Achieving Energy Balance at the Population Level Through Increases in Physical Activity

Abstract

Objectives. We estimated the amount of physical activity required for individuals to expend an additional 418.4 kJ (100 kcal) per day with the goal of achieving energy balance at the population level.

Methods. Data on total daily energy expenditures were derived from a random sample of adults residing in Geneva, Switzerland, who completed a self-administered physical activity frequency questionnaire. These data were used to simulate the effects of typical physical activity pyramid recommendations on average population energy expenditures for various activity intensities and rates of population compliance with pyramid recommendations.

Results. If an average 418.4 kJ (100 kcal) per day increase in energy expenditures is to be achieved, assuming 100% compliance with physical activity pyramid recommendations, the bottom tier of the pyramid must correspond to everyday activities performed at moderate to high intensity levels (e.g., moderate walking or biking). Expected population gains in energy expenditures would be only 167.4 to 251.0 kJ (40 to 60 kcal) per day at a 50% compliance rate.

Conclusions. Achieving population-level energy balance through increasing energy expenditures with physical activity increases alone would require profound structural and environmental changes promoting more active lifestyles.

Physical inactivity increases people’s risk of obesity, ill health, and premature mortality.1 It is therefore paramount to determine the levels of physical activity that should be recommended to substantially reduce such risks at the population level.2^–4

Hill et al.2 estimated that, on average, an extra 418.4 kJ (100 kcal) per day must be expended to restore energy balance and thus eliminate weight gain in Western populations. To achieve this goal, energy expenditures must be increased or energy intakes must be decreased. It remains unclear whether this goal can be achieved at the population level solely through increasing physical activity levels. For example, consider a hypothetical mass public health campaign intended to increase walking; expending an extra 100 kcal per day would require approximately 60 minutes of slow walking, or 30 minutes of moderate or brisk walking, with 50% compliance with the campaign’s recommendations at the population level.3

To promote physical activity, national agencies have issued activity recommendations,5^,6 often translated into a pictorial model such as a “pyramid”7^–15 or a “rainbow.”16 Figure 1 ▶ depicts a generalized model of a physical activity pyramid comprising 3 energy expenditure tiers. The bottom tier recommends engaging in everyday activities of at least moderate intensity (energy expenditures of 4184 kJ [1000 kcal] per week and above), such as climbing stairs, walking or bicycling to work, and gardening, for 210 minutes per week or more. The middle tier recommends engaging in endurance and cardiorespiratory recreational or competitive exercises and sports for at least 90 minutes per week in combination with strength training and flexibility exercises for at least 30 minutes per week.

FIGURE 1—

Generalized physical activity pyramid model quantified for simulation purposes.

Note. Only the bottom and middle tiers were considered in the simulations. Each of the 4 bottom-tier scenarios defining “moderate” physical activity in terms of minimum recommended basal metabolic rate (BMR) multiples of energy expenditures was examined in the simulations.

Individuals already complying with the bottom-tier recommendations are urged to move to the middle tier and then the top tier to further increase their health benefits. For example, individuals already complying with the bottom- and middle-tier recommendations could, as recommended in the top tier, either engage in further higher-intensity exercise and sports or reduce the extent to which they engage in sedentary activities (e.g., watching television, talking on the telephone, playing or working at the computer).

We simulated the potential population gains in total energy expenditure achievable by adults complying with the bottom- and middle-tier recommendations outlined in the pyramid depicted in Figure 1 ▶. These simulations were based on extensive, real data on the total (i.e., not only leisure time) energy expenditures of a random sample of 8528 urban adults obtained with a survey instrument developed and validated in a separate random sample from the same population.17 We estimated potential intervention effects after accounting for those adults already complying with the recommended level of physical activity.

METHODS

Description of Survey and Sample

The Bus Santé, initiated in 1993, is an ongoing, community-based surveillance project designed to monitor the prevalence of chronic disease risk factors among the approximately 100 000 men and 100 000 women who were noninstitutionalized residents of Geneva, Switzerland (city and canton), aged 35 to 74 years.18 Eligible participants, primarily French speaking, were identified from a computerized residential list established by the local government through a standardized protocol. The list included all potentially eligible individuals other than those living in the country illegally.

These population-based surveys are administered to independent, cross-sectional, gender-and age-stratified random probability samples (600 women and 600 men per year) that are proportional to corresponding population distributions. Each participant in our study completed 1 of the 7 surveys conducted from 1997 (when the physical activity questionnaire, described subsequently, was first administered) through 2003.

