¡Mi Vida Saludable! A Randomized, Controlled, 2×2 Factorial Trial of a Diet and Physical Activity Intervention among Latina Breast Cancer Survivors: Study Design and Methods

Abstract

Background:

Most Latina breast cancer survivors do not meet diet and physical activity (PA) guidelines for cancer survivors and effective lifestyle interventions to adopt and maintain these recommendations are limited, especially among underserved populations. Here we describe the design, methods and enrollment of a 2×2 factorial-designed trial testing the separate effects of the ¡Mi Vida Saludable! (My Healthy Life!) intervention program on changes in diet and PA behaviors among Latina breast cancer survivors.

Methods:

Latinas with a history of stage 0-III breast cancer, no evidence of recurrent/metastatic disease, and >90 days post-treatment were primarily identified via cancer registries and physician referral. Participants were randomized to four arms: 1) 4 weeks of in-person group sessions plus 11 months of eHealth communication, 2) in-person group sessions alone, 3) eHealth alone, or 4) control. All participants received a Fitbit to self-monitor PA. Assessments at baseline, 6 and 12 months include diet, PA, anthropometrics, predictors and mediators of behavior change, psychosocial and quality of life outcomes, and blood draw.

Results:

Of 884 women screened between January 2016 and September 2018, 27% were eligible. Primary reasons for ineligibility included not being willing/able to participate due to work/life responsibilities, health reasons, or transportation. Of 241 eligible women, 167 completed baseline assessment and enrolled.

Conclusions:

We successfully enrolled a diverse group of breast cancer survivors representing more than 15 Latin American nationalities to a diet and physical activity trial. If effective, the ¡Mi Vida Saludable! program can be implemented by community groups and medical centers.

Trial registration:

ClinicalTrials.gov, NCT02780271, registered May 2016.

1. Introduction

As breast cancer survival rates have increased, there is a large and growing population of breast cancer survivors in the US.[1] Despite improved cancer treatments, many breast cancer survivors suffer from chronic diseases with modifiable risk factors, such as obesity, diabetes and cardiometabolic diseases,[2–10] which can reduce breast cancer survival.[2, 4, 11–16] An estimated 80% of Latinas in the US are overweight/obese, which is associated with poorer breast cancer outcomes and overall survival.[17] Furthermore, survival rates disproportionally affect Latina/Hispanic women who are 20% more likely to die of breast cancer when compared to non-Latina/Hispanic white women, after adjusting for stage at diagnosis and access to medical care.[18–21]

To reduce breast cancer recurrence and improve overall survival, breast cancer survivors are recommended to achieve and maintain a healthy weight, eat a high-quality diet, and engage in regular physical activity (PA).[22–23] Modifiable health behaviors such as consuming a healthy diet and engaging in physical activity are associated with lower risk of breast cancer recurrence.[24–28] In observational studies of post-diagnosis diet and breast cancer survival, a healthy diet (i.e., rich in fruits, vegetables and dietary fiber) compared to an unhealthy diet (i.e., high in red/processed meats, refined grains, and alcohol) was associated with lower breast cancer recurrence.[27, 29–30] Data from the Women’s Health Initiative Dietary Modification randomized, controlled trial also reported lower mortality after breast cancer in women randomized to the low fat dietary pattern.[31] Similar health benefits were reported for physical activity; breast cancer survivors engaging in high compared to low levels of physical activity have lower cancer-specific mortality.[24, 32–34]

Based on these recurrence and survival benefits, breast cancer survivors are recommended to consume at least 5 servings of fruits and vegetables per day, decrease intake of energy-dense foods and beverages, and engage in at least 150 minutes per week of moderate-to-vigorous physical activity (MVPA).[35–36] However, most breast cancer survivors do not meet these recommendations.[37–39] Furthermore, lifestyle interventions focused on increasing fruit and vegetable intake and physical activity are limited among Latina breast cancer survivors, especially among underserved populations.[19, 40–43] Of the few studies that have addressed diet and physical activity behavior change in Latina breast cancer survivors, most have been small in size, and participants have been primarily of Mexican descent.[44–45]

We previously conducted a pilot randomized controlled trial of an in-person 9-session intervention with nutrition education, hands-on skills-building cooking and food shopping field trips to target diet behavior among 70 Latina breast cancer survivors.[46–48] At 3 months, compared to controls, participants in the intervention arm reported higher intake of fruits and vegetables, which persisted at 6 and 12 months. Though these results were promising, further investigation into the most effective and cost-effective way to achieve and maintain diet and physical activity behavior change over time was warranted.

As an extension of our work, we designed a shorter, standardized program to achieve and maintain diet and physical activity behavior change consistent with current guidelines by the American Institute of Cancer Research and the American Cancer Society. The ¡Mi Vida Saludable! (My Healthy Life!) study is a 12-month randomized, controlled, 2×2 factorial-designed trial testing the separate effects of in-person and eHealth components of the program on change in diet and physical activity behaviors among Latina breast cancer survivors. Here we describe the study protocol and provide data on enrollment and baseline characteristics.

2. Materials and methods

2.1. Overview

The ¡Mi Vida Saludable! study is a 12-month randomized, controlled, 2×2 factorial-designed trial testing the separate effects of an in-person and eHealth components of the program on change in diet and physical activity behaviors among Latina breast cancer survivors who have completed breast cancer treatment. The ¡Mi Vida Saludable! intervention program includes 4 weeks of hands-on in-person diet and physical activity group sessions and 11-months of eHealth communication. Eligible participants are randomized to one of four study arms: 1) four weekly in-person group sessions plus 11 months of eHealth communication, 2) in-person group sessions alone, 3) eHealth alone, or 4) control. Randomization is stratified by preferred language (English or Spanish) and current use of endocrine therapy (yes or no).

All participants receive written materials summarizing national guidelines by the American Institute of Cancer Research (AICR) and the American Cancer Society (ACS) on diet and physical activity for cancer survivors[35–36], a brief verbal overview of the guidelines by study staff, and a wearable physical activity tracker (Fitbit Zip) to self-monitor physical activity during the 12 month study period. Participants in the in-person group sessions receive 4 weekly sessions with classroom education on diet and physical activity for cancer survivors, hands-on skills-building cooking, physical activity sessions and field trips to local, accessible food markets. Participants who are randomized to the eHealth communication receive 11 months of weekly motivational and facilitating action text messages on diet and physical activity for cancer survivors, newsletters via emails twice per month and access to an interactive nutrition website tailored to cancer survivors (cookforyourlife.org). All participants are encouraged to increase their daily intake of fruits and vegetables, reduce intake of energy-dense foods and beverages, and increase minutes per week of MVPA.

