Engaging Men of Diverse Racial and Ethnic Groups With Advanced Prostate Cancer in the Design of an mHealth Diet and Exercise Intervention: Focus Group Study

Introduction

Background

Healthy diet and exercise have been shown in numerous observational studies and randomized controlled trials to improve quality of life, treatment-related adverse effects, and prognosis among men with prostate cancer [1-7]. However, the ability to initiate and sustain healthy diet and exercise habits is contingent on contextual factors, skills, preferences, and perceptions, which are further constrained by patients’ time and resources [8]. Consequently, there are numerous barriers to the effective design and implementation of interventions to improve the quality of life for men with advanced disease [9].

Mobile health (mHealth) interventions, defined by the World Health Organization as “Medical and public health practices supported by a mobile device, such as mobile phone, patient monitoring devices, personal digital assistants and other wireless devices” [10], are becoming increasingly common and are a promising approach for increasing physical activity and modifying dietary behaviors by supporting goal setting, self-monitoring, and instruction and providing feedback about lifestyle changes [11]. However, most of the participants in the studies conducted so far identified as White. More studies are needed to assess the feasibility of and preferences for mHealth interventions that include underrepresented populations. Qualitative studies are uniquely equipped to identify barriers to care and areas of concern for patients, particularly those from vulnerable populations. A recent qualitative study in Taiwan explored the experience of men undergoing androgen deprivation therapy (ADT), which ultimately concluded the need for great emphasis on the provision of topically relevant educational materials, avenues for emotional support, and opportunities to gain improved coping mechanisms [12]. Another recent study, including participants with prostate cancer, explored the role of partner support in cancer survivorship [13]. Studies such as these highlight the complexity of survivorship experience and the need for further qualitative studies.

Objective

Given the importance of healthy lifestyle habits, well-documented disparities in prostate cancer care, and need for remote mHealth interventions, we conducted a qualitative study exploring diet and lifestyle behaviors among a racially and ethnically diverse cohort of men with advanced prostate cancer, to guide the development of an educational intervention focused on men treated with ADT. Findings from this qualitative study may also inform the design and delivery of future mHealth interventions in diverse populations.

Methods

Design

There was a cross-sectional mixed methods study designed to examine preferences, attitudes, and health (PATH) behaviors in men with advanced prostate cancer via a web-based exercise and food habit survey and focus groups. Sampling was purposive to ensure that men from diverse racial and ethnic groups were included. English-speaking and Spanish-speaking participants (n=104) were recruited between July 6, 2019, and November 11, 2020, at the University of California, San Francisco (UCSF); Zuckerberg San Francisco General; and San Francisco Veterans Affairs hospitals. The study was introduced by the study clinician (principal investigator; HTB), clinician (TF), or clinical research coordinator (SZ or ET). The clinical research coordinator screened potential participants for eligibility by reviewing oncology clinic schedules and electronic health records and then approached these patients in the clinic, by phone, or by email to participate in the study. Clinicians also introduced patients to the clinical research coordinator in the clinic, who then introduced them to the study. Participants were aged ≥18 years, diagnosed with hormone-sensitive prostate cancer, on hormone therapy, able to read English or Spanish, and able to understand written informed consent. Participants had metastatic hormone-sensitive prostate cancer if recruited from UCSF; we allowed participants in the community to have metastatic or nonmetastatic hormone-sensitive prostate cancer and did not verify metastasis status for these participants. Any man with any self-reported cognitive or neurologic condition that, in the opinion of the study team, would prohibit the ability to read and navigate the internet or follow a diet or exercise prescription independently were excluded. Recruitment in the community setting occurred through Facebook and Google advertisements; through oncologists at Kaiser Oakland hospital; and at community-based events including church events, support groups, and so on, by a community health educator and outreach or engagement coordinator to increase sample size and include a wide range of perspectives.

Overall, 36 PATH study participants consented to be further contacted by the research team regarding optional study procedures (African American or Black: 14/28, 50%; Asian American or Native Hawaiian or other Pacific Islander or other: 7/9, 78%; Hispanic or Latino: 9/22, 41%; and White: 6/40, 15%). These participants were invited to a focus group via phone or email. All patients provided informed consent. Focus groups were stratified according to self-identified race and ethnicity. Overall, 14 participants—3 ( 21%) Asian American participants, 3 (21%) African American or Black participants, 3 (21%) Hispanic or Latino participants, and 5 (36%) White participants—attended focus groups between April and November 2020. Each participant received a gift card worth US $50 for participation in the PATH study, and focus group participants received an additional gift card worth US $50.

