Behavioral Cancer Pain Intervention Using Videoconferencing and a Mobile Application for Medically Underserved Patients: Rationale, Design, and Methods of a Prospective Multisite Randomized Controlled Trial

Abstract

Background:

Women with breast cancer in medically underserved areas are particularly vulnerable to persistent pain and disability. Behavioral pain interventions reduce pain and improve outcomes. Cancer patients in medically underserved areas receive limited adjunctive cancer care, as many lack access to pain therapists trained in behavioral interventions, face travel barriers to regional medical centers, and may have low literacy and limited resources. mHealth technologies have the potential to decrease barriers but must be carefully adapted for, and efficacy-tested with, medically underserved patients. We developed an mHealth behavioral pain coping skills training intervention (mPCST-Community). We now utilize a multisite randomized controlled trial to: 1) test the extent mPCST-Community reduces breast cancer patients’ pain severity (primary outcome), pain interference, fatigue, physical disability, and psychological distress; 2) examine potential mediators of intervention effects; and 3) evaluate the intervention’s cost and cost-effectiveness.

Methods/Design:

Breast cancer patients (N=180) will be randomized to mPCST-Community or an attention control. mPCST-Community’s four-session protocol will be delivered via videoconferencing at an underserved community clinic by a remote pain therapist at a major medical center. Videoconference sessions will be supplemented with a mobile application. Participants will complete self-report measures at baseline, post-intervention, and 3- and 6-month follow-ups.

Conclusions:

mPCST-Community has the potential to reduce pain and disability, and decrease barriers for cancer patients in medically underserved areas. This is one of the first trials to test an mHealth behavioral cancer pain intervention developed specifically for medically underserved communities. If successful, it could lead to widespread implementation and decreased health disparities.

Introduction

Women diagnosed with breast cancer experience unique pain challenges due to disease stage, tumor pathology, and/or treatment regimens.^1–9 The pain burden of breast cancer can be even higher for women in medically underserved areas.¹⁰ Cancer patients receiving care in medically underserved areas are more likely than those in other regions to be racial minorities and have low literacy levels.¹¹ Studies show that women of racial minority background experience delays in breast cancer treatment and/or inadequate treatment.¹² Later stage diagnosis and delayed treatment increase women’s risk for disease- and treatment-related pain.^13–17 Further, women in medically underserved areas are more likely to have comorbid health problems (e.g., obesity, arthritis) that worsen pain severity and interference.^13–18

Behavioral interventions help alleviate cancer pain. In a review of behavioral cancer pain management strategies, cognitive behavioral coping skills training with education showed the strongest evidence for decreasing pain through mediators of improved self-efficacy for pain management and pain coping.¹⁹ One such intervention is Pain Coping Skills Training (PCST), which was developed by our team and has been shown to significantly decrease cancer pain.^20,21 NIH guidelines recommend integrating behavioral pain interventions into cancer treatment; however, such interventions are largely absent in medically underserved areas.

Behavioral pain management interventions are typically in-person with pain therapists at a major medical center. Barriers to accessing these interventions include cost, time, transportation, and lack of well-trained pain therapists. Additionally, behavioral pain management interventions are rarely adapted for low literacy populations, further limiting pain management options. Disparities among women with breast cancer and low literacy in medically underserved areas are maintained by a lack of cost-effective, accessible, and appropriately adapted supportive care opportunities found at more resourced medical centers.^11,22 This results in a large symptom burden (i.e., pain) for patients, providers, and healthcare systems.

Studies show that mobile health (mHealth) technologies (e.g., videoconferencing, mobile applications) are feasible, acceptable, and can improve pain.^23–26 Yet, these strategies are rarely linked to validated theories (e.g., Social Cognitive Theory),²⁷ not developed by pain experts, and not adapted for and empirically tested in medically underserved populations.^28,29 Advances in mHealth technologies could be leveraged to exploit theory-based mechanistic variables (e.g., self-efficacy, pain coping), and reduce cost, logistical, and literacy barriers limiting medically underserved patients’ access to behavioral pain interventions.