Selected Bus Santé participants are mailed an invitation to take part in the survey, and if they do not respond, up to 7 follow-up telephone attempts are made at different times on various days of the week. If telephone contact is unsuccessful, 2 additional letters are mailed. The recruitment process for each participant extends over a minimum of 2 weeks to a maximum of 2 months, after which no further attempts are made to contact the participant. Those not reached because they are no longer residents of the canton (this was the case among 15% of men and 19% of women between 1993 and 2003) are not eligible for the study, and those who refuse to participate are not replaced. Participants completing the survey in a given year are excluded from future surveys. Annual participation rates have ranged from 57% to 65%.

Physical Activity Frequency Questionnaire

Since 1997, the Bus Santé survey has included a validated, self-administered quantitative physical activity frequency questionnaire (PAFQ) designed to measure respondents’ total and activity-specific energy expenditures, with special attention to light- and moderate-intensity activities.17 A 2-step approach was used in developing the PAFQ.

First, 24-hour recall data from telephone interviews administered by trained public health technicians to 919 randomly selected Geneva adults in 1994 were used to establish, separately by gender, the physical activities that contributed 95% of total energy expenditures. Activities that were performed infrequently by the sample as a whole but that contributed 10% or more to any single individual's total energy expenditure were also included. The final PAFQ incorporated 70 activities grouped according to general types (occupational, daily living, getting to work or elsewhere, leisure time, and recreational or competitive sports) and typical durations.

Second, the PAFQ was validated against heart rate monitoring data. Results showed that PAFQ-estimated total energy expenditures among 41 volunteers correlated equally well (r= 0.76) with heart rate monitor estimates and with 24-hour recalls of physical activity.17

Physical Activity Intensity Levels

A physical activity can be categorized as more or less intense according to the energy expended when it is performed during a specific time period. An activity’s intensity level can be expressed as a multiple of the gender-, age-, height-, and weight-specific basal (resting) metabolic rate (BMR) of the individual engaging in the activity.19 The BMR for a typical Western adult is 4.184 kJ (1 kcal) per minute. For example, walking slowly expends 3.1 times one’s BMR. Hence, someone with a BMR of 4.184 kJ (1 kcal) who walks slowly for 15 minutes expends 194.6 (3.1 × 4.184 × 15) kJ [46.5 (3.1 × 15) kcal]. Thus, to expend 418.4 kJ (100 kcal) in 15 minutes, that person must walk at an intensity of 6 to 7 BMR, which corresponds to athletic walking.

Examples of the energy expenditure intensities of the physical activities included in the PAFQ follow (the French-language questionnaire, augmented with the BMR intensity of each activity, is available from the authors). In a previous study, we reported that engaging in fewer high-intensity activities (expending at least 4.0 times the BMR), as opposed to engaging in fewer moderate-intensity activities (expending 3.0–3.9 times the BMR), was associated with higher odds of being overweight or obese than being of normal weight.4

Physical activities found to expend at least 4.0 times the BMR (“4.0+ BMR group”) were occupational activities such as heavy construction work (7.7 BMR), gardening at work (4.1 BMR), and brisk or uphill walking (4.5 BMR); leisure time activities such as pitch forking (5.7 BMR), splitting logs (7.0 BMR), stacking firewood (5.0 BMR), and climbing stairs (6.0 BMR); activities associated with getting to work or another destination, such as brisk/uphill walking (4.7 BMR), climbing stairs (6.0 BMR), and bicycling quickly (15 km per hour; 5.7 BMR); and sports activities (4.0 to 12.0 BMR). Activities (which include those just listed) expending at least 3.9 BMR (“3.9+ BMR group”) were moderate walking and carrying or pulling loads while walking (in the getting to work or elsewhere category).

In addition, activities expending at least 3.7 BMR (“3.7+ BMR group”) were moderate walking and carrying or pulling loads while walking at work (from the occupational category), along with bicycling slowly (9 km per hour; from the getting to work or elsewhere category). Finally, activities expending at least 3.5 BMR (“3.5+ BMR group”) were the occupational activities of caring for babies or toddlers and engaging in automobile work and the leisure time activities of caring for babies or toddlers at home and gardening at home.

