Journal Logo

APPLIED SCIENCES

A Randomized Trial to Increase Physical Activity in Breast Cancer Survivors

ROGERS, LAURA Q.1; HOPKINS-PRICE, PATRICIA1; VICARI, SANDY2; PAMENTER, RICHARD1; COURNEYA, KERRY S.3; MARKWELL, STEPHEN4; VERHULST, STEVEN4; HOELZER, KAREN5; NARITOKU, CATHERINE1; JONES, LINDA6; DUNNINGTON, GARY7; LANZOTTI, VICTOR8; WYNSTRA, JAMES9; SHAH, LISA10; EDSON, BILLIE1; GRAFF, ASHLEIGH1; LOWY, MICHELLE1

Author Information
Medicine & Science in Sports & Exercise: April 2009 - Volume 41 - Issue 4 - p 935-946
doi: 10.1249/MSS.0b013e31818e0e1b
  • Free

Abstract

Adopting a healthy lifestyle after a cancer diagnosis is important for improving health and quality of life (11). Receiving a diagnosis of cancer may motivate an individual to adopt healthier choices, providing an optimal time to initiate lifestyle behavior change interventions (11). Focusing such interventions on physical activity adherence in breast cancer survivors is warranted because physical activity may reduce breast cancer mortality, independent of diet and body mass index (BMI) (23,24). Moreover, physical activity interventions may not be generalizable across cancer types because physical activity barriers, correlates, and preferences may differ on the basis of cancer site and treatment status (9,10,37).

Change in physical activity behavior as the primary study outcome identifies a study as "behavior change" rather than exercise dose or efficacy study. Among the currently available randomized trials of physical activity behavior change studies in cancer survivors, five focused on breast cancer, one on prostate cancer, and four on mixed cancer types (3,4,7,12,13,28,33,34,42,45). It is encouraging that seven of these studies demonstrated significant improvements in physical activity and two of the three without activity increases reported improvements in the exercise stage of change. Despite this initial success, as many as 79% of participants in physical activity behavior change interventions failed to meet public health recommendations (28,33) and only one study has demonstrated improvements in objective measures of physical activity (33).

Behavior change programs can improve effectiveness by basing program elements and techniques on changing the underlying cognitive variables known to be associated with the behavior of interest (19). The social cognitive theory has proven useful in identifying such variables for physical activity behavior in breast cancer survivors (35,40,41). Although self-efficacy is the central construct of the social cognitive theory, other important constructs exist [e.g., expectations of outcomes, importance of the expected outcomes, environment (including social support), etc.] (19). Several prior behavior change studies have used the social cognitive theory but have often focused on self-efficacy rather than comprehensively considering multiple social cognitive theory constructs (7,12,13,34).

Because physical activity behavior is the primary outcome of interest for behavior changes studies, there is a potential for public health and medical professionals to assume that if an intervention increases activity, it will also improve the health of its participants. Unfortunately, inconsistent results suggest that this assumption is not always correct. When previous physical activity behavior change studies in cancer survivors have included health outcomes such as quality of life, functioning, symptoms, and body composition (3,4,7,12,13,33,34,42,45), intervention effects have been inconsistent owing to the differences in study power, activity adherence rates, and exercise program goals. Including health-related outcomes in physical activity behavior change intervention studies validates the health benefits that are assumed to occur with increased activity.

Importantly, several health outcomes particularly relevant for breast cancer survivors have not been reported in physical activity behavior change intervention trials. Specifically, sleep dysfunction may affect more than half of breast cancer survivors and may improve with regular exercise (27). Cognitive impairment, albeit a less prevalent complaint, may occur during chemotherapy for breast cancer with symptoms persisting after treatment completion (44). Hormonal therapy (i.e., aromatase inhibitors and estrogen receptor modulators) can cause vasomotor symptoms and joint pain, which may interfere with regular exercise participation (8,18,25). Although it is unknown whether regular exercise will improve these symptoms in breast cancer survivors, studies in postmenopausal women and individuals with osteoarthritis suggest that such symptoms may improve with regular physical activity (15,30).

Physical activity behavior change interventions are especially germane for breast cancer survivors on hormonal therapy because exercise may counteract the potential deleterious effects of therapy on serum lipids and/or vascular thrombosis risk (25). Moreover, survivors with hormone receptor positive tumors may have the greatest potential for reducing cancer recurrence and mortality with exercise (24). Although several studies have focused on breast cancer survivors in general (3,28,33,34,42), no study has focused primarily on breast cancer survivors receiving hormonal therapy.

Owing to the paucity of data supporting effective physical activity behavior change interventions in breast cancer survivors on hormonal therapy and the inconsistent or not previously reported impact of such interventions on health-related outcomes, we carried out a pilot randomized controlled trial. Our study aim was to determine the feasibility and preliminary effectiveness of a physical activity behavior change intervention for breast cancer survivors receiving hormonal therapy, which attempted to comprehensively address multiple social cognitive theory constructs. Program effectiveness in changing physical activity, fitness, muscle strength, body composition, perceived health, sick days missed from work, quality of life, fatigue, endocrine symptoms, cognitive function, sleep dysfunction, and joint pain were assessed.

METHODS

Setting and Participants

Participants were recruited through newspaper advertising, public service announcements, physician referrals from local academic and private practice oncology offices, and placement of flyers in clinical waiting rooms and beauty salons. The study protocol was approved by the local institutional review board, and informed consent was obtained from all participants before implementing study procedures. Eligibility criteria included English-speaking female breast cancer survivors between the ages of 18 and 70 yr with a diagnosis of stage I, II, or IIIA. Participants were currently taking aromatase inhibitors or selective estrogen receptor modulators and were expected to remain on hormonal therapy for the duration of the study (i.e., at least 8 months). Medical clearance for participation provided by primary care physician or oncologist was required. If the patient had undergone a surgical procedure, enrollment was delayed until at least 8 wk after the procedure. Participants with the following criteria were ineligible: diagnosis of dementia or organic brain syndrome; medical, psychological, or social characteristic that would interfere with the ability to fully participate in program activities and assessments (e.g., psychosis, schizophrenia, etc.); contraindication to participation in a regular physical activity program (e.g., unstable angina, debilitating arthritis pain); breast cancer recurrence or metastatic disease; inability to ambulate; planning to relocate out of the study area during the 8-month study period; and engaged in ≥60 min of vigorous physical activity or ≥150 min of moderate plus vigorous activity per wk during the past month (based on self-report).

