Long-Term Exercise Adherence following Participation in a 10-week Campus Wellness Program based on the Disconnected Values Model.
Brinthaupt, Thomas M. ; Anshel, Mark H.
Initiating and maintaining regular exercise behavior has challenged researchers and practitioners for many years (Anshel, 2014). To encourage a greater commitment to exercise, researchers have usually examined the efficacy of selected cognitive and behavioral strategies and interventions on exercise initiation and both short-term and long-term exercise adherence. Despite general success in helping unfit individuals to initiate an exercise program, the degree to which exercisers maintain (i.e., adhere to) their exercise program has been disappointing, with disappointing percentages of exercisers adhering to their new program beyond six months (e.g., Buckworth, Dishman, O'Connor, & Tomporowski, 2013; Findorf, Wyman, & Gross, 2009).
There are several possible reasons for people's failure to adhere to an exercise program. One possible reason, according to Hunt and Hillsdon (1996), is that most wellness and exercise programs are restricted to changing only one behavior at a time (e.g., fitness, diet), rather than multiple related behaviors that could affect adherence. Another reason is that previous studies have lacked specific components that ensure treatment fidelity and long-term changes in various health behaviors (Klesges, Estabrooks, Dzewaltowski, Bull, & Glasgow, 2005). Failure to control for program components that might mediate or moderate adherence represents another plausible reason for poor exercise adherence. Examples include the inclusion or exclusion of personal coaching, the nature of the participant-fitness coach relationship, the exerciser's current level of fitness, attitudes toward exercise, available travel to the exercise location, private versus public exercise settings, initial differences in fitness, and the development of new habits or routines that support regular exercise behavior (Anshel, 2014; 2008; Anshel & Kang, 2007a, b).
Another possible issue that compromises exercise adherence is that, while certain individuals enjoy various types of physical activity (e.g., sports, competitive games), exercise, per se, may not be particularly enjoyable. Non-adherence to regular exercise also may be due to a variety of exercise barriers (Conaway, Rejeski, & Miller, 2009). For example, people may not adhere to an exercise routine because they are too busy (i.e., the perception of "no time"), they forget to incorporate exercise into their daily schedule, or they feel stress, physical discomfort, or some other negative mood state. In summary, while attempts at promoting exercise participation have been relatively successful, similar favorable outcomes related to short-term and long-term exercise adherence have been both understudied and less effective.
Effects of Interventions on Exercise Adherence
There has been a paucity of research on the effectiveness of cognitive-behavioral interventions on long-term exercise adherence. Reviews of related literature (e.g., Dishman & Buckworth, 1997; Muller-Riemenschneider, Reinhold, Nocon, & Willich, 2008) have produced mixed results. Programs that apply self-regulatory (SR) strategies can effectively promote exercise adherence. Researchers have shown, for example, that the development of SR skills, such as self-monitoring, goal-setting, and time management, can improve exercise participation and adherence (e.g., Evers, Klusmann, Ziegelmann, Schwarzer, & Heuser, 2012; Winett, Anderson, Wojcik, Winett, & Bowden, 2007). Other researchers (e.g., Tobi, Estacio, Yu, Renton, & Foster, 2012) suggest that long-term adherence to exercise and other forms of physical activity may improve if exercise interventions are individually tailored for meeting personal needs.
More recently, in a review of limitations of the extant exercise intervention literature, Buckworth et al. (2013) lament the absence of a theoretical framework or model to examine the efficacy of an intervention intended to promote exercise participation and adherence. Another limitation of previous exercise intervention research is that it has focused on outcomes (e.g., changes in attitude toward exercise, extent of exercise initiation) rather than mechanisms and processes on changes in exercise-related attitudes and behavior. Not addressed in these studies are the effects of cognitive and behavioral strategies on exercise performance. For instance, Anshel and Kang (2007a) found that the combination of educational materials, personal coaching and social support significantly improved exercise adherence.
In summary, previous intervention research has not focused on the mechanisms, processes, and mediators by which changes in exercise-related attitudes and behavior occur. A valid model-based intervention model is needed that incorporates these suggestions. Given the limitations of previous intervention studies in successfully changing health behavior, there is a need to examine adherence using different models for replacing unhealthy behavior patterns with positive, healthier routines. One approach to promoting exercise adherence that overcomes several limitations of extant published research and has received increased attention in the research literature in recent years is the Disconnected Values Model (Anshel, 2008, 2013).
The Disconnected Values Model (DVM)
The DVM is a values-based cognitive-behavioral intervention for replacing unhealthy behaviors with more desirable, healthy routines (Anshel, 2008, 2013; Anshel & Kang, 2007a, b; Brinthaupt, Kang, & Anshel, 2010, 2013). The primary objective of the model is to help participants identity one or more unhealthy behaviors that are inconsistent with their values, and to overcome the misalignment, or disconnect, between undesirable behaviors and values in establishing new, healthier routines by developing an action plan and working with a program coach. The inconsistency, or disconnect, between values (e.g., health, family, faith, integrity, performance excellence) and leading a sedentary lifestyle, and to acknowledge that the costs and long-term consequences of this disconnect are unacceptable ostensibly provides incentive for changing behavior patterns. In addition to poor stress management and unhealthy eating, lack of exercise is among the more common and neglected unhealthy behaviors that the model addresses and attempts to change. A more thorough explanation of the DVM goes beyond the scope of this study. See Anshel (2008, 2010, 2013), Anshel and Kang (2007a, b), and Brinthaupt et al. (2010, 2013) for more complete information and a review of past studies that support the model.
