Effects of a Personal Trainer and Financial Incentives on Exercise Adherence in Overweight Women in a Behavioral Weight Loss Program

Transcription

1 Effects of a Personal Trainer and Financial Incentives on Exercise Adherence in Overweight Women in a Behavioral Weight Loss Program Rena R. Wing, Robert W Jefsery*, Nicolaas Pronk, Wendy L. Hellerstedt* Abstract WING, RENA Ft, ROBERT W JEFFERY, NICOLAAS PRONK AND WENDY L HELLERSTEDT. Effects of a personal trainer and financial incentives on exercise adherence in overweight women in a behavioral weight loss program. Obes Res. 1996;4: Exercise is important for long-term weight loss, but few studies have examined ways to improve exercise adherence in overweight subjects participating in a behavioral weight loss program. This paper presents two studies, one conducted at the University of Pittsburgh and one at the University of Minnesota, that sought to improve exercise adherence by exerting more direct control over the environmental antecedents and consequences controlling exercise. Study 1 investigated the use of a personal trainer who called participants regularly and met them at their home or office at scheduled times for a walk. Study 2 investigated the effect of a lottery incentive for exercise adherence. In both studies, the effect of these manipulations was examined in the context of a 24-week standard behavioral weight control program with three supervised exercise sessions per week. Neither intervention achieved statistically significant improvements in exercise adherence compared to control conditions, perhaps due in part to the limited statistical power of the studies. Future studies should focus on better understanding the Submitted for publication May Accepted for publication March 7, From the Psychiatry Department, University of Pittsburgh School of Medicine, Pittsburgh, PA and the *Division of Epidemiology, University of Minnesob School of Public Health, Minneapolis, MN. Reprint requests to Dr. Wing, University of Pittsburgh School of Medicine, 3811 O HaraSlreet.Pittsburgh,PA Copyright NAASO. barriers to exercise and designing behavioral interventions that address these barriers. Key words: exercise, weight loss, obesity, behavioral treatment Introduction Numerous studies show that the combination of diet plus exercise is most effective for long-term maintenance of weight loss (17). In randomized controlled trials, long-term weight losses are better with diet plus exercise than with either diet or exercise alone (4,15,16,18) and in retrospective studies, successful weight maintainers are found to be those who adopt and maintain a physically active lifestyle (8,lO). Adding exercise to a weight loss program also increases health benefits, minimizing the loss of fat-free mass, increasing aerobic capacity, and improving serum lipid levels and glucose metabolism over and above that achieved with diet alone (4,7,19). Despite the benefits of exercise, the dropout rates from exercise programs are very high (13) and degree of overweight is one of the most consistent predictors of attrition from exercise programs (5). It is important therefore to develop strategies to promote exercise adherence in obese individuals. We present here results of two pilot studies evaluating behavioral srrategies to improve exercise adherence among obese individuals participating in a weight loss program. These studies are based on the proposition that exercise behavior is strongly influenced by cues and reinforcers in the environment; therefore, interventions to increase exercise should focus on changing these cues and reinforcers. In typical behavioral weight control programs, participants are given suggestions about ways to strengthen cues for exercise (e.g., by buying exercise OBESITY RESEARCH Vol. 4 No. 5 Sept

2 shoes; setting an alarm clock to remind them to exercise) and ways to reinforce its occurrence (e.g., by purchasing small gifts for themselves if they have successfully adhered to exercise). We reasoned that better adherence might be achieved if treatment programs exerted more direct control over the environmental antecedents and consequences, i.e. changing the environment for the participants, rather than relying on the participant to make the changes themselves. Indirect evidence supporting this suggestion comes from a recent study by Jeffery, Wing, and colleagues (9) that applied the same principles to eating behavior. Subjects who were actually given the food they should eat in appropriate portion sizes lost more weight and had lower dietary fat intake than subjects who were taught strategies for changing their own eating environment. However, direct reinforcement in the form of financial payments for weight loss did not improve outcome. Thus more direct control of environmental antecedents through food provision, but not the more direct control of consequences, was shown to be helpful in improving dietary adherence. One way to exert more direct control over the antecedents for exercise is to use therapist-supervised exercise instead of