Simulation Assumptions and Calculations

We initially used the 1997 to 2003 PAFQ data to estimate the existing total energy expenditure distributions (kJ [kcal] per day) in the sample, as well as the energy expenditure distributions associated with the physical activities making up the 3.5+, 3.7+, 3.9+, and 4.0+ BMR intensity activity groups just described. We then simulated the potential effects on total energy expenditures of increasing various combinations of these physical activities according to the pyramid recommendations in Figure 1 ▶.

In the simulations, pyramid recommendations were quantified as follows. The bottom-tier recommendation was 210 minutes per week (7 × 30 minutes per day or the equivalent) of physical activity at the 3.5+, 3.7+, 3.9+, or 4.0+ BMR level (4 different scenarios). The middle-tier recommendation was 90 minutes per week (3 × 30 minutes per day or the equivalent) of exercise or sports activity at the 4.0+ BMR level in combination with 30 minutes per week (2 × 15 minutes per day or the equivalent) of “strength training and flexibility exercises” at the 4.0 BMR level.

As indicated by the “or the equivalent” designations, only total time spent per week was held fixed in the simulations. For example, individuals could also achieve the recommended 90 minutes per week of endurance and cardiorespiratory exercise sports activities (middle tier) by engaging in these activities 45 minutes per day twice a week. Research has suggested that total time spent per week captures the health benefits of physical activity provided that each session is a minimum of 10 minutes in duration.20^,21 For practical purposes, the additional top-tier recommendations of increasing high-intensity exercise and sports activities and limiting sedentary activities were ignored. Thus, the simulation results represented “minimal” compliance scenarios in that they were based only on compliance with the pyramid’s bottom- and middle-tier recommendations.

Two subgroups of adults were established. First, “compliers” were defined as adults who were already engaging in the bottom-and middle-tier activities at least at the recommended intensity levels and durations and would continue doing so. Second, “eligible adults” were defined as all individuals who were not compliers. In the simulations, these individuals were assigned to engage in activities at exactly the minimum recommended bottom- and middle-tier durations and intensity levels. In other words, adults who were not compliers were assumed to increase their level of activity as needed to comply exactly with the minimum recommendations of the bottom and middle tiers of the pyramid.

We calculated the maximum potential effects of the bottom-tier pyramid recommendations separately for each of the 4 BMR intensity groups assuming 100% participation by eligible adults. We then repeated these calculations assuming 50% participation by eligible adults. We simulated this 50% participation rate by generating a pseudo-random value between 0 and 1 for each eligible adult, who was then coded as a complier if the corresponding random value was 0.50 or below and as a noncomplier if the value was above 0.50.

RESULTS

Detailed characteristics of the Geneva adult population, such as physical activity levels and prevalence of overweight and obesity, hypercholesterolemia, diabetes treatment, and hypertension, have been reported elsewhere.3^,4^,22 For example, in earlier studies we found that 58% and 70% of Geneva men and women, respectively, were sedentary (less than 10% of total energy expended in activities at the 4.0+ BMR intensity level)3^,4; 44% and 13% are overweight; and 24% and 9% are obese.22 Thus, Geneva adults were less sedentary and overweight than were US adults (according to data from the Centers for Disease Control and Prevention1). Moreover, the Geneva adult population is among the leanest of those that participated in the Europe Alimentation project.22

Figure 2 ▶ illustrates the range of expected shifts in the population distributions of total energy expenditures corresponding to bottom-tier pyramid recommendations of 30 minutes per day of physical activity at either the 3.5+ or 4.0+ BMR intensity level at a 50% or 100% compliance rate on the part of eligible adults. In the case of 100% compliance, even bottom-tier pyramid recommendations corresponding to a 3.5+ BMR would produce a noticeable population change. Assuming a more realistic 50% compliance rate, however, not even a bottom-tier recommendation corresponding to a 4.0+ BMR would produce a dramatic distributional shift at the population level.

FIGURE 2—

Examples of simulated changes in total energy expenditure (EE) distributions for physical activity pyramid bottom-tier recommendations: 30 minutes per day of 3.5+ basal metabolic rate (BMR) activities (100% compliance; a), 30 minutes per day of 4.0+ BMR activities (100% compliance; b), 30 minutes per day of 3.5+ BMR activities (50% compliance; c), and 30 minutes per day of 4.0+ BMR activities (50% compliance; d).