Design and Procedures

The study piloted a two-armed, randomized controlled trial comparing an intervention to usual care group. Interested participants were provided an explanation of the study and screened by telephone for eligibility using a structured interview format. If participants were interested and eligible, an appointment was made with the study coordinator for completion of the consent and other related study forms (e.g., HIPAA authorization). After consent was obtained, the participant was given specific instructions related to scheduling and appropriate completion of each study measure: submaximal treadmill test, dynamometer testing, anthropometric measures, accelerometer, diet record, self-administered survey, and dual-energy x-ray absorptiometry. Measurements were obtained at baseline (preintervention) and after 3 months (immediately postintervention). A small monetary incentive was paid after completion of each set of study measures.

Random Assignment

Participants were randomized after completion of all baseline assessments. Randomization was computer generated and kept in sealed envelopes until randomization to prevent bias in group allocation by study personnel.

Physical Activity Behavior Change Intervention

The 12-wk physical activity behavior change intervention was entitled the BEAT Cancer program (i.e., Better Exercise Adherence after Treatment for Cancer) and was based on the social cognitive theory, other physical activity correlates, and program preferences determined from prior needs assessments among breast cancer survivors (e.g., face-to-face counseling from knowledgeable staff, home-based, private setting, walking, moderate intensity, self-efficacy, important exercise benefits, fatigue, time management, social networking, role modeling, exercise enjoyment, and importance of avoiding risk) (38,40). The goal of the intervention, based on current public health recommendations (22), was to gradually increase all participants to 150 min of moderate walking per wk. Intervention participation occurred in four cohorts or "waves" to facilitate the group format.

The intervention consisted of group and individual sessions with Table 1 providing the number of each type of session per week during the 12-wk intervention. Participants attended six discussion group sessions with a clinical psychologist who encouraged social support, provided breast cancer survivor exercise role models, and covered the following topics: journaling, time management, stress management, dealing with exercise barriers, and behavior modification. The specific social cognitive theory constructs addressed by the group sessions included self-efficacy, emotional coping, reciprocal determinism, perceived barriers, outcome expectations, behavioral capability, goal setting, environment, observational learning, and self-control. Participants also attended 12 individual supervised exercise and 3 individual "face-to-face" update counseling sessions with an exercise specialist that tapered to a home-based program by the end of the intervention. The supervised and update sessions were based on exercise prescriptions derived from fitness testing described below and used a structured interview format based on a standardized encounter form. In addition to walking, flexibility exercises and exercise barriers were discussed. The specific social cognitive theory constructs addressed by the individual sessions included self-efficacy, outcome expectations, behavioral capability, perceived barriers, and goal setting with self-monitoring. To further enhance self-monitoring, participants were encouraged to "convert" the minutes spent in physical activity recorded on their weekly exercise logs into "miles" (i.e., 1 min = 2 miles), which were graphed on a map indicating travel across the country to the west coast (i.e., "Race across America"). This activity was intended to provide participants with a feeling of accomplishment and a possible competitive opportunity. The exercise specialists were certified by the American College of Sports Medicine (or certification eligible) with training, supervision, and periodic monitoring for quality control performed by an exercise physiologist with experience in exercise programming for cancer survivors. Participants met with different exercise specialists during the intervention because assignment was made on the basis of scheduling logistics.

T1-25
TABLE 1:
Number of each type of intervention session per week for the Better Exercise Adherence after Treatment for Cancer (BEAT Cancer) program.

Usual Care Group

The usual care group was provided written materials related to physical activity obtained from the American Cancer Society (published pamphlets and Web site resources: www.cancer.org). These materials were considered "usual care" because of their availability to the general public. No specific instructions were given to the usual care group concerning physical activity behavior change. Participants randomized to the usual care group were given the opportunity to receive the intervention at no charge once the postintervention assessments were complete.

Assessment of Primary and Secondary End Points

Feasibility framework

Feasibility was on the basis of the RE-AIM (Reach, Efficacy/effectiveness, Adoption, Implementation, and Maintenance) framework (14). Specifically, "Reach" was assessed by the percent recruitment (or accrual) rate and comparison of participant characteristics to public domain cancer registry data for the study community. Program adherence rates, participant program evaluation, physical activity, health outcomes, and adverse events assessed "Efficacy/effectiveness." The number of oncology clinical offices referring participants assessed "Adoption," whereas process evaluation and cost assessed "Implementation." Finally, "Maintenance" was assessed by attrition rates, barriers to participation (i.e., reasons for refusals and withdrawals), and likelihood that the intervention would be repeated.

Process and program evaluation

Successful implementation of intervention activities as planned was monitored by the investigators, study coordinator, exercise physiologist, and psychologist. A 21-item program evaluation form with primarily open-ended and yes/no questions was administered to intervention participants during the postintervention assessment.

Objective activity monitoring

All participants wore a GT1M accelerometer (ActiGraph, Pensacola, FL) for seven consecutive days; the accelerometer was removed when bathing or asleep. Four valid days were required for an adequate assessment with valid days determined according to published protocols by reconciling data printouts with a record of time in and out of bed completed by the participant (32). The accompanying computer software used a combination of the work energy theorem and Freedson equations to calculate total activity counts, total steps, and minutes of light, moderate, hard, and very hard activity (User's Manual, www.theactigraph.com). Although all outcomes were explored, the primary study end point was total activity counts.

Self-reported leisure time physical activity

The Godin Leisure-Time Exercise Questionnaire, a valid measure of self-reported physical activity behavior (20), was used to assess leisure time activity. Participants were asked to report the average weekly duration and frequency of light, moderate, and vigorous leisure time activity for the past month. Reported duration was multiplied by frequency to determine the minutes per week spent in each of the three intensity levels (light, moderate, or vigorous).