Baseline data collection includes 1) in-person clinic visits to collect self-reported questionnaire data on demographics, acculturation, medical history, predictors and mediators of diet and physical activity behavior change, psychosocial and quality of life outcomes, and technology use, anthropometric measures, and an optional fasting blood draw; and 2) telephone-based dietary assessments (three 24-hour dietary recalls). Data collection is repeated at 6 and 12 months. Participants are contacted monthly via telephone calls to collect brief data on Fitbit use and syncing, and to help troubleshoot any issues with the device, as needed. These monthly calls also serve as retention support. The study has been approved by the Columbia University Medical Center (CUMC), Teachers College, and Fred Hutchinson Cancer Research Center Institutional Review Boards (ClinicalTrials.gov NCT02780271). All participants provide written informed consent.

2.2. Specific aims

The primary aim is to test the separate effects of the ¡Mi Vida Saludable! intervention program components from baseline to 12 months on changes in daily servings of fruits and vegetables and total energy density. Secondary aims are to 1) examine the effects of the intervention on changes in physical activity (as assessed via minutes per week of MVPA), and measured anthropometrics; and 2) examine predictors (e.g. demographics, acculturation, locus of control, stressful life events) and mediators (e.g. stage of change, self-efficacy, social support, procedural knowledge and cooking/nutrition skills) of diet and physical activity behavior change. Exploratory analyses will 1) examine the synergistic effects of the intervention program components on changes in daily servings of fruits and vegetables, energy density, physical activity, and measured anthropometrics; 2) examine the role of cognitive function as a predictor diet and physical activity change; and 3) examine the effects of the intervention on biomarkers of dietary intake (panel of plasma carotenoid and tocopherol concentrations). Blood samples will be stored for future analyses of changes in biomarkers hypothesized to be associated with breast cancer recurrence risk, including biomarkers of inflammation (e.g., high sensitivity C-reactive protein, interleukin-6, tumor necrosis factor-α), oxidative stress (e.g., isoprostane), DNA methylation, and metabolic biomarkers (e.g., insulin, glucose, insulin growth factor, lipid panel).

2.3. Eligibility criteria

Inclusion criteria include 1) female, 2) 18 years old or older, 3) self-identification as Latina or Hispanic, 4) self-reported history of stage 0-III breast cancer, 5) no evidence of recurrent or metastatic disease, 6) at least 90 days post final chemotherapy, biologic therapy, or radiation treatment and/or breast surgery (current use of endocrine therapy permitted), 7) nonsmoker (within the past 30 days), 8) intake of <5 daily servings of fruits and vegetables and/or engaging in <150 minutes per week of MVPA, as assessed by brief screening questionnaires;[49–50] 9) willingness and ability to receive text messages via cellphone/smartphone and email newsletters via computer, tablet, or smartphone, and 10) willingness and ability to attend four 4-hour in-person group sessions and to travel to complete study activities and clinic visits at baseline, 6- and 12-month timepoints.

2.4. Recruitment

Potential participants were identified through: 1) Columbia University Medical Center (CUMC) breast cancer patient database registries; 2) direct referrals by physicians in the breast oncology outpatient clinic at CUMC; 3) breast cancer community events, support groups, and organizations; 4) database requests to other collaborating medical institutions; 5) study brochures and flyers; 6) word of mouth; and 7) previous study participants who provided consent to be contacted for future studies. Lists of potential participants identified via databases or medical center physicians were sent to breast oncologists or breast surgeons seeking permission to contact the individual patients regarding study recruitment. Patients for whom permission was granted were mailed a recruitment letter signed by their physician and a study brochure, followed by a phone call from study staff querying interest and eligibility. Patients who were identified via other recruitment methods were contacted via telephone to query their interest and eligibility to participate in the study. Study staff were fully bilingual (English and Spanish) and all study materials were available in Spanish and English and provided to participants based on their preferred language.

Once a patient expressed interest by phone in learning more about the study, study staff completed a brief telephone screening interview to assess eligibility. If eligibility criteria were met, participants are assigned to study staff and were scheduled for two in-person clinic visits at CUMC and telephone-based dietary assessments to be completed during a “run-in” period prior to randomization (Figure 1). The run-in period was intended to assess ability to complete study assessments and data collection procedures.

Figure 1. Study schema.

The study questionnaire included study-specific questions on predictors and mediators of diet and physical activity change and use of technology, and questions from validated tools to measure psychosocial and quality of life outcomes.

2.5. Procedures

2.5.1. Baseline data collection

Baseline data collection includes two in-person clinic visits and telephone-based dietary assessment. During the first baseline clinic visit, participants complete: 1) written informed consent form, 2) medical records release form, 3) study questionnaire with questions on demographic and socioeconomic characteristics, acculturation, medical history, predictors and mediators of diet and physical activity change, psychosocial and quality of life (QOL) outcomes, and use of technology. Initially, all study participants were asked to complete a staff-administered neuropsychological assessment using the NIH Toolbox.[51] This procedure was dropped mid-way through the study due to high participant burden. On average, the first baseline clinic visit takes 2.5–3 hours (with neuropsychological assessment) and 2 hours (without neuropsychological assessment). Participants are reimbursed for parking and transportation costs for all clinic visits.

Following the first clinic visit, study staff contact study participants via the phone to complete three 24-hour dietary recalls (2 weekdays, 1 weekend day). During the second baseline clinic visit, participants complete: 1) staff-administered 7-day physical activity recall; 2) measured anthropometrics (height, weight, waist and hip circumference); and 3) blood draw (optional).

