Breast Cancer Survivors’ Exercise Preferences Change During an Exercise Intervention and are associated with Post-Intervention Physical Activity

Purpose Exercise program preferences are important for designing physical activity (PA) interventions; yet may change following an intervention. Further, the relationship between preferences and PA behavior change is unclear. This study evaluated exercise program preferences among breast cancer survivors (BCS) before and after a behavioral intervention and associations between program preferences and PA change. Methods BCS were randomized to the BEAT Cancer intervention (n = 110) or written materials (n = 112). Questionnaires assessed exercise program preferences. Minutes per week of moderate-to-vigorous PA (MVPA) were accelerometer-measured and self-reported at baseline (M0), post-intervention (M3), and 3-month follow-up (M6). Results At M0, the majority of intervention group participants preferred exercising with others (62%) yet shifted to preferring exercising alone (59%) at M3 (p < 0.001). Furthermore, preferring exercising with others at M0 was associated with greater increases in self-reported MVPA between M0 and M6 (124.2 ± 152 vs. 53.1 ± 113.8, p = 0.014). BCS preferring facility-based exercise decreased after the BEAT Cancer intervention (14% vs. 7%, p = 0.039) and preferring exercising at home/had no preference at M0 had greater improvements in accelerometer-measured MVPA from M0 to M3 (74.3 ± 118.8 vs. −2.3 ± 78.4, p = 0.033) and M0 to M6 (44.9 ± 112.8 vs. 9.3 ± 30.4, p = 0.021). Exercise program preferences regarding mode of counseling, training supervision, and type of exercise changed from M0 to M3 but were not associated with changes in MVPA. Conclusion Findings suggest BCS exercise program preferences may change after an intervention and be associated with changes in MVPA. Understanding the role of PA preferences will better inform the design and success of PA behavior change interventions.

As the number of women living beyond their breast cancer diagnosis continues to increase due to advancements in treatment [1], survivors often incur side effects and symptom burden from treatments, such as chemotherapy, that impact their quality of life [2]. Despite strong evidence of physical, psychological, and functional bene ts from physical activity [3][4][5][6][7][8], breast cancer survivors are not participating in the recommended, 150 + minutes of moderate-to-vigorous physical activity per week (MVPA) [3,8,9]. In an effort to increase physical activity within the breast cancer survivor population, factors that may improve engagement in regular physical activity are being investigated, including programming preferences that could potentially be used to tailor behavioral change physical activity interventions.
Exercise programming preferences have been assessed within various cancer populations [10][11][12] and more speci cally breast cancer survivors [13,14]. However, the majority of these studies in cancer survivors have been cross-sectional, with very few longitudinal studies involving an exercise or physical activity intervention and none comparing preference changes in an intervention vs. control group [12]. This is a notable gap in the scienti c literature because such preferences may change after engaging in a physical activity intervention [11,15], which could potentially impact longer term physical activity behavior maintenance.
The only two available reports of pre/post intervention change in exercise program preferences within the cancer population suggest that changes in preferences regarding location, social support, supervision, trainer education and counseling source may occur while other ndings were inconsistent between the two studies [11,15]. Head and neck cancer survivors recruited into a 12-week group resistance training intervention reported a change in preferred exercise counseling source, i.e., changed from preferring a personal trainer at baseline to preferring an instructor with experience working with cancer survivors following the intervention [11]. In contrast, women with metastatic breast cancer preferred their exercise counseling source to be a physical activity specialist before and after a 6 month home-based, unsupervised, aerobic training intervention [15]. However, both studies saw changes in preferred location and social support in which, after the interventions, participants switched to preferring exercising at a facility and with others who have a similar cancer diagnosis. It is important to note that these studies were not conducted within the non-metastatic breast cancer survivor population and implemented different types of physical activity behavior change interventions. Moreover, no prior study to our knowledge has compared preference changes in an intervention vs. control group, which is particularly important to account for natural evolution of preferences that may occur over time [16]. Therefore, our study purpose was to examine breast cancer survivors' exercise program preferences before and after the 3-month BEAT Cancer physical activity behavior change intervention, which was comprised of 12 supervised exercise sessions tapered over six weeks to exclusively home-based sessions and six social cognitive theory-based exercise counseling sessions. By doing so, we can gain a better understanding of preference changes and how such changes relate to physical activity behavior post-intervention.
Based on the BEAT Cancer intervention design, we hypothesized breast cancer survivors would change their exercise program preferences following the BEAT Cancer intervention, and those who preferred programs that were home-based and unsupervised would have a greater increase in physical activity levels following the BEAT Cancer intervention.