patient directed exercise. In a supervised program, both the specific setting of the exercise and the leaders may serve as cues for exercise. Supervised exercise was shown in one study to improve weight loss results (3) but not in another study (12). Another approach to cuing exercise is to use a "personal trainer," an individual who calls participants at regular intervals to remind them to exercise and/or meets them at their home or office to lead them in exercise. Study 1 evaluated this approach. Although directly controlling environmental consequences did not improve outcome in our prior study, exercise is different from eating; it is a behavior to be increased rather than decreased and experimenters can more easily observe the behavior they wish to modify and can reinforce it directly rather than reinforcing weight loss. Direct reinforcement for exercise has been evaluated in several previous studies (6,20). Subjects who received lottery tickets or monetary rewards for exercise had better exercise adherence over a 5-week program than subjects who were merely instructed to exercise (6). No studies have evaluated longer term effects of financial incentives for exercise adherence. Study 2 evaluated this approach. These two studies were conducted in different locations (University of Pittsburgh and University of Minnesota, respectively). They are presented together because of their conceptual similarity, since both studies tested ways to increase environmental control over exercise behavior. Moreover, the two studies were conduct- ed concurrently, and both utilized the same subject eligibility criteria, standard behavioral treatment program, and assessments. Study 1 Hypothesis Study 1 tested the hypothesis that use of a "personal trainer," who came to the participants' homedoffices on a preset schedule and led the participants on a walk, would increase exercise attendance over a standard supervised exercise program. Subjects Study 1 was conducted in Pittsburgh, PA. Newspaper advertisements were used to recruit 35 women. Eligibility requirements were age 25 to 55.9 to 36 kg over ideal body weight based on the Metropolitan Life Insurance norms (1983), ability to walk for exercise, having a home or work place located within 30 minutes of the treatment site, no history of heart disease or hypertension, no use of pharmacological agents known to affect weight, and currently non-smoking. All subjects signed an informed consent document approved by the Biomedical Institutional Review Board of the University of Pittsburgh. Subjects were not paid for their participation in the study, and there were no costs to them (free parking was available at the exercise and treatment sites). Method Subjects were randomly assigned to a Control Group (n=16) or a "Personal Trainer" condition (n=19). Both groups participated in standard behavioral weight control program with weekly group meetings for 24 weeks, and three supervised exercise sessions/weeks. Meetings included an individual weigh-in and a lecture/discussion on a topic related to diet, exercise, or behavior modification. Stimulus control, relapse prevention, and problem solving were included among the treatment lessons. Subjects were instructed to reduce daily intake to 1000 to 1200 kcal/day according to baseline weight, with 20% of calories from fat, and to selfmonitor their calorie intake, fat intake (in grams) and calories expended in exercise daily throughout the 24- week program. Both groups were also required to participate in three outdoor-walking sessions/week. Walking distance was gradually increased from.5 miledsession initially to 2 miles/session at week 9. There were no explicit goals for intensity and participants were encouraged to walk at a speed that they found comfortable and enjoyable. The two treatment conditions differed only in the procedures used to cue exercise adherence. Subjects in 458 OBESITY RESEARCH Vol. 4 No. 5 Sept. 1996

3 Table 1. Study 1: Comparisons between Personal Trainer and Control Conditions (Mean +SD). N Age (years) Weight (kg) BMI (kg/m2) Percent Attendance at Walking Sessions Weeks 1-6 Weeks 7-12 Weeks Weeks Total Percent Attending >SO% of Sessions weeks 1-12 Weeks Overall Changes from Preto Posttreatment Weight Loss (kg) Change in Activity Change in Calorie Intake Change in Percent of Calorie from Fat Change in Barriers "Personal Trainer" Control (8.0) 85.1 (8.5) 31.2 (3.2) 88.3 (12.4) 78.9 (27.2) 62.8 (26.8) 45.2 (21.0) 68.8 (21.0) 90% 68% 84% 7.0 (4.6) 1352 (2706) -336 (687) (10.9) -9.1 (9.2) (6.0) 86.1 (7.9) 32.6 (2.8) 88.9 (17.8) 72.8 (28.5) 61.4 (41.4) 37.9 (35.2) 65.2 (27.0) 94% 50% 69% 11.0 (5.5) 1988 (2808) 504 (740) -5.6 (1 1.7) -5.1 (9.9) P value for comparison of Personal Trainer vs Control was significant at p=.054; all others p-values were >.lo. the Control Group participated in three exercise sessions at a local track. One of these sessions was held immediately after the treatment meeting and the other two were held on two other evenings in the week. Subjects checked in with the exercise leader