Stage of motivational readiness for physical activity

Although the intervention was based on the social cognitive theory, stage of change [a component of the transtheoretical model (19)] was assessed to describe the characteristics of cancer patients willing to participate in exercise studies (i.e., the "Reach" potential) and ensure similarities between the study groups about readiness for change. All participants were asked to report their physical activity stage of readiness (i.e., stage of change) before learning about the study and postintervention using a previously validated scale (31). Responses were classified into one of five stages: precontemplation (do not exercise regularly and do not intend to do so in the next 6 months), contemplation (do not exercise regularly but intend to do so in the next 6 months), preparation (exercising some but not regularly), action (regular exerciser for less than 6 months), and maintenance (regular exerciser for 6 months or more) (31).

Fitness

A submaximal treadmill test using the Naughton protocol with the end point of 85% of predicted maximal heart rate was used to determine fitness on the basis of a published regression equation estimating total oxygen cost of walking at the treadmill grade and speed achieved (2). The Naughton protocol was chosen because of its lower starting speed and more gradual progression, which is preferred when testing older sedentary participants who may be fatigued and/or who suffer balance difficulties related to breast cancer treatment. Also, maximal oxygen consumption testing is uncomfortable for participants and may increase study drop-outs (15,39).

Muscle strength

A back/leg extensor dynamometer (Takei Back-A Model No. TKK5002, Tokyo, Japan) and handgrip dynamometer (Lafeyette Model No. 78010, Latayette, IN) were used to measure muscle strength. At each assessment period, the participant completed three trials with 1-min rest between trials. The maximum reading (best of the three efforts) provided the absolute strength measure in kilograms.

Body composition and anthropometrics

Body mass index was calculated [weight per height squared (kg·m−2)] using scale and stadiometer. Waist and hip circumference was measured using a nonstretching tape measure with the participant standing with abdomen relaxed and arms at sides. The waist circumference was measured at the smallest circumference of the torso and end of exhalation with no clothing at the waist. Hip circumference was measured at the level of the maximum extension of the buttocks with only undergarments at the hip. At each testing, three measurements were obtained and the results averaged before calculation of the waist-to-hip ratio. To determine percent body fat and bone mineral density, dual-energy x-ray absorptiometry (DXA) was performed using a Lunar Prodigy with Version 5.60.003 software (Lunar GE, Madison, WI).

Patient-rated health-related outcomes

The remaining health-related variables were determined by self-administered survey. Perceived health was assessed by asking participants to rate their general health on a 5-point Likert scale from 1 = poor to 5 = excellent. The participants were asked to report the number of sick days missed from work in the past month using a single, fill-in-the-blank question. Using the sum of 5-point Likert scales, quality of life was measured with the 37-item Functional Assessment of Cancer Therapy-Breast (FACT-B) (5), which includes the subscales of physical well-being (PWB), social well-being (SWB), emotional well-being (EWB), functional well-being (FWB), and additional concerns (AC). PWB, SWB, EWB, and FWB are summed for the FACT-General (FACT-G) with total FACT-B score including FACT-G plus AC. The 13-item FACT-Fatigue (FACT-F) (46), the 19-item FACT-Endocrine Symptoms (FACT-ES) (17), and the 42-item FACT-Cognitive (FACT-Cog) (43) were used to assess fatigue, endocrine symptoms, and cognitive function, respectively. All FACT scales are self-report and have proven validity and reliability (5,17,43,46). Sleep dysfunction was assessed using the Pittsburgh Sleep Quality Index (PSQI) with scoring according to published protocol so that a higher score indicates greater sleep dysfunction (i.e., habitual sleep efficiency, sleep latency, sleep duration, subjective sleep quality, use of sleeping medication, daytime dysfunction, and global score) (6). Owing to limited survey space, the sleep disturbances subscale was not included requiring that the global score be on the basis of the sum of six rather than the usual seven domains. Because the intervention primarily encouraged a lower extremity weight bearing activity (i.e., walking), joint pain, stiffness, and physical function were assessed using the 5-point Likert scale version (i.e., 1 = none to 5 = extreme) of the 24-item Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC), a measure of lower extremity pain and function (36).

Covariates, intervention adherence, and adverse events

With the exception of beta-blocker use and comorbidity score, the demographic characteristics, medical variables, and prior physician exercise counseling information provided in Table 2 were assessed by self-administered survey (multiple choice, yes/no, or fill-in-the-blank items). Owing to the potential influence on the results of the submaximal treadmill test, participants were asked to provide a list of all current medications that were reviewed by the research staff to identify beta-blocker use. The comorbidity score was calculated by summing responses to the Functional Comorbidity Index, which asks participants to report history of 18 different health conditions (1 = yes, 0 = no) (21). Diet intake was assessed with the 3-d diet record (i.e., 1 weekend and 2 weekdays) and was analyzed with Diet Analysis Plus software, version 7.0.1 (www.Thomsonedu.com; Florence, KY). Intervention participation and adverse events were monitored by the investigators, study coordinator, psychologist, and exercise specialists.

T2-25
TABLE 2:
Baseline characteristics of participants overall and by group allocation.

Statistical Analysis

Participants randomized to the intervention group were compared with the usual care group on demographic, medical, diet, physical activity, and other health-related outcomes at baseline using independent t-test, chi-square, or Kruskal-Wallis procedures depending on the nature and distribution of the variable. To achieve the stated study aim, the change for postintervention minus baseline for the intervention versus usual care group was assessed with independent t-test, chi-square, or Kruskal-Wallis procedures depending on variable and distribution characteristics. An intent-to-treat analysis was done, and a P value < 0.05 was set as the level of significance. All differences are provided with calculated effect sizes regardless of statistical significance owing to the pilot nature of the study and the usefulness of the effect sizes for designing future studies assessing physical activity behavior change interventions.

