Development and pilot testing of an mHealth behavioral cancer pain protocol for medically underserved communities

Abstract

The purpose of this study was to refine and test a mobile-health behavioral cancer pain coping skills training protocol for women with breast cancer and pain from medically underserved areas. Three focus groups (Phase 1) were used to refine the initial protocol. A single-arm pilot trial (Phase 2) was conducted to assess feasibility, acceptability, and changes in outcomes. The intervention was delivered at a community-based clinic via videoconferencing technology. Participants were women (N = 19 for Phase 1 and N = 20 for Phase 2) with breast cancer and pain in medically underserved areas. Major themes from focus groups were used to refine the intervention. The refined intervention demonstrated feasibility and acceptability. Participants reported significant improvement in pain severity, pain interference, and self-efficacy for pain management. Our intervention is feasible, acceptable, and likely to lead to improvement in pain-related outcomes for breast cancer patients in medically underserved areas.

Introduction

Patients with cancer report pain to be their most distressing and feared symptom.^1,2 Patients with more pain report higher levels of physical disability, physical symptoms, distress, financial difficulty, and lower overall health-related quality of life.^3,4 Higher levels of pain are also related to decreased survival time in breast and other cancers.⁵

Patients who live in medically underserved areas are particularly likely to experience high levels of pain and pain-related disability.⁶ Medically underserved areas are defined as US regions where there is a relative or absolute deficiency of healthcare resources,⁷ and many patients in these areas have low literacy levels.⁸ Women with breast cancer in medically underserved areas experience high levels of pain due to later stage at diagnosis⁹ and higher likelihood of having comorbid health problems that contribute to pain.¹⁰

Increasing evidence suggests interventions that teach cancer patients behavioral (e.g., relaxation) and cognitive skills (e.g., cognitive restructuring) to manage their pain are an efficacious adjuvant therapy to analgesics.^11–14 These protocols are traditionally delivered to patients through in-person sessions with an expert pain therapist at a major medical center. Recent guidelines recommend that behavioral pain interventions be integrated into cancer treatment.¹ Yet, behavioral cancer pain interventions are often absent in medically underserved areas¹⁴ due to persistent access barriers. Even for patients in areas with high resources there are persistent barriers to accessing in-person behavioral interventions including time constraints, cost, transportation difficulties, and distance from the medical center.^15–17 Patients in medically underserved areas face additional barriers to accessing these interventions, including lack of pain experts at community-based clinics, prohibitive travel to major medical centers, and lack of intervention protocols adapted for their needs (e.g., low literacy).

Advances in mHealth technologies provide new opportunities to implement behavioral cancer pain interventions in medically underserved areas. mHealth technologies can decrease critical barriers (e.g., costs & travel) that limit access to in-person behavioral interventions.^15–17 Studies examining mHealth strategies to manage chronic medical illnesses (e.g., diabetes) and psychiatric disorders (e.g., obsessive-compulsive disorder) have found that mHealth delivery is feasible and acceptable to patients and can improve outcomes.^18–20 However, mHealth advances in behavioral pain protocols have not yet been extended to underserved populations of cancer patients who may face additional technology-related barriers, including poor Internet access.

The objective of this study was to refine and test an mHealth behavioral cancer pain coping skills training protocol to reduce cancer pain and disability in women with breast cancer who live in medically underserved areas. We developed an initial version of mPCST-Community that was adapted from a Pain Coping Skills Training protocol developed by our team and typically delivered through in-person sessions.¹⁴ mPCST-Community was created with the input of a nationally recognized expert (author B.T.) in the development and/or adaptation of behavioral pain interventions for chronic pain conditions in patients living in rural areas with low literacy and low income.^21–23

Study aims were to (1) conduct focus groups with women who have breast cancer and pain in medically underserved areas to refine the initial mPCST-Community protocol and (2) pilot test the fully developed mPCST-Community protocol for feasibility (i.e., meeting target accrual, at least 80% completing the protocol, and at least 75% completing all sessions), acceptability (i.e., high satisfaction ratings), and pre- to post-intervention changes in pain and pain-related symptoms (i.e., pain severity, pain interference, pain catastrophizing, psychological distress, fatigue, physical functioning, & self-efficacy for pain management) in 20 women with breast cancer receiving care in medically underserved areas.

