growing popularization and adoption of motorized private transportation among Japanese have lead to an increasingly car-dependent lifestyle and may be one reason of this apparent population trend in steps per day. The fourth Nationwide Person Trip Survey, which monitors the travel behavior of Japanese, reported the modal share (proportion of trips taken by a particular mode of transportation) of cars (25). The survey results indicated that the modal share of cars has increased from 38.7% in 1992 to 42.1% in 2005. In contrast, the modal share of walking decreased from 24.1% to 20.3% during the same period. One more potential contributor to the observed decreasing trend is the increased diffusion of personal computer and Internet use during a similar period. According to the Communications Usage Trend Survey of Japan, Internet use in households has dramatically increased from 3.3% in 1996 to 91.3% 2007 (23).
In 2004, Tudor-Locke and Bassett (32) defined a sedentary lifestyle as taking <5000 steps per day. The most comparable category used by the NHNS-J is that taking <4000 steps per day. So defined, it seems that the proportion of Japanese that can be classified as sedentary has increased in recent years. This increase was more pronounced among women (+8.2% in the age-adjusted percent from peak to 2007) than among men (+4.8%). Although we can only speculate on reasons for this observation, women with a relatively low employment rate (50.3% among women ≥20 yr old vs 76.7% among men ≥20 yr old in Japan) (24) may suffer more from recent neighborhood environmental changes leading to an increased car-dependent lifestyle (26,27).
There are limitations to this study that must be acknowledged. First, this was a secondary analysis of government-collected and published data. Thus, the description of methods and original analyses must be accepted as is. For example, other researchers have categorized step-determined physical activity levels in 2500-step increments and have defined a sedentary lifestyle as taking <5000 steps per day (32). However, categorization of activity levels herein was possible only by using the NHNS-J results as published. Therefore, we necessarily defined a sedentary lifestyle as taking <4000 steps per day. Second, this survey is routinely conducted using just a single individually selected day of monitoring during a designated period. As we indicate above, Craig et al. (4) have shown that a single day of pedometer monitoring may be sufficient for estimation of group-level physical activity. However, the single day of monitoring in the NHNS-J survey was not randomly assigned. NHNS-J participants were instructed to choose a typical day for pedometer self-monitoring from the assigned survey period, and the specific date was left up to individual choice. It remains possible that the selected date represents a reactive measure on the individual level. Third, the NHNS-J was conducted in November every year. There have been some reports regarding seasonal differences in steps per day (7,12). November is the end of fall in Japan, with no extreme weather; however, it is associated with slightly lower temperatures compared with the annual average. Therefore, it is plausible that a year-round average of steps per day would be different from that collected only in November. However, we can be more confident in the evidence for time trends because the surveillance has been consistently administered in the same month, using the same protocol, and with the same instrument for many years.
Despite these limitations, these Japanese data represent a unique opportunity to examine time trends of step-determined physical activity level of Japanese adults. No other similar data source exists in the world. According to the NHNS-J, Japanese men took 7321 steps per day and women took 6267 steps per day on average in 2007. The population's mean steps per day have decreased by 500-900 steps per day between 1998-2000 and 2007. The increase in the percent taking <4000 steps per day was especially noticeable among women.
This study was supported by a grant-in-aid from the Ministry of Health, Labour and Welfare of Japan (Comprehensive Research on Prevention of Cardiovascular Diseases and Other Lifestyle Related Diseases: H20-Junkankitou-Ippan-001 and H21-Junkankitou-Ippan-007) and a grant-in-aid for scientific research from the Japan Ministry of Education, Culture, Sports, Science and Technology.
All authors have no other conflicts of interest, including related directorships, stock holdings, or contracts.
The results of the present study do not constitute endorsement by the American College of Sports Medicine.
