Patients at Memorial Sloan Kettering Cancer Center (MSKCC) who had completed their primary treatment for either breast or prostate cancer no more than 5 years previously and had an intake scheduled at the Survivorship Clinic were identified via the clinical database and sent an invitational letter signed by the director of the clinic describing the study and procedures. Two weeks later they were contacted by telephone. Once contacted, they were screened for additional eligibility criteria: the presence of at least one behavioral risk factor (engaging in <150 min of physical activity per week or eating less than 5 fruits or vegetables per day), English-speaking, and able to provide informed consent. During this call, the research assistant answered any questions about the study and scheduled a time to meet the patient immediately before their first Survivorship Clinic appointment to complete informed consent, the baseline survey, and randomization. Patients completed the survey on their own. The research assistant was present to clarify any questions when needed. Following the survey completion, the research assistant contacted the research office who provided the group assignment using the permuted block method stratified by the disease type (breast or prostate). If randomized to the intervention arm, the research assistant gave the participant the DVD and explained how to view it. One month following the clinic appointment, patients in the intervention arm were mailed a reminder letter to use it. Three months following the clinic appointment, all patients were mailed follow-up surveys and given the option to complete the surveys by phone or mail. At the completion of the study, patients in the control arm were offered the intervention components. The study was approved by the MSKCC and New York University Medical Center Institutional Review Boards. As a pilot, it was not a registered trial.

All intervention group participants completed use and evaluation items at the 3-month follow up [35]. Qualitative interviews (n=12) were also conducted with a random sample of intervention group participants who used the intervention. These were used to further investigate the acceptability and feasibility of the intervention and to inform improvements to future iterations of the program. The interviews were conducted over the phone by a qualitative methods specialist (Ms Shuk) and were limited to 45 min. Audio recordings were transcribed by an independent transcription company (RA Fisher, Inc).

The primary goal of this pilot study was to examine patient interest in and feasibility of the intervention in order to guide the development of a larger trial. We therefore detail screening, exclusion, and refusal reasons. For each primary outcome, we report means and standard deviations at baseline and 3-month follow up along with effect sizes (Cohen d). This was calculated as difference in the change scores for intervention versus control divided by the pooled standard deviation. For dietary intake, we reported the number of fruit servings, vegetable servings, and combined fruit and vegetable servings. Between-group differences were not analyzed as the pilot was not powered to detect statistically significant differences. All analyses were conducted in SAS version 9.3 (SAS Institute, Inc).

The qualitative data were reviewed using inductive thematic text analysis, an iterative process of transcript review, interpretation, and consensus discussions [36-38]. An initial set of 3 interview transcripts were coded by 2 independent reviewers (Ms Shuk and Ms Williams), in which each reviewer read the same transcript, highlighting important content and recording reflections on the transcript in a process known as margin coding [39], prior to completing a written analysis template with supporting participant quotations. The reviewers subsequently met to generate collective findings for the transcript. Once key thematic findings had been identified for the first 3 transcripts, the reviewers subsequently read and coded the remaining transcripts through the same process, both exploring the themes that had been established and identifying additional salient findings. As per standard procedures, the final analytic phase entailed generating higher-order descriptive and interpretive themes that represented the most frequent concepts observed across all interviews.

A total of 466 individuals were screened for eligibility. Of those, 259 could not be contacted via phone and 25 had no risk factors. Additionally, 10 individuals reported inability to operate a DVD, 8 were non-English speaking, and 1 was still under treatment. Of the 163 eligible individuals, 86 consented to participate for a 52.7% participation rate. The main reasons for refusal were time constraints (n=32) and lack of interest in making health behavior changes (n=29). At the 3-month follow up, the retention rate was 73% (32/44) and 86% (36/42) for the intervention and control groups, respectively (Figure 1). Recruitment and data collection were conducted from August 2013 to March 2014.

Demographic characteristics of the sample are shown in Table 1. Participants were predominantly non-Hispanic white (81%, 69/86), and female (96%, 82/86), with a mean age of 59.8 (standard deviation, SD 11.4). Recruitment of prostate cancer survivors was limited due to clinic scheduling and change in staffing during the study period.