Methods

Setting and participants

This study included focus groups (Phase 1) and pilot testing (Phase 2). Participants were women who had received a diagnosis of breast cancer in the last three years and were treated in a medically underserved area. Additional eligibility criteria included: (1) age ≥21 years; (2) life expectancy ≥12 months; and (3) a pain rating ≥3 (0 to 10 scale) gathered at clinic visits, or reported pain ≥3 at least three days in the last 2 weeks at the time of screening. Exclusion criteria included: (1) cognitive impairment; (2) brain metastases; (3) severe psychiatric condition (e.g., psychosis) identified by provider or medical chart review; and (4) current or past (<6 months) engagement in pain coping skills training.

Study procedures received institutional review board approval (Pro00065621; Clinicaltrials.gov NCT02783755). Recruitment took place between April 4, 2016 and May 26, 2017 at three community-based clinics in rural, medically underserved communities (Laurinburg, NC; Henderson, NC; Smithfield, NC). Local staff used electronic medical records to identify participants using consecutive sampling. A letter signed by the treating oncologist and study PI was sent to the patient introducing the study. Patients were contacted via telephone after the letter was mailed. Interested patients met with the study staff to complete the consent process. All participants provided written informed consent.

Phase 1: Focus groups

Phase 1 utilized focus groups (three groups; N = 19). Information obtained from participants was used to refine the mHealth pain coping skills training (mPCST-Community) protocol. Ninety minute focus groups were held in community clinics. Groups were moderated by the PI (author T.S., PhD Clinical Psychology, female) and a member of the research team (authors C.D. or R.S., PhDs Clinical Psychology, females) using a semi-structured qualitative interview guide developed for this study. The team had previous experience conducting qualitative interviews/analyses with cancer patients and had great clinical interest in developing an appropriately adapted intervention. The interviewers had no previous relationships with participants prior to study commencement. The participants were informed about the purpose of the focus groups (i.e., to refine a behavioral pain management protocol). Only the researchers and participants were present during the groups. Groups were audio-recorded, and facilitators took field notes. Participants discussed their pain experiences and provided input on intervention content and components. Groups were conducted sequentially to allow for iterative updates to the protocol. Repeat interviews were not conducted. Recordings were analyzed using methods from qualitative content analysis.^24,25 Key findings were used to refine the protocol (Table 1).

Table 1.

Qualitative findings on intervention content and associated intervention changes.

Key content	Summary of patient text related to key content	Guide to action: intervention refinement
Bothersome pain and fatigue	Participants acknowledged that while pain was bothersome, fatigue was also bothersome. Many patients reported that both pain and fatigue had been persistent since their treatments.	Added session content on the relationship between pain and fatigue
Lack of psychosocial services	Participants acknowledged that there were few psychosocial services available to them. Participants consistently reported that they would have used psychosocial services during their diagnosis and treatment. Participants could identify one/two medical providers at their clinic who provided support, but there were few/no structured psychosocial programs available.	Incorporated and emphasized supportive material throughout
Some hesitancy about mobile technology for psychosocial services	Participants desired clear guidance about using the videoconferencing technology in their community clinic. They were enthusiastic about receiving the intervention in this way, if the technology would work. Some were hesitant about whether videoconferencing would work at their home. Participants requested explicit instructions and study team contact information in case they were not sure how to use the smartphone components.	Decided on clinic-based videoconferencing Created specific and detailed instructions

