The health benefits associated with physical activity are well established (16), and the magnitude of health improvements generally increases as the volume of activity is increased (17). However, a central issue for understanding the effect of physical activity exposure, or dose, on fitness and health outcomes is the ability to empirically evaluate and compare the actual dose to the recommended or prescribed dose of physical activity (21). Factors that influence biological adaptability to exercise may also influence exercise tolerance (4) and may help to define which exposures will be effective for sustaining participation. For example, higher intensity has been associated not only with greater improvements in health and fitness and increased longevity (7,13,19) but also with potentially poorer adherence (6,18). As stated in the 2008 Physical Activity Guidelines for Americans, studies that evaluate the effects of intensity, frequency, duration, and multiple exercise bouts are needed to fill the gaps in our knowledge about dose response (16). Recent research has provided evidence that up to 7% of individuals exposed to physical activity may actually experience deleterious health outcomes (3). Thus, the ability to quantify total exercise exposure is a critical factor for identifying individuals most likely to respond, positively or negatively, to exercise training.
The intensity, frequency, and duration of physical activity and exercise sum up to a total volume of exposure (9,16). Although about 60% of US adults report meeting the recommended volume of physical activity (approximately 500 MET·min·wk−1) (24), less than 5% meet the physical activity criterion when physical activity volume is estimated objectively by an accelerometer (23). Moreover, similarly, objective estimates show that less than 2.5% of Americans meet recommendations for vigorous physical activity necessary to improve fitness (24).
Population studies have largely relied on participant recall and crude estimates of intensity derived from physical activity types (1) or from motion detectors that do not indicate the intensity of exercise relative to a person’s level of fitness. Early clinical trials of exercise commonly reported that nearly half of the participants dropped out before healthful adaptations could occur or be identified (15). Many efficacy trials report high adherence rates (e.g., 75%–85% of the prescribed number of sessions in trials lasting 6–24 months), but trials rarely report the extent to which participants who attend also comply with the prescribed intensities and durations of exercise, despite early recommendations that this be done (14). The use of objective measures of intensity, duration, frequency, and total volume of physical activity exposure in population studies has been identified as a key research need (16).
HR monitoring is an objective, relatively inexpensive measure that has been used to successfully evaluate exercise intensity and duration among both healthy and medically ill adults (5,11,20); it thus provides information about the volume of exercise that can be used to judge compliance beyond that provided by measures of session attendance or frequency of exposure (11). Here, we report the development and application of an empirical measure of exercise compliance using HR monitor-based measures of intensity, duration, and frequency in a large prospective study of young adults undergoing 15 wk of aerobic exercise training. HR monitor data were used to compute total exercise dose in the form of an HR physical activity score (HRPAS) (described in the succeeding part of this article), which was used to objectively assess compliance with the exercise prescription. Health-related outcomes were examined for association with the HRPAS and compared between compliant and noncompliant individuals and with adherence defined by attendance alone.
Study design and sample
The Training Interventions and Genetics of Exercise Response (TIGER) Study is a prospective cohort study with the goals of introducing sedentary college-age adults to regular exercise and identifying genetic factors that influence physiological responses to exercise training and exercise adherence. The target participant for the TIGER study is a sedentary (i.e., <30 min of activity per week for the previous 30 d before enrollment) individual who was not restricting energy intake for weight loss. Exclusion criteria included having a physical contraindication to aerobic exercise (e.g., cardiomyopathy), a metabolic condition that may alter body composition, and/or pregnancy. All participants provided written informed consent, and the study protocol was approved by the Institutional Review Boards at the University of Houston, Baylor College of Medicine, and the University of Texas Health Science Center at Houston.
To meet contemporary guidelines for vigorous activity (16), prescribed aerobic exercise training included three 30-min exercise sessions per week at 65%–85% of age- and sex-predicted maximum HR reserve (HRR) for 15 wk using the subject’s choice of treadmill, elliptical trainer, stair stepper, or exercise bike. Participants were permitted to exercise more frequently and for longer durations (up to 60 min) than the prescribed 30 min. During each exercise session, participants wore portable HR monitors (Polar Electro, Lake Success, NY), and mode of exercise was documented. The monitors recorded minute-to-minute HR, date, time, and duration for each exercise session and have been shown to provide a valid measure of exercise HR (12). Data from the HR monitors for each participant were downloaded into the manufacturer’s software program (E-Series, Polar Electro) and merged with attendance files to formulate a comprehensive database of exercise parameters. Participants were required to complete a minimum of 25 min within their target HR zone (THRZ) for an exercise session to be considered valid.
