Factors Predicting Adherence to 9 Months of Supervised Exercise in Healthy Older Women

Journal of Physical Activity and Health, 2011, 8, 104-110 2011 Human Kinetics, Inc. Factors Predicting Adherence to 9 Months of Supervised Exercise in Healthy Older Women Amanda J. Visek, Erin A. Olson, and Loretta DiPietro Background: Little is known about factors affecting adherence to highly-structured and supervised exercise programs in older people. Methods: Healthy, inactive older ( 65 y) women (N = 30) were randomized into a 1) higher- (AT H 80% VO 2 peak); 2) moderate- (AT M 65% VO 2 peak) intensity aerobic; or 3) lower-intensity resistance (RT L ; 50% VO 2 peak) group. All 3 groups exercised 4 days week -1 for an average of 45 to 70 min session -1 over 9 months. Adherence (%) was defined as the proportion of prescribed sessions (N = 144) in which subjects achieved their 1) prescribed heart rate (intensity adherence) and 2) their prescribed duration (duration adherence). Primary determinants of adherence included prescribed intensity (METs) and prescribed duration (min), as well as age, body composition, VO 2 peak, and exercise self-efficacy score. Results: Intensity adherence was nearly 100% for all 3 groups, while duration adherence was 95%, 91%, and 85% in the RT L, AT H, and AT M groups, respectively. Prescribed exercise duration was the strongest determinant of duration adherence (r = 0.72; P <.0001), independent of prescribed METs, age, VO 2 peak, and body composition. Conclusions: Due to competing lifestyle demands, exercise intensity may be less of a factor in adherence among older women than is exercise duration. Keywords: attendance; duration; intensity; physical activity The determinants of physical activity participation among older persons have been well-documented. Demographic and physiological factors such as age, sex, health status, and level of aerobic fitness are important predictors of physical activity, 1 as are a number of environmental factors related to safety, transportation, and accessibility to facilities. 1,2 Moreover, a myriad of psychosocial factors, such as perceived enjoyment, self-efficacy, and social support are not only key modulating factors for the adoption and adherence to a regimen of regular physical activity among the general population, but among older adults as well. 3 10 Less is known, however, about factors affecting adherence specifically to highly structured and supervised exercise training programs. Older persons responding to advertisements for exercise studies may, in fact, represent a healthier and more motivated sample compared with their peers who simply wish to begin a regular program at home or in the community. Nonetheless, even within this self-selected group there is likely to be variation in adherence to various exercise prescriptions. Presumably, a similar set of the aforementioned demographic, physiological, environmental, and psychosocial factors would remain important in attracting and keeping older people engaged in a more rigid training schedule, but so would characteristics of the exercise program (eg, setting, The authors are with the Dept of Exercise Science, The George Washington University School of Public Health and Health Services, Washington D.C. individual- vs. group-based, length of the training) and of the exercise itself (eg, type, frequency, intensity, and duration). The term exercise adherence is defined inconsistently throughout the literature, and currently, 4 measures are used: 1) completion (ie, retention), 2) attendance, 3) duration adherence, and 4) intensity adherence. 11 Indeed, these measures are distinct, as one could attend every exercise session for the entire length of the study, but not achieve either the intensity or the duration prescription on every occasion. Further, there is recent evidence 11,12 that program and exercise characteristics each relate distinctly to different aspects of adherence to structured exercise in older people. Because adherence to the exercise prescription is such an important factor in the integrity of any intervention study, the knowledge of its determinants will maximize the validity of the scientific findings. The purpose of this study was to determine the factors related to attendance and adherence to a 9-month randomized-controlled exercise trial among healthy older women. Details of this study have been published previously. 13 Because many of the program characteristics were held constant, we were particularly interested in the influence of the exercise characteristics themselves on adherence. We hypothesized that: 1) attendance and adherence would be greater with resistance training relative to aerobic training; 2) the prescribed intensity of exercise would be the strongest negative predictor of attendance and adherence in this older, female study sample; and 3) these relations would be independent of 104

