We adapted the American Clinical Society of Oncology (ASCO) SCP template [27] for low-literacy, Spanish-speaking Latinas, simplifying the layout and translating it into Spanish. Adaptations were based on iterative review by a Latina psycho-oncologist, 2 oncologists, a bilingual oncology nurse, and 3 Spanish-speaking breast cancer survivors. Participants signed a medical release form, and study personnel extracted the information from medical records to complete the SCP. Completed SCPs were reviewed by the project director and the patient’s oncologist or oncology nurse and scanned into the patient’s electronic health record. This written bilingual SCP was given to participants at the second home visit.

We selected the “ASCO Answers: Cancer Survivorship” guide because it was comprehensive, easy to understand, and available in English and Spanish [28]. The guide covers what to expect after active treatment, including psychological, physical, sexual, reproductive, financial, and work-related challenges.

The Spanish-language mobile app (trackC) was designed to contain women’s breast cancer diagnostic and treatment history and provide information on potential side effects, healthy lifestyles, and survivorship resources. An activity tracker was integrated with the app to display progress toward a personalized daily steps goal. We selected the Fitbit Zip wireless activity tracker, henceforth referred to as activity tracker, based on cost, simplicity, and availability of an application programming interface (API, for integrating the Fitbit with other software applications). The mobile app home page contained 4 section tabs (Figure 1): Daily walks (caminatas diarias), treatment (tratamiento), follow-up care (cuidado de seguimiento), and managing symptoms (manejo de los síntomas). Content was based on ASCO treatment guidelines at the time. We summarize each section briefly:

In phases, we developed mock-ups, a detailed wire frame, and a prototype of the app employing user-centered testing [29] with iterative review and pretesting by 3 Spanish-speaking Latina breast cancer survivors; a Latina psycho-oncologist (breast cancer survivor); an oncologist serving ethnically diverse, low-income cancer patients; and 6 bilingual-bicultural study staff members. The prototype was developed in English and then translated into Spanish using rigorous forward translation and team reconciliation methods.

The coaching protocol was based on evidence-based motivational interviewing and health coaching techniques, which seek to actively engage patients in managing their health within their social contexts [30]. The health coach encouraged use of trackC, walking, reporting symptoms to clinicians, and calling clinicians to ask questions. Communication with clinicians was emphasized because of evidence that Latino patients often lack the confidence to report symptoms or ask questions, especially when the physician speaks a different language [31,32]. The health coach reinforced cancer survivorship information. The health coach was a bilingual-bicultural Latin American–trained internist with extensive health coaching experience. Coaching consisted of 5 weekly phone calls with the following structure: (1) review of progress toward daily steps goal and working through any barriers, (2) daily steps goal setting for the coming week, and (3) information on a weekly health topic. The 5 health topics paralleled the trackC content and included: (1) walking and nutrition, (2) breast cancer follow-up care, (3) signs of recurrence, (4) treatment side effects, and (5) resources and review of content from the first 4 calls. The health coach used a manual, but tailored the content based on participants’ needs.

Eligibility criteria consisted of: (1) self-reported Spanish-speaking Latina, (2) diagnosed with nonmetastatic breast cancer, and (3) within 1 year of termination of active treatment (except for hormonal therapy). Exclusion criteria included walking more than 30 min on 5 days per week or more. Using lists of potentially eligible participants provided by the hospital sites, we mailed them bilingual initial contact letters and postage-paid refusal postcards. A total of 2 weeks later, women who had not returned a refusal postcard were contacted in person or on the telephone by trained bilingual-bicultural RAs to conduct eligibility screening, ask about mobile phone usage, and schedule an appointment to visit the participant’s home within 1 week.

The RA conducted the enrollment visit (45-60 min) at the clinic site or the participant’s home during which the study was explained in detail, written informed consent was obtained, participants signed a medical release form, and the baseline survey was completed. This marked the start of the 1-week run-in period. Women were provided with a masked activity tracker (hidden daily steps display) and instructions to wear it every day for a minimum of 10 hours per day and not to change their usual activity levels. This run-in period was used to establish participants’ baseline average daily steps and personalized goal (average daily steps during run-in period + 2000 steps).