who monitored their exercise laps. Subjects in the "Personal Trainer" condition were divided into groups of three to four based on the location of their worksite. Each small group was assigned a "Personal Trainer." The personal trainers were collegeaged students who were actively involved in exercise. The personal trainer met hislher small group of participants at the track for one exercise sessiodweek (held after their group meeting). For the other two exercise sessionslweek, the personal trainer came directly to the subject's office at the end of the work day; after assembling the three to four participants in the small group, the personal trainer led these individuals on a walk in the vicinity of their offices. Thus, the personal trainer directly cued exercise by arriving at the participant's office at the scheduled time for exercise. At the beginning and end of the 24-week treatment program, all subjects were weighed and completed assessments of eating and exercise behaviors. The Block Food Frequency (2) measure was used to obtain an estimate of caloric intake and percent of calories from fat. The Paffenbarger Activity Questionnaire (14) was administered as an interview to determine overall level of physical activity. Subjects were also asked to rate each of 12 potential barriers to weight control on a scale of 1 ("not a problem for me") to 5 ("a very important problem for me"). These 12 items were summed to form a total barriers score. Results There were no baseline differences between subjects assigned to the Control Group vs. the Personal Trainer Condition (Table 1). Subjects in the Control Group attended 65.2% of the scheduled exercise sessions over the 24-week program, and subjects in the Personal Trainer Condition attended 68.8%. a nonsignificant difference. The percent of the exercise sessions attended decreased significantly over time in the study (p<.oool), and the rate of decline was similar for the two treatment groups. Analyses were also done to determine whether use of Personal Trainers increased the proportion of subjects with "good adherence" to exercise, defined as attending 50% or more of the prescribed sessions. In the Control Group, 94% of subjects met the criterion for "good adherence" to exercise during weeks 1 to 12 of the program, but only 50% of the subjects met this criterion for weeks 13 to 24 (p.006). In contrast, the proportion of subjects with "good exercise" adherence did not decline significantly over time in the Personal Trainer Group; 90% of subjects in the Personal Trainer Condition had good adherence during weeks 1 to 12 and 68% during weeks 13 to 24 (p=.14). Thus, there was some evidence that use of a personal trainer helped to maintain the number of subjects with "good adherence." Weight loss data were available for 13 of the 16 subjects in the Control Group and 15 of the 19 subjects in the Personal Trainer group. The average weight losses from baseline to 24 weeks were 11.0 f 5.5 kg in the Control Condition, and 7.0 f 4.6 kg for the Personal Trainer. This difference between groups approached significance with a parametric analysis (p.054). but was not significant (p=. 12) when a non-parametric analyses (Wilcoxon Rank Sums) was used. Changes in overall level of physical activity, dietary intake, and barriers to weight control did not differ between the two OBESITY RESEARCH Vol. 4 No. 5 Sept

4 these incentives. Table 2. Study 2: Comparisons between Incentive subjects and Control Conditions (Mean f SD). Study 2 was conducted in Minneapolis, MN. Eligibility criteria were identical to those used in Study Incentive Control 1. Again, women were recruited by newspaper advertisements and had to be aged 25 to 55 and 9 to 36 kg N overweight. This study was approved by the Age (years) Weight &g) Percent Attendance at Walking Sessions Weeks 1-6 Weeks 7-12 Weeks Weeks Total Percent Attending >SO% of Sessions Weeks 1-12 Weeks Overall Changes from Preto Posttreatment Weight Loss (kg) Change in Activity Change in Calorie Intake Change in Percent of Calories from Fat Change in Barriers 43.9 (7.2) 45.1 (5.9) Institutional Review Board at the University of 85.6 (9.4) 80.0 (10.9) Minnesota. TW-seven women entered Study (3.8) 30.7 (3.1) Method Subjects in Study 2 were randomly assigned to a Control Group (n=16) or the Incentive Group (n=21) (20.6) 71.7 (28.6) For both groups, the treatment paralleled that described 60.0 (34.7) 50.7 (35.7) above for Study 1, and included weekly group meetings 53.3 (36.4) 45.6 (33.4) for 24 weeks, and three supervised exercise 49.3 (37.5) 37.1 (33.1) sessions/week. These exercise sessions were held at a 60.7 (29.2) 52.2 (30.7) local park with one of the three sessions held immediately after the group meeting, which was held at a community center on the park grounds. 