Predictors of Exercise Adherence 105 several influential variables namely, age, body composition, peak aerobic capacity (VO 2 peak), and exercise self-efficacy. Participants Methods Older ( 65 y) women were recruited via advertisement from private older adult residential communities in Connecticut and from local community senior centers. Eligible study participants included those who reported no regular physical activity for the previous 6 months, were nonsmokers and not on hormone replacement therapy or glucose-lowering medication, and without class I obesity (ie, BMI <30 kg m -2 ). Older volunteers were screened by their personal physician for cardiovascular disease, neuroendocrine disorders, and other uncontrolled chronic disease. Participants meeting the inclusion criteria completed all baseline assessments before being randomized into their training group (N = 30). The protocol was approved by the Human Investigations Committee of Yale University School of Medicine and all eligible study subjects gave written informed consent before their participation. It is important to note that study subjects were not paid for their participation in the exercise training portion of the study. Subjects were compensated monetarily only for clinical procedures at baseline and follow-up requiring overnight stays in the Hospital Research Unit (HRU) of the Yale Center for Clinical Investigations (YCCI). Exercise Training Protocol Subjects were randomized into either one of 3 arms: 1) a higher- (AT H : 80% VO 2 peak); 2) moderate- (AT M : 65% VO 2 peak) intensity aerobic; or 3) a lower-intensity resistance training (RT L : 50% VO 2 peak) group. Exercise intensity was based on a heart rate necessary to achieve this relative intensity during the graded exercise challenge. Thus, the average (range) heart rate targets for the 3 groups were 123 bpm (107 134 bpm) for the AT H group; 104 bpm (97 110 bpm) for the AT M group; and 81 bpm (75 90) for the RT L group. All exercise was supervised by trained research assistants. Further, heart rate was monitored continuously during exercise (Polar Electro, Finland) and was recorded every 10 min during each session. All 3 groups were identical with regard to exercise frequency (4 days wk -1 ) throughout the 9-month exercise intervention. Participants in the AT H and AT M groups exercised on treadmills, while those in the RT L group participated in a program of 15 min of lower-intensity treadmill walking and stretching, and 30 to 45 min of strengthening using Thera-bands, Thera-balls, and hand weights. After an initial 4 to 6 week lead-in period that allowed participants to gradually attain their given exercise duration and intensity, exercise volume was held constant for those women in both the AT H and the AT M groups. Therefore, participants randomized to the 2 aerobic training groups exercised for a duration necessary to expend 300 kcal per session (estimated from VO 2 peak and body weight). 14 If we assume a VO 2 peak of ~1.40 L min -1 [20 ml (kg min) -1 for a 68 kg older women], this required about 55 and 65 min of exercise for AT H and AT M groups, respectively. Subjects in the RT L group exercised for 45 to 60 min per session. Procedures to Reduce Barriers to Exercise Lack of time, transportation, and social support are often cited as 3 key obstacles to exercise participation in older people. 15 To minimize these and other perceived barriers, exercise training was provided on-site at the different residential communities and senior centers. Transportation to/ from the training sites and to/from the University Hospital Research Unit (HRU) in New Haven were provided whenever necessary. Participants were also given a choice of a morning or afternoon exercise class to best accommodate their schedules, and they exercised in groups of 4 to 5 to enhance social support. In the event of a missed exercise session, participants were contacted immediately and their exercise session was rescheduled. Adherence Over the 9-month training period, all subjects were expected to complete a total of 144 exercise sessions. Thus, attendance was operationally defined as the number of actual sessions attended divided by the expected number and was expressed as a proportion (%). We also defined adherence more specifically as the proportion of the prescribed exercise sessions in which subjects met their prescribed intensity for at least 80% of the session (intensity adherence (%)) and the proportion of exercise sessions in which subjects met their prescribed exercise duration (duration adherence (%)). 