The DVM has received empirical support in the extant research literature. For instance, Anshel, Brinthaupt, and Kang (2010) and Anshel and Kang (2007a) applied the model over a 10-week period with university faculty and staff. The collective results indicated significantly improved cardiovascular and strength fitness scores, and a reduction in low density lipoproteins (LDL, or "bad" cholesterol) and triglycerides. In addition, approximately 70-76% of the participants adhered to both the aerobic and strength training components of the program. The Anshel et al. (2010) study also included a measure of mental well-being. The researchers found significant gains in selected dimensions of mental health among participants.
In summary, the DVM overcomes many of the limitations in earlier fitness intervention research such as imposing exercise locations, programs, or schedules, failing to develop social support (e.g., coach-client relationships), and not addressing lifestyle change that integrate exercise with other healthy behavioral patterns such as nutrition, sleep, and stress management. The DVM is specifically designed to encourage lasting changes in healthy behaviors and it has received good empirical support. However, no empirical research has examined long-term adherence among adult exercisers in which the DVM formed the intervention.
Purpose of Study
The purpose of our study, then, was to examine the long-term exercise adherence rates of adults who participated in a DVM-based wellness intervention located on a university campus. Because of limitations to the nature of the data collected, we were unable to assess the effects of specific DVM components on adherence. We predicted that the intervention, that included all components of the DVM, would lead to long-term adherence to the participants' exercise program. To test this general hypothesis, we examined how several aspects of program participation and completion related to long-term adherence.
Our working hypotheses were that: participants who showed larger changes in fitness would be more likely to adhere later than those who made smaller changes in fitness during the program (HI); individuals who completed the program more recently would report higher levels of adherence than exercisers who completed it less recently (H2); program participants who reported lower long-term adherence would be more likely than higher adherence participants to report getting less out of the wellness program (H3); and lower-adherence participants would report experiencing more barriers to adherence since their program participation ended compared to higher-adherence participants (H4).
An additional and particularly important research question that has received minimal attention in the literature is the effect of program "dosage," or the degree of repeated exposure to an exercise and wellness program, on long-term adherence. In their review of longterm exercise adherence literature among cancer patients, Kampsholf et al. (2014) found moderate support for a positive association between exercise history and exercise adherence. Specifically, longer exercise experience was positively associated with superior exercise adherence. Unknown, however, is the effectiveness of engaging in multiple exercise programs on long-term exercise adherence. That is, there is an apparent absence of research on the extent to which multiple "doses" of participating in one or more exercise programs affects long-term adherence. This issue was addressed in the present study. Our final working hypothesis was that participating in multiple DVM programs would be associated with greater long-term adherence, as compared to participating in a single program (H5). In particular, we expected that multiple program completers would report higher scores on a variety of long-term adherence measures, as compared to single-program completers.
Method
Participants
Participants, self-described as non-exercisers (i.e., people who were not engaging in a regimen of vigorous exercise or other forms of physical activity) at the start of this study, were full time academic faculty and staff (i.e., non-academic employees) who were employed by a large public university located in the southeastern U.S. A 10-week wellness program was offered to all campus employees over each of five consecutive semesters at a cost to the person of $25.00 per semester. Program costs were heavily subsidized by an internal university grant. The program, explained more fully later, consisted of fitness testing and coaching, blood testing (i.e., lipids profile), seminars and one individual session with a registered dietician, and seminars with a "life skills coach" who discussed common psychological issues that contribute to poor health (e.g., eating disorders, depression, anxiety, low self-esteem, irrational thinking). No individual counseling sessions were provided, however, participants were referred to a licensed psychologist if desired.
The employees could participate in the program for as many semesters as they wished over a span of 1.5 years, or one 10-week program in each of five semesters. Approximately one year following the final program (i.e., the end of the fifth-semester program) participants who completed at least one program (N = 261, 187 women, 74 men; 123 faculty, 138 staff members) were contacted and invited to complete a brief follow-up online survey. Of the original program completers, 127 participants (49%) completed the survey. They included 92 women and 35 men (63 faulty, 64 staff members) ranging in age from 27 to 64 yrs (M= 46.67, SD = 9.48). The personal characteristics of these participants did not differ significantly from the total sample of program completers on major post-program measures (e.g., gender, weight, BMI, waist-to-hip ratio, percent body fat, V[O.sup.2] max).
Materials and Procedure
As indicated earlier, the primary purpose of this study was to determine long-term adherence through self-report of program participants. We assessed the extent that participants continued to engage in the "healthy" behaviors they learned from the program a minimum of one year following program completion. The major sections of the survey included an assessment of participants' general experiences with the program, as well as its impact, that is, how much the program influenced their current overall lifestyle changes, current exercise and eating behaviors (e.g., the extent to which participants are adhering to the daily and weekly behaviors they developed during the program), barriers to engaging in regular exercise (e.g., reasons why they have difficulty maintaining their program); and open-ended items (e.g., the main difficulties with adherence and suggestions for improvement).
Program experiences and impact. Participants first indicated, using a yes/no format, whether they had established a schedule of "regular" aerobic exercise and lifting weights during the program, and also whether they were currently participating in an ongoing program of regular aerobic and weight training. Using a 5-point Likert-type scale, ranging from 1 (strongly disagree) to 5 (strongly agree), they also rated the extent to which completing the wellness program helped them to become more generally physically active, engage in more exercise, improve their eating habits, and whether they were currently maintaining similar healthy behaviors that they practiced during the program.
Participants also rated the extent to which specific aspects of the wellness program impacted their current overall (healthy) lifestyle (e.g., physical activity and exercise levels, eating habits), using a 5-point scale ranging from 1 (strong negative impact) to 5 (strong positive impact). The assessed program components included the general orientation session, collection of fitness data, support from the program website, fitness coaching received, and exercise and nutrition knowledge provided by the program. The complete list of program components rated by the participants is provided in Table 2.