Once baseline data collection is completed, and prior to randomization, all participants receive a brief verbal overview of the ACS and AICR diet, nutrition and physical activity recommendations for cancer survivors. The ACS and AICR recommendations for cancer survivors include eating a diet high in fruits and vegetables, low in energy-dense foods and beverages, and increasing physical activity.[23, 52] In addition, each study participant receives a wearable physical activity tracker device, a Fitbit Zip, to self-monitor their physical activity. Study staff help study participants download the Fitbit app in their smartphones or tablet. When a participant does not have a smartphone or tablet, they are provided with an iPod touch to be paired with the Fitbit Zip. The Fitbit Zip is linked to the device (i.e., smartphone, tablet or iPod) using a de-identified anonymous user profile. On average, the second baseline visit takes 90 minutes to complete.

2.5.2. Randomization

After completion of baseline data collection, participants are randomized to one of four study arms using a 2×2 factorial design with permuted blocks of variable sizes (Figure 2).

Figure 2.

Randomization schema 2×2 factorial design

Randomization is stratified by preferred language (English or Spanish) and current use of endocrine therapy (yes or no). Participants are assigned to receive either in-person group sessions plus eHealth (Arm A), in-person group sessions alone (Arm B), eHealth alone (Arm C), or control (Arm D).

2.5.3. Monthly retention calls

Brief monthly follow-up retention calls to all study participants are conducted by their assigned study staff during months 1–5 and 7–11. During these calls, study staff ask about Fitbit use and Fitbit syncing, and if applicable, query participants about reasons for not wearing the device and any difficulties syncing. These calls are used as an opportunity to help participants troubleshoot any Fitbit issues and to remind participants of upcoming study data collection visits and phone calls, when appropriate. Study staff were assigned to participants to establish rapport and promote retention.

2.5.4. Follow-up data collection at 6 and 12 months

Baseline data collection is repeated at 6 and 12 months. Participants are first contacted to complete three 24-hour dietary recalls via the phone. After at least two dietary recalls are completed by phone, participants have an in-person clinic visit to complete: 1) staff-administered 7-day physical activity recall; 2) measured anthropometrics; 3) optional blood draw; and 4) an exit questionnaire on intervention acceptability that was developed specific to the trial (follow-up visit at 12 months only). The exit questionnaire uses selective questions depending on women assignment to each one of the study arms and whether they receive in-person and/or eHealth components of the program. For example, participants are queried on how helpful each of the intervention program components were helpful to them by the end of the trial (e.g., classroom education, cooking sessions, physical activity sessions, recipes, website, text messages, electronic newsletters, etc.). On average, each visit takes 90 minutes to complete.

2.6. Interventions

Based upon the AICR and ACS lifestyle recommendations for cancer survivors and our previous work,[48] ¡Mi Vida Saludable! focuses on the following specific behavior change goals: 1) increase daily fruit and vegetable intake to 5 servings per day or more (2 ½ cups of fruits and vegetables), and focus on non-starchy vegetables, 2) reduce intake of energy-dense foods and beverages, and 4) increase weekly minutes of MVPA to 150 minutes per week or more.

2.6.1. In-person group sessions (Study Arms A and B)

Participants randomized to Study Arms A and B are assigned to attend four 4-hour in-person group sessions during the first month. Each in-person group session includes: 1) classroom education on diet and physical activity for cancer survivors (1 hour), 2) hands-on skills-building cooking (1 hour); 3) either a physical activity session (i.e., dance class, walking around local park) or a field trip to a local and affordable grocery store or farmer’s market (1 hour each); and 4) sharing a meal prepared by the group at the end of each class (1 hour). The class curriculum and delivery were based on Contento’s Nutrition Education DESIGN procedure to adapt and develop the four-session group education and parallel electronic communication program for Latina breast cancer survivors.[46, 53] Details on the targeted specific behaviors and class curriculum components have been previously described (unpublished results currently under review by peer-reviewed journal). Trained study staff, not including any staff collecting study data, deliver the intervention sessions. Intervention staff include a trained chef, a nutrition and physical activity educator, and a dance class instructor. Other supporting staff help facilitate the delivery and organization of the sessions. As an incentive and to acknowledge the time and effort they contribute to the study, at the end of each session participants receive kitchenware or exercise gear. To reduce participant burden, the study provides on-site childcare during the in-person group sessions. Participants received transportation reimbursement for travel by car (cost of tolls) or public transporation (metrocard).

2.6.2. eHealth communication (Study Arms A and C)

Participants randomized to Study Arms A and C are assigned to receive: 1) weekly motivational and facilitating action (practical tips) unidirectional and bidirectional text messages on increasing fruits and vegetable intake, reducing consumption of energy-dense foods and beverages, and increasing physical activity, 2) emailed newsletters twice per month with text on increasing fruits and vegetable intake, reducing consumption of energy-dense foods and beverages, and increasing physical activity along with tips and recipes that were linked to the Cook for Your Life website (cookforyourlife.org). The Cook for Your Life website is a cancer patient-facing interactive website offering free nutrition and healthy cooking information, recipes, and cooking videos while disseminating evidence-based nutrition information for cancer survivorship. The website is intended to be used as a source of nutrition information and culinary skills-building. Text messages and emailed newsletters provide links to the Spanish- or English-language version of the Cook for Your Life website, depending on participants’ language preference.

For 11 months, participants receive 2–3 motivational and facilitating action text messages per week and are asked to respond to a text message about their diet (50%), and physical activity (50%) goals at least once per week. Text messages and emailed newsletters are sent using the eHIP (eHealth & Intervention Platform) platform developed by Arizona Research Labs at the Data Science Institute at the University of Arizona (UA). The eHIP platform is a web-based software application that allows researchers to easily integrate and manage integration of SMS/text messaging and email communications into their research projects. eHIP includes functionality to notify study personnel of incoming SMS messages so that study personnel can then monitor and respond back to study participants via SMS message from within the web application with real-time tracking of responses from both participants and study staff. Study staff at UA manage and implement the eHIP platform component of the program.

2.6.3. Control (Study Arm D)

Immediately following randomization, study participants assigned to the control arm receive the brief verbal overview of the ACS and AICR recommendations and Fitbit Zip as described above.

2.6.3.1. Receipt of incentives and study materials at study completion

To promote retention, participants are informed that they will receive all written materials and incentives, regardless of their randomization arm. At the 6-month clinic visit, participants in Arms C (eHealth alone) and Arm D (control) receive half of the incentives provided to participants during the in-person sessions (kitchenware or exercise gear). The other half of the incentives (kitchenware or exercise gear) are provided at the 12-month clinic visit. At study completion (12-month clinic visit), participants in study Arms B, C, and D receive the following: Arm B (group sessions alone) receives all emailed newsletters via email; Arm C (eHealth alone) receives printed recipes; and Arm D (control) receives printed recipes and all emailed newsletters via email, and are directed to the Cook for Your Life website for additional information.