Physiological testing was completed at baseline and after 15 wk of training. Stature was measured to the nearest centimeter using a freestanding stadiometer (SECA Road Rod; Seca GmbH, Snoqualmie, WA), and body mass was measured to the nearest 0.1 kg using a digital scale (SECA 770; Seca GmbH, Hanover, MD). Percent body fat (%Fat) was estimated using dual-energy x-ray absorptiometry (DXA) (Hologic, Bedford, MA), and body mass index (BMI) was calculated from stature and body mass (kg·m−2). Resting blood pressure (BP) was measured three times using a digital BP monitor (Omron HEM 907; Omron Healthcare, Inc., Bannockburn, IL) and calculated from the average of the second and third measures taken after the participant had been sitting quietly for at least 5 min. Phlebotomy was performed after an overnight fast, and blood samples were collected from a peripheral arm vein into evacuated tubes treated with ethylenediaminetetraacetic acid. Plasma was separated from the packed cells by centrifugation, aliquoted, and frozen (−80°C) until further analysis. Plasma was analyzed for total cholesterol using a portable analyzer (CardioChek, Brooklyn, NY) and standard chemistry on the basis of the Trinder method (22), and glucose was determined by using a calibrated glucose analyzer (YSI 2300 Stat Plus; YSI Life Sciences, Yellow Spring, OH).
Calculation of exercise dose
Approximately 83.2% (27,883/33,473) of the attendance records documented during the semester had usable HR observations. The remaining 16.8% of attendance records had either missing or unusable HR data due primarily to technical errors with the monitors. Because the participants were known to have exercised during the sessions that were missing HR data, values for the missing/excluded exercise data were imputed on the basis of the within-participant distributions of duration and average HR across all nonmissing exercise sessions. For each documented exercise session (based on attendance records) for which HR monitor data were unavailable, a randomly generated z-score value was converted to an imputed parameter value on the basis of the within-subject mean and SD of the exercise parameter values (e.g., average HR and duration), under the assumption that the data were missing at random. Imputation of missing values required that valid data be available for at least 60% of all possible exercise sessions for each subject. Imputed values that were outliers from the original exercise parameter distributions across all participants were eliminated, and imputed values were recalculated. After imputation, the distributions of average HR and duration with and without imputed values were statistically compared; this imputation approach was demonstrated to produce unbiased values with respect to means, variances, homoscedasticity, and kurtosis.
To quantify total exercise dose, duration for each session was adjusted for average exercise intensity (%HRR) to create a measure of intensity minutes for each workout, which were summed over all exercise sessions to formulate the HRPAS. Exercise duration was recorded as the time the monitor was stopped minus the time the monitor was started. Because resting HR is difficult to measure accurately and consistently across subjects, percent of predicted HRR (%HRR) was estimated by dividing average exercise HR by each individual’s age- and sex-predicted maximum HRR, using sex-specific constants for resting HR, as described elsewhere (2,10). The HRPAS was calculated in two steps. A workout HRPAS (W-HRPAS) was first calculated for each exercise session by adjusting exercise duration in minutes (ExMin) by exercise intensity:
Exercise frequency was determined by the number of exercise sessions attended across 15 wk, and the HRPAS was then calculated as the sum of each W-HRPAS across the 15-wk program:
Total available sessions for each cohort varied because of differences in semester length and weather-related university closures and consisted of 34, 30, 36, and 37 sessions, respectively, for each of the four cohorts examined in this study. These exercise sessions occurred over a period of 15 wk. To compare values across cohorts, a normalized HRPAS, adjusted for the different numbers of possible workouts in each cohort, was also calculated.
HRPAS as a measure of compliance
Compliance was evaluated by comparing observed HRPAS values to prescribed HRPAS values on the basis of a minimum %HRR of 65% for at least 30 min per session. For example, the prescribed HRPAS for a total of 34 sessions was equal to the following:
Compliance was defined as an observed HRPAS equal to or greater than this prescribed HRPAS.