To address these problems, we designed a theory-driven, mHealth Pain Coping Skills Training protocol (mPCST-Community). Patients use videoconferencing at their community-based clinic to receive mPCST-Community sessions from a well-trained pain therapist at a major medical center. mPCST-Community includes an adjunct mobile application informed by Social Cognitive Theory (Figure 1) designed to extend the intervention into the patient’s home. We present the rationale, design, methods, and analytic plan for a prospective multisite randomized controlled trial (RCT) that compares mPCST-Community to an education attention control condition (mHealth-Ed). The RCT aims to: 1) test the extent to which mPCST-Community reduces breast cancer patients’ pain severity, pain interference, fatigue, physical disability, and psychological distress; 2) examine potential mediators (e.g., self-efficacy, pain coping) of mPCST-Community’s effects; and 3) evaluate mPCST-Community’s cost and cost-effectiveness.

Figure 1.

mPCST-Community Mobile Application Features from Social Cognitive Theory (SCT)

Methods

This study was approved by the necessary Institutional Review Boards and registered with ClinicalTrials.gov (NCT04175639). Recruitment procedures comply with HIPAA guidelines.

Participants

Eligible participants are English-speaking women (≥ 18 years old) with a diagnosis of stage 0-IV breast cancer (N = 180) within the last three years and pain severity ≥ 4 out of 10 on at least ten days in the past month. Participants will be recruited from eight cancer clinics that are part of the Duke Cancer Network (DCN). Each clinic is in a medically underserved area, as evidenced by < 62 Index of Medical Underservice (IMU) from Health Resources and Services Adminstration. Sites are mostly in rural areas of North Carolina (e.g., Wilson [IMU = 53.3], Smithfield [IMU = 46.30]), South Carolina (e.g., Spartanburg [IMU = 60.30], West Columbia [IMU = 57.80]), and Georgia (e.g., Augusta [IMU = 36.6]). Exclusion criteria include: 1) cognitive impairment as indicated by a baseline Folstein Mini-Mental Status Examination of < 25⁵, 2) brain metastases and/or life expectancy of less than 12 months, 3) presence of a severe psychiatric condition or a psychiatric condition (e.g., suicidal intent) that would contraindicate safe involvement in the study, or 4) current or past (< 6 months) engagement in PCST for cancer pain elsewhere.

Study Design

Eligible participants are identified from a pain report in their electronic medical records and screened at recruitment to assess for a pain severity rating of ≥ 4 out of 10 on at least ten days in the past month (see Figure 2). A baseline assessment is conducted prior to randomization and consists of psychosocial questionnaires assessing pain, fatigue, physical disability, and psychological distress. Randomization is blocked by study site. Participants are assigned with equal allocation to either: 1) mHealth behavioral cancer pain intervention (mPCST-Community); or 2) mHealth education attention control (mHealth-Ed). The mHealth-Ed attention control condition allows us to control for nonspecific intervention effects and is designed to be informative, engaging and acceptable to participants. Participants are re-assessed postintervention, and at 3- and 6-month follow-ups. Randomization and all study assessments are completed by study coordinators who are blinded to group assignment.

Figure 2.

Study Flow

Both study conditions involve four 50-minute individual sessions conducted over eight weeks via videoconferencing at the participant’s community-based clinic. Once the video connection is established, the participant and therapist complete a remote session that mimics an in-person, face-to-face session. The therapist delivering the mPCST-Community intervention is a doctoral-level clinical psychologist specializing in behavioral pain and symptom management located at Duke University. mHealth-Ed content is delivered by a nurse who is not familiar with pain coping skills training. Both mPCST-Community therapists and mHealth-Ed nurses are trained in the manualized protocol by the study PI (Somers, PhD). Therapists and nurses role play each session with an actor; audio recordings of sessions are rated for adherence and competence by the PI. Therapists achieving 80% adherence and competence on each session are certified for intervention delivery.

Sessions are conducted in a private room to ensure confidentiality. Clinic staff are present nearby, outside the room, to assist with any technical concerns. Sessions are scheduled weekly with some flexibility for scheduling conflicts (e.g., sickness, holidays). All sessions will be audio recorded and a subset will be reviewed with the therapist or nurse by the PI (or a senior clinician) during weekly supervision meetings to ensure treatment fidelity. Study staff will assist participants in both mPCST-Community and mHealth-Ed conditions with downloading the mobile application to extend the intervention to their home environment (i.e., home, work). If participants do not have a smartphone, they will be provided one for the study duration to access the mobile application.