In this 1-hour visit, participants received materials and verbal instructions on the use of the written SCP; survivorship booklet; iPhone and charger with trackC app installed; unmasked activity tracker (with visible daily steps and goal graph); and a step-by-step illustrated guide on how to use the iPhone, app, and activity tracker devices. The RA reviewed the SCP, survivorship booklet, device guide, and the individualized average daily steps goal to be achieved within 2 months. Women were instructed on synchronizing the tracker and mobile app at the end of every day to update the app’s average daily steps graph. The RA helped participants enter diagnostic and treatment information from the written SCP into trackC.

At this visit, the RA conducted the final assessment and a brief satisfaction survey, synchronized the activity tracker with the Fitbit app to update the final daily steps data, and collected the mobile phone and charger. Participants were allowed to keep the tracker and encouraged to continue to maintain a daily exercise routine. Upon returning to the office, the RA logged in using the participant’s study Fitbit account credentials and downloaded the Fitbit steps data to the study computer.

An electronic database (REDCap) was developed to track usability issues [33]. This system contained data from multiple sources, including phone calls from participants, issues reported by the health coach, daily review of the mobile app back-end database, RA and project director tracking forms and notes, and timing of software updates for the activity tracker. Mobile app data were sent to the study’s secure database via encrypted transmission. If the mobile phone or app lost connectivity, data were transmitted the subsequent time the app was connected to the internet.

The health coach recorded attendance and duration for the 5 calls. At every call, women were asked how many times in the past week they had synced their activity tracker with the trackC app and checked the app’s average daily steps graph and if they had experienced any problems doing this. On calls 1 and 3, women were asked how difficult they found it to use the graph (5-point response scale was 0=not at all to 5=difficult). After every session, the health coach was asked to rate how much of the material she felt the participant had understood (5-point response scale was 0=none to 5=all).

Semistructured debriefing interviews were conducted with a subset of participants to ask about their study experiences and suggestions for improvement. Selection of women was stratified to include those who had an iPhone versus other type of mobile phone or none, aged <50 versus ≥50 years, and from the 2 study sites. A trained bilingual-bicultural Latina interviewer (not the RA who conducted home visits) used an interview schedule that asked about their experience using the app (eg, what they did and did not like), ease of use, perceived usefulness for managing their cancer, and facilitating factors.

A 5-min satisfaction survey was administered at the final home visit after downloading participants’ activity tracker data for the study period and the final assessment. The survey asked them to rate the program’s perceived quality, ease of use, and usefulness. Overall quality of the app was assessed using a 5-level response set of “poor,” “fair,” “good,” “very good,” or “excellent.” Ease of use was assessed by asking about the overall difficulty of using the trackC app, syncing, and using the treatment summary, with response options of “not at all hard,” “a little hard,” “somewhat hard,” “quite hard,” or “very hard.” Perceived usefulness was assessed by asking participants to rate how useful the app and health coach were for helping them gain a sense of control over their health and how useful the app was for keeping their cancer treatment information in one place and knowing about cancer symptoms and treatment side effects to monitor. Response options for the usefulness ratings were “not at all,” “a little useful,” “somewhat useful,” “quite useful,” or “very useful.”

We measured 6 self-reported primary outcomes: 2 on symptoms, 3 on knowledge of cancer survivorship care, and 1 on self-efficacy for managing their cancer follow-up health care and self-care.