To compare the HRPAS as a measure of compliance with a simple attendance criterion of adherence, participants who attended at least 80% of their scheduled exercise sessions were assigned as adherent and participants who attended less than 80% of their exercise sessions as nonadherent. Because the TIGER protocol is administered within the structure of a college course for credit, this attendance criterion was based on the minimum class participation needed to pass the course.
All statistical analyses were performed using Stata version 11.1 (StataCorp Inc., College Station, TX). Histograms, normal probability plots, central tendency and variability measures, and zero-order product–moment correlation coefficients were used to examine the distributions and bivariate associations, respectively, of exercise frequency (i.e., attendance), exercise HR, %HRR, ExMin, and HRPAS. On the basis of HRPAS as the dependent variable, exercise frequency, %HRR, and ExMin were included in multiple linear regression analysis adjusted for age, sex, race/ethnicity, and BMI to determine the variance in HRPAS accounted for by the exercise components independent of the covariates. Multiple regression was also used to examine the association between HRPAS and changes in physiological outcomes, including BMI, %Fat, waist and hip circumferences, resting systolic and diastolic BP, resting HR, estimated aerobic capacity, fasting glucose, and fasting total cholesterol. Each dependent variable was analyzed using a separate model adjusted for age, sex, race/ethnicity, and baseline value of the measure. The tests of the respective regression coefficients were evaluated to determine whether HRPAS was associated with changes in the outcome variables independent of baseline values and the other covariates. Logistic regression was used to examine predictors of exercise compliance and adherence. For all tests, statistical significance was set at P < 0.05.
HR monitor data
Participants were excluded from analysis if no exercise sessions were completed (n = 132) or if more than 40% of HR observations were missing (n = 147). The final sample consisted of 1150 participants (447 men and 703 women) whose ages ranged from 18 to 35 yr (Table 1). The racial/ethnic groups most frequently represented by the majority of participants in this study were non-Hispanic white (28.5%), African American (27.3%), Hispanic (23.7%), and Asian (7.4%). Participants completed an average of 29.1 (SD = 6.9) exercise sessions at an average duration of 38.4 min (SD = 3.7 min). Average HR was 156 bpm (SD = 8 bpm), and mean %HRR was 67.9% (SD = 5.8%).
The mean calculated normalized HRPAS was 739 (SD = 202) intensity minutes. The HRPAS was positively related to all three components of exercise, most strongly to frequency (r = 0.87, P < 0.001) and also significantly to duration (r = 0.34, P < 0.001) and intensity (r = 0.38, P < 0.001). The three components together explained 98% of the variability in the HRPAS (i.e., R 2 = 0.98, P < 0.001). After accounting for these components, the HRPAS was unassociated with age (P = 0.570), sex (P = 0.220), race/ethnicity (P = 0.395), or BMI (P = 0.156).
A total of 868 participants (75.5%) were identified as compliant and 282 (24.5%) were identified as noncompliant on the basis of the HRPAS, whereas 885 participants (77.0%) were defined as adherent and 265 (23.0%) were defined as nonadherent on the basis of attendance. Concordance between adherence defined by attendance and compliance defined by HRPAS is depicted in Table 2. The HRPAS classified 9.1% (n = 105) of the participants differently for the adherence criterion than for the compliance criterion. Forty-four participants (3.8%) who were identified as compliant by the HRPAS criterion were classified as nonadherent by attendance alone. These participants exceeded the prescribed HRPAS despite attending fewer sessions by exercising for a longer duration per session (40.5 ± 2.8 min) and/or at a higher relative intensity (70.1% ± 4.2%) than prescribed. By contrast, 61 participants (5.3%) identified as noncompliant by the HRPAS criterion were classified as adherent by attendance records alone. These participants failed to meet their prescribed HRPAS despite attending most of the sessions, primarily because of noncompliance with the exercise intensity prescription (i.e., average relative intensity was lower (60.0% ± 5.5%) than the prescribed intensity). After controlling for age, sex, race, and BMI, both mean duration and average intensity (%HRR) were significantly predictive of both noncompliance (P < 0.001) and nonadherence (P < 0.01).