In order to simplify analyses of the impact measures, a program impact index was created that consisted of the sum of each participant's ratings of the 12 items listed in Table 2. Possible scores on this index ranged from 12-60, with higher scores indicating a more positive impact of the program on one's current health lifestyle. The index showed an acceptable internal consistency (Cronbach's alpha) value, r = .84; for the sample, M= 48.06, SD = 5.08.
Current exercise and eating behaviors. In the next section of the survey, participants rated items pertaining to their continuation of exercise routines and eating habits they learned in the program. In particular, they reported their current number of days per week (0-7) engaging in cardio exercise and strength training. Next, participants rated the option that best described their current exercising, degree to which they were currently physically active, and extent that they were maintaining "healthy" eating habits over recent weeks, using a 3-option response format: 1 = have not been regularly, 2 = have been occasionally, and 3 = have been often (similar to the wellness program).
Barriers to continued post-program exercise. In the next survey section, participants rated how often a wide range of possible barriers currently applied to them. The 21 exercise barriers included factors such as lack of time, coaching, equipment, information, confidence, support from others, injury or fear of injury, and lack of results and motivation. Participants rated these items as reasons why they have had difficulty maintaining an exercise program, using a 5-point frequency scale ranging from 1 (never) to 5 (very often).
In order to simplify analyses of the barriers measures, an exercise barriers index was created that consisted of the sum of each participant's ratings of the barrier items. Possible scores on this index ranged from 21-105, with higher scores indicating more frequency encountering the barriers to maintaining an exercise program. The index showed an acceptable internal consistency (Cronbach's alpha) value, r = .87; for the sample, M= 37.53, SD = 11.09.
Major adherence challenges and difficulties. Finally, participants answered open-ended questions about three issues: (1) their main difficulty in maintaining changes in their exercise and physical activity habits since the program ended, (2) primary areas of difficulty in maintaining changes in their eating habits and diet since the program ended, and (3) what could have been offered during or soon after the program to improve their longterm adherence.
Fitness data. Pre- and post-program fitness data for all participants based on the most recent program that they completed also were analyzed. These data (all measured at the campus recreation center) included weight, body mass index (BMI), percent body fat (using skinfold calipers), V[O.sup.2] max (a measure of cardiovascular fitness), and lower and upper body strength. More details about these measures and how they were collected during the program are available in Anshel et al. (2010).
DVM Program
Participants completed a 10-week wellness program that was based on the DVM. All participants received a workbook that reflected the model's components. The program began with a 3-hour orientation given by the program's "performance coach" that consisted of lecture on relevant concepts, viewing a DVD, completing written segments of a workbook, and group member interaction guided by the performance coach in which participants shared workbook information. Immediately following the seminar, participants met the fitness coach and registered dietician with whom they would work during the program.
During the program's first and final weeks, participants completed pre- and post-fitness tests. Pre-and posttests were also obtained for blood samples, obtained by a campus nurse, to determine changes in blood lipids (i.e., lipids profile) as a function of program participation. The participants also developed an "action plan" and worked with their fitness coach to implement this plan on a weekly basis. Additional details of the approach and implementation of the DVM wellness program are available in other publications (e.g., Anshel, 2008; Anshel et al., 2010; Anshel & Kang, 2007a, b).
While fitness and dietary behaviors were not assessed during the program, fitness coaches monitored participant exercise activity and obtained adherence data at the end of each week. Participants were asked to engage in cardio-type activities a minimum of three times a week, preferably on alternate days, and strength training activities at least twice a week, also not on consecutive days. The fitness coach explained that desirable exercise outcomes are more likely if a person included at least one day to recover from aerobic and strength activities (Dunn, Andersen, & Jakicic, 1998). In addition, the fitness coaches reviewed important concepts from the orientation, which reflected and reinforced the DVM. Examples include reviewing participants' values, the need to reduce their "disconnects" between their unhealthy habits and their values, and remaining consistent with their exercise prescription. Coaches also discussed ways in which participants could make small changes in their dietary habits and encouraged each individual to maintain involvement in the program through motivational statements and positive feedback on performance. All of these behaviors are inherent in effective fitness coaching (Chenoweth, 2007). The registered dietician provided weekly seminars on topics related to maintaining proper nutrition and following a healthy diet, presented in group settings.
Among the survey responders, 94 participants (74%) completed one wellness program, with 19 participants (15%) completing two, seven (6%) completing three, three (2%) completing four, and four participants (3%) completing all five of the wellness programs. In order to simplify analyses and provide adequate statistical power for testing the program dosage hypothesis, two groups were created: one-time (n = 94) and multiple (n = 33) program participants. Prior to data collection this study was approved by the university's Institutional Review Board.
Results
Descriptive Statistics
Table 1 provides the descriptive statistics for the program experiences and habits data from the sample. As the table indicates, 65% of the sample was currently participating in a cardio/aerobic exercise routine. This reflects a drop of approximately 28% in long-term adherence. As indicated in Table 1, 38% of the sample reported that they were currently participating in a strength/weight training exercise routine. Long-term adherence to this component of the program dropped by 53%. Additional data in Table 1 concerning the habits developed during and maintained after the program showed that participants were generally positive about the program, tending to agree rather than disagree that completing the program helped them to become more physically active, exercise more often, and improve their eating habits. However, participants did not differ from the scale midpoint (neither disagree nor agree) for the items that ascertained their behaviors related to maintaining exercise (and other forms of physical activity), and their eating habits that they learned during the 10-week program.
Table 2 presents descriptive statistics from the items assessing the extent to which components of the wellness program impacted participants' current overall healthy lifestyle. As the table shows, participants rated all of the program components as having more of a positive than a negative impact on their current health-related lifestyle. Items rated as particularly influential included learning about proper exercise techniques, receiving fitness coaching, low cost of the current program, meeting "regularly" at the campus recreation center, and obtaining exercise knowledge from the program.