Phase 2: Pilot testing

Phase 2 was a single-arm pilot trial (N = 20) testing the fully developed mPCST-Community protocol (Table 2). Sample size was determined based on similar pilot studies. Participants received five 50-minute individual sessions delivered weekly at the community clinic where they received their cancer care via videoconferencing technology. Study therapists were PhD-level, licensed clinical psychologists with experience in pain management. Study therapists delivered the intervention remotely, about 40–70 miles from the community clinic. Participants received a smartphone to extend the intervention into their own environment. It provided daily access to session content, video-clips modeling coping skills, stories about pain experiences from example participants, and other materials (e.g., relaxation audio). Participants were prompted to use the smartphones to send daily pain ratings and skills practice experiences to their therapist. These data were used to tailor text messages to promote self-efficacy based on the social cognitive theory²⁶ concepts (e.g., modeling). Participants could use videotext to share their skills practice experience with their therapist and receive feedback.

Table 2.

Description of the fully developed protocol.

Session content	Examples of adaptations/refinements
Session 1. The therapist presented a rationale for the intervention based on a simplified version of the gate control theory of pain, which describes pain as an interaction of thoughts, feelings, and behavior. Progressive muscle relaxation (PMR) and guided imagery were taught by modeling and guiding participants through practices. The therapist reinforced practice, problem solved issues, and encouraged participants to identify benefits from the practice. Participants were given an audio-recording of PMR and guided imagery on their smartphone, and asked to practice daily.	• Simplified gate control theory rationale • Increased graphics, decreased text for complex interactions of thoughts, feelings, and behavior • Smartphone audio with stories from other patients • Face graphics added to pain rating scale • Smartphone audio with session summaries
Session 2. Participants were taught activity-rest cycling to remain productive/active without overexerting and exacerbating pain. Pleasant activity planning was also taught to increase enjoyable activities. Participants received goal-directed instruction, encouraging them to participate in chosen activities.	• Graphics of rural scenes used for activity-rest cycles • Changed wording from pleasant activities to fun activities • Included activities mentioned in focus groups (e.g., church)
Session 3. Cognitive restructuring was taught, which included recognizing how thoughts can negatively influence pain and coping with pain. Negative pain-related thoughts not only impact pain, but also physical ability and well-being. Common examples of negative thinking were presented.The therapist and participant worked to identify negative thoughts, define negative consequences, and generate a more neutral/positive thought with better consequences.	• Referred to cognitive restructuring as “positive thinking” • Added graphics to illustrate concepts • Eliminated other session topics given the complexity of cognitive restructuring
Session 4. Guided imagery was reinforced as a central skill. The gate control theory was reviewed. Participants were encouraged to brainstorm about memories from pleasant scenes.Participants were guided through imagery practice, and asked to continue daily practice.	• Increased graphics for imagery • Focused on modeling to increase self-efficacy
Session 5. Participants were taught problem solving skills related to pain and goal setting to help them continue skills practice.	• Simplified goal-setting worksheet completed collaboratively • Smartphone audio with overall summary

Study measures

Demographic and medical information was collected via medical record review and self-report using a secure web-based assessment (i.e., Qualtrics). Health Literacy was assessed using four items.²⁷ Participants rated their ability to read and understand medical information and forms. Items were summed, with higher scores indicating worse health literacy. The mean score was 6.74 (SD = 2.45, range = 4–13). Ninety percent of participants reported difficulty with health literacy at least “a little of the time” on one or more items.

Participants in the pilot trial (Phase 2) completed the following measures pre- and post-intervention. All outcome measures have been well-validated for use in cancer populations.

Pain Severity and Pain Interference were assessed with the Brief Pain Inventory.²⁸ Pain Severity was computed as the average of four items that assess patients’ pain at its worst, least, and average in the past week as well as current levels of pain. Pain Interference was computed as the average of seven items that assess the degree to which pain interfered with patients’ daily activities and other aspects of life in the past week. Internal consistency was good for both subscales (Cronbach’s α = 0.84 and 0.92, respectively).

