Increasing physical activity is a common goal for programs addressing obesity, but success has been limited. Walking is the most common form of physical activity for Americans (14), and some success has been achieved in increasing walking in small-scale studies using electronic pedometers (5,7,16,17). Electronic pedometers are inexpensive, minimally invasive, and easy to use. They provide an easy way both to quantify amount of physical activity performed and to provide a clear and measurable goal for physical activity (19). It is likely that they will be used increasingly as tools to promote physical activity.
Colorado on the Move™ is a statewide weight-gain prevention program that was launched in early 2002. The program is based on data suggesting that most of the population has experienced gradual weight gain over the past several years, and that this could likely be prevented with small changes in energy intake and energy expenditure (4). Colorado on the Move inspires people to make two small behavior changes to stop weight gain. The behavioral changes are: 1) choosing one behavior each day that eliminates about 100 kcal (e.g., drink a diet soda instead of a regular soda), and 2) walking 2000 steps (about 100 kcal) more than usual. Colorado on the Move™ uses electronic step counters to set individual goals for physical activity. For the physical activity intervention, participants are asked to establish their usual steps per day during a 7-d baseline period, and to immediately increase this by 2000 steps per day. Participants are then encouraged, as they feel comfortable, to continue to make small increases to their daily step goal.
We have previously shown that Colorado on the Move is effective in significantly increasing physical activity in small groups (21). Our intent is now to promote the program widely across the state of Colorado to evaluate its impact on walking in the general population. Thus, the major purpose of this study was to provide baseline values for current levels of walking in the state population. Secondary objectives were to determine whether measured steps per day varied with subjects’ own evaluation of their overall level of physical activity or physical inactivity, and to identify potential predictors of walking based on subject characteristics and answers to questions about other self-reported lifestyle factors.
We collaborated with Harris InteractiveSM to design and administer a short telephone survey on a statewide representative sample of adults that would serve as an accurate baseline measurement of the entire state, and that could be repeated using the same methods at a future date. The objective of the survey was to collect self-reported physical, behavioral, and attitudinal characteristics concerning health and physical activity followed by a 4-d baseline assessment of walking using a step counter. Harris InteractiveSM identified the representative sample population and conducted the interviews from a telephone research center in Rochester, NY. Those individuals who gave verbal consent to participate in the study were administered a short computer-aided telephone interview (CATI) to gather demographic, behavioral, and attitudinal measures.
The CATI program permits online data entry and editing of telephone interviews. The interviewing staff asked a series of questions prompted from a computer screen with a standard script that was continuously monitored by supervisory staff. The verbal answers to the questions by the participants were recorded by the interviewer directly into the computer. The CATI system reduces clerical error by eliminating the need for keypunching, because the interviewer enters the respondent’s answer directly into the system. For answers with precoded responses, the system only permits answers within a specific range. All data are tabulated, checked for internal consistency, and processed by the computer. The edit program lists any errors by case number, question number, and type. These are then resolved by senior personell, who inspect the original file and make appropriate corrections.
The interviews averaged 9 min in length. A total of 1098 adults aged 18 and older from Colorado participated in the telephone-based survey after giving consent.
The telephone survey consisted of a series of questions to which each participant responded. The answers to the questions were recorded and put into a database from which these analyses were performed. Questions related to participant characteristics (height, weight, age, income, etc.), self-reported physical activity (e.g., How often do you exercise strenuously? How physically active are you?) or physical inactivity (How many hours do you sit each day? How many hours of TV do you watch per day?), and desire to change (lose weight, increase physical activity, increase walking). Most questions were constructed to be similar to those used in other surveys, such as the behavioral risk factor surveillance system. Our intent was to provide descriptive statistics for walking, and to examine how walking differed based on subject characteristics and on responses to questions about physical activity and eating habits.
After completing the initial telephone interview, step counters (Yamax Model SW-200, Yamasa Corporation, Tokyo, Japan) were mailed to the 1098 participants. Along with the step counter, individuals were sent a protocol describing how and when to use the step counter, a self-administered reply form to record steps, a business reply envelope, and an 800 number to call for support or questions. Participants were asked to wear the step counter for four consecutive days (including one weekend day), and to record their behavior using either the provided reply form or the 800 number. The reply form asked for the date and time the participant started wearing the step counter, and the date, time, and step number displayed on the step counter when they finished wearing it. Those choosing to call the 800 number were administered a short follow-up CATI interview to determine start and stop date and time of the step-counting period, as well as the total number of steps. Reminder letters were mailed to participants 1 wk after the initial mailing.