Focus Groups

Focus groups were conducted by researchers with expertise in urologic cancers, lifestyle, and associated disparities (SAK: non-Hispanic Native Hawaiian, Asian American, White female associate professor of Urology and Epidemiology & Biostatistics; HTB: non-Hispanic Middle-Eastern female assistant professor of Hematology/Oncology; SLW: non-Hispanic African American male assistant professor of Urology and Epidemiology & Biostatistics; and SZ: non-Hispanic Asian American female research coordinator). SZ was the primary contact for study participants. Focus groups were conducted in English and recorded via Zoom (Zoom Video Communications) video software. Participants were asked about their experience with and perceptions regarding various lifestyle tools (website, wearable technology, etc). For the interview guide, refer to Multimedia Appendix 1. Focus groups lasted 60 to 90 minutes and were transcribed using an external service. Data were deidentified. To optimize transferability, we also explored how diet and exercise were affected by the COVID-19 pandemic.

Grounded Theory Analyses

We used a grounded theory approach [14,15]. The grounded theory methodology is well suited for investigating topics without substantial previous qualitative literature owing to its characteristic emphasis on open or data-driven coding versus theory-driven analysis. EYW conducted the initial paragraph-by-paragraph open coding manually and the subsequent coding in ATLAS.ti (version 9). Open codes were refined into axial codes and selective codes (categories) using embodied categorization [16] and constant comparison methods [17]. Codes and categories were finalized with other investigators (HTB, SZ, SLW, and SAK). We report findings in adherence with COREQ (Consolidated Criteria for Reporting Qualitative Research) [18].

Data Saturation

The number of focus group participants required to reach data saturation is debated and largely dependent on the scope of the topic of interest [19]. We designed this study to balance privacy and data saturation. ADT can have a wide range of side effects, including hot flashes, loss of muscle mass, increased fat mass, weight gain, lowered libido, erectile dysfunction, and reduced quality of life. To respect the potentially sensitive and culturally specific aspects discussed in the focus groups related to the cancer diagnosis, cancer treatment, and diet and lifestyle habits, we used small groups and assigned men to racially and ethnically concordant focus groups. Given the narrow and focused nature of the research question (Multimedia Appendix 1), few participants were required to reach saturation. In consideration of the small number of participants within each focus group, the transcripts were analyzed as a whole and presented together. Codes that were only represented in a subset of focus groups are specified.

Ethics Approval

The study was conducted in accordance with the Declaration of Helsinki and approved by the institutional review board (or ethics committee) of UCSF (protocol number 19-27137; March 18, 2019).

Results

Overview

Self-reported characteristics of focus group participants are presented in Table 1. The mean age was 67 (SD 8.9) years, with racial and ethnic composition of 21% (3/14) African American or Black, 21% (3/14) Asian American, 21% (3/14) Hispanic or Latino, and 36% (5/14) White. Most participants were retired (10/14, 71%), had Medicare insurance (11/14, 79%), and had a 4-year college degree or higher (11/14, 79%). Approximately half (8/14, 57%) of the participants were married. All participants (14/14, 100%) were found to have adequate health literacy based on a validated survey [20]. These men were diagnosed with prostate cancer an average of 4 years before enrollment in the study, and many (8/14, 57%) had Gleason grades of 8 to 10.

Analyses yielded 67 open codes, 25 axial codes, and 7 selective codes (categories), which are presented in Figure 1. These seven categories include (1) context (home environment, access, competing priorities, and lifestyle programs), (2) motivation (accountability, discordance, feeling supported, fear, and temptation), (3) preparedness (health literacy, technological literacy, technological preferences, trust, readiness to change, identity, adaptability, and clinical characteristics), (4) data-driven design (education, psychosocial factors, and quality of life), (5) program mechanics (communication, materials, customization, and being holistic), (6) habits (eg, dietary habits), and (7) impressions (regarding the intervention; Figure 1). Each code represents an actionable component, as demonstrated by the participant quotes in the following sections. Illustrative quotes are organized according to 7 categories (column 1 in Figure 1) and open or axial codes (green or blue boxes, respectively, in Figure 1) for the design and delivery of mHealth interventions. Codes represented in all focus groups are bolded, and codes not represented in all focus groups are italicized (Figure 1). Quotes have been edited for clarity and to illustrate the breadth of responses representing selected codes (Table 2).