Pain Catastrophizing was assessed using the 6-item pain catastrophizing subscale of the Coping Strategies Questionnaire.²⁹ These items ask about patients’ tendency to catastrophize when faced with pain. Items were averaged and higher scores indicate worse pain catastrophizing. Internal consistency was good for this sample (Cronbach’s α = 0.90).

Psychological Distress was assessed using the 4-item Patient Reported Outcomes Measurement Information System (PROMIS)-Depression Scale,³⁰ which is a self-report measure of depressive symptoms experienced in the past week. Items were summed and converted to T-scores. Internal consistency was good (Cronbach’s α = 0.95).

Fatigue was assessed using the 4-item PROMIS-Fatigue Scale,³¹ which is a self-report measure of fatigue symptoms experienced in the past week. Items were summed and converted to T-scores. Internal consistency was good (Cronbach’s α = 0.89).

Physical Functioning was assessed using the 7-item Functional Assessment of Cancer Therapy-General Physical Well-Being Scale,³² which assesses a variety of physical symptoms and concerns experienced in the past week. Responses were reverse-coded and averaged, such that higher scores indicate better physical functioning. Internal consistency was fair in this sample (Cronbach’s α = 0.79).

Self-Efficacy for Pain Management was assessed using the self-efficacy for pain management subscale of the Chronic Pain Self-Efficacy Scale.³³ This subscale contains five items that ask about patients’ confidence or certainty in their ability to manage pain. Items were averaged and higher scores indicate greater self-efficacy. Internal consistency was good (Cronbach’s α = 0.86).

Feasibility was assessed by examining participant accrual, attrition, and adherence. Acceptability was assessed post-intervention with the 10-item Client Satisfaction Questionnaire,³⁴ which assesses degrees of satisfaction with treatment. Internal consistency was good for this sample (Cronbach’ s α = 0.83).

Data analysis plan

For Phase 1, focus group data were evaluated in a systematic format using methods from qualitative content analysis with the goal of creating a practical guide for action (i.e., refining the intervention protocol).^25,35 Audio-recordings of focus groups and field notes were reviewed by two members of the research team. A deductive process was used to identify key concepts most relevant to the study aims²⁴; the unit of analysis was the whole interview and analyses were limited to manifest content.²⁵ The study team agreed that saturation was reached after three focus groups.

For Phase 2, feasibility and acceptability were characterized using descriptive statistics. Missing data were consistent with what would be expected in an intervention trial, ranging from 10% to 20% on some pre- and post-intervention pairs of variables.¹² Simulation studies demonstrate that the accuracy of parameter estimates can be improved by imputing missing data with auxiliary variables (e.g., demographics, medical).^36,37 Ten datasets were imputed using a Markov Chain Monte Carlo algorithm that included demographic and medical variables and scores on other outcomes as predictors. Pre- and post-intervention scores were compared using pooled, paired-samples t-tests.³⁸ Effect sizes were computed using Hedges’ g_av, which is recommended for repeated-measures designs with small samples.³⁹ Hedges’ g_av is interpreted similarly to Cohen’s d (i.e., 0.2 = small, 0.5 = medium, & 0.8 = large effects).

Results

Results from phase 1: Focus groups

Forty-two patients were approached, and 34 were deemed eligible. Of those 34, 15 declined participation due to: (1) lack of time, scheduling conflicts, or feeling overwhelmed (n = 4); (2) distance/transportation concerns (n = 3); (3) family concerns (n = 3); (4) lack of interest (n = 2); (5) health reasons (n = 1); (6) no pain (n = 1); and (7) no reason given (n = 1). Overall, 19 women attended one of three focus groups. Participants were on average 60 years old (SD = 9.96), and most were married (63%), African-American (53%), and did not have a college degree (75%). The average pain score was 3.2 (SD = 2.2) in the last week.

Qualitative content analyses identified three key content areas that were used to create a guide to action to refine the mPCST-Community protocol: (1) bothersome pain and fatigue; (2) a dearth of psychosocial services; and (3) some hesitancy about using a smartphone (Table 1). Table 2 provides a summary of the developed session content.