The descriptive statistics for the participants' current exercise and eating behavior measures appear in Table 3. These data indicate relatively infrequent current daily cardio and strength training. This table also indicates there were three groups of participants for each of three adherence measures (not regular, occasional, and often or similar to during the wellness program). These groups were compared using a set of 1-way ANOVAs to examine the program impact index as the dependent variable. For the current exercise measure, this analysis revealed no significant group effect, F(2, 118) = 1.59, p = .209. For the current physical activity measure, the analysis revealed a significant effect, F(2, 118) = 3.88, p = .023, for the groups: not regular (M= 46.35, SD = 5.23; 95% CI [43.90-48.80]), occasional (M= 47.10, SD = 4.90; 95% CI [45.51-48.69]) and often (M= 49.31, SD = 4.89; 95% CI [48.06-50.55]). Bonferroni post hoc paired comparisons did not reach statistical significance.
Finally, for the current eating habits measure, the analysis also revealed a significant group effect, F(2, 118) = 3.52, p = .033, with post hoc tests indicating that the "not regular" adherence group scored significantly lower (M= 45.25, SD = 5.30; 95% CI [42.43-48.07]) than the "often" (M= 49.00, SD = 5.23; 95% CI [47.57-50.43]) group on the program impact index, however, not for the "occasional" (M= 48.06, SD = 4.56; 95% CI [46.78 49.34]) group. In other words, participants with lower adherence rates for physical activity and proper eating habits reported a less-positive impact of the program components on their current health lifestyle than participants with higher adherence rates.
Table 4 presents the descriptive statistics from the specific barriers to exercise that currently applied to the participants. As this table shows, most of the barriers were infrequently encountered as impediments to maintaining a persistent exercise behavior. The "not enough time" item was significantly above (p = .001) the scale midpoint (3 = sometimes). All remaining means except the lack of motivation (lazy) item were significantly below (p < .001) the scale midpoint.
Hypothesis 1: Program Fitness Changes and Adherence
The first hypothesis was that larger changes in fitness measures experienced during the program would be associated with more superior adherence values. This prediction was tested by creating change scores based on the difference between pre- and post-program fitness scores. In particular, we created difference scores for the participants' pre- and post-program weight, BMI, percent body fat, V[O.sup.2] max, lower body strength, and upper body strength measures. Examination of the correlations between change scores and adherence measures indicated no significant relationships. Thus, there was no support for the prediction that larger pre/post changes in fitness scores would be related to better long-term adherence.
Hypothesis 2: Recency of Wellness Program and Adherence
The second hypothesis was that individuals who completed the wellness program more recently would report higher levels of adherence than those who completed it less recently. This prediction was tested by associating the last program participants completed with the various adherence measures. Chi-square analysis comparing last program completed with their current participation in a regular cardio routine (yes/no) failed to find a significant difference, [X.sup.2] (4) = 2.37, p = .668. Last program completed also was not significantly related to currently participating in a weight training program (yes/no), [X.sup.2](4) = 2.07, p = .724. The current frequency of exercise behavior (1-3 programs) was not related to last program completed, [X.sup.2](8) = 4.90. p = .769. Current frequency of physical activity levels (1-3) also was not related to last program completed, [X.sup.2](8) = 12.58, p = .127. Finally, current frequency of eating habits (1-3) was unrelated to the last program completed, [X.sup.2] (8) = 8.57, p = .380.
The last program completed was not significantly correlated with participants' current level of general physical activity, exercise, and eating habits that they developed during the program (all rs < .14). Last completed program was also not significantly related to the current weekly participation in cardio exercise, r(119) = -. 113, p = .217, or strength training exercise, r(121) = .120, p = .187.
More recently completed programs were positively correlated with scores on the program impact index, r(120) = .296, p < .001. That is, participants who completed the program more recently reported a more positive impact of the program on their current health lifestyle than those who completed it less recently. More recent program completers also reported higher scores on the exercise barriers index, r(120) = .228, p < .012. That is, the more recent the program participation, the higher the number of barriers the participants reported encountering while trying to maintain their exercise habits. Although these relationships were statistically significant, the amount of variance accounted for by the correlations was small. Taken together, the results provided little support for the recency hypothesis.
Hypothesis 3: Perceived Wellness Program Impact and Adherence
Our third hypothesis stated that program participants reporting lower long-term adherence would be more likely than higher adherence participants to report receiving fewer benefits from the wellness program. Results indicated that scores on the program impact index were significantly and positively correlated with maintaining program habits of physical activity, r(120) = .229, p = .011, exercise, r(120) = .193, p= .033, and healthy eating habits, r(120) = .466, p < .001. Program impact scores also were significantly correlated with current weekly strength training exercise levels, r(119) = .265, p = .003, but not with current weekly cardio exercise levels, r(117) = .071, p = .440. In summary, these data, coupled with the descriptive statistics described earlier, provided good support for the program impact hypothesis.
Hypothesis 4: Perceived Barriers to Exercise and Adherence
For the fourth hypothesis, we predicted that low adherence participants would report experiencing more current exercise barriers, as compared to high adherence participants. The exercise barriers index was significantly and negatively correlated with several of the adherence measures, including current maintenance of physical activity habits, r(120) = -.238, p = .008, exercise habits, r(120) = -.194, p = .032, eating habits, r(120) = -.207, p = .022, frequency of current weekly cardio exercise, r(120) = -.298, p < .001), and frequency of current weekly strength exercise, r(120) = -.303, p < .001. Higher scores on the exercise barriers index were also negatively correlated with scores on the program impact index, r(118) = -.234, p = .010.