Intervention Conditions

mHealth Behavioral Cancer Pain Intervention (mPCST-Community).

Content of the mHealth behavioral cancer pain intervention (mPCST-Community) is based on Pain Coping Skills Training (PCST).^30,31 PCST was originally developed by Francis Keefe, PhD, and subsequently adapted for and tested in cancer patient populations.^32–34 In general, PCST enhances patients’ abilities to cope with their pain by teaching skills to improve behaviors, thoughts, and feelings about pain. For the current study, the research team carefully designed the mPCST-Community intervention to include content relevant to cancer patients with low literacy and pain in medically underserved areas.²⁰

mPCST-Community session content was adapted in several ways, including: removal and/or replacement of concepts not well-received by low literacy pain patients (e.g., in-depth Gate Control Theory^35,36 review replaced by brief graphic focused on Gate Control Theory education); increased use of graphics; changing text amount, size, and placement shown to benefit low literacy patients; explicit directions for home practice; addition of session audio summary and clear instruction and encouragement to contact the study team for procedural help.²⁰ The mPCST-Community protocol was further refined through focus groups (3 groups; n=19) and a small trial of breast cancer patients (n=20) from three of the eight medically underserved, rural community clinics included in the present study.²⁰ Additional changes made to the mPCST-Community intervention included: more content on cancer-related fatigue and its relationship to pain; focused instructions on clinic-based videoconferencing; increased supportive material in the protocol.

The final mPCST-Community protocol is brief (4 sessions) and delivered through videoconferencing to the breast cancer patient with pain at an underserved community clinic by a remote, well-trained pain therapist at Duke University. This adaption from our home-based prior work to clinic-based is critical, as patients in medically underserved areas often lack access to personal internet or data networks to support videoconferencing in their home. To retain the home-based components of the initial PCST videoconferencing protocol, mPCST-Community extends to the patients’ daily life through use of a simple mobile application that serves as an adjunct to the videoconference sessions and provides modeling videos and stories from other cancer patients with pain, relaxation audio for home practice (e.g., progressive muscle relaxation [PMR], imagery, etc.), and low literacy text/audio protocol summaries (e.g., study handouts, audio recording of “Session Tips”).

The mobile application for mPCST-Community also prompts participants to enter ratings (on a 0 to 10 scale) of pain, fatigue, and stress, as well as daily coping skills practice (number of days each skill was practiced during the past 7 days) three times per week; these data are used to send personalized texts with advice and skills use reminders. For example, if a participant obtains a symptom (pain, fatigue, and/or stress) average between 0–3, they could receive a tailored text such as: “Looks like you’re using good coping today” or “Think of a coping skill that has provided you with some energy in the past.” If a participant’s average symptom rating falls between 8–10, they might receive a tailored text such as: “Try one of the relaxation audios now” or “Try picking one of your favorite pleasant activities that is easy to do.” Caring mesaages, for example “Take time today to focus on what matters most to you” (i.e., encouragement not linked to assessments), will also be sent through the mobile application two times per week; in our past work we have found patients to be responsive and appreciative of caring messages.³⁷ The therapist will have access to the participant dashboard via their laptop prior to each session to view the participant’s weekly self-reported symptom ratings and coping skills practice data. These data will be used to help the therapist tailor the session content.

Recent literature examining behavioral change mobile applications has found these components (i.e., information, modeling, feedback, planning, symptom tracking) to be critical mobile applications components to enhance behavior change.³⁸ Based on these data, the therapist will tailor parts of the session to the participants’ needs. Decision algorithms will be used to guide the therapist in personalizing content (e.g., problem solving around increased pain of 30% from one day to the next; activity planning for increased use of PMR audio).

Each mPCST intervention session has two key components: 1) a 50-minute, videoconferenced session with a well-trained pain therapist conducted in the patient’s community-based clinic; and 2) extension of intervention content into the patient’s home environment via an adjunct mobile application. Our use of videoconferencing plus mobile application is unique in that it retains a “human touch” component that is absent in many mHealth behavioral pain interventions. These components are summarized below and in Table 1.