Results from phase 2: Pilot testing

Eighty-five patients were approached, and 60 were deemed eligible. Of those 60, 40 declined due to: (1) lack of interest (n = 13); (2) no reason given (n = 15); (3) distance/transportation concerns (n = 3); (4) family concerns (n = 3); (5) health reasons (n = 2); (6) lack of time, scheduling conflicts, or feeling overwhelmed (n = 2); and (7) no pain (n = 2). Overall, 20 patients were consented for the study, and 18 completed post-intervention assessments. Participant characteristics are presented in Table 3.

Table 3.

Pre-intervention participant characteristics (N = 20).

	N or mean (SD)	%
Age	57.85 (11.72)
Race
African American	15	75
Caucasian	5	25
Marital Status
Married	14	70
Not married	6	30
Education
Some high school or less	1	5
High school graduate	9	45
Some college or advanced vocational training	5	25
College degree or higher	5	25
Employment
Working full time for pay	7	35
Retired or homemaker	5	25
Self-employed	2	10
Disability	4	20
Other	2	10
Income
Less than $10,000	7	35
$10,000 to $19,999	4	20
$20,000 to $39,999	2	10
$40,000 to $59,999	4	20
$60,000 to $100,000	3	15
Time since diagnosis in years	1.35 (0.69)
Stage
0	2	10
I	5	25
II	6	30
III	7	35
Number of comorbidities
0	2	10
1	5	25
2 or more	13	65

The fully developed mPCST-Community protocol was feasible and acceptable. There was good accrual, as we reached our recruitment goal of 20 participants in 10 months with limited staffing. Attrition was low and adherence was excellent: 90% of participants completed all intervention sessions and assessments. Acceptability was demonstrated by high client satisfaction ratings (mean = 3.7, SD = 0.27).

All outcomes demonstrated changes in the hypothesized directions (Table 4). Based on pooled results (N = 20) from paired samples t-tests, there were significant pre- to post-intervention improvements in pain severity, t=−2.52, p = 0.01, 95% confidence intervals (CI) for mean difference [−2.13, −0.27]; pain interference, t =−2.62, p = 0.01, 95% CI [−2.72, −0.39]; and self-efficacy for pain management, t = 3.57, p = 0.0004, 95% CI [9.01, 30.90]. The effect size estimates for these differences were classified as medium for pain severity, g_av =0.57, 95% CI [0.10, 1.08]; medium for pain interference, g_av=0.62, 95% CI [0.12, 1.16]; and large for self-efficacy for pain management g_av =0.98, 95% CI [0.37, 1.66]. There were small (i.e., g_av<0.50), non-significant changes in pain catastrophizing, t =−1.02, p = 0.31, 95% CI [−1.12, 0.36]; psychological distress, t =−1.85, p = 0.06, 95% [−7.98, 0.24]; fatigue, t =−1.58, p = 0.11, 95% CI [−6.61, 0.71]; and physical functioning, t = 1.38, p = 0.17, 95% CI [−0.12, 0.66].

Table 4.

Study outcomes at pre- and post-intervention.

	Score mean (SD)		Effect size and 95% CI
Outcome	Pre-intervention	Post-intervention	gav	Lower	Upper
Pain severity	4.12 (1.92)	2.97 (2.20)	0.57	0.10	1.08
Pain interference	4.38 (2.40)	2.79 (2.47)	0.62	0.12	1.16
Pain catastrophizing	1.54 (1.51)	1.13 (1.43)	0.26	−0.23	0.76
Psychological distress	52.97 (10.13)	49.17 (7.42)	0.44	−0.04	0.94
Fatigue	56.64 (8.77)	53.68 (7.37)	0.36	−0.11	0.85
Physical functioninga	2.67 (0.79)	2.95 (0.74)	0.34	−0.11	0.81
Self-efficacy for pain managementa	50.40 (21.20)	70.33 (18.70)	0.98	0.37	1.66

Note. ns = 18–20 for outcome scores due to missingness. Effect sizes are Hedges’ g_av for paired samples, which were calculated using pooled values from 10 imputed datasets.