We also examined the participants who chose different frequencies (i.e., not regular, occasional, or often) of current behavior with respect to their exercise barriers index scores. One-way ANOVAs indicated that groups did not differ significantly on the barriers score for the current exercise behavior item, F(2, 119) = 2.54, p = .083. However, groups did differ significantly on the barriers score for the current physical activity level measure, F(2, 119) = 3.35, p = .038. Bonferroni post hoc comparisons showed that those indicating often (M= 35.03, SD = 10.98; 95% CI [32.24-37.82]) reported significantly lower barriers scores than those indicating occasional (M= 40.33, SD = 10.16; 95% CI [37.08-43.57]), p = .05. Participants also differed significantly for the current eating habits measure, F(2, 119) = 3.22, p = .042. Post hoc comparisons showed that those indicating often (M = 35.15, SD = 11.16; 95% CI [32.13-38.16]) reported significantly lower barriers scores than those indicating not regular (M= 42.41, SD = 9.32; 95% CI [37.62-47.20]), p = .043. In summary, the exercise barriers hypothesis received good support.
Hypothesis 5: Multiple Program Participation and Adherence
The final hypothesis predicted that participating in multiple DVM programs would be associated with greater long-term exercise adherence, as compared to participating in a single program. In accordance with the grant that supported this program, the number of programs completed ranged from 1-5 (M= 1.46, SD = .94). The majority of participants (n = 94; 74%) completed one program. As noted earlier, because so few participants completed more than one 10-week program (n = 33; 26%), two groups--single and multiple program completers--were created to test this hypothesis. Table 5 presents the descriptive statistics for the major adherence measures. As the table shows, the multiple program completers reported significantly higher levels of currently maintaining eating habits, t(125) = 3.01, p = .003, and of the program impact index scores, t(125) = 3.62, p < .001, than single program completers. However, the two groups did not differ significantly on the exercise adherence measures (p > .05). Thus, there was only partial support for this hypothesis.
Discussion
The primary purpose of this study was to determine the long-term exercise adherence of participants who completed one or more 10-week values-based adult fitness programs based on the Disconnected Values Model (DVM; Anshel, 2008, 2013). A secondary purpose was to investigate the differences between participating in a single wellness program, as compared to experiencing multiple similar programs with respect to long-term adherence. Among the study participants, 65% reported currently participating in cardiovascular exercise and 38% reported currently participating in resistance/strength training. While studies have examined the influence of interventions on exercise participation, relatively few investigations have investigated long-term exercise adherence of these programs (Anshel, 2014; Muller-Riemenschneider et ah, 2008).
Descriptive statistics provided several important findings on the participants' program and post-program experiences. First, participants rated all of the program components as having a positive impact on their current health lifestyle. In other words, they evaluated the overall program quite favorably. Second, participants did not agree that they were currently maintaining the habits that they developed during the program. This result reflects a drop in exercise adherence, as reported earlier. Third, participants had a difficult time adhering to the strength training component of the program. This result was likely related to the importance and limited availability of the fitness coaches in their role as instructors on the proper use of the weight equipment, as well as access to equipment. It is plausible to surmise that one coaching session per week was insufficient to provide the proper amount of instruction and motivational support, particularly during the early part of the 10-week program before strength training skills were fully developed (Findorf et al., 2009).
Another finding from the descriptive data was that participants who reported they were not currently exercising or not eating properly also reported that the program had a less positive impact on their current health behavior, as opposed to exercise and eating adherers. This trend likely reflects the fact that unhealthy habits are difficult to break, even if people know that their habit is unhealthy (Loehr & Schwartz, 2003). Program participants who possessed a higher rate of unhealthy habits before the program began may have felt less buy-in of the program's values-based content than participants who had relatively healthier habits at the program's start. It is also possible that those with a relatively high rate of unhealthy habits did not adhere to the program to the same extent than their healthier program peers. This may have been due to their need for high structure and monitoring (i.e., coach presence) which they received during the program but not post-program. This finding implies that less healthy individuals may need more structure, coaching, and motivation than their fitter, healthier counterparts. The extent to which participants succeeded in transferring program components to adopt permanent lifestyle changes requires additional research.
Hypothesis 1: Program Fitness Changes and Adherence
We predicted, in the first hypothesis, that improved program fitness scores would be associated with greater long-term exercise adherence. Results indicated non-significant correlations between fitness change scores and adherence rates. There are several possible interpretations for this result. Perhaps exercisers who experienced acceptable fitness changes during the program might have been inclined to discontinue adhering to their post-program habits to maintain those changes because they reached their primary program goal of improving their fitness (such as losing some weight). After reaching their fitness goal, particularly within the program's 10-week time frame, there was no intention to develop long-term behavioral changes. This explanation becomes increasingly likely based on marked changes in scheduling regular exercise sessions and the sudden removal of social support after the program ended.
Another plausible reason for low post-program adherence is that some participants did not experience sufficient improvement in fitness scores. It would have been desirable for these individuals to adhere to the program components after its completion in order to continue improving their fitness. Additionally, the participants with larger program changes may have begun the study with poorer overall fitness as compared to more-fit individuals. An important next step in future research would be to determine which of these interpretations is valid in examining the link between initial fitness levels, changes in fitness, and long-term adherence rates. In particular, tracking participants' perceptions of the nature and extent of their fitness changes during and at the completion of a wellness program, and how these perceptions of change relate to long-term adherence, warrants future study.
Hypothesis 2: Recency of Wellness Program and Adherence
The second hypothesis was that more recent participation in the 10-week wellness program would be associated with greater exercise adherence rates. We felt that compared to distant program experiences, exercise experienced more recently would be associated with learned routines that are still easily remembered and performed, less time for exercise barriers to reduce program adherence, and less opportunity for unhealthy behavior patterns to have returned. More recent programs should also be associated with the continued presence of motivational incentives, such as social support, observed improvements in fitness and exercise performance, and improved general well-being. The results indicated that length of time since program completion had relatively little relationship to long-term adherence.