Figure 3 ▶ summarizes the sample’s expected median gains in energy expenditures corresponding to various levels of compliance by eligible adults with bottom-tier physical activity recommendations based on a 3.5+, 3.7+, 3.9+, or 4.0+ BMR. Even at compliance rate of 100%, the bottom tier of the pyramid needs to comprise daily activities of at least 3.9 BMR to achieve an increase of 418.4 kJ (100 kcal) per day in daily energy expenditures at the population level. If the compliance rate is 50%, the expected median gains in energy expenditures are only between 167.4 and 251.0 kJ (40 and 60 kcal) per day.

FIGURE 3—

Sample median gains in energy expenditure (kJ per day), stratified according to level of compliance with 3.5+, 3.7+, 3.9+, and 4.0+ basal metabolic rate (BMR) bottom-and middle-tier physical activity pyramid recommendations.

DISCUSSION

Our simulations showed that adults need to engage in high levels of physical activity (i.e., in terms of intensity, frequency, duration, and compliance) if an average increase in energy expenditures of 418.4 kJ (100 kcal) per day is to be achieved at the population level. This study had 3 primary strengths: (1) it was based on extensive real data covering the total spectrum of physical activities engaged in by adults in a noninstitutuionalized European urban population; (2) the physical activity frequencies and durations used in the calculations were measured with a survey instrument developed and validated with this same adult population; and (3) the potential effects of various hypothetical mass campaigns designed to increase physical activity levels were estimated after realistic consideration of the percentage of adults already complying with typical physical activity recommendations such as those quantified in the present simulations.

Three possible limitations of this study were practically unavoidable. The first limitation involved the specific manner in which the physical activity pyramid was constructed and its recommendations quantified to generalize across the many different pyramids or other models used by various countries; however, these study features facilitate comparisons between the generalized pyramid model assessed and specific national guidelines (e.g., 210 minutes per week of 3.9 BMR activities in our generalized model correspond to the 150 minutes per week of 5.5 metabolic equivalent activities recommended by the Centers for Disease Control and Prevention23). A second limitation was our assumption that engaging in a given physical activity in a single “marathon” session and doing so in several sessions of shorter but equal total duration are equivalent in terms of their health benefits.

A final limitation was the assumption that people will increase their physical activity levels while keeping their dietary energy intakes constant. Given that people who become more physically active are also likely to feel more hungry—and hence may overcompensate for their additional energy expenditure by increasing their dietary intake—this issue merits further attention as an additional objective of health promotion efforts designed to achieve energy balance at the population level.

Several key issues remain to be addressed in terms of the potential impact of a mass public health campaign aiming to achieve energy balance24^–26: which physical activities (e.g., walking, bicycling, climbing stairs) are most likely to be adopted by the general public? At which intensities, frequencies, and durations should these activities be engaged in to achieve benefits? How will different patterns of participation in the campaign overall and among various subgroups (e.g., subgroups defined according to gender and age) affect its success at the population level? Which urban environment changes are likely to facilitate widespread adoption of the targeted physical activities by the general public?

The present results, which indicate the difficulty involved in increasing the average energy expenditure of a given population by 418.4 kJ (100 kcal) per day through increases in physical activity alone, emphasize the importance of urging populations as a whole to decrease their energy intakes, increase their energy expenditures, or both. Expending more energy and eating less may not be realistic for most people, but controlling energy intakes may be a viable alternative for people unwilling or unable to exercise more. Achieving energy balance through diet alone might not be as feasible as it appears, however. For example, most snack bars and 12- or 16-oz (360- or 480-mL) cans of most regular sodas contain more than 418.4 kJ (100 kcal) per serving. Thus, although reducing daily energy intakes by 418.4 kJ (100 kcal) may appear to involve less significant lifestyle changes than expending 418.4 kJ (100 kcal) per day by engaging in additional moderate to vigorous physical activity, it may prove to be no less challenging for individuals who regularly consume these “single serving” amounts either without being aware of or ignoring the energy content.

Public health professionals use models such as the generalized physical activity pyramid shown in Figure 1 ▶ to guide the public toward becoming more physically active. However, mass public health campaigns designed to increase physical activity levels need to distinguish between short-term and middle- or long-term gains in view of specific targeted health outcomes and should consider pyramid recommendations as lower rather than upper limits for the activity in question. Effectively dealing with the latter issues in designing successful physical activity interventions will have major societal consequences. For example, if the objective of full compliance with pyramid guidelines is to be met, safer and better-equipped urban environments are needed to encourage individuals to be physically active in their everyday lives.26