Participants
A sample size of 116 subjects (58 by study arm) was determined a priori as being necessary to detect an effect size of 0.25, assuming a power of 0.90 and a significance level of 0.05. Sample size calculations were done using G*Power 3.1.3. Promotional materials were disseminated through 1) media: newspaper advertisement, email, and Facebook, 2) personally by recruiters through visits to physicians' offices, support groups, cancer survivorship annual events, and forums, and 3) snowballing.
Inclusion criteria were 50 years of age or above, diagnosed with breast cancer in stages 0 to III according to the American Joint Committee on Cancer TNM Staging System [20], and have completed adjuvant curative therapy at least two months to 5 years before recruitment. It has been found that breast cancer survivors, even after 5 years of treatment completion, experience side effects associated with cancer treatment [21, 22]. Exclusion criteria were based only on safety issues and included contraindications to exercise, such as extreme fatigue, unstable cardiac disease, uncontrolled diabetes, uncontrolled hypertension, coagulopathies, or inability to ambulate independently; and stage IV breast cancer or secondary cancers [23, 24, 25, 26]. Participants were expected to be able to read and write in Spanish and live in Puerto Rico. Participants were screened using the Spanish version of the Rapid Assessment of Physical Activity Questionnaire [27] to exclude those who already met the national standard for exercise (≥ 150 minutes per week of moderate exercise). Participants needed to obtain from their physician a referral for physical therapy to participate, as required by the Physical Therapy Practice Act in Puerto Rico.
Interventions
Seven face-to-face sessions were provided and distributed in six months. The two study groups differed only in terms of exercise intensity for the endurance (aerobic) component. The intervention was based on the recommendations for exercise prescription by the American College of Sports Medicine (ACSM) Roundtable, published in 2010 and updated in 2019 [4, 25], and the American Heart Association [28]. Heart rate (HR) max was calculated using the formula: HR max= 206 – (0.88 x age) [29]. Group intensity was ensured both by the percentage of heart rate max and rate of perceived exertion on the original Borg scale of 6 to 20 [30, 31]. Participants in the low-intensity group exercised at 57-63% of heart rate max and 9-11 at the Borg scale, while participants in the moderate-intensity group exercised at 64-75% of heart rate max and 12-14 at the Borg scale. Participants were provided with a heart rate monitor to exercise at the intensity according to the group assignment (Polar Electro, Lake Success, NY). For deconditioned participants in both groups, the endurance component required multiple shorter periods of exercise (5-10 minutes) with rest intervals, progressing to 30 minutes.
Participants were instructed to walk in their communities for 30 minutes per session; five times a week, to accumulate 150 minutes at their assigned group intensity. Participants performed 7 to 10 minutes of stretching after each workout. Gentle stretching of large muscles was performed, holding each stretch for 20-30 seconds, two repetitions of each stretch. Therapeutic bands were used to exercise large muscles, in addition to core exercises. For participant education in the flexibility and balance components, we used the free National Institute of Aging publication ‘Exercise & Physical Activity: Your Everyday Guide from the National Institute on Aging!’ [33]. We also provided participants with weekly exercise logs to record their exercise duration (minutes) in the endurance component; this information was analyzed for compliance with endurance exercise.
The interventions were guided by the Theory of Planned Behavior [TPB; 17, 18, 2007]. The TPB banks on understanding predictors of behavior change to guide the development of health interventions [33]. Motivational strategies to increase participation, based on the TPB constructs [34] included: education on the benefits of exercising, the structure of an exercise program, and injury prevention; health professionals providing the exercise intervention, including practicing under supervision; follow-up and feedback by health professionals; setting goals; self-monitoring; time-management skills; and problem-solving for adapting the routine when complications arise such as sickness or taking care of others. The implementation of the TPB-based booklet for breast cancer survivors, developed by Vallance et al. [35] aided in integrating these behavioral strategies as it includes homework material for participants to apply the skills and strategies learned in the individualized interventions beyond the structure of the intervention. With the author's permission, this booklet was translated into Spanish and adapted to use culturally specific terminology and images for Puerto Rico, following the guidelines provided by Beaton and colleagues [36, 37]. Table 1 details each session’s content. This study incorporated a methodology to promote treatment fidelity [38, 39, 40, 41, 42]. Each session has a content list that the physical therapist used to check that all content was delivered as planned; modification needs were recorded as well.
[Place Table 1 around here.]