HRPAS and health-related risk factors
After adjusting for age, sex, race, and baseline values of each measure, HRPAS exercise dose was significantly associated with positive change in BMI, body mass, waist and hip circumferences, resting HR, %Fat, systolic BP, and fasting glucose and cholesterol (Table 3). Standardized regression coefficients, which represent the amount of change in each physiologic variable associated with change in each of the measures of compliance/adherence in SD units, allow for comparison across analyses (see Table, http://links.lww.com/MSS/A304, Supplemental Digital Content: Regression analyses for absolute change in health-related outcomes for HRPAS, HRPAS-based compliance, and attendance-based adherence). Only change in diastolic BP was not associated with HRPAS, possibly because of the narrow range of this variable in this young, healthy cohort. Conversely, compliance (defined by HRPAS) and adherence (defined by attendance) cut-points were only associated with changes in body mass, hip circumference, %Fat, and resting HR. HRPAS-based compliance was also associated with BMI and waist circumference, whereas attendance-based adherence was associated with cholesterol change (Table 3). Unadjusted absolute differences in health-related outcomes by compliance status are summarized in Figure 1. Compliance with the prescribed protocol across 15 wk was associated with an average decrease of 1.4 kg in body mass, a mean of 4.5 bpm decrease in resting HR, and approximately 1-cm decrease in waist and hip circumferences independent of age, sex, and race. Mean changes in physiological parameters by combined compliance/adherence cut-points are summarized in Table 4. Standardized mean differences and confidence intervals, along with effect sizes (Hedges d), by compliance/adherence cut-points are provided in the Supplemental Table, http://links.lww.com/MSS/A304. Combined noncompliance and nonadherence was associated with the poorest 15-wk outcomes (Table 4). Individuals who were compliant with the exercise protocol but nonadherent based on attendance did not differ significantly for any physiologic parameter measured from those who were classified as both adherent and compliant by their attendance and HRPAS.
In this study, exercise dose was quantified using objective HR monitoring as a measure of exercise intensity, duration, and frequency. The effects of HRPAS on biomarkers of health risk were compared with the effects when adherence was defined on the basis of either compliance or attendance cut-points. Larger values of HRPAS were significantly associated with improvements in multiple physiologic parameters, regardless of whether participants met a definition of adherence on the basis of attendance only. The positive association between HRPAS and these health-related measures is consistent with the expected effect of greater amounts of physical activity improving health and fitness (9).
On the basis of an empirical measure of exercise dose, compliance to an exercise prescription can be achieved by modifying any of the components of the prescription, including frequency, duration, and intensity. In addition to the direction (i.e., adoption of an exercise program) and persistence (e.g., attendance or completion of a program) of behavior, motivation theory incorporates the intensity of behavior as an important component of physical activity. Individuals in this study who were compliant but nonadherent based on attendance did so by increasing the intensity and/or duration of each session performed. Although the average duration of each session was similar in compliant and noncompliant participants (38.8 ± 3.6 vs 37.0 ± 3.7 min, respectively), exercise intensity was substantially higher in compliant participants compared with those who were noncompliant (68.8% ± 5.2% vs 65.2% ± 6.4%, P < 0.001). Intensity has often been neglected in clinical trials of the determinants and outcomes of exercise training, which are generally limited to activity type in most population surveys (6). Importantly, compliant individuals who were nonadherent based on attendance did not differ significantly on any physiologic response parameter measured from those who were classified as both adherent and compliant, suggesting that total exercise dose may be more informative and predictive of change in health-related outcome than cut-point measures.
Exercise dose is a complex stimulus involving not only attendance but also actual duration and intensity of exercise in each exercise session accumulated across all sessions of an intervention. By using session attendance as a proxy for exercise exposure, the true effect of exercise may be over- or underestimated because of unknown exposure to the active feature of the exercise stimulus. Here, we quantified exercise exposure by adjusting the duration of each session by the intensity of the exercise performed. Thus, qualitatively different 30-min exercise bouts performed at 50% and 70% intensity can be quantified using this approach (i.e., W-HRPAS for these sessions would be 15 and 21, respectively, whereas attendance measures would give equal credit to the exercise bouts).