One possible explanation for this outcome was that the participants, regardless of when they completed the program, struggled with overcoming their own, personal exercise barriers. The sudden removal of social support (e.g., exercise buddies or program coach) may have reduced the incentive for participants to go from developing an exercise behavior pattern to creating an exercise habit in which exercise and other forms of physical activity would become an integrated, scheduled, and automatic part of their day (Loehr & Schwartz, 2003). When the program ended, participants were less likely to have coaching unless they solicited self-funded coaching on their own.
It is possible that other factors not directly related to time of program participation (e.g., changes in problem-solving skills, improved exercise skills, reduced physical exertion, improved self-efficacy beliefs) might account for the failure to detect a recency effect. Research into the ways that participants are able to replace essential program components after completion of structured interventions, and how these activities are related to long-term adherence, warrants further study.
Hypothesis 3: Perceived Wellness Program Impact and Adherence
The third hypothesis addressed the relationship between the exercisers' perceived quality of the wellness program and their adherence rate to the exercise program. Participants generally reported a highly favorable attitude toward the wellness program. Top ranked reasons for their favorable attitude included, in descending order: (a) learning new exercise techniques, (b) availability of fitness coaching, (c) low program cost, (d) meeting at the (conveniently located) campus recreation center, and (e) improved fitness knowledge. Participants who reported that the program components had a relatively less effective impact on their health-related lifestyle behavioral patterns also did not adhere to their exercise program or maintain proper eating habits consistently, as compared to participants who rated the program components more positively.
Whether negative attitudes toward the program directly resulted in non-adherence to the program's components (i.e., fitness, dieting, lifestyle change), or that non-adherence generated these unpleasant feelings is uncertain. It is also possible that current physical activity and eating habits ratings are affected by other variables that we did not measure. Clearly, program components that are more compatible with participant expectations will result in stronger adherence to effective program outcomes than if individual needs are ignored (Goldstein, DePue, & Kazura, 2009).
The DVM extols the virtues of developing rituals that are planned, specific, detailed, and executed accordingly. It is key to replace a person's need for immediate gratification with delayed gratification and creating health-protective behaviors for the long-term (Anshel, 2008, 2013; Hall & Fong, 2003; Loehr & Schwartz, 2003). The current results suggest that participants who made improved health protective behaviors and reduced health-damaging behaviors based on the DVM program were able to engage better in long-term adherence than participants who did not (see also Anshel, Kang, & Brinthaupt, 2010; Brinthaupt et al., 2013).
Hypothesis 4: Perceived Barriers to Exercise and Adherence
The present results supported our prediction that participants reporting more exercise barriers would report lower exercise adherence. This finding is consistent with previous research showing that people who maintain their exercise habits also report experiencing fewer exercise barriers (e.g., Brinthaupt et al., 2010; Kruger, Yore, Bauer, & Kohl, 2007). The inability to effectively deal with exercise barriers, however, does provide a possible explanation for poorer exercise adherence, at least among some of the participants. Future research is needed to assess the exact nature and extent of the various exercise barriers and how each barrier might affect adherence, an issue not addressed in the present study.
Hypothesis 5; Multiple Program Participation and Adherence
Hypothesis 5 stated that participants who experience multiple 10-week wellness programs would maintain superior adherence on eating and exercise habits than single program completers. It is important to note that testing this prediction was limited by a relatively small sample size for participants who engaged in two or more programs. Still, the results indicated that multiple program completers did, in fact, report significantly better eating habits, but not significantly better exercise adherence, as opposed to single program completers. Participants were not asked to reveal their reasons for engaging in more than one DVM program.
There are several possible interpretations for why completing multiple programs was not associated with higher levels of exercise adherence than completing only one program. First, some participants who found it difficult to adhere to their first program might have been more likely to participate in additional programs and overcome earlier limitations. Second, people who participated in more than one program may have been relying on the program to learn proper exercise techniques and dietary habits to improve their fitness and control body weight, with no intention to continue receiving fitness and nutrition coaching after the program ended. In addition, it may take more time (i.e., beyond one program) to establish new, healthy habits. Some studies have shown that habits can take 12 to 18 weeks to become established and routine (Loehr & Schwartz, 2003). Future research that determines the number of wellness programs that will lead to optimal benefits and long-term post-program adherence is clearly warranted.
Limitations and Implications for Future Research
It is important to note some of the major limitations of the present study. For instance, it would have been advantageous to more closely monitor coach behaviors and the quality of the relationship with their clients to at least partially explain exercise adherence rates. While it was not feasible to monitor coach behaviors regularly and to provide ongoing feedback on coach performance, participant feedback about their coach was generally very favorable. What aspects of the coaching they received were associated with long-term adherence rates is a topic for future research.
A challenge to exercise researchers and practitioners is to find a valid means of determining short-term and long-term exercise adherence. A limitation of the current study was that exercise adherence was obtained entirely from participant self-reports. This might partly explain why several of the reported correlations, while statistically significant, accounted for relatively low levels of variance. Adherence also can be measured with the use of pre- and post-intervention testing which could corroborate self-report data (Anshel, 2014; Kampshoff et al., 2014). In addition, a wide range of physical devices, such as pedometers, accelerometers, and heart rate monitors may be used to measure exercise adherence (Conaway et al., 2009).
Future research also needs to examine the variety of psychological barriers to exercise adherence not examined in the current study. Some of these mental barriers include anxiety, depression, unrealistically high self-expectations, negative self-talk, and negative or neurotic perfectionism (see Anshel, 2014, and Lox, Martin, & Petruzzello, 2014, for reviews of literature). Examining the nature and frequency of these mental barriers might help account for post-program exercise non-adherence.