Table 1.

mPCST-Community Intervention and Mobile Application

Intervention Component	Sessions
	Session 1	Session 2	Session 3	Session 4
Videoconference Session	• Introduction; study overview • Review participant’s cancer history • Gate Control Theory of Pain • Progressive muscle relaxation (PMR) • Set home practice	• Review home practice • Activity-rest cycle • Fun activity scheduling • Mini-relaxation • Set home practice	• Review home practice • Positive thinking • Calming self-statements • Set home practice	• Review home practice • Imagery • Goal setting • Coping skill review problem-solving for continued practice • Wrap-up and goodbye
Mobile Application Features For Each Session	• Video explanation of Gate Control Theory of Pain • Video demonstration of PMR and audio of full and brief versions of PMR • Audio of patients describing experience with PMR • Participant can share video and/or photograph of where they intend to practice PMR via mobile application; therapist will offer feedback and problem-solving as necessary	• Video explanation of activity-rest cycle • Video of actors engaging in pleasant activities • Audio of mini-relaxation • Participant can share video and/or photograph of where they intend to practice activity-rest cycle, fun activities, and mini-relaxation via mobile application; therapist will offer feedback and problem-solving as necessary	• Video demonstration of positive thinking • Video demonstration of calming self-statements • Reminder of calming self-statement from session • Participant can share video and/or photograph of positive thinking practice and calming self-statements via mobile application; therapist will offer feedback and problem-solving as necessary	• Audio of imagery • Video explanation of goal setting • Reminder of goal discussed during session • Participant can share video and/or photograph of where they intend to practice imagery via mobile application; therapist will offer feedback and problem-solving as necessary • Video to wrap-up study and offer congratulations
Mobile Application Features Across All Sessions	• PDFs of session handouts • Audio summaries of each session and “Session Tips” review • “Caring Messages” with general encouragement and advice (e.g. “Have you practiced your coping skills today?”) • Brief assessments on Monday, Wednesday, and Saturday of daily pain, fatigue, and stress ratings • “Tailored Texts” on Monday, Wednesday, and Saturday based on responses to brief assessments (e.g., “Try PMR or another relaxation practice!”) • Mobile application homescreen will display a “To Do List” with skills for practice, caring messages, tailored texts, as well as reminders to complete brief assessments and full assessments (baseline, post-intervention, 3-month follow-up, 6-month follow-up) • Virtual pet or meter to track skills practice, completion of assigned home practice, brief and full assessments etc.

Session 1.

Videoconference component:

At the beginning of the videoconferenced session, participants are provided with a brief overview of the session format and given an opportunity to describe their cancer experience. Participants are presented with an educational rationale for the intervention based on a simplified version of Melzack and Wall’s Gate Control Theory of Pain,^35,36 which demonstrates pain as an interrelated experience affected by thoughts, feelings, and behaviors. Participants are taught progressive musle relxation (PMR) by observing the therapist, watching a video clip in which an actor demonstrates PMR, and hearing stories from patients with cancer. Then, the participant practices using a guided PMR recording, and receives verbal feedback and encouragement. They are asked to practice PMR daily to develop skill mastery. Mobile application component: Participants are able to access a demonstration and guided recording of PMR on the mobile application. They are to use a smartphone to take a photograph showing their therapist where in their home they will practice PMR. The photograph is shared with the therapist through the mobile application; he/she reviews the photograph and provides feedback to the participant regarding their practice of PMR as needed. A brief video review of the Gate Control Theory is also available through the mobile application.

Session 2.

Videoconference component:

Participants are taught to manage their pain while increasing their general activity level and to increase their pleasant activities. Patients with cancer often find their normal activity levels decrease due to their disease and its treatments. Skills to cope with pain related to activity are particularly important for patients with persistent pain, since pain may worsen as they attempt to increase activity. Paradoxically, inactivity to avoid pain can also lead to more pain. The skill of activity-rest cycling teaches participants to schedule activity so that they can increase their activity level and productivity, but avoid the increased pain from too much activity without breaks. The second skill taught is pleasant activity planning (i.e., “fun activity planning”), which refers to increasing participants’ engagement in enjoyable activities.³⁹ Increasing involvement in pleasant activities can distract from pain, improve physical condition, enable patients to be more involved in experiences that are valued and meaningful, and provide a sense of mastery. Engagement in pleasant activity planning is closely associated with decreasing distress.³¹ Participants view video clips of actors engaging in enjoyable activities and are encouraged to participate in their own pleasant and meaningful activities. The activity list provided to participants is tailored to reflect commonly reported activities of those living in underserved areas (e.g., community church event vs. a movie theater). Finally, participants are taught a brief applied relaxation – the mini-relaxation practice. Mobile application component: After the videoconferenced session, participants are asked to take a photograph using their smartphone of different areas in their home where they can practice activity-rest cycling (e.g., kitchen, yard); this photograph is sent to the therapist. A similar exercise can be completed for scheduled pleasant activities. The therapist sends a message to the participant providing supportive and corrective feedback, as needed. Simple explanation and demonstration videos of activity-rest cycle, pleasant activity scheduling, and mini-relaxation techniques are available for patient review through the mobile application.

Session 3.

Videoconference component:

Participants learn the skill of cognitive restructuring (i.e., “positive thinking”) to recognize how their thoughts, such as pain catastrophizing, can negatively influence their pain and their ability to cope with pain. Through the Gate Control Theory of Pain, participants are taught how negative pain-related thoughts not only impact their pain, but also their physical ability, psychological well-being, and fatigue.³¹ The therapist works through three common examples of negative thoughts having a detrimental impact on patients with pain. Next, participants are encouraged to think of an example from their own life. The therapist and participants work together to identify the unhelpful thought, define its negative consequences, and generate a more helpful positive thought (i.e., “calming self-statement”) that will have better consequences. Mobile application component: Several days after the session, participants receive a reminder through the mobile application of a calming self-statement (e.g., “This too shall pass”) determined with her pain therapist during her videoconferenced session. Brief videos explaining the skills presented during this session are available through the mobile application. Likewise, audio clips from actors who have benefited from these skills are accessible to the participant at home through the smartphone.

Session 4.

Videoconference component:

Participants are instructed in the skill of pleasant imagery, and are invited to explore The Gate Control Theory of Pain^35,36 in the context of this skill (i.e., closing the “pain gate”). Participants identify personal pleasant scenes and the therapist guides the participant through several practices using an audio recording. The therapist offers problem-solving (as necessary) around any negative reactions, and provides feedback, visual modeling, and encouragement. Participants are asked to practice imagery several times a week. To conclude, participants review pain coping skills learned and their benefits. Goal setting is used to plan for continued coping skill use. Mobile application component: The smartphone is equipped with videos of pleasant scenes and accompanying audio to assist the participant extend the imagery to their own home environment. Participants are to use a smartphone to take a photograph showing their therapist where in their home they will practice imagery. The photograph is shared with the therapist through the mobile application; he/she reviews the photograph and provides feedback to the participant regarding their practice of imagery, as needed. A simple and brief video review of goal setting is available to participants through the mobile application.

mHealth Educational Intervention (mHealth-Ed).

Each mHealth-Ed session also has a videoconference and a mobile application component (Table 2). Participants randomized to the mHealth-Ed condition receive comparable time and attention (i.e., videoconferencing, mobile application) to the mPCST-Community condition. Participants randomized to the mHealth-Ed condition will not receive any information specific to cancer pain coping skills.

Table 2.

mHealth-Ed Control and Mobile Application

Intervention Component	Sessions
	Session 1	Session 2	Session 3	Session 4
Videoconference Session	• Introduction; study overview • Review participant’s cancer history • General cancer pain information	• Common medical cancer treatments • Benefits and side-effects of cancer treatments	• Communicating with your healthcare team	• Healthy eating • Physical activity • Wrap-up and goodbye
Mobile Application Features Across All Sessions	• PDFs of session handouts • Audio summaries of each session and “Session Tips” review • “Caring Messages” with general encouragement and advice (e.g., “Make sure to fuel your body with proper nutrition.”) • Brief assessments on Monday, Wednesday, and Saturday of daily pain rating • Mobile application homescreen will display a “To Do List” with caring messages and reminders to complete brief assessments and full assessments (baseline, post-intervention, 3-month follow-up, 6-month follow-up)

Videoconference component:

Participants complete four, 50-minute video-sessions at their community-based clinic, with the assistance of staff as needed. Session 1 primarily focuses on the patients’ cancer experience and a brief review of general cancer pain information (i.e., prevalence). Session 2 focuses on common medical treatments and their benefits and side effects. Session 3 includes information on communicating with the healthcare team (e.g., list of questions). Session 4 provides basic information on healthy eating and activity levels.