1. Adabonyan I, Loustalot F, Kruger J, Carlson SA, Fulton JE. Prevalence of highly active adults-Behavioral Risk Factor Surveillance System, 2007. Prev Med
2. Bassett DR Jr, Wyatt HR, Thompson H, Peters JC, Hill JO. Pedometer-measured physical activity and health behaviors in U.S. adults. Med Sci Sports Exerc
3. Craig CL, Cameron C, Griffiths JM, Tudor-Locke C. Descriptive epidemiology of youth pedometer-determined physical activity: CANPLAY. Med Sci Sports Exerc
4. Craig CL, Tudor-Locke C, Cragg S, Cameron C. Process and treatment of pedometer data collection for youth: the Canadian Physical Activity Levels among Youth study. Med Sci Sports Exerc
5. Crouter SE, Schneider PL, Bassett DR Jr. Spring-levered versus piezo-electric pedometer accuracy in overweight and obese adults. Med Sci Sports Exerc
6. Crouter SE, Schneider PL, Karabulut M, Bassett DR Jr. Validity of 10 electronic pedometers for measuring steps, distance, and energy cost. Med Sci Sports Exerc
7. Dasgupta K, Joseph L, Pilote L, Strachan I, Sigal RJ, Chan C. Daily steps are low year-round and dip lower in fall/winter: findings from a longitudinal diabetes cohort. Cardiovasc Diabetol
8. Hagströmer M, Oja P, Sjöström M. Physical activity and inactivity in an adult population assessed by accelerometry. Med Sci Sports Exerc
9. Hagströmer M, Troiano RP, Sjöström M, Berrigan D. Levels and patterns of objectively assessed physical activity-a comparison between Sweden and the United States. Am J Epidemiol
10. Haskell WL, Lee IM, Pate RR, et al. Physical activity and public health: updated recommendation for adults from the American College of Sports Medicine and the American Heart Association. Med Sci Sports Exerc
11. Inoue S, Ohya Y, Odagiri Y, et al. Characteristics of accelerometry respondents to a mail-based surveillance study. J Epidemiol
12. Kang M, Bassett DR, Tudor-Locke C, Barreira TV, Ainsworth B. Measurement effects of seasonal and monthly variability on pedometer-determined data. J Phys Act Health
. >(in press)>.
13. McCormack G, Giles-Corti B, Milligan R. Demographic and individual correlates of achieving 10,000 steps/day: use of pedometers in a population-based study. Health Promot J Austr
15. Ministry of Health, Labour and Welfare of Japan. The Report of the Japan Health and Nutrition Survey 1995
. Tokyo (Japan): Dai-ichi Shuppan; 1997. p. 121.
16. Ministry of Health, Labour and Welfare of Japan. The Report of the Japan Health and Nutrition Survey 1996
. Tokyo (Japan): Dai-ichi Shuppan; 1998. p. 116.
17. Ministry of Health, Labour and Welfare of Japan. The Report of the Japan Health and Nutrition Survey 1997
. Tokyo (Japan): Dai-ichi Shuppan; 1999. p. 119.
18. Ministry of Health, Labour and Welfare of Japan. The Report of the Japan Health and Nutrition Survey 1998
. Tokyo (Japan): Dai-ichi Shuppan; 2000. p. 118.
19. Ministry of Health, Labour and Welfare of Japan. The Report of the Japan Health and Nutrition Survey 1999
. Tokyo (Japan): Dai-ichi Shuppan; 2001. p. 116.
20. Ministry of Health, Labour and Welfare of Japan. The Report of the Japan Health and Nutrition Survey 2000
. Tokyo (Japan): Dai-ichi Shuppan; 2002. p. 114.
21. Ministry of Health, Labour and Welfare of Japan. The Report of the Japan Health and Nutrition Survey 2001
. Tokyo (Japan): Dai-ichi Shuppan; 2003. p. 129.
22. Ministry of Health, Labour and Welfare of Japan. The Report of the Japan Health and Nutrition Survey 2002
. Tokyo (Japan): Dai-ichi Shuppan; 2004. p. 134.
26. Owen N, Humpel N, Leslie E, Bauman A, Sallis JF. Understanding environmental influences on walking; review and research agenda. Am J Prev Med
27. Saelens BE, Handy SL. Built environment correlates of walking: a review. Med Sci Sports Exerc
. 2008;40(7 suppl):S550-66.
28. Sallis JF, Saelens BE. Assessment of physical activity by self-report: status, limitations, and future directions. Res Q Exerc Sport
. 2000;71(2 suppl):S1-14.
29. Schneider PL, Crouter SE, Bassett DR. Pedometer measures of free-living physical activity: comparison of 13 models. Med Sci Sports Exerc
30. Troiano RP, Berrigan D, Dodd KW, Mâsse LC, Tilert T, McDowell M. Physical activity in the United States measured by accelerometer. Med Sci Sports Exerc
31. Trost SG, McIver KL, Pate RR. Conducting accelerometer-based activity assessments in field-based research. Med Sci Sports Exerc
. 2005;37(11 suppl):S531-43.
32. Tudor-Locke C, Bassett DR Jr. How many steps/day are enough? Preliminary pedometer indices for public health. Sports Med
33. Tudor-Locke C, Brashear MM, Johnson WD, Katzmarzyk PT. Accelerometer profiles of physical activity and inactivity in normal weight, overweight, and obese U.S. men and women. Int J Behav Nutr Phys Act
34. Tudor-Locke C, Johnson WD, Katzmarzyk PT. Accelerometer-determined steps per day in US adults. Med Sci Sports Exerc
35. US Department of Health and Human Services. 2008 physical activity guidelines for Americans, be active, healthy, and happy! [Internet]. [cited 2011 Jan 7]. Available from: www.health.gov/paguidelines
37. Yoshioka M, Ayabe M, Yahiro T, et al. Long-period accelerometer monitoring shows the role of physical activity in overweight and obesity. Int J Obes (Lond)