About the missing data, 1 of the 38 participants completing follow-up measures refused to answer four of the endocrine symptom items and two of the cognitive function items on the follow-up survey. Because these were above our a priori thresholds for the minimum number of missing values allowing imputation (e.g., missing values were not imputed if >2 of the endocrine scale items or >1 of the cognitive subscale items were missing), the following three scores could not be calculated for this participant: endocrine symptoms, cognitive function interference with functioning, and the total cognitive function. There were no missing values related to non-survey measures with the exception of missing daily steps on one participant because of accelerometer malfunction. Also noteworthy, DXA results are available on 39 rather than 38 participants because one of the participants not completing the follow-up measures completed the follow-up DXA measurement.

An a priori sample size of 20 participants per study group was based primarily on logistical and budgetary constraints. When the study power for this feasible sample size was assessed, information related to the expected between-group difference for physical activity counts was not available. However, power calculations based on daily kilocalorie expenditure (which is derived from both frequency and intensity of activity similar to activity counts) demonstrated that enrolling 20 intervention participants would provide 81% study power to detect an increase of 200 kcal expended per day using a paired t-test with a significance level of 0.05 in the intervention group. Because 100 min of walking per week is equivalent to 150 to 200 kcal expended daily, we felt that our sample size, although logistically restricted, was adequate for detecting meaningful differences.

RESULTS

Participants were recruited from April 2006 to April 2007. Flow of participants through the study is provided in Figure 1. All of the community and academic oncology clinics involved in the study referred participants for possible enrollment. Of the 119 women responding to a physician referral, study flyer, public service announcement, or newspaper advertisement, 49 (41%) were consented. Because 8 withdrew between consenting and randomization owing to time (n = 7) or to being physically unable (n = 1), 41 (34%) completed baseline testing and were randomized, achieving our a priori sample size estimate of 20 participants per study group. Attrition after randomization for the trial overall was 7% (3/41) with 1 not completing follow-up in the intervention group owing to unrelated illness, 1 not completing follow-up in the usual care group owing to distance, and 1 providing incomplete follow-up data (i.e., DXA only) in the usual care group owing to unrelated illness.

F1-25
FIGURE 1:
Participant recruitment, allocation, and retention by study group.

The only deviation from protocol was the enrollment of a participant who did not initiate hormonal therapy until several weeks into the intervention owing to miscommunication with study staff. Although not a protocol deviation, one participant in the intervention group completed all intervention activities and meetings but did not exercise at home during the final 6 wk of the study because of activity restrictions related to receiving elective cosmetic surgery. Although intent-to-treat analysis was the primary analyses, post hoc supplementary analysis was done excluding both of these participants for the following reasons: 1) initiating hormonal therapy during the trial may unduly influence changes in endocrine symptoms and joint pains and 2) lack of exercise adherence was not a volitional choice of the participant receiving cosmetic surgery. Because a similar pattern of change in the study outcomes was seen with exclusion of these two participants, only the intent-to-treat analyses are reported.

Table 2 provides the baseline characteristics overall and by group assignment. Study participants were primarily white (93%) with a mean age of 53 ± 9 yr. Half (51%) had stage II disease with 27% taking estrogen receptor modulators and 73% taking aromatase inhibitors. An oncologist had advised 39% of the participants to exercise, and most (80%) were in the contemplation or preparation exercise stage of change. As described in Table 2, the two study groups did not significantly differ in age, race, education, income, medical factors, prior physician counseling about exercise, and readiness for physical activity behavior change. Because only two participants dropped out of the study after randomization, comparison with those who did not drop out was not made.

Process evaluation and adherence

All components of the intervention were successfully implemented with the exception of a group "competition" (or "Race across America" activity) involving conversion of activity minutes to miles for participants to track progress during the group sessions, which many participants did not find helpful. About adherence, the intervention participants completed 100% (252/252) of the individual exercise sessions, 95% (60/63) of the individual update sessions, and 98% (123/126) of the group sessions for an overall 99% adherence to all possible intervention sessions (435/441). Of the 63 individual update sessions with the exercise specialists, 4 (6%) were administered by telephone rather than face-to-face owing to logistical reasons and participant preference.

Of the 21 intervention participants, 19 (90%) completed a program evaluation with 89% feeling that the group sessions were helpful, 79% stating the number of sessions was appropriate, and 84% reporting that the behavioral modification plan was helpful. About the individual exercise sessions, 100% felt the exercise specialists were knowledgeable and helpful with 95% feeling that the sessions were tailored to their needs. When asked if their exercise behavior had changed during the program, 89% indicated "yes" with 11% being uncertain. When asked which part of the program was most helpful regarding their exercise change, the following were listed: interactions with exercise specialists (47%), experiencing the beneficial effects of exercise (16%), "everything" (11%), keeping the exercise log (11%), developing a routine (11%), and beginning slowly (5%). Most (89%) would recommend the BEAT Cancer program to others with the program ranked as excellent by 21% and very good by 37%. Participants who rated the program as good (37%) or fair (5%) represented the full range of age, education, and income; however, these individuals recommended that the program be revised to include strength training, diet counseling (for a more "complete" program), and a variety of exercise options.

Physical activity, fitness, and muscle strength

Table 3 provides the change in physical activity, fitness, muscle strength, and body composition from baseline to postintervention for the usual care versus intervention group. Significant improvements were noted in total physical activity counts for the intervention versus usual care group (+51,744 vs −20,359; mean difference = 72,103; 95% confidence interval (CI) = 25,383-119,000; effect size (d) = 1.02; P = 0.004). Although not statistically significant, large effect size increases were noted in moderate and vigorous (i.e., hard + very hard) activity minutes. A small nonsignificant effect size increase in self-reported moderate + vigorous activity was noted (79.4 vs 62.8; mean difference = 16.6; 95% CI = −53.1 to 86.2; d = 0.16; P = 0.63). Consistent with the physical activity improvements, a medium to a large effect size increase was noted for stage of change (+1.45 vs +0.5; mean difference = 0.95; 95% CI = 0.75-1.83; d = 0.71; P = 0.034).