There has been continued interest in the relative merits of 1) accumulation of total volume, 2) plausible differences between short exercise bouts versus long, and 3) continuous exercise bouts. Recently, Glazer et al. (8) reported that the total number of minutes of moderate-to-vigorous physical activity was significantly related to blood lipids, BMI, waist circumference, and the Framingham risk score (P < 0.0001), regardless of whether the physical activity was accumulated in bouts shorter or longer than 10 min. A wide range of duration may provide equivalent benefits when similar total daily and weekly volumes are accumulated (17). The results reported here confirm that objective assessment of exercise dose using HR monitoring provides unique associations with health-related outcomes beyond those observed when adherence is defined only by session attendance. As long as the product of intensity and duration is at least the prescribed value (for each session and the total across sessions), the health benefits appear to be similar. In other words, an individual can choose to exercise at a lower intensity for a longer duration or for a shorter duration but higher intensity and get similar health benefits. One could even “skip” sessions (frequency) and “make it up” by going longer and/or harder the next time—from a public health perspective, this is an important message.
HR monitoring allows for calculation of HRPAS as both a measure of total exercise dose and a measure of exercise compliance during an intervention. In addition, most HR monitors provide audible feedback to the participants when they are in their target training zone, serving as an excellent tool for teaching participants how to exercise in an HR range most likely to elicit change. Compliant participants in this study spent significantly more time each session in their THRZ (i.e., 65%–85% maximum HRR) compared with noncompliant participants (28.5 ± 4.1 vs 25.1 ± 6.6 min, respectively; P < 0.001) despite exercising for similar total duration. The ability to efficiently reach one’s THRZ may play an important role in compliance to the exercise prescription. Given expert opinions that “just showing up” to exercise is not enough (17,25), it is particularly noteworthy that HRPAS provided a conceptually more comprehensive measure of exercise response than using only attendance records, the most common method of evaluating exercise adherence.
Although a defined exercise prescription naturally creates a cut-point for compliance, our results demonstrate that increasing levels of total exercise dose are associated with greater health and fitness benefits, as outlined by contemporary guidelines (9,16,17), beyond simply achieving the minimum prescribed exercise dose. These results also suggest that healthy young adults are willing and able to exercise at higher levels than those currently recommended and that higher intensity and/or longer duration of each exercise session are both associated with better exercise adherence.
This study had several methodological and conceptual strengths that represent improvements over previous exercise adherence research. These include 1) the recruitment of a large, ethnically diverse cohort of young adults; 2) the use of a standardized, supervised exercise program designed in accordance with guidelines for vigorous physical activity (16); 3) the use of objective measures of frequency, intensity, and duration of physical activity; 4) the use of individualized target heart prescriptions; and 5) the ability to download data from the HR monitors to a computer database. Future studies that include a full assessment of HRPAS over a broader range of exercise intensities and durations will provide information that is more complete regarding the role of the components of exercise dose in adherence and compliance.
HR monitoring is an objective, practical measure of physical activity and presents low interference with normal activities. Importantly, HR monitors can be used to teach individuals how to exercise at a level most likely to elicit physiologic change. The HRPAS is a quantifiable measure of exercise dose that was associated with improvement in health indicators beyond that observed when adherence is defined as session attendance.
Support for this work was provided by the National Institute of Diabetes and Digestive and Kidney Diseases/National Institutes of Health grant DK062148.
The authors have the following conflicts to disclose: F. L. Miller (none), D. P. O’Connor (Nimbic, Inc., Centers for Disease Control and Prevention, NASA, JW King Orthopedic Institute, SLACK, Inc., Association of Bone and Joint Surgeons), M. P. Herring (none), M. H. Sailors (none), A. S. Jackson (none), R. K. Dishman (none), and M. S. Bray (none).
Results of the present study do not constitute endorsement by the American College of Sports Medicine.
1. Ainsworth BE, Haskell WL, Herrmann SD, et al. 2011 Compendium of Physical Activities: a second update of codes and MET values. Med Sci Sports Exerc. 2011; 43 (8): 1575–81.
2. Astrand PO, Ryhming I. A nomogram for calculation of aerobic capacity (physical fitness) from pulse rate during sub-maximal work. J Appl Physiol. 1954; 7 (2): 218–21.
3. Bouchard C, Blair SN, Church TS, et al. Adverse metabolic response to regular exercise: is it a rare or common occurrence? PLoS One. 2012; 7 (5): e37887.