‎

Special Cases

Codes only represented in a subset of focus groups are presented in Table 3. Safety was mentioned in all groups, but notably, police were noted only in the African American or Black group. Identity contributed to preparedness in the Hispanic or Latino, African American or Black, and Asian American groups, with some clinical characteristics affecting preparedness across all groups. White and Asian American groups generated similar codes for data-driven design and program mechanics, including relevance, efficiency, and tailoring.

The relationships among these codes (Figure 2) represent actionable pathways to increase program intuitiveness for survivors of prostate cancer engaged in mHealth interventions that could occur via multiple strategies. For example, we might increase motivation by performing a detailed intake assessment using an intake form to characterize participants’ preparedness that can be used to provide a tailored step-wise program, understanding the participants’ home environment, assessing the influence of other family members on diet and exercise and involving them in lifestyle goals and plans, understanding the participants’ preferences for communication for better participant engagement, and tailoring educational material and behavior change plans to the participant using a customized approach. These and other grounded theory–based solutions (Multimedia Appendix 2) may result in a more engaging and integrated intervention for survivors of prostate cancer, which could improve benefits. Broad themes noted in Multimedia Appendix 2 focus on assessment at the individual, household, and neighborhood levels to support a tailored intervention; tailoring of the intervention to the patient where possible (eg, considering the individual’s health and technological literacy, communication preferences, baseline responses and major concerns, home or neighborhood environment, and support network); and implementing strategies to foster engagement during the intervention (eg, feedback systems, routine check-ins, earning, and sustaining trust).

‎

Discussion

Principal Findings

The purpose of this paper was to elucidate the perspectives and attitudes surrounding lifestyle change in racially or ethnically diverse men with advanced prostate cancer, as this has not previously been studied. Our results suggest that lifestyle-related preferences, needs, and limitations of men with prostate cancer from diverse racial and ethnic backgrounds are affected by multiple inherent, learned, and contextual dimensions, precluding a one-size-fits-all approach to intervention design for men of any given race and ethnicity. Lifestyle interventions may be improved and tailored to the individual by leveraging these components and their interrelationships. The findings from this study are informing a digital platform that provides lifestyle resources and support for men receiving ADT (supportive therapy in androgen deprivation–technology; ClinicalTrials.gov NCT05324098).

So far, few studies have qualitatively explored the experiences of diverse groups of survivors of prostate cancer. The recent COVID-19 pandemic has presented novel challenges, especially among minoritized racial or ethnic populations [21,22] and an increased urgency to optimize remote interventions, particularly for patients from minoritized racial or ethnic groups who have been underrepresented in clinical trials [11,23]. Given the highly social nature of both diet and exercise, race, ethnicity, and other factors related to social determinants of health also likely influence the implementation of lifestyle interventions. Although lifestyle interventions will not mitigate the negative effects of systemic and policy-driven contributors to racial disparities, a design that incorporates the multilevel nature of these issues should address an individual’s experience of detrimental systemic and societal influences.

Many codes under “context” (home environment and access) and “preparedness” (literacy, identity, and adaptability) represent the downstream effects of social determinants of health [24,25] or “factors that involve a person’s relationships to other people” including race, ethnicity, socioeconomic status, and gender identity [24]. To add to the Fundamental Causes Theory, Riley [26] has challenged researchers to take a more nuanced “systems of exposure” approach and to blend theories such as spatial polygamy, intersectionality, systems theory, and the life course perspective. The theory of intersectionality proposes that social identities interact at multiple levels of oppression to collectively influence health outcomes [27,28]. Applied to lifestyle interventions, the interactions among each participant’s various social identities need to be understood at baseline and again at incremental time points, and the intervention needs to be comprehensively tailored to participants’ evolving identities and social environment. The importance of a comprehensive and tailored approach is further illustrated by the breadth and interconnectedness of the codes we observed, demonstrating the intersectionality of the multiple facets of participants’ lives and perspectives influencing behavior change over time.