T3-25
TABLE 3:
Physical activity, fitness, muscle strength, and body composition at baseline and postintervention for usual care versus group receiving intervention.

Accelerometer data were classified into meeting or not meeting current public health recommendations (i.e., ≥150 min·wk−1 of moderate or vigorous activity). From baseline to follow-up, the proportion meeting recommendations in the usual care group increased from 28% to 33% (McNemar, P = 0.65) with an increase from 35% to 60% in the intervention group (McNemar, P = 0.06).

A nonsignificant but moderate-to-large effect size increase in fitness (4.9 vs 2.0 mL·kg−1·min−1; mean difference = 2.9; 95% CI = −0.1 to 5.8; d = 0.64; P = 0.058) was noted for the intervention versus usual care group. Although right handgrip did not significantly change, increases were noted in the intervention group for left handgrip (0.1 vs −1.9; mean difference = 2.0; 95% CI = 0.3-3.9; d = 0.76; P = 0.03) and back/leg extensor strength (+8.2 vs −4.1 kg; mean difference = 12.3; 95% CI = 0.4-15.9; d = 0.81; P = 0.017).

About body composition, no difference was noted for BMI or percent body fat. However, there was a significant decline in the waist-to-hip ratio for the intervention versus the usual care group postintervention (−0.03 vs 0.02; mean difference = −0.05; 95% CI = −0.08 to −0.01; d = −0.77; P = 0.02). Also, no difference was noted for femoral neck and lumbar (i.e., L2-L4) bone mineral density (i.e., −0.01 change for each measure in both groups). Relevant to body composition, the intervention group demonstrated a nonsignificant decline in daily kilocalorie intake when compared to the usual care group (i.e., −302 vs −206; mean difference = −95; 95% CI = −449 to 259; d = −0.18; P = 0.59).

Patient-rated outcomes

No significant difference in mean change was noted for perceived health (usual care = 0.17 vs intervention = 0.15, P = 0.94) or sick days missed from work (usual care = −0.20 vs intervention = 0.20, P = 0.80). Also, no difference for the usual care versus intervention group was noted for fatigue (−4.2 vs −1.9, P = 0.53) or endocrine symptoms (−2.9 vs −0.9, P = 0.45).

Table 4 provides the change in quality of life from baseline to postintervention for the usual care versus intervention group. Social well-being demonstrated significantly greater improvements in the intervention group versus usual care group (i.e., 0.8 vs −1.2, P = 0.03).

T4-25
TABLE 4:
Quality of life at baseline and postintervention for usual care versus group receiving intervention.

Table 5 provides the changes in cognitive function, sleep dysfunction, and joint symptoms from baseline to postintervention for the usual care versus intervention group. The effect sizes were variable in size and direction depending on the component of cognitive function or sleep dysfunction measured with no significant differences note. For joint symptoms, no improvement was noted with the intervention but, unexpectedly, joint stiffness was greater among the intervention versus usual care group (i.e., 0.4 vs −0.72, P = 0.04).

T5-25
TABLE 5:
Cognitive function, sleep dysfunction, and joint symptoms at baseline and postintervention for usual care versus group receiving intervention.

No adverse events related to the intervention or other study procedures occurred. The following nonserious, nonrelated events were recorded: wheezing requiring physician evaluation for asthma, cholinergic urticaria, herpes zoster, sinusitis, back pain related to falling, and elective cosmetic reconstructive surgery.

DISCUSSION

Our randomized controlled trial is unique in its focus on breast cancer survivors receiving hormonal therapy and comprehensive assessment of health-related outcomes. Our physical activity behavior change intervention (BEAT Cancer program) was feasible and well-accepted by the participants. The BEAT Cancer program significantly improved physical activity counts measured by accelerometer, muscle strength, and social well-being with large effect size increases in moderate activity, vigorous activity, and fitness that were not statistically significant owing to limited study power. Although no improvements in BMI and percent body fat were noted, a significant decrease in waist-to-hip ratio was noted in the intervention group. Unexpectedly, the intervention group reported a significant increase in joint stiffness. The intervention did not significantly change perceived health, sick days missed from work, fatigue, endocrine symptoms, cognitive function (based on self-report), or perceived sleep dysfunction.

Our intervention was highly feasible from the standpoint of efficacy and implementation. Most intervention sessions were successfully completed, most participants felt the intervention was worthwhile and helpful, and no related adverse events were reported. It is noteworthy that our "competition" among the participants was the only intervention component not well received suggesting that this activity be dropped from the intervention. Because such an activity has proven helpful in other populations (e.g., physicians in training [39]), this occurrence reinforces the importance of considering the specific preferences of the target population when designing interventions. It is also interesting that the individual sessions with the exercise specialists were felt to be the most helpful intervention component. This is consistent with prior reports that face-to-face counseling is the preferred physical activity delivery channel among breast cancer patients (38) but is contrary to our expectation that the group sessions would provide the most assistance in changing behavior by emphasizing issues such as behavioral modification.

The high intervention completion rates suggest that our intervention has short-term "maintenance" potential within the RE-AIM framework with the fact that 89% would recommend the program to others suggesting long-term "maintenance" of program dissemination, as well. Although these results are encouraging, the intervention required significant staff time to ensure that each session was scheduled and completed according to protocol. Furthermore, distance, lack of time, and the development of unrelated combordities were the most frequent reasons for refusing to participate in or complete the intervention and study follow-up. Although a multidisciplinary intensive intervention may be effective in changing behavior, its complexity carries the risk of refusal to participate owing to these identified barriers. Future studies to determine the most effective components of the intervention to simplify implementation, reduce cost, and enhance participation are warranted. Such studies should encourage oncologists to recommend the study to patients, a recruitment strategy we found effective because all oncology clinics referred participants for possible enrollment demonstrating broad "adoption" by local oncologists.