Mobile application component:

Similar to mPCST-Community, mHealth-Ed participants are provided with a smartphone to use in their home environment that has a mobile application (i.e., the Pocket Cancer Care Guide App) with text and audio content relevant to each session topic. The smartphone provides activities comparable to mPCST-Community, but conceptually distinct and without pain coping skills components. To control for reporting effects, mHealth-Ed participants will complete pain assessments three times per week. Likewise, mHealth-Ed participants will receive three weekly caring messages related to the education content.

Measures

Primary Outcomes.

Pain.

Pain severity and pain interference are measured using the Brief Pain Inventory-Short Form (BPI-SF). The BPI-SF consists of four items assessing pain severity (worst, least, average, and current pain) and seven items assessing pain interference in the past week. The BPI-SF uses an 11-point response scale, with options ranging from 0 (no pain or no interference) to 10 (pain as bad as you can imagine or completely interferes). Separate composite scores are computed for pain severity and pain interference by averaging items, with high scores indicating greater pain severity and pain interference. The BPI-SF has demonstrated good reliability and validity in prior studies with women with breast cancer.⁴⁰

Fatigue.

The PROMIS six-item fatigue scale is used to assess the frequency, duration, and intensity of several fatigue symptoms (e.g., tiredness, exhaustion). Items are rated on a five-point scale ranging from 1 (not at all) to 5 (very much). Items are summed and converted to standardized T-scores; higher T-scores represent greater fatigue. The PROMIS fatigue scale has good reliability and validity in healthy samples, as well as those diagnosed with cancer.⁴⁰

Physical Disability.

Physical disability is assessed using four items from the Patient Care Monitor (PCM) v2.0.^41,42 These items ask about patients’ ability to run, do light physical work or fun activities, do hard physical work or fun activities, and ability to function normally in the last seven days using an 11-point response scale range from 0 (not a problem) to 10 (as bad as possible).

Psychological Distress.

Psychological distress is assessed using the four-item psychological distress scale of the PCM. These items ask about crying or feeling like crying, being worried, feeling nervous, tense, or anxious, and feeling sad or depressed. Items reference the last seven days and are scored using an 11-point response scale range from 0 (not a problem) to 10 (as bad as possible). The PCM has been validated with other standard symptom and QOL inventories.^41,42

Secondary Outcomes.

Cost Effectiveness.

To estimate expected costs and health outcomes associated with mPCST-Community and mHealth-Ed conditions, four types of information are collected throughout the trial.

1. All-cause medical resource use. Information on medical resource use is focused on expected cost-drivers and resources that may be differentially affected by the interventions, including hospitalizations, emergency department and urgent care visits, outpatient visits and use of pain medication. Data are collected at three time points: post-intervention and 3- and 6-month follow-ups. Participants are prompted to specify whether or not the medical resource use is specifically related to pain. We use standard cost assignment methods to estimate direct medical costs from the societal perspective.⁴³

2. Participant time. The societal perspective⁴³ also necessitates valuation of participants’ time spent on training and practicing skills learned in mPCST-Community. At the initial videoconference session, participants are asked how much time it takes them to travel from their home to the community clinic and whether a caregiver provided transportation. Participant time costs are valued using the average hourly wage in the US ($25.72, May 2019).⁴⁴ The smartphone will automatically record the participants’ interaction time with the mobile app. The same metrics are collected for the Health-Ed condition.