16 Determinants of Adherence Since exercise intensity varied among the 3 groups, prescribed intensity was considered a primary predictor variable and was characterized in metabolic equivalents (METs). Exercise duration also varied widely among the participants and the groups because it was calculated from individual VO 2 peak and body weight based on the time necessary to achieve 300 kcal of energy expenditure in the AT groups and was set at 45 to 60 min for women in the RT L group. Therefore, prescribed minutes of exercise was also considered a primary predictor variable and was expressed in minutes. We further considered several physiological variables as potential confounding variables, as they have demonstrated a strong association with both exercise prescription and exercise adherence in the literature 17 age, aerobic capacity (VO 2 peak), body weight, lean mass, and body fat and the change in these physiological variables

106 Visek, Olson, and DiPietro over 3, 6, and 9 months. Peak VO 2 [ml (kg min) -1 ] was determined on a treadmill at baseline, 3, 6, and 9 months according to methods previously described by our Laboratory. 13 Height and weight were measured on a balance-beam scale and the body mass index (BMI; kg m -2 ) calculated as a measure of weight status. Overall body composition [whole body muscle (kg) and fat mass (kg, %)] scans were obtained using dual energy x-ray absorptiometry (DXA). Finally, exercise self efficacy was assessed immediately following randomization (to assess accurately the subject s confidence in the task at hand) using the Stanford-Sunnyvale Health Improvement Project (SSHIP) II Questionnaire. 18 The SSHIP questionnaire provides a confidence scale ranging from 0% (not at all confident) to 100% (totally confident) and asks subjects to rate their confidence in exercising under a number of different conditions (eg, when tired, during bad weather, when there are competing interests). The SHHIP also queries overall confidence in completing the prescribed number of exercise sessions over the 9-month training period. We used this latter overall score as the measure of exercise self-efficacy. Statistical Analysis Univariate statistics (mean ± SE) first were generated on all study variables. Differences in the mean levels of the study variables among the 3 training groups were tested using analysis of variance (ANOVA), while simple correlations were tested using the Spearman Rank Order Correlation Coefficient. Significant findings from the simple analyses were then entered into multivariable linear regression models to determine the strongest independent predictors of exercise attendance and adherence. Statistical significance was set at an alpha level of 0.01 to correct for the multiple comparisons. We note that both baseline and 9-month followup assessments were completed on 26 of the 30 older women, resulting in 87% retention. The full sample of older women (N = 30) was used in the denominator of all calculated measures of adherence, however, to provide the most conservative estimates. 11,19 Among the 3 groups, retention was 100% in the AT H, 82% in the AT M, and 82% in the RT L. Older women who completed the 9-months of training were similar to those who did not with regard to age, VO 2 peak, and BMI. Results Baseline characteristics of the study participants are presented in Table 1 according to their assigned exercise group. Although women assigned to the AT H group appeared slightly younger and more aerobically fit than women in the RT L group, these differences were not statistically significant. Level of lean mass and body fat were also similar among the groups. As expected, the exercise self-efficacy score was highest in the RT L group (96%) and lowest in those women assigned to higherintensity exercise (88%). Table 2 contains information on the prescribed exercise frequency (days wk -1 ), intensity (METs), and the prescribed minutes of exercise according to group assignment. Although the mean values for prescribed minutes of exercise appear similar for women in the AT M and AT H groups, exercise times ranged from 45 to 100 min in the AT M group and from about 44 to 86 min in the AT H group, depending on the body weight and the VO 2 peak of the individual study participant. Table 1 Subject Characteristics of the Study Population by Group Assignment RT L (n = 11) AT M (n = 11) AT H (n = 8) Age (y) 75 ± 2 73 ± 1 71 ± 1 BMI (kg m -2 ) 27.7 ± 1.8 29.0 ± 1.7 25.8 ± 1.0 Body fat (%) 38.4 ± 2.3 37.8 ± 2.0 37.4 ± 1.7 Lean mass (kg) 38.7 ± 2.0 38.0 ± 1.7 40.0 ± 2.6 Peak VO 2 (ml (kg min) -1 ) 18.3 ± 1.6 20.7 ± 1.3 21.4 ± 1.3 Self-efficacy score 96.4 ± 2.3 92.1 ± 2.4 87.5 ± 5.6 Note. Values are mean ± SE. Table 2 Exercise Prescription by Group Assignment RT L (n = 11) AT M (n = 11) AT H (n = 8) Frequency (d wk) -1 4 4 4 Intensity (METs) 2.61 ± 0.23 3.85 ± 0.24 4.89 ± 0.29 Duration (min) a 51.0 ± 3.2 68.2 ± 5.3 63.4 ± 5.4 a Subjects in the AT M and AT H groups exercised for a duration necessary to expend 300 kcal per session. Values are mean ± SE.