Sixty-eight percent (or 742 of the 1098 subjects) completing the initial telephone survey also provided step count information. Twelve of the 742 individuals returned step counter information that was unreadable; therefore, we report step data for 730 individuals (344 men and 386 women).
Table 1 shows descriptive characteristics for the subjects completing the self-reported telephone survey (N = 1098) and the subgroup (N = 730) that returned readable step count data. We received more step count responses from older individuals and less step responses from Hispanics in the step subgroup compared with the larger telephone survey population. BMI and gender, however, were similar in both groups. All participants were allowed to keep their step counters after completing the study.
Interviewing, data collection, and data returns took place between January 22 and March 25, 2002. This study was approved by the Colorado multiple institutional review board.
Statistical analyses of the Harris survey were carried out using the survey data commands in Stata 8 (15). For the survey data, sampling weights had been calculated that took into account unequal probabilities of selection resulting from sampling design, nonresponse, and planned oversampling. All analyses took into account these probability weights. This included estimation of proportions, tests of unadjusted mean step counts, and regression analyses that adjusted for demographic variables. P values in Table 3 are based on analysis of variance F-statistics. In Table 6 we report P values based on tests of coefficients equal to zero.
The average adult in Colorado reported taking 6804 steps per day. Figure 1 illustrates the mean steps per day for men and women in the state of Colorado. Table 2 depicts a suggested classification of activity by steps per day proposed by Tudor-Locke and Bassett (19). The percent of Coloradans falling in each classification is also included in Table 2. Based on this classification, 33% of Coloradans have a sedentary lifestyle, 29% are low active, 22% are somewhat active, 9% are active, and 7% are highly active. Table 2 also depicts how this activity classification varies based on current self-reported BMI level.
Table 3 reveals how the number of steps per day differed based on subject characteristics collected in the self-reported telephone survey. Steps per day did not differ significantly between men and women, or as a function of education level or race. In general, steps per day declined with increasing age (P < 0.0001). The exception was the 40- to 49-yr-old group, which reported more steps per day than individuals 30–39 yr of age. Steps per day were directly related to income (P < 0.0001), with steps per day increasing with income level.
Marital status had a significant effect on steps per day. Single individuals walked more than married or divorced individuals, and the lowest walking was seen in widowed individuals. The latter group, however, tended to be significantly older than the other groups, and steps per day were also found to decrease with age.
There was also a significant inverse relationship between body mass index (BMI) and steps per day (P < 0.0001). In this sample, 47% of the sample were overweight (BMI > 25 kg·m−2) and 13% were obese (BMI ≥ 30 kg·m−2). This is similar to information obtained about overweight and obesity prevalence from the Colorado Behavioral Risk Factor Surveillance System (BRFSS) (http://apps.nccd.cdc.gov/brfss/Trends/trendchart.asp?qkey=10010&state=CO). Overweight individuals reported 555 fewer steps per day than normal-weight individuals, and obese individuals reported 2393 fewer steps per day than normal-weight individuals.
We also asked participants if they had been weight stable or had gained or lost ≥5 lb of weight during the last year. Twenty-two percent of individuals surveyed reported gaining more than 5 lb during the last year, and 22% reported losing more than 5 lb during the last year. Although differences in steps per day did not reach statistical significance (P < 0.09), it is interesting to note that, in general, those who reported they had been weight stable over the past year reported higher step per day values than those who either lost or gained weight (Table 4).
Participants were asked whether they were actively trying to lose weight. About half of those surveyed (49%) answered yes to this question. Individuals who answered yes reported an average of 6409 steps per day versus an average of 7218 steps per day for those who answered no to the question (P < 0.018).