2.7. Measures

2.7.1. Primary outcome measures

2.7.1.1. Fruit and vegetable intake and total energy density

Dietary intake is assessed via three 24-hour dietary recalls using the Nutrition Data System for Research (NDSR) developed by University of Minnesota (versions 2016–2019) (Table 1).

Table 1.

Timeline of study assessments

Assessment by study outcome	Measure	Baseline	Months 1–5	6 months	Months 7–11	12 months
Primary outcomes
Servings per day of fruits and vegetables	3 24-hour dietary recalls	X		X		X
Energy density, kcal/g per day	3 24-hour dietary recalls	X		X		X
Secondary outcomes
Minutes per week of MVPA	7-day Physical Activity Recall (7DPAR)	X		X		X
Measured Anthropometrics	Height, weight, hip, waist circumferences	X		X		X
Dietary intake	(3) 24H dietary recalls	X		X		X
Physical Activity	7-day Physical Activity Recall (7DPAR)	X		X		X
Predictors & Mediators of dietary and PA change
Nutrition knowledge	General Nutrition Knowledge	X		X		X
Self-efficacy in changing diet	Self-efficacy for Eating Behaviors Scale	X		X		X
Confidence in changing diet	Eating Habits Confidence Survey	X		X		X
Access to healthy food options	Morland Local Food Environment	X		X		X
Preferences and self-efficacy	PSEDPAL*	X		X		X
Psychosocial & quality of life outcomes
Perceived control of events	Multidimensional Health Locus of Control	X		X		X
Perceptions of general health	PROMIS Global Health	X		X		X
Perceptions of mental & physical health	PROMIS-43	X		X		X
Major stressful life events	Social Readjustment Rating Scale	X		X		X
Social support from close ones	Scale of Perceived Social Support	X		X		X
Social isolation	De Jong Gierveld Loneliness Scale	X		X		X
Exploratory outcomes
Neuropsychological function	NIH Toolbox Cognition Batteries	X
Physiological (future blood biomarker analyses)	Blood draw (optional)	X		X		X
Other assessments
Demographic and socioeconomic status	Study questionnaire	X
Acculturation status	Short Acculturation Scale for Hispanics	X
Medical history	Study questionnaire	X
Use of technology	Study questionnaire	X		X		X
Clinical characteristics
Self-reported	Study questionnaire	X
Abstracted from medical records	Medical records	X
Fitbit use and syncing	Retention monthly calls		X		X
Fitbit troubleshooting	Retention monthly calls		X		X
Physical activity	Fitbit data		← Daily →

^*PSEDPAL, Preferences and Self-Efficacy of Diet and Physical Activity Behaviors Questionnaire for Latina

This validated tool has been used frequently to measure dietary intake in a variety of Hispanic/Latino populations.[47, 54] A certified NDSR data collector (certified by the Nutrition Coordinating Center, University of Minnesota) trained study staff over a two-to-three-week period before data collection started. Participants are instructed by their assigned study staff on how to estimate portion size during the first assessment while completing the first in-person clinic visit with food models and the NDSR Food Amounts Booklet. Each participant receives the NDSR Food Amounts Booklet to take home for subsequent telephone-based dietary assessments. The study protocol states that two of the assessments were to be conducted during weekdays and one on a weekend day. The NDSR-certified study staff also assess all 24-hour dietary recalls for quality assurance, examining portion size and food types selected for accuracy in reporting.

2.7.2. Secondary outcome measures

2.7.2.1. Minutes per week of moderate-to-vigorous physical activity

Physical activity is assessed using a staff-administered and validated 7-day Physical Activity Recall (7DPAR).[55–56] This is a brief assessment designed to query the amount of time spent engaged in various types of physical activity and sleep during the past 7 days. Using standard protocols, study staff administer the 7DPAR, which yields duration and intensity of physical activities participants engaged in during the past week. During data collection, physical activities are categorized as “moderate” (e.g., yoga, sweeping, walking 3–4 mph), “hard” (e.g., aerobic dance, fast walking, scrubbing floors), or “very hard” (e.g., circuit training, jumping rope, cross-country running). Total weekly minutes or hours spent in each activity category, as well as category type, are then derived.

2.7.2.2. Measured anthropometrics

Height, weight, waist and hip circumferences are measured by trained study staff using a standardized protocol during the in-person clinic visits. A calibrated electronic scale (SR Instruments SR Scale, Tonawanta, NY) and stadiometer (Genentech Accustat Stadiometer, San Francisco, CA) are used to measure weight and height, respectively. Study participants wear light clothing and are without shoes. Height is measured using a stadiometer with a fixed vertical backboard and an adjustable head piece. A Gulick (Country Technology, Inc., Gays Mills, WI) anthropometric measuring tape is used to measure waist and hip circumferences. Participants are instructed to gather their upper clothes above the waist, cross the arms, and place the hands on opposite shoulders. When necessary, participants are also asked to lower the pants and underclothing to slightly below the waist. After locating the right ilium of the pelvis, a line is marked with the measuring tape to collect the measurement. Similar procedures are used to measure hip circumference. For all anthropometric measures of weight and waist/hip circumference, to improve the accuracy of the measurements a total of three measurements are collected and the final value is derived by averaging the three measurements.