At the 3-month assessment, all intervention group participants (n=32) completed survey items to provide feedback on their experiences with the intervention. Results are summarized in Table 3. Among the intervention group, 72% (23/32) viewed the DVD, with 50% (16/32) completing the entire DVD. Of those who used it, 60% did so more than once. More than half (14/23) rated it as easy to use, whereas a third (35%, 8/23) found it neither easy nor difficult. Most reported that it kept their attention at least somewhat (87%, 20/23), and looked professional (96%, 22/23). All the participants (100%, 23/23) stated it did a good job of presenting health information, found nothing offensive in the material, and was culturally appropriate, with only 1 person stating it made her feel uncomfortable. Users found it relevant for them as cancer survivors (91%, 21/23). Overall, they felt it was the right length (83%, 19/23), were satisfied (30%, 7/23), or extremely satisfied with it (61%, 14/23), and would recommend the intervention to others (91%, 21/23).

Qualitative interviews were also conducted with 12 of the intervention group participants who reported using the intervention. Three key themes emerged from the qualitative analysis of transcripts regarding engagement, content, and usability. Key themes and select quotes are presented below:

Theme 1: The program engaged patients’ interest as cancer survivors . Patients liked that it came from a reputable information source, found it to be engaging and interesting, liked the positive, encouraging tone, and found the survivor stories to be inspiring. Participants suggested greater racial, ethnic, and socioeconomic diversity of the survivors who were interviewed, and 1 user did not like that a survivor interview mentioned cancer as a “blessing in disguise.”

I think the whole DVD is a very positive, you know, approach, and it had a nice balance of survivors and then professionals.

Theme 2: Patients made suggestions for adding specific physical activity and healthy eating content. Nutrition and physical activity information was perceived to have been presented with an appropriate level of detail, and participants appreciated the focus on setting small goals and how to incorporate changes into their daily routines. They found it to be motivational and a good reminder of what their goals should be. In terms of preferences not included, patients requested more information on how fruits and vegetables and activity help the body physiologically, specific recipes, more exercises to perform, and how to tailor exercise to various needs such as living in urban areas or for older persons.

Take it to the next level. That was my biggest complaint about it, that I wanted more.

Theme 3: Patients liked the interactive structure and suggested usability improvements. Patients reported that the self-assessments were engaging and helped determine their current eating and activity habits. Patients generally found the menu structure easy to use on screen, but wanted it to be easier to go back and review sections. Patients also found it difficult to access the additional resources listed as they had to write them down. Patients noted they would want to have a follow up with greater detail and more specifics, with some noting they would like to have it available in a mobile app version they could access more readily on a mobile phone or tablet.

They didn’t seem in depth enough for me to—you know, again, I think the disconnect I have is you sit, you watch it, and then you’re left to your own devices. So, if there was something that I could—you know, again, that I could take with me to refer to during the day, I think that would impress me more, you know, impress upon my life more.

This pilot study examined the potential of a theory-based, eHealth intervention designed to assist adult cancer survivors make improvements in healthy diet and physical activity. Results indicate that recruitment and retention was feasible for this older adult survivorship population and that they had high interest in the intervention. Results on behavior change outcomes should be interpreted with caution as the study was not powered to detect reliable differences. Observed power for the effect sizes ranged from .06 to .21. Findings indicate small effects on dietary outcomes, primarily fruit intake, and suggest that additional modifications would be necessary to increase efficacy of the physical activity component.

Results should be interpreted with regard to feasibility of pursing a larger powered trial. The following criteria were assessed as indicators of whether to pursue follow-up work: recruitment of the target sample size in the allotted timeframe, acceptance rate of at least 50% [40], retention of at least 80% [40], at least small effect sizes (d=0.2) on primary outcomes, minimal adverse events, and patient report of interest in and acceptability of the intervention. In light of these criteria, recruitment goals were met within the allotted time frame of 7 months. The study intended to recruit equal numbers of prostate and breast cancer survivors, but a loss of clinic staff resulted in no new survivorship visits for men during the recruitment period. Thus, unfortunately, we do not know how men would respond to the intervention. Consent rates among those eligible were good (53%) and consistent with or higher than diet and activity studies among cancer survivors [41]. Response bias is always a possibility, but given our inclusion criteria, the identified and final sample was likely to differ little by demographics or treatment characteristics. Follow-up rates in the control group met the criteria (86%), but could slightly be improved in the intervention group (73%). No adverse effects were reported by the participants. The majority of the intervention group (72%, 23/32) used the DVD and rated it highly in terms of engagement and usability. Interviews with participants indicated that they found it to be helpful.