Importantly, our study is among only a few behavior change studies in cancer survivors to use an objective measure of physical activity. Because significant improvements in self-reported physical activity may occur without significant increases in objectively measured physical activity (34,42), it is unfortunate that only three of the studies have reported objectively measured physical activity (33,34,42). We also demonstrated a greater between-group difference for self-report versus the accelerometer (i.e., 16.6 min of moderate + vigorous for self-report vs 9.2 min of moderate + vigorous for accelerometer), further supporting the importance of using an objective measure. However, the effect size was smaller for self-report physical activity in our study owing to the relatively large SD of the between-group difference (i.e., 16.6 ± 105.7). The small percentage of individuals in the action or maintenance exercise stage of change at baseline despite the study participant criteria may be related to participants incorrectly considering occupational rather than leisure exercise when answering the stage of change items or initiating exercise between the screening interview and completion of baseline study measures. Fortunately, there was no difference between the study groups concerning stage of change, and the inadvertent inclusion of more active individuals at baseline would bias the results toward a null finding, further emphasizing the potential effectiveness of our intervention.

It is notable that significant improvements in objective physical activity were found with our intervention because only one of the three prior behavior change studies assessing objective physical activity has demonstrated significant improvements in the intervention group (33). Furthermore, the large effect size increase in fitness is important because only two of the prior four behavior change studies assessing fitness have demonstrated improvements (3,34). Despite significant improvement in activity counts, the absolute between-group differences for moderate and vigorous activity were below the recommended 150 min·wk−1 of moderate activity with only 60% of the individuals receiving the intervention meeting current recommendations based on objective monitoring. This suggests that future interventions should focus on assisting participants in adequately assessing and achieving at least moderate intensity activity.

Several reasons may explain the counterintuitive greater increase in daily step count in the usual care group, a result noted in a prior behavior change study as well (42). Because activity counts are based on intensity and frequency of movement, a shift into higher intensity activity by participants in the intervention group could contribute to an increase in activity counts without a comparable increase in daily steps. In addition, the usual care group may have increased their activity because of a readiness to change without increasing their intensity contributing to an increase in steps without a comparable increase in activity counts. Finally, some participants in the intervention group may have transitioned to exercise types not measured in a vertical plane (as with the step function) during the home-based portion of the program but exercise logs indicated that all were participating in walking with the exception of two who were using the elliptical machine.

About study limitations, physical activity and social well-being improvements may have been due to study staff attention and not specifically related to intervention activities because study staff contact time was not similar for the intervention and usual care group. Therefore, definitive conclusions related to social well-being cannot be made without confirmation in future studies providing equal staff attention to both the usual care and intervention groups. Although study outcomes may have been influenced by different exercise specialists leading the encounters, the sessions were standardized by consistent training, periodic quality control monitoring, and use of structured interview format. Moreover, the fact that the intervention was effective despite using different individuals suggests a greater intervention translational potential. Also, exercise specialists performing the treadmill and muscle strength testing were not blinded to participant group allocation owing to staffing availability. Although the exercise specialists were carefully trained and monitored for objective, standardized administration of the protocol, measurement bias may have occurred if participants in the intervention group were provided additional encouragement, enhancing treadmill or dynamometer test performance. Although blinding of the testing personnel should not have influenced the primary outcome measure (i.e., accelerometer measurement) or the remaining study measures, the inability to blind participants to group allocation is an inherent limitation of exercise studies, which may bias participant responses and behaviors because of their awareness of the study purpose and desire to please the study staff. Also, exercise logs were used as a part of the intervention materials but were not used concurrent with the accelerometer to identify activities that are not measured by the accelerometer (e.g., swimming, bicycling). Nevertheless, the two intervention participants logging activities other than walking reported use of an elliptical machine, an activity which can be detected by the accelerometer. Finally, our sample size precluded analyses of the potential moderating effects of demographic and medical factors (e.g., age).

Although our recruitment rate of 34% is similar to rates reported for randomized clinical trials in cancer survivors (16), this percentage may reduce the generalizability (or "Reach") of our study. For example, on the basis of public domain data from the state cancer registry, our mean age was slightly lower than the mean age of 61 yr for all breast cancer survivors in the state, suggesting that our program may be more attractive to the younger survivor. In contrast, it is encouraging that the percent minority representation is similar to that of the breast cancer survivor population in the study community. Also relevant to "Reach," the low percentage of participants in the precontemplation stage of change suggests that different interventions will be required to attract survivors in this stage.

Because body composition influences breast cancer risk before and after diagnosis, the improvements in waist-to-hip ratio may be clinically important (1). The lack of change in percent body fat or BMI may be caused by the short follow-up period and the intervention focus on physical activity rather than weight loss. Our findings are consistent with a recent review suggesting that physical activity interventions may alter body composition rather than body weight (26). The significant reductions in waist-to-hip ratio are especially important because abdominal obesity is associated with greater risk of breast cancer recurrence and mortality (1). Although conclusions about whether the intervention could possibly reduce cancer recurrence cannot be made, the waist-to-hip ratio changes may serve as a biomarker suggesting that the increases in physical activity are sufficient for achieving important health benefits.

Effects on quality of life and fatigue have been inconsistent in other randomized exercise behavior change studies with improvements noted in three of the seven studies assessing quality of life (3,42,45) and in two of three studies assessing fatigue (34,42). The lack of effect in our study may have been related to relative high quality of life and low fatigue scores at baseline causing a ceiling or floor effect (29). Because the greatest improvements in fatigue and quality of life can be expected immediately after treatment, the large variation in cancer stage and time since treatment may have prevented the study's ability to detect differences. Future studies should consider focusing inclusion criteria or enrolling an adequate sample size to assess potential moderation of the intervention effect on the basis of time since treatment or cancer stage. Furthermore, a single dimension rather than a multidimensional fatigue assessment was used precluding the assessment of differential effects on fatigue dimensions that might occur. Nevertheless, the increase in social well-being suggests that future behavior change trials should assess whether intervention effects vary for the individual aspects of quality of life.