2.7.2.3. Predictors and mediators of diet and physical activity change

Participants are queried about general nutrition knowledge, self-efficacy and confidence related to eating habits and food environment via previously validated instruments, including the general nutrition knowledge[57] and self-efficacy for eating behaviors scales,[58] the eating habits confidence survey,[59] and the Morland local food environment assessment tool.[60] For this trial, we developed the Preferences and Self-Efficacy of Diet and Physical Activity Behaviors Questionnaire for Latinas (PSEDPAL) to measure determinants of preferences and self-efficacy for four diet and physical activity behaviors: fruits and vegetables; dietary fat; added sugar; and physical activity (8 scales). The tool was developed by our research group and was based on feedback from focus groups conducted prior to the randomized controlled trial described here and test re-tested for internal validity and reliability during the trial baseline data collection.[61]

2.7.2.4. Psychosocial and Quality of Life (QOL) outcomes

To assess psychosocial and QOL outcomes we used validated tools, including, the Multidimensional health locus of control (MHLC) scale;[62–63] 2) Patient-reported outcome measurement information system (PROMIS^®) for physical, mental, and social health;[64–67] 3) Social readjustment rating questionnaire;[68–69] 4) Multidimensional scale of perceived social support;[70] and 5) De Jong Gierveld loneliness scale to measure social isolation.[71] The health locus of control as measured by the MHLC is defined as the individual’s tendency to perceive the events of influence as a result of personal abilities and behaviors or as a result of external circumstances. To measure physical, mental, and social health, we used the PROMIS Scale v1.2 – global health and PROMIS-43 profile v2.1.[64–67] The PROMIS global health specifically captures perceptions of general health that cut across domains. PROMIS-43, specifically assesses domains of physical function, anxiety, depression, fatigue, sleep disturbance, ability to participate in social roles and activities, and pain interference and intensity. The social readjustment rating scale measures major life events over the past 12 months.[69] The multidimensional scale of perceived social support measures social support from friends, family and significant other, whereas the loneliness scale measures social isolation.

2.7.3. Exploratory outcome measures

2.7.3.1. Neuropsychological function

To assess study participants’ cognitive function, we used the Cognition Batteries from the NIH Toolbox^®, which is a comprehensive set of neuropsychological tests that assess cognitive, emotional, sensory, and motor functions from the convenience of the NIH Toolbox iPad app.[72] For this trial, executive function, attention, episodic memory, language, processing speed, and working memory were assessed with the NIH Toolbox Cognition battery. The NIH Cognition Batteries, which include the following measures: 1) Dimensional Change Card Sort Test, 2) Flanker Inhibitory Control and Attention Test, 3) Pattern Comparison Processing Speed Test 4) Picture Sequence Memory Test, and 5) List Sorting Working Memory Test.[72] The cognition batteries also includes two tests to measure language, however they were not used in this study. Although the neuropsychological tests were in Spanish, when they were implemented in the first group of enrolled study participants (n=66), many of the women expressed frustration and immense feelings of inadequacy during the testing. Therefore, the study protocol was amended to remove this measure due to unacceptable participant burden. Exploratory analyses using these measures will be restricted to only focus on cognition as a predictor of change in diet and physical activity in the subset of participants who provided these data.

2.7.3.2. Physiological (blood biomarkers)

An optional fasting blood sample is collected at baseline, 6 and 12-month follow-up visits. A certified phlebotomist follows standard protocols for blood collection. Briefly, blood is drawn via venipuncture using a butterfly needle and processed into serum and plasma. Samples are then stored in −80°C freezers until the study is completed and future analyses are performed. Biomarkers will be stored for future analyses of changes in biomarkers associated with breast cancer recurrence risk.

2.7.4. Other outcome measures

2.7.4.1. Demographic, socioeconomic, acculturation, and medical history

Questions are included on age, race/ethnicity, nationality, education level, annual household income, marital status, Zip code, participation in programs targeting food insecurity (Supplemental Nutrition Assistance (SNAP) and Women, Infants, and Children (WIC)). Of note, many women in this population live in multi-generational households and may benefit from both SNAP and WIC. Acculturation is measured by the Short Acculturation Scale for Hispanics (SASH),[73] which assesses language preferences based on personal use, media use, and social relationships and has been validated in many Hispanic/Latino populations. Medical history is assessed with questions on food allergies, dietary restrictions, health insurance, history of breast cancer and treatment history, existing medical conditions, and medication use.

2.7.4.2. Use of technology

Use of technology is assessed with study developed questions querying participants on 1) use of Internet, email, text messages, and social media; 2) comfort level with use of Internet, email, text messages, and social media; 3) device of choice to access Internet, and 4) source of news and information; 5) preference for social media; and 6) people with whom they are connected via social media.

2.7.4.3. Clinical characteristics

Participants’ clinical characteristics related to cancer diagnosis and treatment are collected via self-report using study questionnaire and via electronic and requested paper-format medical records assessed by a trained Nurse Practitioner. Cancer stage, tumor characteristics and type of breast surgery data are assessed via medical records.

2.8. Statistical analyses

2.8.1. Hypotheses

The a priori primary aim is to test the separate and synergetic effects of the ¡Mi Vida Saludable! intervention program components from baseline to 12 months on changes in daily servings of fruits and vegetables and total energy density in women in the four arms. We hypothesize that women receiving the in-person group sessions plus eHealth will have greater change in increased intake of daily servings of fruits and vegetables and decreased total energy density compared to (in descending order) women receiving the in-person group sessions alone, eHealth alone, and control. In the secondary aims, we hypothesize that there will be a detectable dose response on changes in increased physical activity and decreased weight, following the pattern hypothesized in the primary aim. In the exploratory aims, we hypothesize that the predictors of diet and physical activity change will determine who is successful at making change; and the mediators of diet and physical activity change will determine the degree of change and whether the change is sustained.

2.8.2. Approach

Descriptive statistics will be presented for all variables. Shapiro-Wilk’s normality test and histograms will be applied to check for normality assumptions. In cases where the normality assumption is required, log transformation will be considered for those variables that do not appear to adhere to this assumption. If the log transformation does not sufficiently address the issue, nonparametric techniques (e.g., Wilcoxon rank sum tests, Spearman’s rank correlation coefficients) will be applied.

One- and two-way ANOVA models will be used to test our primary and secondary hypotheses for change in daily servings of fruits and vegetables, total energy density and minutes per week of MVPA across study arms. These models will be adjusted for baseline values and stratification variables (preferred language and use of endocrine therapy at time of enrollment). We will use chi-square tests to determine whether each important potential confounder was balanced during the randomization process; if variables are not balanced, we will adjust for these in the models. These same methods will be employed to address the aims related to 6- and 12-month changes in diet, physical activity, weight, predictors and mediators of diet and physical activity change, and blood biomarkers.