As a first step, the codes generated in this study may serve as a preliminary guide for designing a comprehensive intake form. The breadth and interrelatedness of codes generated by participants signaled the need for a holistic and integrated mHealth intervention design; for example, our recommendations include providing education about normal adverse effects of prostate cancer treatments and the evidence surrounding diet and exercise recommendations as they relate to energy, strength, and motivation and gaining an understanding of participants’ current habits to identify realistic and priority areas for change (Multimedia Appendix 2). Future interventions should focus on increased tailoring that could include prioritizing information to disseminate based on participants’ major concerns, health literacy, and technological preferences; prescribing personalized educational materials and interventions based on individuals’ baseline responses; and incorporating responsive and relevant feedback systems to aid participant decision-making and behavior change in real time. In addition, high-technology interventions may pair well with high-touch aspects such as a patient navigator model for patients with limited technological literacy. The navigator role could be reimagined to provide digital intervention–related support to patients, such as assistance with using internet-based resources (eg, a study web portal), setting up and using app-based devices (eg, Polar heart rate monitors and Fitbit devices that connect to smartphone apps), and setting up video visits (eg, Zoom-based coaching visits).

Every category was constructed with input from all focus groups, but certain codes were not represented in every focus group (Table 3). These variations should not be overinterpreted to signify differences between racial and ethnic groups; however, certain themes appeared in groups for which those themes are most prevalent and relevant. Additional studies are needed to identify unique combinations of themes across groups and to assess which themes are most relevant for different groups. Race and other social constructs are dynamic, and certain intersections will be most salient based on the research focus and the population studied [29]. The patients’ context, motivation, and preparedness that may be associated with race; ethnicity; and other factors associated with social determinants of health such as income, access to nutritious foods, and neighborhood characteristics should be considered when formulating an individualized plan for each patient and when discussing the barriers and solutions that will help them to make and maintain healthy behavior changes.

Limitations and Strengths

Limitations of the study include the small subgroup sample size. Overall, 13.5% (14/104) of the eligible participants were both interested and available to participate in the focus groups at scheduled times. Although our sample size is acceptable because our objectives were to explore themes using a grounded theory approach, the absence of theoretical sampling precluded certainty of data saturation. However, open coding minimized researcher assumptions. Our focus groups of 3 to 5 participants provided a more intimate environment for people to share their experience with cancer, their treatments, side effects, and so on and thus was effective for eliciting responses to potentially sensitive research questions such as ours [30]. The corroboration of our findings with other previous studies of prostate cancer, which similarly highlighted important themes related to context (eg, identity), motivation, preparedness (eg, competencies), and mechanics (eg, tailored feedback and goal setting) to consider for a successful intervention [31,32]; consistency with prominent public health theories; and inclusive recruiting bolster study validity as defined by Whittemore et al [33] (credibility, authenticity, criticality, and integrity). Interview guides did not explicitly probe how race or culture played a role in lifestyle change, but the diverse focus groups enabled us to identify more specific themes surrounding social environment and individual-level factors influencing receptiveness to lifestyle change compared with a similar study in a less diverse group [8]. Our participants were well educated and demonstrated adequate health literacy, limiting the generalizability of our findings to broad groups. However, the study’s strengths include the inclusion of racially or ethnically diverse participants and researchers, insights during an acute stressor (COVID-19 pandemic), and consistency with previous theories around this topic. This study highlights the need for future ethnographies and in-depth interviews to explore these concepts in participants from diverse racial or ethnic, socioeconomic, and educational backgrounds.

Conclusions

The discussions with focus groups of racially and ethnically diverse patients with prostate cancer about mHealth lifestyle interventions support a tailored approach that leverages the identified components and their interrelationships to ensure that the final intervention will engage and be effective in diverse patients with a cancer diagnosis. Addressing the home environment and patients’ roles related to diet and exercise in the household, access (to food and exercise), competing priorities, health and technological literacy, readiness to change, and clinical characteristics will help to customize the intervention to the participant. This study provides preliminary evidence that multiple dimensions should be considered in behavior change interventions and that each contributes to the totality of an individual’s social identities and contexts that influence dietary and exercise behaviors. Thus, an intersectional approach to tailoring interventions for all men that accounts for their needs based on an assessment of their context, motivation, preparedness, habits, and impressions, while incorporating design and program mechanics preferences of the participant, would most likely enhance prostate cancer survivorship.