Setting
For study assessments and intervention sessions, participants met with trained research staff at the Medical Sciences Campus, University of Puerto Rico, in San Juan, PR. Participants were also met at a hospital setting in Caguas, PR; this setting was unavailable after COVID-19. For exercising, participants along with their physical therapist co-identified an accessible and safe setting in their community.
Procedures
A randomized controlled trial with two groups was conducted, upon approval of the Institutional Review Board (Protocol Number: A0050116, Medical Sciences Campus, University of Puerto Rico). An intervention manual was developed to operationalize TPB constructs, including motivational behavioral skills consistent with the TPB model [17]. The manual included information on elements of the exercise intervention that needed to be consistent for all breast cancer survivors, such as safety parameters and exercise components, and those that needed to be tailored, such as progression and intensity [43]. Based on the intervention manual, essential elements, such as individualization of the exercise routine, teaching safety parameters, explaining goal setting, and how and when to progress the exercise program, were translated into statements of observable behavior used to construct the checklists for assessment of treatment delivery and treatment receipt [44]. The procedures reported by Vallance et al. [45] were followed for content validity: at least three experts in behavioral change theories and exercise judged the theory incorporation in the manuals of the intervention using the TPB Content Assessment Tool. Manuals were revised based on experts’ feedback.
After screening, eligible participants were contacted by the Study Coordinator for an initial appointment, where informed consent was obtained, and baseline data was collected by a blinded evaluator trained in the standardized procedures for data collection. Randomization was done after participants completed the baseline evaluation and submitted a referral with their physician’s approval to exercise. A computer-generated schema, developed by a statistician who did not have contact with participants, was used to randomize participants; random assignment was implemented by the Study Coordinator, using sealed, security envelopes that were opened after completing baseline evaluation.
After randomization, a physical therapist met with the participant for the home exercise program. Three initial sessions per participant were provided for education in the home program during the first month, plus one intervention session every 4 to 6 weeks for the remaining five months, for a total of seven intervention sessions across the 6-month intervention. Soon after, a 6-month evaluation was conducted. A final follow-up evaluation was conducted 12 months after the baseline evaluation. Safety of interventions, adverse events, serious adverse events, and side effects of the intervention were tracked by the response of the patient (passive surveillance of harms) and petition of providers [active surveillance; 46]. A Data Safety and Monitoring Board (DSMB) met at least once a year to assess project progress and advise on recruitment, retention, safety, and adverse events, among other areas. The DSMB was composed of a researcher in breast cancer, an ethics expert, and a woman who is a breast cancer survivor not enrolled in the study. To promote engagement in the exercise program, follow-ups incorporated the 5A's approach in counseling addressing barriers to conducting the exercise program [47, 48].
Personnel
The intervention was provided by physical therapists trained at either master’s or doctoral level, with a minimum of two years of clinical experience that included exercise planning and implementation. After each intervention, the physical therapist completed a self-assessment checklist on intervention delivery, including intervention dose (i.e., length, content). They also used a checklist to assess participant comprehension and ability to perform the components of the exercise intervention, to ensure that participants understood the intervention. The evaluators who administered the assessment questionnaires were psychologists trained at the master’s or doctoral level, blind to group assignment. The evaluators who performed the six-minute walk test were physical therapists with a minimum of two years of clinical experience. The evaluators were blind to group assignments and completed a self-assessment checklist to assure compliance with essential elements of the evaluation.
Outcome measures
Sociodemographic and clinical data was collected at baseline evaluation. Outcome measures were collected at baseline, six-month evaluation, and 12-month evaluation. The following were the primary outcome measures:
- PROMIS Item Bank v. 1.2 Physical Function, Spanish version, consisting of 20 items, assessing mobility (lower extremity function), dexterity (upper extremity function), axial (neck and back) function, and ability to conduct instrumental activities of daily living. The Spanish scale is highly reliable (Coefficient alpha 0.99) and has one dimension [49].
- PROMIS Item Bank v.1.0/1.1 Global Health, Spanish version consisting of two subdomains: Global Physical Health (internal consistency of 0.81) and Global Mental Health (internal consistency of 0.86) [50]. It has ten items (each item can be examined separately) on physical function, pain, fatigue, emotional distress, social health, and general perceptions of health. The final score can be converted to a quality-of-life EUROQoL score. For both PROMIS measures, its value can be transformed to a T-score and compared to the normal distribution of the general population. A score of 50 is the average for the general population of the U.S. (S.D.=10); a higher score represents higher physical and mental health. A person who has T-scores of 60 on any scale is one standard deviation better (healthier) than the general population.