Study Design
This is a secondary analysis of the multicenter two arm randomized controlled BEAT Cancer trial [7,8]. Eligible participants were women, ages 18-70 years old with a history of ductal carcinoma in situ (DCIS) or stage I-IIIA breast cancer, and with a self-reported sedentary lifestyle (engaging in ≤ 30 minutes of vigorous or ≤ 60 minutes of moderate intensity physical activity per week on average over six months). Eligible women must have completed primary cancer treatment, been ≥ 8 weeks post-surgery, English speaking, and received medical clearance from their physician. Exclusion criteria included conditions that contraindicated physical activity or would interfere with assessments, and current participation in another exercise study. Additional details regarding inclusion and exclusion criteria and study design were reported by Rogers et al [8,17]. Participants were recruited into the study between 2010 to 2013. Institutional Review Board (IRB) approval was obtained, and all participants provided written informed consent. Study personnel completing the assessments were blinded to the randomly assigned group allocation.

Beat Cancer Intervention
This 3-month, social cognitive theory-based intervention consisted of 12 supervised exercise sessions tapered over six weeks, followed by three face-to-face exercise counseling sessions every two weeks with an exercise specialist. The exercise prescription aimed to gradually engage participants in 150 weekly minutes of MVPA using a progression previously reported [7]. Additionally, six group sessions led by trained facilitators provided behavioral counseling (time management, stress management, behavioral modi cation strategies etc.). An educational notebook with exercise-speci c information related to intervention goals, exercise safety, nutrition information, exercise log sheets and personal heart rate monitor were provided to participants as part of the intervention. Further details about participant adherence and quality control for delity were previously published [7,8,17].

Usual Care Intervention
The usual care (UC) group received written materials from the American Cancer Society describing physical activity recommendations for cancer survivors (see Rogers et al. [7] for additional details regarding the UC group intervention).

Data Collection
Measurements were taken at baseline (M0), immediately post-intervention (M3) and 3 months after intervention completion (M6). Anthropometric measurements were obtained from a scale and stadiometer (Continental Health-O-Meter #400 DML medical scale [precision to nearest 8th of a pound] and Seca 763 Digital Column Scale [precision to nearest 0.1 lb]). Body mass index (BMI) was calculated from the height and weight [weight (kg)/height (m 2 )]. Demographics, medical history, and cancer history were self-reported. Exercise program preferences were collected via self-administered surveys based on previous work [13,18,19] (Table 1). Participants were asked: whom would you most like to receive exercise counseling? (cancer exercise specialist, personal trainer, health club trainer, cancer survivor, physician, nurse, no preference); how would you prefer to receive exercise counseling sessions? (one-onone, group); and how would you most like to receive exercise counseling? (face-to-face, telephone, video tape, written material, internet, audiotape, interactive workbook, no preference). Additionally, participants were asked about their preferences regarding exercise training related social support (group, with others); trainer supervision (supervised, unsupervised, no preference); location (outdoors, at home, health club, cancer center, at work, no preference); program type (aerobic, resistance training, combination of both, no preference); exercise type (walking, resistance training, water activities, bicycling, yoga, pilates, jogging, other); and structure (scheduled, exible, no preference).
Physical activity was measured using ActiGraph accelerometers (models GT1M and GT3X; protocol previously published) [17]. Additionally, self-reported leisure time physical activity was assessed using the Godin Leisure-Time Exercise Questionnaire [20,21]. The frequency and average duration of exercise bouts per week over the previous month at light, moderate, and vigorous intensity were collected, and minutes of moderate intensity physical activity were calculated. Both accelerometer and self-reported minutes of vigorous intensity activity were multiplied by two and added to moderate minutes.

Statistical Analysis
Self-reported exercise program preferences were described by frequency and percentages at baseline and 3-months. To investigate the differences in program preferences before and immediately after the 3month intervention period, preferences at M0 and M3 were dichotomized for both groups, and the McNemar chi-square test was used to examine changes between the two time periods for paired data. In addition, the Pearson chi-square test, or Fisher's exact test (if the assumptions for the Pearson chi-square were not satis ed), was used for group comparisons (BEAT cancer group, UC group) of categorical variables. Associations between program preferences and change in MVPA were examined using the twogroup t test. Statistical tests were two-sided. Statistical signi cance was set at p < 0.05. Analyses were performed using SPSS software, version 28 (IBM Corp., Chicago, IL).

Participant Characteristics
A total of 222 breast cancer survivors were randomized to either the BEAT Cancer (n = 110) or UC (n = 112) groups. The mean age of participants was 54 ± 9 years old, with the majority being White (84%).
Participants had an average of 16 ± 3 years of education. Cancer stages varied from, DCIS (11%), stage I (42%), stage II (35%), and stage III (12%). The average time since breast cancer diagnosis was 54 ± 55 months, with 58% and 68% reporting a history of chemotherapy and radiation therapy, respectively.