Predictors of Exercise Adherence 107 Factors Predicting Attendance Women in the AT H group attended the greatest number of sessions (94%), whereas women assigned to RT L had the lowest attendance (65%; P <.01) (see Table 3). In the simple analysis, we observed no correlation between age and attendance (r = 0.09; P <.63). Body mass index was significantly and inversely correlated with attendance (r = 0.48; P <.01), while peak VO 2 demonstrated a marginally significant positive association (r =.49; P <.03). Contrary to our study hypothesis, we observed a positive (rather than an inverse) correlation between the prescribed intensity of training (METs) and attendance (r =.48; P <.02); however, prescribed minutes of exercise had little effect on this same adherence measure (r =.14; P <.46) (see Figure 1). For the multivariable analysis, group assignment, BMI, VO 2 peak, and age next were entered into a regression model to determine the independent effects of each of these variables on attendance over 9 months of training. The overall model was statistically significant (F = 5.76; P <.002) and explained 51% of the variance in the percentage of sessions attended. Group assignment maintained marginal statistical significance (P <.03) in the presence of the other variables; however, BMI remained an even stronger independent and negative predictor of attendance (P <.01). Predictors of Duration Adherence Adherence to the prescribed exercise duration tended to be greatest in the RT L (94%), compared with the AT H (91%) and AT M (85%) groups, although these differences were not statistically significant (see Table 3) Neither age (r = 0.04; P <.82) nor BMI (r = 0.34; P <.08) were correlated with adherence to the prescribed exercise duration, although percent body fat was a significant inverse correlate (r = 0.58; P <.002) of this adherence measure. As expected, VO 2 peak also demonstrated a strong positive association with duration adherence (r =.60; P <.01) in the simple analysis. In contrast to our hypothesis, however, prescribed intensity of exercise (METs) was a significant, but positive, correlate of this same measure (r =.53; P <.01), presumably because the higher the prescribed intensity, the shorter the duration of the session. Even though the prescribed minutes of exercise was not associated with attendance, it was a strong, inverse correlate of adherence to that duration of exercise (r = 0.72; P <.0001) (see Figure 1). In the Table 3 Attendance and Adherence Data by Group Assignment RT L (n = 11) AT M (n = 11) AT H (n = 8) Attendance (%)* 65 90 94 Adherence** prescribed time (%) 95 85 91 Adherence** prescribed HR (%) 97 100 100 * Percentage of total classes (n = 144) attended. **Percentage of classes in which subjects adhered to the prescribed duration (min) and intensity of exercise (HR). Figure 1 The association of prescribed exercise time with exercise attendance and with exercise adherence (N = 30).