Geometric means will be reported for log transformed data. Following this, reduced F-tests will be used to test the null hypotheses that there are no differences in change in daily servings of fruits and vegetables and decreasing total energy density; or physical activity from baseline to 12 months between study arms. If we reject the above null hypotheses, we will employ the Jonckheere-Terpstra test, a non-parametric, rank-based trend test that can be used to determine the significance of a trend whether an increase in one variable results in an increase or decrease in another variable. Generalized Estimating Equations (GEE) methods will be used to assess the change in secondary outcomes from baseline to 6 and 12 months. Mediation analysis using bootstrap method will be used to examine potential predictors and mediators of diet and physical activity change as previously conducted by our group.[74] All statistical models will be adjusted for baseline values and stratification variables.

2.8.3. Sample size

With an anticipated 40 evaluable participants in each of the 4 arms at 12 months, the trial was powered to detect a range of meaningful differences in change in the primary endpoints (daily intake of fruits and vegetables and total energy density) for a range of standard deviations (SD). For change in daily intake fruits and vegetables, with 40 evaluable participants per arm we would have over 80% power to detect a difference in 0.6 servings (2.0–2.5 SD), 0.7 servings (2.0–3.0 SD), and 0.8 servings (2.0–3.0 SD). For change in total energy density (grams/kilocalories), we would have over 80% power to detect a difference in 4% (15% SD), 5% (15–20% SD), and 6% (15–25% SD). In order to have an analytic sample of 160 women, our goal was to recruit 250 women to be eligible to begin baseline data collection procedures. Based on prior experience, we conservatively anticipated 20% of the women recruited not completing baseline in-person and telephone-based data collection prior to randomization, and an additional 20% dropout rate was estimated post-randomization, resulting in an anticipated 160 evaluable participants.[47, 75] As the trial did not meet our a priori recruitment goals, the primary analyses will focus on the main effects of the two interventions (in-person group classes vs. no classes, and eHealth vs. no eHealth), which will have ample power. The analyses of the synergistic effects of the interventions will be considered exploratory analyses.

3. Results

3.1. Recruitment and enrollment

Participants were screened between January 2016 and September 2018. The primary sources of recruitment were breast cancer patient databases (n=530), followed by database request to outside medical institutions (N=141) and direct referrals from physicians (N=115). Though distributed, recruitment flyers and brochures were not an effective recruitment strategy (n=9). Of 884 women assessed for eligibility through the staff-administered screening questionnaire, 241 (27%) were eligible (Figure 3).

Figure 3.

CONSORT flow diagram

Nearly three-quarters (73%) of women screened for eligibility were ineligible. Primary reasons for ineligibility were not being able to participate due to work or life responsibilities (17%), health reasons (12%), not being able or willing to participate (17%), transportation barriers (9%), not being able to receive email or text messages (6%), or meeting both the diet and MVPA recommendations (4%). Of the 241 women eligible to participate, 17% did not show to the first in-person baseline clinic visit, and an additional 14% failed to complete the run-in period to collect baseline data, either not completing the telephone-based dietary assessments or did not show to the second baseline clinic visit. From July 2016 to October 2018, a total of 167 women were randomized into the trial and of those, 77.2% provided baseline blood draws.

3.2. Characteristics of study participants

On average, women were 56.8 ± 9.9 years old (Table 2). There were no statistically significant difference across study arms for all but one demographic characteristic. Women in the control group (Arm D) trended to be younger in age (53 years) compared to women in study arms A (58 years), B (59 years) and C (57 years) (p=0.07). The majority (65.3%) of the women were Dominican (65%), followed by Puerto Ricans (13%) and Ecuadorians (5%). About a third (28%) of women reported high school education of less, 38% working full- or part-time, 17% being disabled, and 53% an annual household income of ≤$15,000. Over half (51.5%) of study participants reported current participation in food assistance programs (SNAP and/or WIC). Proxies of acculturation were also measured, including language preference, place of birth, and duration in the US. Most women preferred Spanish (89%) and were foreign-born (75%). Among foreign-born women, mean duration in the US was 28 ± 13 years.

Table 2.

Baseline characteristics of 167 Latina breast cancer survivors enrolled into the trial overall and by randomization group