The 2 symptoms assessed were cancer-related fatigue and health distress. We adapted the Patient-Reported Outcomes Measurement Information System (PROMIS) Cancer-Fatigue Scale, which assesses the extent of fatigue and its impact on daily life over the past 7 days [34]. We dropped 1 item (“enough energy to exercise strenuously”) and added 2 items from the PROMIS Cancer Fatigue Short Form [35]: “felt tired when hadn’t done anything” and “limited social activities because of fatigue.” The final 7 items assess 4 aspects of severity (frequency that they felt tired, tired even when hadn’t done anything, extreme exhaustion, run out of energy) and 3 aspects of interference with daily life (frequency with which fatigue limited work, thinking clearly, taking bath or shower). To assess health distress, we selected 4 items from the Medical Outcomes Study Health Distress Scale [36] that asked how much of the time during the past month they felt discouraged, fearful, worried, or frustrated by their health problems. Response options for both fatigue and distress scales were as follows: “never,” “rarely,” “sometimes,” “often,” or “always.” Scale score were the mean of nonmissing items, with higher scores indicating greater fatigue effects (Cronbach alpha=.85) or health distress (Cronbach alpha=.91).

The 3 knowledge measures consisted of 2 global single item measures and 1 6-item scale. The 2 single items of global knowledge of survivorship care asked how true the following statements were for them: “you know what to expect now that your initial treatment has finished” and “you know how to take care of yourself after cancer.” The new scale assessed knowledge of follow-up care and ease of finding information. A sample item is “How true is the following statement for you: you know the possible side effects of your cancer treatment?” Response options for the 3 knowledge measures were 0=not at all true to 4=completely true. The scale was scored as the mean of nonmissing items with higher scores indicating greater knowledge (Cronbach alpha=.82).

A new 8-item self-efficacy for managing cancer care scale assessed confidence in ability to do what is needed to manage health care and health after cancer. A sample item is “How confident are you that you will be able to call your doctor if you have a question about a symptom that might be related to your cancer or treatment?” with response options of 0=not at all confident to 4=completely confident. The scale was scored as the mean of nonmissing items with higher scores indicating greater confidence (Cronbach alpha=.90). These new measures assessing women’s sense of control over their survivorship care drew on published questionnaires [37,38].

Secondary outcomes included emotional well-being, depressive and somatic symptoms, and average daily steps as recorded by the activity tracker.

Emotional well-being was assessed with the 6-item Emotional Well-Being Scale from the Functional Assessment of Cancer Treatment-General [39]. Scores range from 0 to 24, with higher scores indicating more well-being (Cronbach alpha=.77). We used the Patient Health Questionnaire 8-item version to assess depressive symptoms [40]. Scores range from 0 to 24, with higher scores indicating more depressive symptoms (Cronbach alpha=.64). We used the 6-item Brief Symptom Inventory Somatization Scale, which assesses the extent to which they were bothered by symptoms such as faintness and dizziness, pains in heart or chest, nausea, trouble getting their breath, numbness or tingling, and feeling weak [41]. Scores range from 0 to 4, with higher scores indicating more symptoms (Cronbach alpha=.76).

Baseline steps were calculated as the average daily steps during the 1-week run-in period (total steps divided by number of days) before the intervention start date. Postintervention steps were calculated as the average daily steps during the last week of the 2-month study period. Pre-post changes in average daily steps were calculated as the postintervention average daily steps minus the preintervention average daily steps.

Nonscheduled home visits by the RA to all participants became necessary because participants requested help with the trackC app, activity tracker, or phone, or study staff noticed a lack of data transmission from trackC to the app backend database. A total of 63 nonscheduled visits occurred (mean=3 per participant, SD 1.9; range 1-7), during which the RA would troubleshoot technical and user issues and provide additional support and instruction. Most issues were related to technical (46 instances because of the app host site expiring, activity tracker software updates, or the app and tracker not syncing) or hardware-related problems (22 instances of activity tracker needing a new battery or the iPhone locking them out). Some were related to user issues (28 instances of forgetting to sync or how to do it, not knowing how to swipe out of an app section, or losing the activity tracker).