4. Bray MS, Hagberg JM, Perusse L, et al. The human gene map for performance and health-related fitness phenotypes: the 2006–2007 update. Med Sci Sports Exerc. 2009; 41 (1): 35–73.
5. Church TS, Earnest CP, Skinner JS, Blair SN. Effects of different doses of physical activity on cardiorespiratory fitness among sedentary, overweight or obese postmenopausal women with elevated blood pressure: a randomized controlled trial. JAMA. 2007; 297 (19): 2081–91.
6. Dishman RK. Psychological factors and physical activity level. In: Bouchard C, Katzmarzyk P, editors. Physical Activity and Obesity. Champaign (IL): Human Kinetics; 2000. pp. 89–93.
7. Duncan GE, Anton SD, Sydeman SJ, et al. Prescribing exercise at varied levels of intensity and frequency: a randomized trial. Arch Intern Med. 2005; 165 (20): 2362–9.
8. Glazer NL, Lyass A, Esliger DW, et al. Sustained and shorter bouts of physical activity are related to cardiovascular health. Med Sci Sports Exerc. 2013; 45 (1): 109–15.
9. 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. Circulation. 2007; 116 (9): 1081–93.
10. Jackson AS, Blair SN, Mahar MT, Wier LT, Ross RM, Stuteville JE. Prediction of functional aerobic capacity without exercise testing. Med Sci Sports Exerc. 1990; 22 (6): 863–70.
11. King AC, Haskell WL, Taylor CB, Kraemer HC, DeBusk RF. Group- vs home-based exercise training in healthy older men and women. A community-based clinical trial. JAMA. 1991; 266 (11): 1535–42.
12. Laukkanen RM, Virtanen PK. Heart rate monitors: state of the art. J Sports Sci. 1998;( 16 Suppl): S3–7.
13. Lee IM, Paffenbarger RS Jr. Associations of light, moderate, and vigorous intensity physical activity with longevity. The Harvard Alumni Health Study. Am J Epidemiol. 2000; 151 (3): 293–9.
14. Martin JE, Dubbert PM. Adherence to exercise. Exerc Sport Sci Rev. 1985; 13: 137–67.
15. Oldridge NB. Compliance and exercise in primary and secondary prevention of coronary heart disease: a review. Prev Med. 1982; 11 (1): 56–70.
16. Physical AGAC. 2008 Physical Activity Guidelines for Americans. Washington (DC): U.S. Department of Health and Human Services; 2008. p. 64. Available from: U.S. GPO, Washington.
17. Powell KE, Paluch AE, Blair SN. Physical activity for health: What kind? How much? How intense? On top of what? Annu Rev Public Health. 2011; 32: 349–65.
18. Sallis J, Owen N. Physical Activity and Behavioral Medicine. Thousand Oaks (CA): Sage Publications; 1999. pp. 110–34.
19. Schnohr P, Scharling H, Jensen JS. Intensity versus duration of walking, impact on mortality: the Copenhagen City Heart Study. Eur J Cardiovasc Prev Rehabil. 2007; 14 (1): 72–8.
20. Sigal RJ, Kenny GP, Boule NG, et al. Effects of aerobic training, resistance training, or both on glycemic control in type 2 diabetes: a randomized trial. Ann Intern Med. 2007; 147 (6): 357–69.
21. Thompson D, Batterham AM, Markovitch D, Dixon NC, Lund AJ, Walhin JP. Confusion and conflict in assessing the physical activity status of middle-aged men. PLoS One. 2009; 4 (2): e4337.
22. Trinder P. Oxidase determination of plasma cholesterol as cholest-4-en-3-one using iso-octane extraction. Ann Clin Biochem. 1981; 18 (Pt 2): 64–70.
23. Troiano RP, Berrigan D, Dodd KW, Masse LC, Tilert T, McDowell M. Physical activity in the United States measured by accelerometer. Med Sci Sports Exerc. 2008; 40 (1): 181–8.
24. Tucker JM, Welk GJ, Beyler NK. Physical activity in U.S.: adults compliance with the Physical Activity Guidelines for Americans. Am J Prev Med. 2011; 40 (4): 454–61.
25. Wilmore JH. Aerobic exercise and endurance: improving fitness for health benefits. Phys Sportsmed. 2003; 31 (5): 45–51.