In summary, we found that participants who perceived more of a positive impact from the program and who experienced fewer barriers post-program reported higher rates of longterm adherence. There was little support that long-term adherence was related to extent of program changes in fitness, the recency with which participants completed the program, or the completion of multiple programs. Future research on the DVM and similar interventions is needed to examine the robustness of these findings. Additional research is also needed on the causes and mechanisms of long-term adherence of newly acquired healthy habits, particularly related to wellness programs that promote a healthy lifestyle.
References
Anshel, M. H. (2014). Applied health fitness psychology'. Champaign, IL: Human Kinetics.
Anshel, M. H. (2013). A cognitive-behavioral approach for promoting exercise behavior: The Disconnected Values Model. Journal of Sport Behavior, 36, 107-129.
Anshel, M. H. (2008). The Disconnected Values Model: Intervention strategies for health behavior change. Journal of Clinical Sport Psychology, 2, 357-380.
Anshel, M. H., & Kang, M. (2007a). An outcome-based action study on changes in fitness, blood lipids, and exercise adherence based on the Disconnected Values Model. Behavioral Medicine, 33, 85-98.
Anshel, M. H., & Kang, M. (2007b). Effect of an intervention on replacing negative habits with positive routines for improving full engagement at work: A test of the Disconnected Values Model. Consulting Psychology Journal: Practice and Research, 59, 110-125.
Anshel, M. H., & Kang, M. (2008). Effectiveness of motivational interviewing techniques on changes in fitness, blood lipids, and exercise adherence of police officers: An outcome-based action study. Journal of Correctional Health Care, 14, 48-62.
Anshel, M. H., Kang, M., & Brinthaupt, T. M. (2010). A values-based approach for changing exercise and dietary habits: An action study. International Journal of Sport and Exercise Psychology', 8, 413-433.
Brinthaupt, T. M.,Kang, M., & Anshel, M. H. (2010). A delivery model for overcoming psycho-behavioral barriers to exercise. Psychology of Sport and Exercise, 11, 259-266.
Brinthaupt, T. M., Kang, M., & Anshel, M. H. (2013). Changes in exercise commitment following a values-based wellness program. Journal of Sport Behavior, 36, 3-22.
Buckworth, J., Dishman, R. K., O'Connor, R J., & Tomporowski, P. (2013). Exercise psychology (2nd ed.). Champaign, IL: Human Kinetics.
Chenoweth, D. H. (2007). Worksite health promotion (2nd ed.). Champaign, IL: Human Kinetics.
Conaway, M. R., Rejeski, W. J., & Miller, M. E. (2009). Statistical issues in measuring adherence: Methods for incomplete longitudinal data. In S. A. Shumaker, J. K. Ockene, & Riekert. K. A. (Eds.), The handbook of health behavior change (3rd ed.; pp. 375-391). New York: Springer Publishing Company.
Dishman, R. K., & Buckworth, J. (1997). Adherence to physical activity. In W. P. Morgan (Ed.), Physical activity & mental health (pp. 63-80). Washington, D.C.: Taylor & Francis.
Dunn, A. L., Andersen, R. E., & Jakicic, J. M. (1998). Lifestyle physical activity interventions. History, short- and long-tenn effects, and recommendations. American Journal of Preventive Medicine, 15, 398-412.
Evers, A., Klusmann, V., Ziegelmann, J. P., Schwarzer, R., & Heuser, I. (2012). Long-term adherence to a physical activity intervention: The role of telephone-assisted vs. self-administered coping plans and strategy use. Psychology and Health, 27, 784-797.
Findorf, M. J., Wyman, J. F., & Gross, C. R. (2009). Predictors of long-term exercise adherence in a community-based sample of older women. Journal of Women s Health, 18, 1769-1776.
Goldstein, M. G., DePue, J., & Kazura, A. N. (2009). Models of provider-patient interaction and shared decision-making. In S. A. Shumaker, J. K. Ockene, & Riekert, K. A. (Eds.), The handbook of health behavior change (3rd ed.; pp. 107-125). New York: Springer Publishing Company.
Hall, P. A., & Fong, G. T. (2003). The effects of a brief time perspective intervention for increasing physical activity among young adults. Psychology and Health, 18, 685-706.
Hunt, R, & Hillsdon, M. (1996). Changing eating & exercise behaviour: A handbook for professionals. London: Blackwell.
Kampshoff, C. S., Jansen, F., van Mechelen, W., May, A. M., Brag, J., Chinapaw, J. J. M., & Buffart, L. M. (2014). Determinants of exercise adherence and maintenance among cancer survivors: a systematic review. International Journal of Behavioral Nutrition and Physical Activity, 11, 80. doi: 10.1186/1479-5868-11-80
Klesges, L. M., Estabrooks, P. A., Dzewaltowski, D. A., Bull, S. S., & Glasgow, R. E. (2005). Beginning with the application in mind: Designing and planning health behavior change interventions to enhance dissemination. Annals of Behavioral Medicine, 29, 66-75.
Kruger, J., Yore, M. M., Bauer, D. R.. & Kohl, H. W., Ill (2007). Selected barriers and incentives for worksite health promotion services and policies. American Journal of Health Promotion, 21, 439-447.
Loehr, J., & Schwartz, T. (2003). The power of full engagement: Managing energy, not time, is the key to high performance and personal renewal. New York: The Free Press.
Lox, C. L., Martin, K. A., & Petruzzello, S. J. (2014). The psychology of exercise: Integrating theory and practice (4th ed.). Scottsdale, A Z: Holcomb Hathaway.
Muller-Riemenschneider, F., Reinhold, T., Nocon, M., Willich, S. (2008). Long-term effectiveness of interventions promoting physical activity: A systematic review. Preventive Medicine, 47, 354-368.