	No. (%) or mean (SD)
Characteristic	All (N=167)	Arm A: In-person classes + eHealth (n=41)	Arm B: In-person classes only (n=43)	Arm C: eHealth only (n=42)	Arm D: control group (n=41)	P
Age, years, mean (SD)	56.8 (9.9)	57.7 (11.1)	58.9 (9.1)	57.1 (8.4)	53.4 (10.5)	0.07
Race, n (%)						0.58
Black	33 (19.8%)	9 (22.0%)	3 (7.0%)	11 (26.2%)	10 (24.4%)
White	47 (28.1%)	8 (19.5%)	14 (32.6%)	11 (26.2%)	14 (34.2%)
Native American	5 (3.0%)	2 (4.9%)	2 (4.6%)	1 (2.4%)	0 (0.0%)
More than one race	78 (46.7%)	21 (51.2%)	23 (53.5%)	18 (42.9%)	16 (39.0%)
Nationality, n (%)						0.65
Argentinian	1 (0.6%)	0 (0.0%)	0 (0.0%)	0 (0.0%)	1 (2.4%)
Colombian	8 (4.5%)	2 (4.9%)	1 (2.3%)	4 (9.5%)	1 (2.4%)
Cuban	1 (0.6%)	0 (0.0%)	1 (2.3%)	0 (0.0%)	0 (0.0%)
Dominican	109 (65.3%)	26 (63.4%)	26 (60.5%)	29 (69.1%)	28 (68.3%)
Ecuadorian	9 (5.4%)	2 (4.9%)	5 (11.6%)	1 (2.4%)	1 (2.4%)
Salvadorian	1 (0.6%)	0 (0.0%)	1 (2.3%)	0 (0.0%)	0 (0.0%)
Honduran	1 (0.6%)	1 (2.4%)	0 (0.0%)	0 (0.0%)	0 (0.0%)
Mexican	6 (3.4%)	3 (7.3%)	1 (2.3%)	1 (2.4%)	1 (2.4%)
Nicaraguan	1 (0.6%)	1 (2.4%)	0 (0.0%)	0 (0.0%)	0 (0.0%)
Peruvian	1 (0.6%)	0 (0.0%)	0 (0.0%)	0 (0.0%)	1 (2.4%)
Puerto Rican	21 (12.6%)	4 (9.8%)	5 (11.6%)	6 (14.3%)	6 (14.6%)
Other	8 (4.8%)	2 (4.9%)	3 (7.0%)	1 (2.4%)	2 (4.9%)
Education, n (%)						0.33
Less than high school	46 (27.5%)	9 (22.0%)	9 (20.9%)	18 (42.9%)	10 (24.4%)
High school graduate or GED	40 (24.0%)	7 (17.1%)	12 (27.9%)	9 (21.4%)	12 (29.3%)
Some college	38 (23.0%)	10 (24.4%)	11 (25.6%)	8 (19.1%)	9 (22.0%)
College degree or higher	43 (25.5%)	15 (36.5%)	11 (25.6%)	7 (16.6%)	10 (24.3%)
Employment, n (%)						0.61
Full-time	42 (25.2%)	9 (22.0%)	10 (23.3%)	9 (21.4%)	14 (34.2%)
Part-time	22 (13.2%)	4 (9.8%)	6 (14.0%)	7 (16.7%)	5 (12.2%)
Retired	37 (22.2%)	14 (34.2%)	8 (18.6%)	8 (19.1%)	7 (17.1%)
Homemaker	18 (10.8%)	2 (4.9%)	8 (18.6%)	5 (11.9%)	3 (7.3%)
Unemployed	17 (10.2%)	6 (14.6%)	4 (9.3%)	3 (7.1%)	4 (9.8%)
Student	2 (1.2%)	1 (2.4%)	0 (0%)	0 (0%)	1 (2.4%)
Disabled	28 (16.8%)	5 (12.2%)	7 (16.3%)	10 (23.8%)	6 (14.6%)
Missing	1 (0.6%)	0 (0.0%)	0 (0.0%)	0 (0.0%)	1 (2.4%)
Annual household income, n (%)						0.80
$0 – $15,000	89 (53.3%)	22 (53.7%)	23 (53.5%)	21 (50.0%)	23 (56.1%)
$15,001 – $30,000	33 (19.8%)	9 (22.0%)	11 (25.6%)	9 (21.4%)	4 (9.8%)
$30,001 – $60,000	22 (13.2%)	4 (9.8%)	6 (14.0%)	4 (9.5%)	8 (19.5%)
$60,001 – $100,000	15 (9.0%)	4 (9.8%)	3 (7.0%)	4 (9.5%)	4 (9.8%)
More than $100,000	5 (3.0%)	1 (2.4%)	0 (0.0%)	3 (7.1%)	1 (2.4%)
Currently on food assistance program, n (%)						0.78
SNAP and/or WIC	86 (51.5%)	21 (51.2%)	23 (53.5%)	21 (50.0%)	21 (51.2%)
None	80 (47.9%)	20 (48.8%)	20 (46.5%)	21 (50.0%)	19 (46.3%)
Language preference, n (%)						0.97
Spanish	149 (89.2%)	36 (83.7%)	38 (88.4%)	38 (90.5%)	37 (90.2%)
English	18 (10.8%)	5 (16.3%)	5 (11.6%)	4 (9.5%)	4 (9.8%)
Place of birth, n (%)						0.98
Foreign-born	125 (74.9%)	31 (75.6%)	32 (74.4%)	32 (76.2%)	30 (73.2%)
US-born	25 (15.0%)	5 (12.2%)	6 (14.0%)	7 (16.7%)	7 (17.1%)
Duration in the US, mean years (SD)	27.8 (12.8)	28.9 (14.3)	28.9 (11.1)	26.5 (11.7)	27.2 (14.6)	0.86
BMI categories, n (%)						0.64
Normal weight, BMI >18.5 to <25	28 (17.2%)	7 (18.0%)	5 (11.9%)	10 (23.8%)	6 (15/0%)
Overweight, BMI 25 to <30	73 (44.8%)	15 (38.5%)	23 (54.8%)	18 (42.9%)	17 (42.5%)
Obese, BMI >=30	62 (38.0%)	17 (43.6%)	14 (33.3%)	14 (33.3%)	17 (42.5%)
Mean time since diagnosis, mean (SD)
Number of months	71.6 (58.6)	79.5 (69.7)	83.1 (58.9)	70.3 (53.1)	52.8 (47.7)	0.09
Number of years	6.0 (4.9)	6.6 (5.8)	6.9 (4.9)	5.9 (4.4)	4.4 (4.0)	0.09
Cancer stage, n (%)						0.08
Atypia / In Situ	41 (24.6%)	11 (26.8%)	9 (20.9%)	15 (35.7%)	6 (14.6%)
I	51 (30.5%)	12 (29.3%)	13 (30.2%)	13 (31.0%)	13 (31.7%)
II	31 (18.6%)	10 (24.4%)	6 (14.0%)	7 (16.7%)	8 (19.5%)
III	18 (10.8%)	5 (12.2%)	4 (9.3%)	6 (14.3%)	3 (7.3%)
Breast surgery, n (%)						0.58
Yes	165 (98.8%)	41 (100%)	42 (97.7%)	41 (97.6%)	41 (100%)
No	2 (1.2%)	0 (0.0%)	1 (2.3%)	1 (2.4%)	0 (0%)
Chemotherapy therapy received, n (%)						0.12
Yes	103 (61.7%)	26 (63.4%)	28 (65.1%)	21 (50.0%)	28 (68.3%)
No	62 (37.1%)	13 (31.7%)	15 (34.9%)	21 (50.0%)	13 (31.7%)
Radiation therapy received, n (%)						0.65
Yes	106 (63.5%)	25 (61.0%)	29 (67.4%)	30 (71.4%)	22 (53.7%)
No	58 (34.7%)	15 (36.6%)	14 (32.6%)	11 (26.2%)	18 (43.9%)
Endocrine therapy, n (%)						0.99
Yes	79 (47.3%)	19 (46.3%)	21 (48.8%)	20 (47.6%)	19 (46.3%)
No	88 (52.7%)	22 (53.7%)	22 (51.2%)	22 (52.4%)	22 (53.7%)
Medical conditions, n (%)
Hypertension	65 (38.9%)	14 (34.2%)	18 (41.9%)	20 (47.6%)	13 (31.7%)	0.43
Diabetes	30 (18.0%)	9 (22.0%)	4 (9.3%)	9 (21.4%)	8 (19.5%)	0.39
Cholesterol problems	53 (31.7%)	19 (46.3%)	12 (27.9%)	12 (28.6%)	10 (24.4%)	0.14
None	47 (28.1%)	10 (24.4%)	11 (25.6%)	11 (26.2%)	15 (36.6%)	0.58
Mean comorbidity index score (0 – 20)1, mean (SD)	1.8 (2.0)	2.2 (2.2)	1.7 (2.3)	2.0 (2.0)	1.4 (1.6)	0.32

NOTE. Not all percentages add to 100% due to missing data. One-way ANOVA tests were used for continuous variables and chi-square tests were used for categorical variables.

