This cohort of American Indians was no more sedentary than other race/ethnic groups in the U.S. surveyed in the Behavioral Risk Factor Surveillance System (BRFSS). Indeed, 18% of the cohort engaged in no leisure-time physical activity, which is far less than figures reported in other studies of American Indian adults based on BRFSS. In one report, 33% of American Indian/Alaska Native adults reported no leisure-time physical activity.17
In another, estimates varied by gender, ranging from 36% to 24%.6
The estimate of no leisure-time physical activity in the present study is even lower than the 28% observed for non-Native individuals in the former report. Similarly, the prevalence of no leisure-time physical activity in the present study was lower than in the Inter-Tribal Heart Project,18
where 33% of women and 21% of men reported no leisure-time activity. Lastly, in the Strong Heart Study,19
total physical activity levels (occupational plus leisure-time activity) were higher in men (37%) and women (85%), and among participants from the Dakotas (83%) and Arizona (13.5%), compared with comparable combined occupational and leisure-time categories by gender and region in this cohort.
These discrepancies are most likely due to differences in sampling procedures, physical activity instruments used, and scoring procedures in the various studies. For example, BRFSS is a random-digit telephone survey whereas EARTH was a large, community-based sample of American Indians. Telephone surveys among American Indians might introduce a bias in that individuals with higher SES are over-represented.20,21
Further, telephone surveys exclude certain segments of this population because telephone coverage varies greatly among American Indian households in both urban and rural areas.22,23
In addition, activity data are based on responses to a single question in BRFSS whereas participants in EARTH completed a comprehensive, computerized questionnaire. Great care was taken to develop a culturally appropriate instrument that included activities common to traditional and contemporary American Indian and Alaska Native peoples. Finally, the estimate of leisure-time physical activity was not restricted to moderate-intensity activities. Instead, a total leisure-time physical activity score was calculated that included a variety of lower-intensity activities such as stretching/hatha yoga (2.5 METs) and bowling (3.0 METs). There were over 25 individual activities within the leisure-time category of the instrument. Despite the differences noted above, activity patterns in this study were consistent with previous reports when examined by demographic characteristics such as gender and age among American Indians.24
Of all categories of physical activities, the most time was spent in the household category, which included activities such as mopping, sweeping, and the like. Using intensity-weighted values, this cohort also derived their greatest amount of energy expenditure by performing household activities. Although unusual, this is consistent with a previous report from another American Indian population.18
Poverty, unemployment or underemployment, low educational levels, and a large number of individuals living in the same household are features common to many Native communities. Many prior reports have linked education, unemployment and/or poverty to leisure-time inactivity.25–30
Unemployed individuals living on a reservation likely spend many hours per day at home, increasing the likelihood that they will perform household activities. High housing density, which is typical of many impoverished settings, may also explain the predominance of household activities. American Indian households tend to have a large number of infants and young children living in large multigenerational families, increasing the need for household activities. Lastly, the widespread gravel/dirt roads, weed-filled yards and lots, and often poorly controlled insects and rodents in these remote, rural, dry, and windswept locations likely create a need for cleaning-related activities.
Merely describing the customary activity patterns in American Indians is not a major reason for obtaining this type of data. Rather, physical activity data are important for linking health behaviors to health outcomes. As an example, the results demonstrate that average time spent in sedentary activities increased systematically with increasing BMI while leisure-time physical activity declined as a function of BMI, although the former was no longer significant after adjusting for covariates. Differences in physical activity in the Northern Plains and Southwest that parallels regional differences in BMI was also observed. Total amounts of time spent in leisure time were higher, and in sedentary activities lower, in the Northern Plains compared to the Southwest. In future studies by the current authors, these physical activity data will be used to examine relationships with a host of health conditions in this population.
A strength of this study includes the comprehensive way in which physical activity was assessed. This approach provided a more complete picture of physical activity patterns among American Indians than is discernable from existing data. The use of a computer assisted personal interview to assess physical activity may have reduced respondent burden and thus increased reliability. The large sample size also distinguishes it from nearly all prior studies of physical activity among American Indians. Finally, in the Northern Plains, the sampling strategy and age-representativeness make this effectively a population-based study.
This study also has several limitations. First and foremost, physical activity assessed with interviews or questionnaires may not accurately capture physical activity levels.31
A pedometer was used as the standard measure in these unpublished observations, which is problematic because a pedometer is not a gold standard, thus raising the possibility that the results are influenced by measurement bias. The analysis also does not take into account other social factors that may influence physical activity, including social support, acculturation, neighborhood safety, and the built environment.24
Third, the EARTH study used mixed sampling methods to achieve the desired overall target sample size in the Northern Plains and a random systematic household sampling strategy in Gila River Indian Community. Several of the tribal governments required that the study be open to all tribal members, not just those randomly selected. Fourth, results from EARTH participants may not be generalizable to nonreservation populations and American Indians living in other regions. Finally, it is not clear whether household activities result in mental and physical health benefits comparable to leisure-time physical activity, specifically exercise.