^aHigher scores represent greater physical functioning and self-efficacy for pain management.

Discussion

This study used focus groups and a small single-arm trial of the intervention to refine and test an mHealth behavioral cancer pain coping skills training protocol designed to meet the needs of women with breast cancer receiving care in medically underserved areas. This intervention was adapted to be cost-effective and sustainable for patients in communities with limited resources and barriers to traditional psychosocial intervention methods. The developed intervention has a high potential to meet the needs of patients in these areas as it reduces barriers to receipt of traditional behavioral cancer pain interventions such as low literacy, distance to a medical center with trained therapists, and lack of locally trained therapists. This work is important as women with breast cancer and pain who receive care and/or live in medically underserved areas are likely to have more pain and fewer options for pain management.^6,9,10 Patients frequently use analgesics as the primary treatment for pain;¹ however, increasing evidence suggests that behavioral pain interventions are an efficacious adjuvant therapy.^11–13

Focus group results found that women were enthusiastic about receiving a behavioral cancer pain intervention. Group participants confirmed the use of an iPad+Skype kiosk at the community clinic as the delivery format given participants’ concerns about internet connectivity for videoconferencing in their homes. Additionally, the focus groups were conducted in the community clinics, providing the study team with important information about participants’ comfort with the clinic environment, and that there were close relationships between patients and providers and other staff in this setting. These relationships and participants’ familiarity with this setting added reassurance that support would be available to the participant for this community clinic, technology-based intervention.

Focus group participants consistently reported that they did not receive the offer for any formal psychosocial support during their cancer diagnosis and treatment. This finding is not surprising given that medically underserved areas often have inadequate numbers of primary care providers let alone any subspecialty care providers (e.g., psychology). Participants also frequently suggested that the intervention address fatigue as this was an ongoing concern for them. Increased attention was given to both distress and fatigue in the context of the pain protocol as many of the coping strategies introduced might be expected to benefit patients in several areas (i.e., pain, distress, & fatigue).

While some women described concerns about their abilities to use mobile health technology, most participants reported being willing to try the mobile technology, especially if they had received training from study staff and detailed instructions for using the equipment. Both written (with pictures) and verbal detailed technology instructions were provided to patients and staff. Responses from focus group participants were also used to refine patient handouts to include more pictures and less text, plain language (e.g., changed “pleasant” to “fun”), and graphic scenes relevant to the community population (e.g., rural vs. city landscapes) as well as eliminate and alter complicated conceptual ideas (e.g., cognitive restructuring changed to thinking patterns).

Evidence from the single-arm pilot trial suggests that the mPCST-Community is an acceptable and feasible intervention for breast cancer patients with pain who are treated in medically underserved areas. Participants found the intervention to be highly acceptable, and 90% completed all sessions. Though results should be interpreted with caution due to the small sample size and lack of control group, participants reported statistically significant improvements from pre- to post-intervention in pain severity, pain interference, and self-efficacy for pain management. Findings for pain catastrophizing, psychological distress, fatigue, and physical functioning, though not statistically significant, were all in expected directions of improvement. Perhaps most striking was the improvement in participants’ reported confidence in their abilities to manage their pain (i.e., pain self-efficacy); reported scores on a well-validated measure of pain self-efficacy improved 40% on average. This may suggest that patients living in medically underserved areas have a dearth of exposure to pain coping strategies and mPSCT-Community provided skills that quickly improved their ability to manage their pain. This finding is particularly important as there is an abundance of literature suggesting that patients with chronic pain who have a high level of pain self-efficacy are much more likely to have less pain and higher overall quality of life.^14,33,40 Importantly, this intervention is highly implementable as therapists at our academic medical center could feasibly provide the intervention to patients in diverse locations including cities without therapists trained to deliver behavioral pain protocols, rural areas with little access to any psychosocial or behavioral services, and urban areas where travel may be challenging (e.g., public transportation).