As with fatigue, a floor effect may have contributed to the lack of improvement seen with endocrine symptoms, cognitive function, and sleep dysfunction. Also, our feasibility study was not powered to assess the potential moderating effect of chemotherapy on cognitive function, which may have been important because cognitive impairment is more often associated with chemotherapy. Nevertheless, our results suggest that future studies, which include these outcomes, should enroll larger sample sizes allowing adequate study power and stratification by treatment type. Such studies, when appropriate, should assess the individual construct domains because of the variability in intervention effect that may occur and should include all seven sleep domains to allow comparison of the absolute global score to other reports. Moreover, objective measures of cognitive function and sleep dysfunction may need to be included because our assessment was based on self-report alone.

The increase in joint stiffness in the intervention group compared with the usual care group was surprising because exercise is a proven nonpharmacologic therapy for arthritis (30). Joint stiffness is not a major adverse event that is generally assessed or reported in physical activity intervention studies. Also, most individuals who are older (e.g., our study population had a mean age of 53 ± 9 yr) consider mild joint symptoms as "normal" and may not report these unless specifically asked about them. Nevertheless, no previous exercise trial has monitored joint stiffness in breast cancer survivors receiving hormone therapy and possibly exercise may exacerbate joint stiffness that results from aromatase inhibitors. Although confirmation of this finding in larger studies is needed, future interventions should monitor this possible adverse event.

Additional research is needed to confirm our intervention effectiveness (short and long term), evaluate dissemination potential, and determine the most cost-effective components. Future interventions, including a refinement of the BEAT Cancer program, should emphasize achieving adequate activity intensity with the provision of personal heart rate monitors and/or rigorous training in accurately rating perceived exertion. Efforts to decrease joint complaints by emphasizing adequate shoe support and the stretching/flexibility exercises prescribed before and after each exercise session are also warranted. Future studies assessing the effectiveness of physical activity behavior change interventions should include an objective physical activity assessment whenever feasible and confirm health benefits by measuring relevant health outcomes. Furthermore, identifying possible differential effectiveness of the group versus individual intervention components will advance current understanding of the most cost-effective modalities for enhancing physical activity behavior change and maintenance.

This project was supported by the following grants: Southern Illinois University School of Medicine Excellence in Academic Medicine Award (No. E200634), Brooks Medical Research Fund, and Memorial Medical Center Foundation Regional Cancer Center.

The authors did not have any relationships to disclose that would cause a conflict of interest. The results of the present study do not constitute endorsement by ACSM.