Table 5 shows how steps per day varied with answers to specific questions about usual physical activity and inactivity. The specific question asked, and the range of possible responses, is shown for each question in Table 5. Steps per day increased with the number of days that participants reported exercising strenuously (P < 0.0001). Strenuous activity was defined to the subject in the survey as a time when you breathe heavily and your heart and pulse rate are accelerated for a period lasting 20 min or more. Participants were asked how physically active they were, with the four possible responses shown in Table 5. Steps per day were highest in those who reported they were very active, next highest in those reporting they were somewhat active, next highest in those who reported they were somewhat inactive, and lowest in those who reported they were very inactive (P < 0.0001).
Interestingly, only 3% of individuals rated themselves as very inactive, and another 11% rated themselves as somewhat inactive (see Table 5). At the same time, 18% of the survey population reported that they never get strenuous physical activity, and another 17% reported exercising strenuously only once a week. Only 52% of the survey reported exercising strenuously on three or more days each week.
Steps per day were significantly negatively related to hours per day of television watching that participants reported (P < 0.001) and to hours of sitting per day reported by subjects (P < 0.001).
About 70% of the respondents reported that they wanted to be more physically active (Table 5). There was a significant negative relationship between steps per day and rating of walking as a good method to increase physical activity (P < 0.04; Table 5). Those individuals who were least active were more likely to rate walking as a good method of increasing physical activity (Table 5).
We performed a multiple regression analysis to see how steps per day related to the combination of a number of subject characteristics and to physical inactivity. Table 6 shows the results of the multiple regression model. Gender was the only nonsignificant factor included in the model. There are two models illustrated. The first is a model with only significant demographic characteristics. The second includes significant behavioral characteristics (TV watching and hours sitting) as well. The coefficients represent the difference in steps per day from the designated reference group in each category, and are adjusted for all other factors in the model. For TV watching, the coefficient is the drop in steps per day for each hour of TV watching. As an example of interpretation, in the first model obese subjects take about 2004 fewer steps per day than normal-weight individuals adjusted for other demographic characteristics. In the second model, after additional adjustment for time spent sitting and watching TV, obese individuals walked an average of 1633 fewer steps per day than normal-weight subjects. Self-reported sitting and TV watching explains only a small part of the difference in walking between obese and normal-weight subjects.
The models illustrate subgroups at highest risk for low activity. The addition of self-reported TV watching and sitting does not explain the demographic factor effects, because each factor remains significant. After adjustment for all of the factors in the second model, the other behavioral and demographic factors (e.g., diet or education) are still not significantly related to steps per day.
To our knowledge, this is the first study to describe levels of walking, quantified by a pedometer, and determinants of walking in a large population in a statewide survey. There are many other reports of average number of steps per day in small populations. These have been reviewed by Tudor-Locke and Myers (18). The results vary with the group surveyed, with higher step per day values in the younger, fitter samples, and lower step per day values in older, less fit samples. These descriptive data of walking in the population in one of the leanest states may be useful in helping determine step per day targets for the general population. For example, a common recommendation is to aim for 10,000 steps per day. Only 9% of the adult population in Colorado currently meets this goal, and 33% do not even achieve half of this goal. Based on this information, 10,000 steps per day may not be an achievable goal, at least initially for a large proportion of the population. If we set the step per day goal too high initially, we risk discouraging people who make significant increases in walking but do not reach this goal.
A major reason for conducting this survey was to provide a baseline level of walking within the state of Colorado in the hopes that we will be able to increase this level by promoting Colorado on the Move, and thus slow the rate of weight gain in the population. One initial concern was that physical activity would already be high in this population and that it might be difficult to increase it further. This was not the case, because 62% of the Colorado population was classified as either sedentary or low active, based on criteria suggested by Tudor-Locke and Basset (19). This suggests that even in one of the leanest states, there is substantial room for increases in physical activity.
A number of methods, most of them self-report, are used to track physical activity in the population (9,20). Our results show that steps per day, which is an objective measure of walking, tracks well with self-reported physical activity. Having immediate and objective information about the amount of physical activity performed could help promote increases in physical activity in the population. Pedometers are accurate and inexpensive, and hold promise as a tool to help increase population physical activity. We have shown that we can use pedometers to produce short-term increases in physical activity (21), but data about their long-term effectiveness are lacking.