The videoconferencing was done via iPad with Skype with the participants at their community clinic. Focus group participants described not having consistent internet access in their homes, and the community clinics provided a familiar environment with reliable internet access. The pain coping skills training therapist was at an academic medical center 40–60 miles away. The experience of the participants and therapists with this mode of delivery was positive; there were no issues with connectivity or difficulty accessing the intervention through the iPad in the clinic. Clinic staff known to the patients were trained by the research staff to assist patients with the technology as needed; therapists were also trained to provide assistance as needed. Participants rated high satisfaction with the intervention and evidenced important pain and pain-related changes. This delivery modality is a highly viable route for implementation of a pain coping skills training protocol for patients living in medically underserved areas. Most pain coping skills training protocols are only available at major medical centers, however, the use of video-conferencing technology allowed the patients to participate in the intervention at their community clinic. This is important as patients live near these clinics and are familiar with the clinic environment and staff making them more likely to engage in the intervention.

This study has several strengths. First, to our knowledge it provides preliminary evidence on the first pain coping skills training protocol adapted to meet the needs of breast cancer patients with pain in medically underserved areas. The protocol was adapted with the input of a nationally recognized expert in the use of behavioral pain protocols in medically underserved areas. Further, we used focus groups comprised of patient stakeholders to further refine the protocol prior to testing. Second, our study results suggest that the intervention protocol is feasible, acceptable, and likely to lead to improvements in pain and pain-related outcomes.

Third, an intervention delivery modality was used (iPad+Skype in community clinic) that could be widely implemented to deliver behavioral pain protocols and other psychosocial or behavioral symptom management (e.g., fatigue & weight loss) protocols. This delivery modality as used in this study was relatively cost-effective – therapists in most medical settings have access to computers with video-conferencing capabilities and medically underserved clinics typically have internet access and would only require the addition of technology hardware (e.g., iPad and smart phone) to deliver this intervention.

Finally, this study provides important clinical implications including: (1) the intervention, as delivered, was relatively cost effective and can be implemented in many diverse settings, (2) cancer patients with pain found this intervention relevant and beneficial, and (3) the brief technology-based intervention was associated with improvements in pain-related outcomes, at least in the short-term. While this work and the results of the present study may also be important for patients in cities without consistent pain coping skills access, rural areas, and urban areas where pain-related barriers exist, it will likely need to be adapted to address these patients’ particular needs and the unique aspects of these environments; future work in these settings is critical to improving the reach of these types of interventions.

Limitations of this study include a small sample size, lack of a control group, and use of only a limited number (N = 3) of community sites. Additionally, the present study did not include a longitudinal follow-up assessment, so whether pre- to post-treatment effects persist over time is currently unknown. Future work is planned that will test mPCST-Community in a larger randomized controlled trial with an active education control group in 8–12 medically underserved areas and will include a longitudinal follow-up. Lessons learned from the current study will inform our future work in this area. For example, the majority of women who declined to participate either did not give a reason or stated that they were not interested (27/40). It is possible that these women were hesitant about a researcher from a major medical center relatively far away calling them at their home about a research study; we plan in the future to have the community providers tell patients about the study and have study team members present in the clinic for recruitment rather than recruiting participants solely over the phone.

In summary, the results of this study suggest that appropriately adapted mobile-health technologies may provide an avenue to reach underserved patients and implement behavioral interventions to improve pain management.

IMPLICATIONS FOR PSYCHOSOCIAL ONCOLOGY PRACTICE.

• Breast cancer patients being treated in medically underserved areas have a dearth of exposure to behavioral interventions that may improve their ability to manage pain.

• Evidence from this single-arm pilot trial suggests that our mobile-health behavioral cancer pain coping skills training protocol is acceptable and feasible in this vulnerable population.

• Appropriately adapted mobile-health technologies may provide an avenue to reach underserved patients and implement behavioral interventions to improve pain management.