108 Visek, Olson, and DiPietro multivariable modeling, duration adherence then was regressed on the study variables prescribed minutes of exercise, BMI, VO 2 peak, percent body fat and age. The overall model was statistically significant (F = 10.25; P <.003) and explained a large proportion of the variance in adherence to the prescribed exercise duration (R 2 =.72). Moreover, prescribed exercise time remained an independent negative predictor of duration adherence in the presence of the other influential covariables. Indeed, as described in Table 4, for every added minute of prescribed exercise time, adherence to that time decreased by nearly one-half of a percent (β = 0.47; 95% CI: 0.73 to 0.20). Age emerged marginally as an independent predictor of duration adherence in the presence of the other variables. For each year of older age, duration adherence decreased by nearly 2% (β = 1.78; 95% CI: 3.11 to 0.44). The further addition of prescribed intensity (METs) to the model did little to alter the parameter estimates, and we observed no statistical interaction between prescribed time and age. As shown in Table 3, intensity adherence was nearly 100% for all 3 groups; however, none of the study variables demonstrated a significant association with adherence to the prescribed intensity, presumably due to this lack of variability among the women. Exercise self-efficacy did not correlate significantly with any of the outcome measures of adherence either. Similarly, we observed no change in VO 2 peak or in body composition at 3, 6, or 9 months, which we have previously reported, 13,20 and therefore, a positive or negative training effect was not a factor in any type of adherence. Discussion Findings from 2 meta-analyses of randomized controlled exercise trials in older populations 11,12 suggest that characteristics of the exercise training program and of the exercise itself relate distinctly to measures of adherence and of completion (ie, retention). Indeed, these comprehensive analyses reported that group-based exercise and resistance training predicted higher attendance, whereas facility-based programs predicted better retention compared with home-based programs. We were interested specifically in how characteristics of the exercise itself (eg, type, prescribed duration and intensity) predicted adherence. Similar to the findings of Martin & Sinden 11 and of Hong et al, 12 we observed distinct relations between several of these study factors and the different measures of adherence. For example, attendance over 9 months was significantly lower in the resistance training group than in either of the 2 aerobic training groups, which was contrary to our study hypothesis, as well as to findings from other studies of older people (see 11,12 for reviews). More notably, prescribed exercise time did not predict attendance or intensity adherence, but was a strong independent predictor of duration adherence. In fact, women assigned to the AT M group had the longest prescribed exercise times (range = 45 100 min) and although their attendance was similar to the AT H group (90% vs. 94%), they demonstrated the lowest level of duration adherence (85%) overall. This suggests that a prescribed exercise duration beyond 60 min per session may not be practical for older women due to physical or emotional fatigue or perhaps due to competing life-style demands. It is not surprising that prescribed exercise time negatively impacts adherence; however, it does refute our hypothesis that the prescribed intensity of exercise would be the stronger predictor of adherence over 9 months in older women. Interestingly, we observed a positive (rather than the predicted inverse) correlation between the prescribed METs and both attendance and duration adherence in the simple analysis, and this is more than likely because a higher prescribed intensity resulted in shorted exercise duration. Prescribed intensity, however, failed to retain an independent relation with any measure of adherence in the multivariable modeling. Other studies of older people also report a null association between exercise intensity and adherence. 11,12 We were also interested in how various participant characteristics, such as age, BMI, VO 2 peak affected adherence. Body mass index was an independent predictor of attendance, while age demonstrated a marginal Table 4 Multiple Linear Regression Estimates Describing the Independent Relation Between Prescribed Exercise Time and Adherence Variable Parameter estimate 95% Confidence Interval P Prescribed time (min) 0.47 0.73 to 0.20 0.004 VO 2 peak (ml (kg min) -1 1.05 0.36 to 2.46 0.16 BMI (kg m -2 ) 0.41 1.51 to 0.69 0.47 Age (y) 1.78 3.11 to 0.44 0.02 Note. Overall model: F = 10.25; P <.0003; R 2 =.72. The regression estimate describes the change in adherence to the prescribed exercise duration (%) per unit change in the study variable.