For fruit and vegetable intake, a small effect was observed for increased consumption for the intervention as compared with the control group at 3-month follow up, primarily attributable to increased fruit intake (d=.33) versus vegetables (d=0.12). The mean difference between intervention and control in combined fruit and vegetable score would be comparable with about 0.28 standard servings per day. Intensive, multisession telephone-counseling studies conducted with survivors have generally observed increases of 0.5-0.9 servings a day at the 3-month follow up [17,18]. Thus, the results here make sense for a low-contact intervention. Nevertheless, effects on fruit and vegetable intake would meet the criteria for pursuing a larger trial, albeit with greater attention to intervention intensity to further improve results.

The results did not meet the criteria for physical activity. Surprisingly, declines in METs were observed in both intervention and control groups with a slightly greater decline in the intervention group compared with the control. We investigated additional analyses to provide further insight into this finding. As participants only had to have at least one risk factor (meaning some were already active) we analyzed results by baseline activity and found an interesting trend—those who were already physically active and who used the intervention had less of a decline in activity compared with those who did not use the intervention; participants who did not meet the physical activity criteria at baseline showed a trend for greater increase in activity compared with those who did not use the intervention (Multimedia Appendix 1). These follow-up findings were similar to the findings of Pinto et al [42], who found that those who were more active at the baseline regressing to the mean at follow up. It should also be noted that in this study, the control group reported more physical activity at the baseline than the intervention group. Other studies have also observed over-reporting of physical activity at the baseline [43], which increases the difficulty of observing changes at follow up. It is noteworthy that study follow-ups were conducted solely during the winter months in the Northeastern United States, such that seasonality may potentially explain the overall mean decreases in activity for an older survivorship population. The intervention’s focus on low-impact activities such as walking could also have led participants to decrease their activity during the winter months.

In terms of informing a larger trial, the recruitment plan and follow up generally went well, a small effect was observed on fruit and vegetable intake, and patients liked the program. Nevertheless, results for physical activity were disappointing. A number of improvements would be indicated prior to pursuing further work. In terms of recruitment, contacting patients via telephone appeared a difficulty given that few people now answer their phones. Although a more time-intensive method, it may be necessary to meet patients first in person to offer study enrollment versus on telephone or through email. Nowadays, we employ an “on-call” research assistant who can quickly come to the clinic when a provider identifies an interested patient. Follow up should be extended to 6 months or a year as time and funding limited the period of study here. In terms of measurement, while standard measures of diet and activity were used, these rely on self-report and recall. Using multiple 24-hour recalls (ie, ASA24) would also improve assessment of dietary practices, and which are now available on the Internet through the National Cancer Institute (NCI) [44]. Measurement of physical activity, in particular, would benefit from use of accelerometers and mobile heart rate monitors now available that can more accurately capture additional activities such as strength training. In addition, depending on location, physical activity studies should account for seasonality by conducting the intervention period over a longer time frame and by stratifying by season of recruitment.

In terms of behavioral endpoints, promoting lifestyle changes among cancer survivors remains a challenge. Studies of fruit and vegetable consumption indicate that it has tended to be an easier behavior to improve than physical activity, likely due to a number of factors; it does not require much additional time or scheduling changes, offers immediate reward, has few if any contraindications due to comorbidities, does not result in physical discomfort, and does not require large increases in knowledge or skills [21]. Physical activity, on the other hand, has been a particularly difficult intervention endpoint in cancer survivorship. A comprehensive review of physical activity interventions in cancer populations found that no studies reported 75% or greater adherence to the 150 min per week guideline, even when they used multisession counseling and supervised training sessions [21]. The most common barriers survivors report are being “too busy” and lack of “willpower,” factors which predict level of activity [45]. As survivors are also concerned about safety and comorbid medical conditions, combining introductory in-person demonstrations [46] along with an interactive self-guided program would better address barriers related to self-efficacy, motivation, and time. Achieving and sustaining robust behavior change will likely require further contacts and specific goal setting and monitoring, enhancements that have been linked to increased behavioral adherence [21,47]. Interviews with participants indicated interest in having the resource available via a mobile platform that would enable additional features such as tracking and goal setting. Indeed, a mobile phone app modality to present information, combined with automated tailored text messaging, to provide ongoing intervention components may be one strategy to integrate these features into an updated intervention. Recent reviews have called for interventions that can be more readily disseminated to a more diverse range of survivors beyond those who can attend inperson sessions at major cancer centers [21]. This study provided important insights that can be integrated into a more intensive mobile-based intervention, which is planned in a future trial.