Tobi, R, Estacio, E. V., Yu, G., Renton, A., & Foster, N. (2012). Who stays, who drops out? Biosocial predictors of longer-term adherence in participants attending an exercise referral scheme in the UK. BMC Public Health, 12, 347. doi: 10.1186/1471-2458-12-347
Winett, R. A., Anderson, E. S., Wojcik, J. R., Winett, S. G., & Bowden, T. (2007). Guide to health: Nutrition and physical activity outcomes of a group-randomized trial of an internet-based intervention in churches. Annals of Behavioral Medicine, 33, 251-261.
Thomas M. Brinthaupt and Mark H. Anshel
Middle Tennessee State University
Address correspondence to: Tom Brinthaupt, Middle Tennessee State University P.O. Box X034, MTSU, Murfreesboro, TN 37132. Email: tom.brinthaupt@mtsu.edu Table 1 Descriptive Statistics for Major Adherence Measures: Program Experiences and Habits Variable Yes No Established a regular cardio/aerobic exercise 118 (93%) 9 (7%) routine during the 10-week program Established a regular weight lifting exercise 116 (91%) 11 (9%) routine during the 10-week program Currently participating in an ongoing program 83 (65%) 44 (35%) of regular cardio (aerobic) exercise Currently participating in an ongoing program 48 (38%) 79 (62%) of regular weight (strength) training Disagree/Agree items (1 = strongly disagree, M SD 5 = strongly agree) Completing the wellness program helped me to 3.74 1.03 *** become more physically active. Completing the wellness program helped me to 3.80 1.86 *** exercise more often. Completing the wellness program helped me to 3.30 0.96 *** improve my eating habits. I am currently maintaining approximately the 3.16 1.14 same habits of physical activity I practiced during the program. I am currently maintaining approximately the 3.02 1.15 same exercise habits I practiced during the program. I am currently maintaining the eating habits 1 3.10 1.01 learned from the program. Note. N= 127. Comparison to the scale midpoint (3), *** p < .001. Table 2 Impact of Program Components on Current Health Lifestyle Variable M SD Range Initial program orientation session 3.78 .61 2-5 Collection of pretest and posttest fitness data 4.02 .59 3-5 Emphasis on developing a lifestyle that was 3.83 .65 2-5 consistent with my values Learning the proper exercise techniques I was 4.31 .60 3-5 taught Receiving fitness coaching 4.42 .73 1-5 Receiving nutrition coaching 3.65 .83 1-5 E-mail educational messages from the wellness 3.60 .79 1-5 program Director Support material from the wellness program Website 3.54 .74 2-5 Low cost of the program 4.44 .64 3-5 Meeting my coach and exercising at the Campus Rec 4.50 .73 1-5 Center Nutrition knowledge provided by the program 3.70 .79 1-5 Exercise knowledge provided by the program 4.29 .60 3-5 Note. N = 124. Respondents rated the items using a 5-point scale (1 = strong negative impact, 5 = strong positive impact). All means were significantly higher than the scale midpoint (3 = neither positive nor negative impact), *** p<.001. Table 3 Descriptive Statistics for Current Exercise and Eating Behavior Variable M SD n How often per week participating 2.65 1.66 121 in cardio exercise (0-7) How often per week participating 1.16 1.22 123 in strength training exercises (0-7) Option best describing your current: (n = 123) not occa- often regular sional Exercise behavior (over recent weeks) 23 31 69 Physical activity levels (over recent weeks) 20 40 63 Eating habits (over recent weeks) 17 51 55 Table 4 Frequency of Currently Experiencing Barriers to Exercise Variable M SD Not enough time 3.37 1.14 Lack of fitness information/knowledge 1.80 0.83 Lack of confidence 1.71 0.87 Intimidated to exercise in public 1.68 0.96 Do not have exercise partner 2.07 1.19 Lack of support from partner and others 1.80 1.11 Have history of giving up 2.15 1.13 Too expensive (for exercise clothing, personal 1.57 0.91 trainer, club membership) No close access to exercise facility 1.45 0.82 Fear of injury 1.53 0.86 Find exercise unpleasant 2.21 1.17 Injury 1.88 1.10 No personal trainer to instruct or motivate me 2.30 1.24 Lack of results (not improving) 2.20 1.15 Too much stress in my life 2.33 1.21 Too uncomfortable exercising (fatigue, sweating, 2.02 0.98 discomfort) Already reached my goals (1 feel healthy) 1.49 0.80 Thinking that my health is in God's hands 1.38 0.73 Cannot affors fitness facility membership (off campus) 1.42 0.85 Lack of motivation (lazy) 2.87 1.30 Do not have proper equipment 1.68 0.92 Note. N= 122. Respondents rated the items using a 5-point scale (1 = never, 5 = very often). All means except the lack of motivation (lazy) item were significantly below (p < .001) the scale midpoint (3 = sometimes)', the not enough time item was significantly above (p = .001) the scale midpoint. Table 5 Adherence Data for Single and Multiple Program Completers Single Program (n = 94) Variable M SD I am currently maintaining approximately 3.10 1.14 the same habits of physical activity 1 during the program. I am currently maintaining approximately the 3.00 1.14 same exercise habits I practiced during the program. I am currently maintaining the eating habits 2.95 1.00 I learned from the program. How often per week participating in cardio 2.60 1.62 exercise (0-7) How often per week participating in strength 1.09 1.16 training exercises (0-7) Program impact index 47.11 5.12 Exercise Barriers index 37.26 11.06 Multiple Programs (n = 33) Variable M SD I am currently maintaining approximately 3.33 1.14 the same habits of physical activity 1 during the program. I am currently maintaining approximately the 3.09 1.21 same exercise habits I practiced during the program. I am currently maintaining the eating habits 3.55 0.94 ** I learned from the program. How often per week participating in cardio 2.78 1.80 exercise (0-7) How often per week participating in strength 1.33 1.37 training exercises (0-7) Program impact index 50.72 3.92 *** Exercise Barriers index 38.27 11.31 Note. ** p < .005; *** p < .001.