Predictors of Exercise Adherence 109 independent relation to duration adherence. Other studies of older people report equivocal findings with regard to the influence of these subject characteristics and others on program adherence. 11,16,21 Recently, Flegal et al 21 reported that self-rated depression, fatigue, and physical components of health-related quality of life were significant correlates of adherence (ie, attendance) to a 6-month exercise intervention in people 65 to 85 years old, whereas age was not. Exercise self-efficacy score did not predict any of the 3 adherence measures in our study, which contradicts findings from several other studies of older people. 6,22 24 We assessed exercise self-efficacy after randomization, but before the start of exercise training, and therefore, subjects may not have assessed their ability accurately due to a lack of familiarity with their specific exercise regimen. Other studies have measured exercise selfefficacy after a 2-week lead-in period, resulting in more realistic and accurate assessments. 22 Also, since neither body composition nor VO 2 peak changed measurably over the course of the training study, negative outcome expectations and decreased motivation could have influenced the relations of interest 25 especially among women in the RT L group, among whom attendance was significantly lower compared with the AT groups. Unfortunately, we collected no longitudinal data on these important psychosocial variables and therefore cannot assess their contributions to the exercise adherence outcomes. Other unmeasured factors that may have influenced adherence to the 9 months of structured exercise in older people include enjoyment, significant life events (serious illness, death of a spouse), or interpersonal relations with the research personnel. Nonetheless, a major strength of this study included the control for several program characteristics and environmental barriers known to influence physical activity participation. We provided an exercise training program that was conducted within on-site facilities and was group-based. These design features have demonstrated a relation to exercise adherence and completion among older people participating in studies of long duration. 12,26 Moreover, the group setting is a key factor in maximizing exercise attendance and adherence to a difficult exercise prescription due to greater social support, 27 motivation, 28 peer support 29 and group cohesion. 30 Exercise frequency and study length also were not factors in predicting adherence in our study as they are in other studies (see 12 for review), as these factors were held constant among the 3 training groups. Exercise training studies remain necessary as we continue to modify practice guidelines to include greater specificity of exercise prescription within various age or risk groups. Our findings suggest that group-based exercise and exercise offered via facility-based programs remain important program characteristics in maximizing attendance and adherence to highly-structured training studies. Contrary to our initial premise, exercise intensity may be less of a factor in adherence among older people than is exercise duration. This may be an indication that due to competing demands of their lifestyles, at least older women may be willing to work harder during exercise training if such greater intensity results in a shorter exercise duration. This information is important in balancing the scientific requirements of a given exercise intervention with what participants are willing to accept. Acknowledgments We thank Jodi L. Crimmins, Anne Marie Cheatham, Jennifer Fawcett, and the nursing and technical staffs of the Hospital Research Unit of the Yale University Center for Clinical Investigations and the Yale University Diabetes Endocrinology Research Center for their technical expertise, and the study subjects for their commitment to this research. This publication was made possible by the CTSA Grant Number UL1 RR024139 from the National Center for Research Resources (NCRR), a component of the National Institutes of Health (NIH), and NIH Roadmap for Medical Research. Its contents are solely the responsibility of the authors and do not necessarily represent the official view of NCRR or NIH. This work was supported by Grants RO1 AG.17163 (LDP) and P30 AR.46032. References 1. Lox CL, Martin Ginis KA, Petruzzello SJ. The psychology of exercise: integrating theory and practice. 2nd ed. Scottsdale, AZ: Holcomb Hathaway Publishers; 2006. 2. Carron AV, Hausenbals HA, Estabrooks PA. The psychology of physical activity. New York, NY: McGraw-Hill Higher Education; 2003. 3. Bean JF, Bailey A, Kiely DK, Leveille SG. Do attitudes toward exercise vary with differences in mobility and disability status? A study among low-income seniors. Disabil Rehabil. 2007;29:1215 1220. 4. Blue CL. The predictive capacity of the theory of reasoned action and the theory of planned behavior in exercise research: an integrated literature review. Res Nurs Health. 1995;18:105 121. 5. Cohen-Mansfield J, Marx MS, Guralnik JM. Comparison of exercise models in an elderly population. Aging Clin Exp Res. 2006;18:312 319. 6. King AC, Blair SN, Bild DE, Dishman RK. Determinants of physical activity and interventions in adults. Med Sci Sports Exerc. 1992;24:221 236. 7. King AC, Rejeski WJ, Buchner DM. Physical activity interventions targeting older adults: a critical review and recommendations. Am J Prev Med. 1998;15:316 333. 8. Resnick B, Palmer MH, Jenkins LS, Spellbring AM. Path analysis of efficacy expectation and exercise behaviors in older adults. J Adv Nurs. 2000;21:1309 1315. 9. Stiggelbout M, Hopeman-Rock M, Tak E, Lechner L, Mechelen W. Drop out from exercise programs for seniors: a prospective cohort study. J Aging Phys Act. 2005;13:409 421. 10. Trost SG, Owen N, Bauman AE, Brown W. Correlates of adults participation in physical activity: review and update. Med Sci Sports Exerc. 2002;34:1996 2001. 11. Martin K, Sinden A. Who will stay and who will go? A review of older adults adherence to randomized controlled trials of exercise. J Aging Phys Act. 2001;9:91 114.