- Six-minute walk test [51] was used to assess exercise tolerance [52] and fitness, as it is highly correlated with oxygen uptake [53]. This test is correlated to physical function in breast cancer survivors [54]. An increase in walking distance indicates a performance improvement. Data was collected on the total distance walked using a 42-meter hallway, the number of rests, and the duration of rest during the 6 minutes. The rating of perceived exertion was monitored [30, 52]. Heart rate and oxygen saturation were examined using a pulse oximeter (Nellcor Puritan Bennett, Pleasanton, CA).
Secondary outcome measures are physical activity level and adherence.
- Accelerometers were used to evaluate the amount of time spent in physical activity. Accelerometry is a valid measure of physical activity, as compared to oxygen uptake and calorimetry [55, 56], and has been used nationally by the National Health and Nutritional Examination Survey (NHANES) study to document, more precisely than self-reported measures, the intensity and time of physical activity [57]. The procedures established by NHANES [58] were followed. Participants were asked to wear an accelerometer (GT3XE-Plus Triaxial Activity Monitor, Actigraph, Pensacola, FL) under or on top of light clothing, over the non-dominant hip on a removable elastic belt for seven days. Participants were asked to wear the device while awake and to take it off for swimming, bathing, and sleeping. Verbal and written instructions on these procedures were given to the participants. Data from the accelerometers were downloaded and analyzed using the ActiLife 6 software (Actigraph, Pensacola, FL). For accelerometer data analysis, validity was determined if, during a seven-day wear, at least four valid days of data were collected; and ten hours of usage data were registered for a valid day (600 minutes). Participants completed a diary to record activities not registered by the accelerometer such as swimming. This activity’s time and intensity were manually added for validation. Cut points in counts minutes used for the intensities were 0-99 for sedentary, 100-1951 for low; 1952-5724 for moderate; and >5725 for vigorous [60].
- The self-administered, long Spanish version, of the International Physical Activity Questionnaire (IPAQ) was used to evaluate the time spent in habitual physical activities. The IPAQ assesses multiple domains of activity such as leisure, occupation, household, and transport, among others [59]. This questionnaire has proved to be reliable [60] and has been validated with the accelerometer, pedometer, physical activity log, and maximal oxygen uptake [61].
- Body appreciation scale-2 (BAS-2) was administered to assess body image [62]. It has 10 items with individual item responses ranging from 1 (never) to 5 (always); the total test score ranges from 10 to 50. The BAS-2 was adapted for the Puerto Rican Spanish (unpublished data) and showed adequate internal consistency (Cronbach’s alpha was .92). Exploratory factor analysis showed the one-dimensionality of the test. Body appreciation was negatively correlated with higher levels of depressive symptomatology, and a higher body mass index, and positively correlated with higher levels of mental health.
- The Beck Depression Inventory-II (BDI-II) was used to measure symptoms, characteristics, and attitudes related to depression. The Spanish version has proper psychometric properties for the assessment of depression: internal consistency Cronbach alfa of 0.89 and has reported known-group validity [63].
- A TPB questionnaire was administered [64] to assess the constructs of attitude, subjective norms, perceived behavioral control (direct and indirect measures of each construct), and intention to exercise. It consists of 84 items, organized into 13 subscales.
Data analysis
The study sample was described, using absolute and relative frequency for categorical variables (e.g.: age, academic degree, employment, civil status, body mass index, cancer stage, treatments) and with mean, median, standard deviation, and range for continuous variables (e.g.: BECK depression score; BAS-2; accelerometer; six-minute walk; PROMIS Global Mental Health, Global Physical Health, and Physical Functioning, among others). Before performing any association, the distribution of all continuous variables was assessed using the Shapiro-Wilk test. Associations among the group's condition (low intensity vs. moderate intensity) and the characteristics mentioned above were performed using Pearson’s Chi-Square, Fisher Exact Test, t-test, or Mann-Whitney test as appropriate. Associations within groups through time (baseline vs. 6-month vs. 12-month) were done for the same characteristics using Repeated Measures ANOVA (for normally distributed variables), and Friedman’s Test or Skilling’s Mack Test (for non-normally distributed variables). The Bonferroni correction was used, and a new specific p-value was generated to determine a true statistically significant association. Analysis was performed using Stata version 16 [65].