100% 100% The only difference between conditions in Study 2 76% 75 % was that subjects in the Incentive Group could earn 71% 56% financial prizes for exercise attendance. Subjects who attended a specific walk session were eligible for a drawing of a $50 gift certificate held after the walk. At 5.9 i f 7.4 the last group session, a drawing for a $2,000 travel cer (2418) 1380 (2376) tificate was held for Incentive subjects. The number of chances for the grand prize drawing was based on the -220 (747) -641 (732) number of walking sessions attended over the previous 24 weeks. There were no other costs or payments to participants in this study (8.9) -6.6 (9.9) -6.3 (8.3) -7.5 (9.7) treatment conditions (Table 1). Correlational analyses were done to examine predictors of exercise attendance. Due to the non-nod distribution of several of the variables, Spearman rank order correlations were used. Exercise attendance during the first 6 weeks of the program significantly predicted later attendance (rho=0.34; p=.05). Exercise attendance week 1 to 24 was significantly associated with weight loss (rho=0.53; lx.005). However, baseline measures of weight, overall physical activity, dietary intake and self-reported barriers to weight loss were unrelated to initial exercise (weeks 1 to 6) or overall exercise (week 1 to 24). Study 2 Hypothesis Study 2 tested the hypothesis that subjects who could earn substantial financial rewards for attending exercise sessions would have better attendance at exercise sessions than subjects who were not eligible for Results No baseline differences were observed between subjects in the Incentive and Control Conditions. As shown in Table 2, the incentive procedure improved attendance at exercise sessions, but this effect did not reach conventional levels of significance. Subjects in the Incentive Group attended an average of 60.7% of the exercise sessions over the 24-week program, whereas Control Group attended an average of 52.2 % of sessions. The difference between conditions was most apparent in weeks 19 to 24 of the program, when Incentive subjects attended 49.3% of the exercise sessions and Control subjects attended 37.1%. The proportion of subjects with "good adherence" (250% of sessions attended) was not significantly affected by the incentive manipulation but again the data favored the Incentive Condition. Seventy-one percent of the subjects in the Incentive Condition met the criterion for "good adherence" across the 24-week program compared to 56% of Controls. Weight loss data were available for 35 of the 37 subjects. Subjects in the Incentive Group lost 5.9 f OBESITY RESEARCH Vol. 4 No. 5 Sept. 1996

5 kg over the 24-week program; those in the Control Group lost 6.7 k 7.4 kg. This difference between groups was not significant. Changes in dietary intake, overall physical activity and the barriers to weight loss also did not differ between groups. Correlational analyses confirmed results from Study 1. Again, attendance at exercise during the initial weeks of the program predicted later attendance (rhw.69; pc.ooo1). Moreover, attendance at exercise sessions was significantly related to weight loss (rhw.75; pc.0001). The only significant predictor of exercise attendance was caloric intake at baseline (rho= -.&, pc.004); participants who reported the greatest intake at baseline attended the fewest exercise sessions. Discussion The importance of exercise for improving longterm maintenance of weight loss has been appreciated for over a decade. Yet, there has been little effort to develop strategies to improve exercise adherence in obese participants in weight loss programs. The studies presented here grow out of social learning theory (1) and examine the effects of directly cuing and rewarding exercise adherence. Although most of the findings in these studies were not statistically significant, this may have resulted in part from the small sample sizes that were used. With 20 subjects/group, we were able to detect effect sizes of 1.0 with a power of 35. Since the standard deviation for percent of sessions attended was 25 to 30 and the control groups attended 52% to 65% of their sessions, the experimental groups would have had to attend 80% to 90% of their sessions for the difference to be detectable. It is also possible that other approaches to modifying the environmental antecedents and consequences would be more successful. We considered several approaches to increasing the environmental cues for exercise. For example, it would be possible to purchase exercise equipment for the patients' home; the sight of equipment such as an exercise bike should cue exercise adherence. Although we have not seen empirical studies evaluating this approach, anecdotal data suggest that such cues remain powerful for only short periods of time. Cuing exercise could also be accomplished by calling patients regularly to remind them to exercise (11). We felt, however, that a personal trainer would provide a stronger cue to exercise. Although the "personal trainer" helped to maintain rates of "good adherence" over time, several problems were encountered. First, some of the small groups, led by a personal trainers, included subjects with very different exercise ability. Thus walking together as a group was not always successful. In addition, the use of small