We obtained some useful information about how steps per day varies with characteristics of the participants. Age, marital status, income, and weight category all influenced the number of steps per day. It is generally appreciated that physical activity declines with increasing age (11), and we saw this trend in steps per day. The highest level of walking was seen in 18- to 29-yr-old subjects, and this declined with advancing age. Those who were greater than 60 yr of age averaged about 2000 fewer steps each day than 18–29 yr olds. Increases of 2000 steps per day in individuals over the age of 30 would be sufficient to increase their physical activity level to that of 18- to 29-yr-old subjects. The decline in physical activity that occurs with age is thought to be one factor that could increase the probability of weight gain with age (2,11). These data also suggest that achieving high step population goals, such as getting 10,000 steps per day, will be harder for older Americans than for younger ones.
The fact that steps per day were highest in single individuals may not be surprising, because these individuals may have fewer barriers, such as time and family responsibilities, to being physically active. For similar reasons, it is not surprising that more walking was seen in higher income levels. Obesity also shows an inverse relationship with income (8).
Those participants who were obese reported the lowest number of steps per day. This is consistent with research suggesting that a low level of physical activity is a risk factor for weight gain and obesity (12). The average steps per day were about 2400 per day lower than those with a healthy body weight, and about 1800 per day lower than those who reported being overweight. Even more striking is that 65% of those who reported being obese were in the most sedentary category based on steps per day.
The state survey also provided some interesting information about the degree to which participants would like to change body weight and physical activity. Although generally recognized as one of the leanest (10) and healthiest states, about half of the people in the survey were trying to lose weight, and 70% reported that they would like to be more physically active. Eighteen percent of those surveyed reported that they never engage in strenuous physical activity, and only about half reported that they achieve the recommended goal of engaging in physical activity at least 3× wk−1.
Although most researchers accept that a decline in lifestyle physical activity has likely contributed to the obesity epidemic, specific data to quantify the decline in lifestyle physical activity are lacking (4). These results, in combination with results recently reported by Basset et al. (1), give some indication of the extent to which daily walking has declined over the past century. Bassett et al. (1) used pedometers to assess lifestyle physical activity in 98 Amish adults in southern Ontario. The Amish lifestyle likely approximates the American lifestyle seen before the technological developments of the past several decades. Amish men reported an average of 18,425 steps per day, and Amish women reported an average of 14,196 steps per day. In contrast, Colorado men report an average of 7028 steps per day, and Colorado women report 6606 steps per day. This decline in walking alone, if not compensated for by changes in energy intake or by the body’s regulation of energy balance, could explain far more weight gain that has been seen in the population. These data are certainly supportive of an important role in the decline in physical activity that has contributed to the weight gain of the population over the past decades.
The steps per day difference was not statistically significant depending on whether people were weight stable or had lost or gained at least 5 lb over the past year (P < 0.09). However, some trends here are worth noting. For both healthy BMI individuals and obese individuals, those who reported that they were weight stable over the past year reported more steps per day than those who had either lost or gained at least 5 lb. The results are consistent with the notion that low levels of physical activity can increase the probability of weight gain. Those who are attempting weight loss also report walking less than those maintain a constant weight. Although we do not have information about their walking level before attempting weight loss, this could reflect the fact that most people who attempt weight loss focus much more on food restriction than on increasing physical activity. This is not encouraging, because increasing physical activity is a good predictor of success in weight loss maintenance (6,12,13). It is not clear why this relationship was not seen in the overweight group.
On a positive note, walking was more likely rated as a good way to increase physical activity among those who walked the least. This suggests that those who need to increase physical activity the most may be receptive to programs seeking to increase walking as a means to increase physical activity level. This could be an advantage in implementing interventions in this group.
Overall, these results provided a baseline for the physical activity level of the population in Colorado on the Move before we began to promote it statewide as a weight-gain prevention program. Colorado on the Move has now been implemented in hundreds of schools, work sites, families, and communities across the state. It is our intent to repeat this survey in a few years to document the impact of the program on walking and body weight in the state’s population.
The results provide encouragement for the intent of Colorado on the Move. The aim of Colorado on the Move is for people to increase walking by 2000 steps. We did not provide an overall goal (e.g., 10,000 steps per day) for the population. The results are supportive of not having a single step goal (such as 10,000 steps per day), because so many people in this survey get fewer than 5000 steps per day. Asking them to double their physical activity may be counterproductive in that they may get discouraged if they are unable to make such a big increase in a reasonable period of time. They may be much more successful by striving to make small, incremental increases in walking over time, gradually building up to an activity level they never would have attempted at first. Second, it was interesting in this study that obese individuals reported about 2000 steps per day less than nonobese individuals. Although this cannot be taken as cause and effect, it does suggest that the aim of Colorado on the Move—to increase walking by 2000 steps per day—could be of a magnitude that could aid in prevention of weight gain.