¹A score of 1, 2, or 3 is then assigned to each condition. Ulcers, diabetes, neurological problems, gastrointestinal problems, respiratory problems (shortness of breath and asthma), and cardiovascular risk factors (hypertension and high blood cholesterol) each received a score of 1; kidney disease, heart problems, chest pain, and physical limitations each received a score of 2; HIV/AIDS, and cancer other than breast cancer each received a score of 3. The score is then computed for each participant by summing the weighted values for each comorbid condition (possible index range 0–20 points).

Abbreviations: BC, breast cancer; BMI, body mass index; EBT, Electronic Benefits Transfer; GED, General Education Development; SASH, Short Acculturation Scale for Hispanics; SD, standard deviation; SNAP, Supplemental Nutrition Assistance; WIC, Women, Infants, and Children program.

Clinical characteristics are also presented in Table 1. There were no statistically significant differences across study arms for all but one clinical characteristic. Time since diagnosis trended to be longer in study arms A (7 years), B (7 years) and C (6 years) compared to the control group (4 years) (p=0.09). Approximately 25% of study participants had been diagnosed with Atypia / In Situ and all but two women reported having had breast surgery (98%). Most women received chemotherapy (61.7%) and/or radiation (63.5%) while about half (47.3%) were on endocrine therapy at time of enrollment. More than a third of the women reported having hypertension (39%) and or cholesterol problems (32%), 18% having diabetes; whereas 28% reported having none of the medical conditions queried.

4. Discussion

We designed a novel, targeted, culturally based 12-month 2×2 factorial-designed trial testing the separate and synergetic effects of the ¡Mi Vida Saludable! in-person and eHealth intervention program components on changes in diet and physical activity behaviors among Latina breast cancer survivors. The ¡Mi Vida Saludable! Trial is among the first randomized controlled trials targeting diet and physical activity behaviors among Latina breast cancer survivors. Latinas are at a greater risk of overweight and obesity[76] and have poorer cancer prognosis and survival when compared to their non-Hispanic/Latino counterparts[18, 77] yet have rarely been exclusively targeted for improving modifiable risk factors in culturally-tailored randomized trials. Here we described the ¡Mi Vida Saludable! trial design, methods, recruitment and enrollment.

The study enrolled a diverse group of breast cancer survivors residing in NYC and representing more than 15 Latin American nationalities. Our primary sources of recruitment included cancer registries, direct referrals by physicians in the local oncology clinics, and various community events, breast cancer support groups and organizations. Primary reasons for ineligibility included not being able to participate due to work or life responsibilities, current health issues, not living in NYC or inability to attend in-person study activities due to transportation barriers. About 12% of women screened refused to participate with no reason given. We also found a small but significant proportion of women screened that were deemed ineligible due to not being willing or able to receive emails and/or text messages (6%). Although, previous studies have reported that US Latinos have a high use of digital technology compared to other ethnic groups,[78] our study suggests that access and use of electronic communication may still be a barrier to eHealth trial participation for Latina breast cancer survivors.

4.1. Study strengths

First, the ¡Mi Vida Saludable! trial addresses a critical gap in behavioral intervention research among Latina breast cancer survivors who are at a greater risk of poor cancer prognosis and cancer recurrence compared to non-Latina/Hispanic breast cancer survivors. Second, we had a successful recruitment and accrual of a diverse group of breast cancer survivors representing more than 15 Latin American nationalities. Third, the trial’s rigorous 2×2 factorial design will allow us to tease out important results on the effects of behavioral interventions using either in-person and/or eHealth modalities to achieve and maintain diet and physical activity change. Fourth, the in-person and eHealth modalities were carefully designed to address the same diet and physical activity behaviors and the theory-based motivating and facilitating action determinants, making them parallel in educational content allowing for testing of the modalities.

4.2. Study limitations

Our study results will be limited by the use of self-reported diet and physical activity, both of which are subject to measurement error and recall biases. Importantly, findings will be limited to urban Latina breast cancer survivors and therefore may not be generalizable to other populations of breast cancer survivors. Second, we did not achieve the stated recruitment goal to enroll 250 women to undergo baseline assessments prior to randomization. Instead a total of 241 women were identified as eligible with 30% failing to successfully complete baseline assessments and therefore, were not randomized into the study. Third, we provided all participants with a Fitbit device to self-monitor PA and its limitation to tracking accurate levels of PA should be noted. Although participants were not instructed or encouraged to use the diet tracking feature of the Fitbit App or linking to the MyFitnessPal application, it is possible that participants used these features and behavior change could be due to these features instead of the specific interventions tested. However, since all participants have access to a Fitbit Zip, the differences should not be significantly different across the four arms. Finally, study staff collecting study data are not blinded to randomization assignments, which has the potential to bias data collection. Instead, each participant is assigned to study staff to complete all data collection starting at the screening timepoint throughout study completion at 12 months. This was decided to assure good rapport and help maintain retention. It is important to note that study staff collecting study data are not involved in the implementation of any of the intervention program components.

Conclusions

The ¡Mi Vida Saludable! trial addresses a critical gap in behavioral intervention research among Latina breast cancer survivors and will provide important information on the effects of an in-person and/or eHealth intervention program components on achieving and maintaining diet and physical activity change. If effective, the ¡Mi Vida Saludable! lifestyle behavioral intervention program has the potential to improve diet and physical activity behaviors. If successful, the intervention program component of greatest effectiveness can be tested in future, statistically well-powered trials to determine the impact on cancer recurrence and secondary cancers among this underrepresented and high at-risk population of Latina breast cancer survivors.