groups may have limited the individualized and "personal" attention that could be given. Second, the personal trainer system provided no flexibility in terms of timing of the walk. If the subject was not available at the preset time for the walk, the subject was unable to exercise with the personal trainer. Make-up sessions were not available. These problems should be considered in future studies with this approach. Similarly, although there is evidence to suggest the benefit of financial incentives for exercise (6,20), several problems were encountered in this study. Most notably, the lottery engendered feelings of unfairness and created tension among subjects, even though the distribution of the winnings in the lottery was clearly random and subjects had to attend the session to be eligible for earnings. A direct reinforcement system where each subject can earn some money for attending the session, is recommended for future research. Finally, these interventions may not have addressed the real barriers to exercise. Other interventions that deal with barriers such as lack of time and inclement weather may be more effective. Future studies should attempt to identify the barriers to exercise and then develop strategies that specifically address these barriers. These two studies also suggested that baseline characteristics, such as age and weight, may not be strong predictors of exercise attendance. However, initial exercise behavior predicted subsequent exercise behavior. That is, subjects with higher exercise attendance during the initial 6 weeks of the program had higher exercise attendance during weeks 7 to 24. In future studies, it would be interesting to develop intervention that can be applied immediately after a subject fails to attend an exercise session during the initial weeks of the program. Such early intervention might help maintain higher levels of exercise adherence. Acknowledgments This research was supported by National Institutes of Health (NIH) Grant KL41332 to Robert W. Jeffery and NIH grant HL41330 to Rena R. Wing. References 1. Bandura, A. Social Learning Theory. Englewocd Cliffs, NJ: Prentice-Hall; Block G, Hartman AM, Dresser CM, Carroll MD, Gannon J, Gardner L. A data-based approach to diet questionnaire design and testing. Am J Epidemiol. 1986; Craighead LW, Blum MD. Supervised exercise in behavioral treatment for moderate obesity. Behuv Ther. 1989;20: Dahlkoetter J, Callahan EJ, Linton J. Obesity and the unbalanced energy equation: Exercise versus eating habit change. J Consult Clin Psychol. 1979;47: Dishman RK, Sallis JE, Orenstein DR. The determi- OBESITY RESEARCH Vol. 4 No. 5 Sept

6 nants of physical activity and exercise. Public Health Rep. 1985; 100: Epstein LH, Wing RR, Thompson JK, Griffin W. Attendance and fitness in aerobics exercise: The effects of contract and lottery procedures. Behav Modif. 1980;4: Hill JO, Schlundt DG, Sbrocco T, et al. Evaluation of an alternating-calorie diet with and without exercise in the treatment of obesity. Am J Clin Nutr. 1989;50: Jeffery RW, Bjornson-Benson WM, Rosenthal BS, Lindquist RA, Kurth CL, Johnson SL. Correlates of weight loss and its maintenance over two years of followup among middle-aged men. Prev Med 1984;13: Jeffery RW, Wing RR, Thorson C, et al. Strengthening behavioral interventions for weight loss: A randomized trial of food provision and monetary incentives. J Consult Clin Psychol. 1993;61: Kayman S, Bruvold W, Stern JS. Maintenance and relapse after weight loss in women: Behavioral aspects. Am J Clin Nutr. 1990;52: King AC, Frey-Hewitt B, Dreon DM, Wood PD. The effects of minimal intervention strategies on long-term outcomes in men. Arch Intern Med. 1989;149: 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. JAM. 1991;266: Oldridge NB. Compliance and exercise in primary and secondary prevention of coronary heart disease: A review. Prev Med. 1982;11: Paffenharger RS, Wing AL, Hyde RT. Physical activity as an index of heart attack risk in college alumni. Am J Epidemiol. 1978; 108: Pavlou KN, Krey S, Steffee WP. Exercise as an adjunct to weight loss and maintenance in moderately obese subjects.am J Clin Nutr. 1989; Perri MG, Lauer JB, McAdoo WG, McAllister DA, Yancey DZ. Enhancing the efficacy of behavior therapy of obesity: effects of aerobic exercise and a multicomponent maintenance program. J Consult Clin Psychol. 1986;54(5): Pronk NP, Wing RR. Physical activity and long-term maintenance of weight loss. Obes Res. 1994;2: Wing RR, Epstein LH, Paternostro-Bayles M, Kriska A, Nowalk MP, Gooding W. Exercise in a behavioural weight control programme for obese patients with type 2 (non-insulin-dependent) diabetes. Diabetologia. 1988;31 : Wood PD, Stefanick ML, Williams PT, Haskell WL. The effects on plasma lipoproteins of a prudent weightreducing diet, with or without exercise, in overweight men and women. N Engl JMed. 1991;325: Wysocki T, Hall G, Iwata B, Riordan M. Behavioral management of exercise: Contracting for aerobic pints. J Appl Behav Anal. 1979;12: OBESITY RESEARCH Vol. 4 No. 5 Sept. 1996