Program Preferences At Baseline
Participant preferences for exercise counseling, training, and programming are provided in Table 1 (see  below). At M0 for all participants combined, the most preferred exercise counseling source was a cancer exercise specialist (41%), with one-on-one sessions (67%) and face-to-face delivery (77%). Other preference options included exercising with others (57%), supervised (69%), and outdoors (28%). Although most breast cancer survivors preferred a combination of both aerobic and resistance training (67%), the most popular speci c type of exercise was walking (54%). With regard to program structure, participants were most interested in a exible rather than scheduled exercise program (42%). When asked to indicate the most they were willing to pay per month for an exercise program, 31% responded ≤$10, with 26% indicating $11-$20, 25% indicating $21-$30, and 18% indicating ≥$31. Traveling up to 15 miles was the furthest most participants were willing to travel for an exercise program (74%), even if gas were paid for by another entity (62%).

Program Preference Changes Within Each Study Group
Within the BEAT Cancer group, ve preferences had signi cant changes in the proportion of participants endorsing them from M0 to M3 (Table 2). A majority of BEAT Cancer participants preferred face-to-face delivery to other options (video tape, written materials, internet, or no preference) but this proportion decreased from M0 to M3 (77% vs. 65%, p = 0.019). At M0, the majority of participants preferred exercising with others (i.e., 62%); however, this decreased to 39% at M3 such that the majority (59%) preferred exercising alone at M3 (p < 0.001). The number of BEAT Cancer participants preferring supervised exercise training decreased from M0 to M3 (67% vs. 41%, p = 0.015). Furthermore, the number of participants who preferred to exercise at a facility decreased after the intervention (14% vs. 7%, p = 0.039). Although a combination of aerobic and resistance training was most preferred at M0 and M3 (83% vs. 69%, p = 0.011), aerobic and resistance training as separate preferences gained interest after the intervention. No other within group differences in program preferences were noted for the BEAT Cancer intervention group.

Associations With Change In Physical Activity Within Beat Cancer Intervention
Baseline (M0) preferences related to training social support and location were found to be associated with changes in MVPA levels (