Coaching calls lasted on average 15 min each (SD 3.4). A total of 19 of 23 participants (83%) completed all 5 coaching calls, 1 woman completed 4 calls, 1 woman completed 1 call, and 2 women completed no calls. Number of times per week that women synced their activity tracker and app ranged from 4.4 to 5.7. Number of times per week that women checked their daily steps graph on the app ranged from 4.2 to 5.9. Ratings of the difficulty with using the daily steps graphs at call 1 and call 3 were almost identical, with most women (12 at call 1, 11 at call 2) rating it as not at all difficult. A total of 3 women reported vision problems interfered with reading the app screens. On the basis of the coach’s ratings, the number of women understanding all of the material ranged from 17 (81%) for call 1 (daily steps and goal-setting) to 20 (100%) for call 3 (signs of recurrence).

A total of 10 semistructured postintervention debriefing interviews were conducted (Table 2). Participants were aged 56 years on average, and most were from Mexico (Mexico=7, Guatemala=2, and Nicaragua=1). All participants reported elementary school completion or less. In general, participants reported positive attitudes toward the program and increased awareness of the importance of walking. Themes emerging from the interviews are described next.

Participants voiced appreciation for the trackC app information about their disease, treatments, side effects, and signs of recurrence, having felt misinformed about cancer survivorship before the study. All the women wanted the written SCP in addition to the app version. They reported feeling motivated and supported by the weekly check-ins with the health coach because she provided them with tailored, detailed, and credible information and support; helping them understand their disease, symptoms, and bodies; and achieve their walking goal. Participants valued the visual and auditory instant feedback provided by the activity tracker and app, for example, applause received after achieving their daily goal, helping them maintain a positive attitude toward walking.

Participants described varied experiences about the effort required to navigate and use the app. Users with mobile phone experience found the app easy to use. However, 4 of 10 participants with little or no mobile phone experience expressed that use of the app required more effort and support at the beginning of the study. Some participants reported difficulties because of poor literacy or poor eyesight. All women reported being satisfied with the app’s interface, fonts, colors, and visuals.

Informants reported positive outcomes related to walking. A total of 7 of 10 women reported enhanced physical health because of their participation in the study, including weight loss, improved digestion and bowel movements, and improved sleep. Participants also reported improved emotional well-being, that is, decreased stress and better mood.

A total of 3 women felt a sense of accountability because they knew their steps were being monitored by themselves and others. Women reported that social support and encouragement from family members and neighbors pushed them to achieve their daily goal. Finally, several women expressed a shift from being extrinsically motivated by the app and coach to increase their walking to being intrinsically motivated because they wanted to do it for themselves.

The majority of the women (17/21; 81%) rated the overall quality of the app as very good or excellent (all rated it as at least “good”). The overall quality of the information received on how to use the trackC app was rated as very good or excellent by 16 women (76%); all rated it as at least good.

Most women (15/21; 71%) rated the ease of syncing the trackC app and activity tracker as being not at all hard (Table 3). Fewer respondents reported it being not at all hard to use the treatment summary found in the trackC app (11/21; 52%).

Regarding their ratings of the usefulness of the SCPP for feeling more in control of their health, all except for 1 woman rated the health coaching calls as quite or very useful, and all women reported the trackC app as quite or very useful. Almost all women (n=19) reported that the trackC app was quite or very useful for keeping their cancer treatment information in one place. Having information on trackC about cancer symptoms and side effects were both reported as being quite or very useful by 18 and 19 respondents.

Regarding primary outcomes, compared with baseline, fatigue (B=–.26; P=.02; Cohen d=0.4) and health distress levels (B=–.36; P=.01; Cohen d=0.3) were significantly lower post intervention (Table 4). Women reported significantly greater knowledge of recommended follow-up care and resources after the intervention (B=.41; P=.03; Cohen d=0.5); self-efficacy for managing cancer follow-up care did not change.

Of the secondary outcomes, emotional well-being improved significantly post intervention (B=1.42; P=.02; Cohen d=0.3). Women’s average daily steps increased significantly from 6157 to 7469 (B=1311.8; P=.02; Cohen d=0.5).