110 Visek, Olson, and DiPietro 12. Hong S, Hughes S, Prohaska T. Factors affecting exercise attendance and completion in sedentary older adults: a meta-analytic approach. J Phys Act Health. 2008;5:385 397. 13. DiPietro L, Dziura J, Yeckel CW, Neufer PD. Exercise and improved insulin sensitivity in older women: evidence for the enduring benefits of high intensity training. J Appl Physiol. 2006;100:142 149. 14. American College of Sports Medicine. ACSM s resource manual for guidelines for exercise testing and prescription. 5th ed. Philadelphia, PA: Lippincott Williams and Wilkins; 2001. 15. King A. Interventions to promote physical activity by older adults. J Gerontol A Biol Sci Med Sci. 2001;56(Spec No 2):36 46. 16. Courneya KS, Segal RJ, Gelmon K, et al. Predictors of supervised exercise adherence during breast cancer chemotherapy. Med Sci Sports Exerc. 2008;40(6):1180 1187. 17. Dishman RK, Sallis J. Determinants and interventions for physical activity and exercise. In: Bouchard C, Shephard RJ, Stephens T, eds. Physical activity, fitness and health: international proceedings and consensus statement. Champaign. IL: Human Kinetics; 1994:214 238. 18. Stewart AL, Mills KM, King AC, Haskell WL, Gillis D, Ritter PL. CHAMPS physical activity questionnaire for older adults: outcomes for interventions. Med Sci Sports Exerc. 2001;33(7):1126 1141. 19. Buchner DM, Cress ME, de Lateur BJ, et al. A comparison of the effects of three types of endurance training on balance and other fall risk factors in older adults. Aging. 1997;9(1-2):112 119. 20. DiPietro L, Yeckel C, Dziura J. Progressive improvements in glucose tolerance following lower-intensity resistance versus moderate-intensity aerobic training in older women. J Phys Act Health. 2008;5:854 869. 21. Flegal KE, Kishiyama S, Zajdel D. Adherence to yoga and exercise interventions in a 6-month clinical trial. BMC Complement Altern Med. 2007;7:37. 22. McAuley E, Katula J, Mihalko S, et al. Mode of physical activity and self-efficacy in older adults: a latent growth curve analysis. J Gerontol B Psychol Sci Soc Sci. 1999;54:283 292. 23. Lee C. Attitudes, knowledge, and stages of change: a survey of exercise patterns in older Australian women. Health Psychol. 1993;12(6):476 480. 24. Plonczynski DJ. Physical activity determinants of older women: what influences activity? Medsurg Nurs. 2003;12(4):213 221. 25. Marcus B, Forsyth L. Motivating people to be physically active. Champaign, IL: Human Kinetics; 2003. 26. Cox KL, Burke V, Gorely TJ, Beilin LJ, Puddey IB. Controlled comparison of retention and adherence in homevs center-initiated exercise interventions in women ages 40-65 years: The S.W.E.A.T. Study (Sedentary Women Exercise Adherence Trial). Prev Med. 2003;36:17 29. 27. Oka RK, King AC, Young DR. Sources of social support as predictors of exercise adherence in women and men ages 50 to 65 years. Womens Health. 1995;1:161 175. 28. Dishman RK, Sallis JF, Orenstein DR. The determinants of physical activity and exercise. Public Health Rep. 1985;100:158 171. 29. Cress ME, Buchner DM, Prohaska T, et al. Physical activity programs and behavior counseling in older adult populations. Med Sci Sports Exerc. 2004;36:1997 2003. 30. Estabrooks PA, Carron AV. Group cohesion in older adults exercisers: prediction and intervention effects. J Behav Med. 1999;22(6):575 588.