Discussion
Breast cancer survivors' exercise program preferences can change during an exercise intervention and may be associated with exercise behavior change from pre-to post-intervention. This is the rst randomized intervention study to our knowledge to elucidate program preference changes for the intervention vs. control condition coupled with exploring the associations between these preferences and post-intervention MVPA. Statistically signi cant changes over the 3-month intervention period were noted among the BEAT Cancer intervention recipients for ve preferences (i.e., delivery, group/alone, supervision, location, type), while only two of the thirteen program preferences (i.e., counseling source and supervision) demonstrated a statistically signi cant change among those in the UC group. Two baseline preferences (i.e., group/alone and location) were associated with the BEAT Cancer intervention success from M0 to M3 and from M0 to M6 in the BEAT Cancer intervention group. Yet, location was the only preference associated with change in physical activity during the 3-month post-intervention period (M3-M6).
After receiving the BEAT intervention, the proportion of participants interested in face-to-face exercise counseling declined; however, it was still the most preferred mode of delivery. Similarly, the proportion preferring supervised exercise training decreased when assessed immediately post-intervention. Although the same change was seen within the UC group, analyses indicated a greater proportion of BEAT Cancer participants changed their preference from supervised to unsupervised training when compared with UC.
These results indicate survivors had less of a desire for professional exercise support following the BEAT Cancer intervention. This could be a result of the intervention design tapering the exercise sessions from supervised in a facility to unsupervised home-based sessions over the 3-month period with the goal of helping participants become more independent with their physical activity behaviors. It seems possible that this design gave the participants the education needed to feel more capable of exercising on their own without needing more professional support.
Also, fewer survivors were interested in exercising at a facility after participating in the intervention, which was expected as the intervention transitioned from facility to home-based exercise sessions over the intervention period. While home-based exercise gained interest, having this preference at M3 was associated with a decrease in self-reported MVPA following the intervention from M3 to M6. Yet, freeliving accelerometer-measured MVPA levels did not decrease among those who preferred to exercise at home. A possible explanation for these results may be that without traveling to a designated exercise location, the participants may feel like they are not being as active, thus self-reporting lower activity levels. Further research is needed to better understand if behavior changes preferences or vice versa and how these relationships may differ for free-living vs. volitional leisure-time physical activity behavior. Survivors who preferred exercise training in a group setting at M0 exhibited greater increase in selfreported MVPA between M0 and M6; however, at M3 the majority changed to preferring exercise training alone. As seen in this study and previous studies, having social support by training with a partner or in a group setting can provide accountability and increase activity levels [22][23][24]; yet, coordinating exercise in a group setting may be more di cult and require more effort than exercising alone, which could have contributed to this preference change. Since the intervention tapered to individual exercises sessions at home the participants may have gained con dence in exercising on their own. The majority of participants preferred a combination of both aerobic and resistance exercise training before and after the BEAT cancer intervention, yet after the intervention a portion of participants changed to no longer preferring the combination of both, enough to be statistically signi cant. Furthermore, when asked about speci c types of exercise, resistance exercise training was least preferred by the majority of participants who reported to prefer another form of exercise (i.e., walking, water activities, cycling, yoga, etc.) over resistance training. Since the BEAT Cancer intervention focused on aerobic exercise training, speci cally walking, participants may have been more inclined to focus strictly on aerobic rather than resistance exercise.
These results are consistent with those reported by Owens et al. [25] examining exercise program preferences in adults with metabolic syndrome. Participants' preferences changed signi cantly from preferring supervised exercise training in a group setting to favoring independent (i.e., alone), Contrasting results were found in women with metastatic breast cancer following a 6-month unsupervised home-based intervention. Most participants favored training at home but following the unsupervised home-based intervention, the majority changed to preferring training at a community-based facility. Further, patients with head and neck cancer had a signi cant increase in preference towards training at a cancer center following a 12-week, group resistance exercise training intervention. While the ndings are not consistent with our results, this preference change could support intervention fatigue, as the observed preference changes were a shift away from the intervention design. It seems possible that these contrasting results could be due to people preferring variety and in uence the desire for the opposite of the intervention.
No statistically signi cant differences were found for counseling source preference in this study; however, previous studies in the cancer patient populations found a statistically signi cant change [11,15]. The majority of head and neck patients changed from preferring a personal trainer to a professional with a kinesiology degree [11], while the majority of metastatic breast cancer patients preferred an exercise specialist with an increase in this preference following their intervention [15]. While both studies can offer potential explanations for the observed results, it is important to note that metastatic breast and head and neck cancer patients have greater risks for side effects or adverse events and cancer/treatmentrelated sequelae, which may in uence the differences in results.
The strengths of this study include the randomized controlled trial design, large sample size (relative to the current literature), high participant retention rate, and assessment of both objective and self-report physical activity within the context of statistically signi cant intervention e cacy at M3 and M6 [7,11,15,25]. This study adds value to the current limited literature reporting pre/post changes in program preferences during an exercise intervention in cancer survivors. Further, no prior study to our knowledge within the cancer population has compared program preference changes in an intervention vs. control group, in an effort to account for natural evolution of preferences that may occur over time in the absence of an intervention [25].

Study Limitations
We acknowledge this study is limited in variability among race and rurality of the sample. Participants were primarily White women residing in metro or small urban areas and may have better access to exercise resources than non-White breast cancer survivors in more rural or inner-city urban areas.
Additionally, these ndings may not be generalizable to other cancer types beyond breast or to resistance training interventions. Further, our ndings should not be used to conclude the prevalence of exercise preferences because it was not a population-based survey. Nevertheless, comparison of our study with a similar survey conducted in a large population-based sample of breast cancer survivors [14], demonstrates similar ndings in that both report the majority of breast cancer survivors preferring faceto-face exercise counseling, exible exercise scheduling, walking as the exercise type, and exercising alone, unsupervised, and at home/no preference [14].

Clinical Implications
Our study suggests that tailoring to social support (i.e., group vs. individual) and location (i.e., homebased vs. facility) preferences may be the key for increasing perceived acceptability of and engagement with an intervention. Further, preference may change over time suggesting periodic reassessment for additional tailoring may be indicated. Future study is warranted to determine why preferences change and whether tailoring an intervention towards preferences is feasible and optimizes exercise behavior change short and long term. Current approaches in the exercise oncology eld acknowledge that a generic exercise program does not work for the general population and interventions are commonly tailored using the Frequency, Intensity, Time, Type (FITT) principles and no other intervention characteristics. Previous research has highlighted that tailoring interventions solely using the FITT principles may not be enough to optimize adherence to a physical activity intervention within the breast cancer population [27]. It is important to note the results in the present study demonstrate that breast cancer survivors have speci c exercise program preferences and could bene t from interventions that are tailored using preferences along with the FITT principles.

Conclusion
The main goal of the current study was to gain a better understanding of program preference changes during an exercise behavior change intervention vs. UC and the associations between preferences and breast cancer survivors' physical activity levels post-intervention. In summary, several preferences changed during the exercise intervention primarily mirroring (or better aligning with) the intervention design with a few, albeit limited, number of preferences being associated with physical activity outcomes.