The success of Colorado on the Move has led to the launch of America on the Move (http://www.americaonthemove.org), a national weight-gain prevention program, and to the establishment of 19 “On the Move” state affiliates. It may be useful to conduct a similar survey of walking in other states, both to assess the current level of physical activity and to use this level as a baseline to evaluate the impact of programs aimed at increasing physical activity.
1. Bassett, D. R., P. L. Schneider, and G. E. Huntington. Physical activity in an old order Amish community. Med. Sci. Sports Exerc
. 36:79–85, 2004.
2. Caspersen, C. J., and R. K. Merritt. Trends in physical activity patterns among older adults: the behavioral risk factor surveillance system, 1986–1990. Med. Sci. Sports Exerc
. 24:S26, 1992.
4. Hill, J. O., Wyatt, H. R., G. W. Reed, and J. C. Peters. Obesity and the environment: where do we go from here? Science
5. Iwane, M., M. Arita, S. Tomimoto, et al. Walking 10,000 steps per day or more reduces blood pressure and sympathetic nerve activity in mild hypertension. Hypertens. Res
. 23:573–580, 2000.
6. Klem, M. L., R. R. Wing, M. T. McGuire, H. M. Seagle, and J. O. Hill. A descriptive study of individuals successful at long term maintenance of substantial weight loss. Am. J. Clin. Nutr
. 66:239–246, 1997.
7. Kosta, E. Make every step count. Walking
8. Kumanyika, S., R. W. Jeffery, A. Morabia, et al. Obesity prevention: the case for action. Int. J. Obes
. 26:425–436, 2002.
9. Melanson, E. L., Jr., and P. S. Freedson. Physical activity assessment: a review of methods. Crit. Rev. Food Sci. Nutr
. 36:385–396, 1996.
10. Mokdad, A. H., B. A. Bowman, E. S. Ford, et al. The continuing epidemics of obesity and diabetes in the United States. JAMA
11. Poehlman, E. T. Physical activity level and weight control in older citizens. In: Physical Activity and Obesity, C. Bouchard (Ed.). Champaign, IL: Human Kinetics, 2000, pp. 263–278.
12. Saris, W. H. M., S. N. Blair, M. A. van Baak, et al. How much physical activity is enough to prevent in healthy weight gain? Outcome of the IASO 1st Stock Conference and consensus statement. Obes. Rev
. 4:101–104, 2003.
13. Schoeller, D. A., K. Shay, and R. F. Kushner. How much physical activity is needed to minimize weight gain in previously obese women? Am. J. Clin. Nutr
. 66:551–556, 1997.
14. Simpson, M. E., M. Serdula, D. A. Galuska, et al. Walking trends among U.S. adults. Am. J. Prev. Med
. 25:95–100, 2003.
15. Survey Data Reference Manual. College Station, TX: Stata Press, 2003, pp. 6-98.
16. Tudor-Locke, C. President’s Council on Physical Fitness and Sports: taking steps toward increases physical activity: using pedometers to measure and motivate. Res. Digest
17. Tudor-Locke, C., A. M. Myers, and N. W. Rodger. Formative evaluation of the First Step Program: a practical intervention to increasing daily physical activity. Can. J. Diabetes Care
18. Tudor-Locke, C., and A. M. Meyers. Methodological considerations for researchers and practitioners using pedometers to measure physical (ambulatory) activity. Rev. Q Exerc. Sport
19. Tudor-Locke, C., and D. R. Basset. How many steps/day are enough? Preliminary pedometer indices for public health. Sports Med
. 34:1–8, 2004.
20. Westerterp, K. R. Assessment of physical activity level in relation to obesity: current evidence and research issues. Med. Sci. Sports Exerc
. 31:S522, 1999.
21. Wyatt, H. R., J. C. Peters, G. W. Reed, et al. Using electronic step counters to increase lifestyle physical activity: Colorado on the Move. J. Phys. Activity Health