REFERENCES

1. Abrahamson PE, Gammon MD, Lund MJ, et al. General and abdominal obesity and survival among young women with breast cancer. Cancer Epidemiol Biomarkers Prev. 2006;15(10):1871-7.
2. Balady G, Berra KA, Golding LA, et al. ACSM's Guidelines for Exercise Testing and Prescription. 6th ed. Philadelphia (PA): Lippencott Williams & Wilkins; 2000. p. 303.
3. Basen-Engquist K, Taylor CL, Rosenblum C, et al. Randomized pilot test of a lifestyle physical activity intervention for breast cancer survivors. Patient Educ Couns. 2006;64(1-3):225-34.
4. Bennett JA, Lyons KS, Winters-Stone K, Nail LM, Scherer J. Motivational interviewing to increase physical activity in long-term cancer survivors: a randomized controlled trial. Nurs Res. 2007;56(1):18-27.
5. Brady MJ, Cella DF, Mo F, et al. Reliability and validity of the Functional Assessment of Cancer Therapy-Breast quality-of-life instrument. J Clin Oncol. 1997;15(3):974-86.
6. Buysse DJ, Reynolds CF 3rd, Monk TH, Berman SR, Kupfer DJ. The Pittsburgh Sleep Quality Index: a new instrument for psychiatric practice and research. Psychiatry Res. 1989;28(2):193-213.
7. Carmack Taylor CL, Demoor C, Smith MA, et al. Active for Life After Cancer: a randomized trial examining a lifestyle physical activity program for prostate cancer patients. Psychooncology. 2006;15(10):847-62.
8. Cella D, Fallowfield LJ. Recognition and management of treatment-related side effects for breast cancer patients receiving adjuvant endocrine therapy. Breast Cancer Res Treat. 2008;107(2):167-80.
9. Courneya KS, Blanchard CM, Laing DM. Exercise adherence in breast cancer survivors training for a dragon boat race competition: a preliminary investigation. Psychooncology. 2001;10(5):444-52.
10. Courneya KS, Friedenreich CM. Utility of the theory of planned behavior for understanding exercise during breast cancer treatment. Psychooncology. 1999;8(2):112-22.
11. Demark-Wahnefried W, Aziz NM, Rowland JH, Pinto BM. Riding the crest of the teachable moment: promoting long-term health after the diagnosis of cancer. J Clin Oncol. 2005;23(24):5814-30.
12. Demark-Wahnefried W, Clipp EC, Lipkus IM, et al. Main outcomes of the FRESH START trial: a sequentially tailored, diet and exercise mailed print intervention among breast and prostate cancer survivors. J Clin Oncol. 2007;25(19):2709-18.
13. Demark-Wahnefried W, Clipp EC, Morey MC, et al. Lifestyle intervention development study to improve physical function in older adults with cancer: outcomes from Project LEAD. J Clin Oncol. 2006;24(21):3465-73.
14. Dzewaltowski DA, Glasgow RE, Klesges LM, Estabrooks PA, Brock E. RE-AIM: evidence-based standards and a Web resource to improve translation of research into practice. Ann Behav Med. 2004;28(2):75-80.
15. Elavsky S, McAuley E. Physical activity and mental health outcomes during menopause: a randomized controlled trial. Ann Behav Med. 2007;33(2):132-42.
16. Ellis PM, Butow PN, Tattersall MH, Dunn SM, Houssami N. Randomized clinical trials in oncology: understanding and attitudes predict willingness to participate. J Clin Oncol. 2001;19(15):3554-61.
17. Fallowfield LJ, Leaity SK, Howell A, Benson S, Cella D. Assessment of quality of life in women undergoing hormonal therapy for breast cancer: validation of an endocrine symptom subscale for the FACT-B. Breast Cancer Res Treat. 1999;55(2):189-99.
18. Felson DT, Cummings SR. Aromatase inhibitors and the syndrome of arthralgias with estrogen deprivation. Arthritis Rheum. 2005;52(9):2594-8.
19. Glanz K, Lewis FM, Rimer BK. Health Behavior and Health Education, Theory, Research, and Practice. San Francisco (CA): John Wiley & Sons, Inc; 1997. p. 60-3, 153-78.
20. Godin G, Shephard RJ. A simple method to assess exercise behavior in the community. Can J Appl Sport Sci. 1985;10(3):141-6.
21. Groll DL, To T, Bombardier C, Wright JG. The development of a comorbidity index with physical function as the outcome. J Clin Epidemiol. 2005;58(6):595-602.
22. Haskell WL, Lee IM, Pate RR, et al. Physical activity and public health: updated recommendation for adults from the American College of Sports Medicine and the American Heart Association. Circulation. 2007;116(9):1081-93.
23. Holick CN, Newcomb PA, Trentham-Dietz A, et al. Physical activity and survival after diagnosis of invasive breast cancer. Cancer Epidemiol Biomarkers Prev. 2008;17(2):379-86.
24. Holmes MD, Chen WY, Feskanich D, Kroenke CH, Colditz GA. Physical activity and survival after breast cancer diagnosis. JAMA. 2005;293(20):2479-86.
25. Howell A, Wardley AM. Overview of the impact of conventional systemic therapies on breast cancer. Endocr Relat Cancer. 2005;12(Suppl 1):S9-S16.
26. Ingram C, Courneya KS, Kingston D. The effects of exercise on body weight and composition in breast cancer survivors: an integrative systematic review. Oncol Nurs Forum. 2006;33(5):937-47; quiz 948-50.
27. Irwin ML, Ainsworth BE. Physical activity interventions following cancer diagnosis: methodologic challenges to delivery and assessment. Cancer Invest. 2004;22(1):30-50.
28. Jones LW, Courneya KS, Fairey AS, Mackey JR. Effects of an oncologist's recommendation to exercise on self-reported exercise behavior in newly diagnosed breast cancer survivors: a single-blind, randomized controlled trial. Ann Behav Med. 2004;28(2):105-13.
29. Linden W, Satin JR. Avoidable pitfalls in behavioral medicine outcome research. Ann Behav Med. 2007;33(2):143-7.
30. Mangani I, Cesari M, Kritchevsky SB, et al. Physical exercise and comorbidity. Results from the Fitness and Arthritis in Seniors Trial (FAST). Aging Clin Exp Res. 2006;18(5):374-80.
31. Marcus BH, Selby VC, Niaura RS, Rossi JS. Self-efficacy and the stages of exercise behavior change. Res Q Exerc Sport. 1992;63(1):60-6.
32. Masse LC, Fuemmeler BF, Anderson CB, et al. Accelerometer data reduction: a comparison of four reduction algorithms on select outcome variables. Med Sci Sports Exerc. 2005;37(11 suppl):S544-54.
33. Matthews CE, Wilcox S, Hanby CL, et al. Evaluation of a 12-week home-based walking intervention for breast cancer survivors. Support Care Cancer. 2007;15(2):203-11.
34. Pinto BM, Frierson GM, Rabin C, Trunzo JJ, Marcus BH. Home-based physical activity intervention for breast cancer patients. J Clin Oncol. 2005;23(15):3577-87.
35. Pinto BM, Maruyama NC, Clark MM, Cruess DG, Park E, Roberts M. Motivation to modify lifestyle risk behaviors in women treated for breast cancer. Mayo Clin Proc. 2002;77(2):122-9.
36. Rogers JC, Irrang JJ. Measures of adult lower extremity function. Arthritis Rheum. 2003;49:S67-84.
37. Rogers LQ, Courneya KS, Robbins KT, et al. Physical activity correlates and barriers in head and neck cancer patients. Support Care Cancer. 2008;16(1):19-27.
38. Rogers LQ, Courneya KS, Verhulst S, Markwell S, McAuley E. Factors associated with exercise counseling and program preferences among breast cancer survivors. J Phys Act Health. 2008;5(5):688-705.
39. Rogers LQ, Gutin B, Humphries MC, et al. A physician fitness program: enhancing the physician as an "exercise" role model for patients. Teach Learn Med. 2005;17(1):27-35.
40. Rogers LQ, McAuley E, Courneya KS, Verhulst SJ. Correlates of physical activity self-efficacy among breast cancer survivors. Am J Health Behav. 2008;32(6):594-603.
41. Rogers LQ, Shah P, Dunnington G, et al. Social cognitive theory and physical activity during breast cancer treatment. Oncol Nurs Forum. 2005;32(4):807-15.
42. Vallance JK, Courneya KS, Plotnikoff RC, Yasui Y, Mackey JR. Effects of print materials and step pedometers on physical activity and quality of life in breast cancer survivors: a randomized controlled trial. J Clin Oncol. 2007;25(17):2352-9.
43. Vardy J, Wong K, Yi QL, et al. Assessing cognitive function in cancer patients. Support Care Cancer. 2006;14(11):1111-8.
44. Wefel JS, Lenzi R, Theriault RL, Davis RN, Meyers CA. The cognitive sequelae of standard-dose adjuvant chemotherapy in women with breast carcinoma: results of a prospective, randomized, longitudinal trial. Cancer. 2004;100(11):2292-9.
45. Wilson RW, Taliaferro LA, Jacobsen PB. Pilot study of a self-administered stress management and exercise intervention during chemotherapy for cancer. Support Care Cancer. 2006;14(9):928-35.
46. Yellen SB, Cella DF, Webster K, Blendowski C, Kaplan E. Measuring fatigue and other anemia-related symptoms with the Functional Assessment of Cancer Therapy (FACT) measurement system. J Pain Symptom Manage. 1997;13(2):63-74.
Keywords:

BODY COMPOSITION; EXERCISE; INTERVENTION; ONCOLOGY; PREVENTION; SURVIVORSHIP

©2009The American College of Sports Medicine