3. Productivity. As now recommended by the 2nd Panel on Cost-Effectiveness,⁴⁵ we account for changes in productivity in the societal perspective.⁴³ To evaluate the impact of mPCST and mHealth-Ed conditions on productivity, we measure employment status, weekly hours worked outside and inside the home at baseline, post-intervention, and 3- and 6-month follow-ups. Productivity costs are valued using the average hourly wage plus fringe benefits (30% of total compensation).⁴⁶

4. Health utilities. To incorporate potential differences in health-related quality of life experienced by patients in the mPCST-Community vs. mHealth-Ed conditions into the cost-effectiveness analysis, we measure health status using the five-level EQ-5D (5L EQ-5D), which has been shown to be more sensitive to changes in health status in comparison to the historic three-level version.^47,48 The EQ-5D is a brief measure of health status that can be linked to five-level or mapped to three-level population-based preference weights to estimate quality-adjusted life-years for use in economic evaluations.^49,50 The EQ-5D assesses five dimensions (mobility, self-care, usual activities, pain/discomfort, anxiety/depression).

Mediators.

Self-Efficacy for Pain Control.

Self-efficacy for pain control is assessed using a subscale of the Chronic Pain Self-Efficacy Scale.⁵¹ This subscale contains five items that assess patients’ certainty about their degree of pain control, pain during daily activities, controlling pain during sleep, and making pain reductions without extra medication. The items are answered on a 10 (very uncertain) to 100 (very certain) scale, and averaged to obtain an overall score. This scale has shown good reliability⁵¹ and has been used in cancer patients.⁵²

Pain Catastrophizing.

Pain catastrophizing is assessed with the six-item pain catastrophizing subscale of the Coping Strategies Questionnaire.⁵³ These items ask patients to rate their tendency to catastrophize when faced with pain (e.g., “When I feel pain it is awful and it overwhelms me”).⁵³ Items are answered on a 0 (never) to 6 (always) scale, and summed for a total score. This scale has good reliability in cancer patients.⁵⁴

Medical and Demographic Variables.

Demographic and medical data are collected by study staff through electronic medical record review. Medical information includes cancer diagnosis date, cancer stage, cancer treatments, medical comorbidities, height and weight, and use of antidepressant, anxiolytics, pain medications (i.e., number of days used in past 7 days and type of medication), and past use of mobile applications for pain management. Demographic data includes date of birth, race, marital status, and education. Electronic patient reported outcomes (ePROS) are completed by participants through the study mobile application at baseline, post-intervention (i.e., approximately 4–8 weeks later), and 3- and 6-month follow-ups. Moreover, at each intervention session, participants are asked about their cancer treatments in the last seven days and how many days in the last seven they have taken pain medication and type of pain medication.

Potential Confounds.

Potential confounding variables are measured at each assessment: 1) Depression is assessed with the 20-item Center for Epidemiologic Studies-Depression Scale (CES-D⁵⁵); 2) Social support is assessed with an eight-item Medical Outcomes Study Social Support Survey (MOS-SS^56,57), that was recently validated in cancer patients; 3) Physical activity (i.e., strength, flexibility, and level and intensity of activity) is assessed with the nine-item Rapid Assessment of Physical Activity (RAPA⁵⁸); 4) Physical performance is assessed with the nine-item physical function subscale of the Functional Status Questionnaire (FSQ⁵⁹) that asks patients to rate activities of daily living using a scale from 1 (usually did not do because of health) to 4 (usually did with no difficulty). All of these scales have demonstrated good reliability and validity.^56,58,59

Intervention Engagement.

Participants’ engagement in mPCST-Community is assessed post-intervention and at 3- and 6-month follow-ups by asking about use of skills learned during the intervention over the last seven days. For both mPCST-Community and mHealth-Ed conditions, session attendance and smartphone use metrics (e.g., number of log-ins, accessing the tips audio/text content, and accessing patient stories) are also collected to measure participants’ engagement.

Intervention Acceptability.

Participants’ acceptability of both intervention conditions is assessed with the 10-item Client Satisfaction Questionnaire⁶⁰ that rates items (e.g.,”How satisfied are you with the amount of help you received?”) using a scale from 1 (quite dissatisfied) to 4 (very satisfied).

Statistical Analyses

We will summarize demographic, medical characteristics, and other potential covariates. We will use an intention-to-treat framework for these analyses.⁶¹

Analysis of Aim 1.

The primary study aim will be to test the extent to which mPCST-Community reduces pain severity (primary outcome), pain interference, fatigue, physical disability, and psychological distress. The primary study hypothesis will be that mPCST-Community will lead to decreases in these pain-related outcomes compared to the mHealth-Ed control condition. Focusing on pain severity as an illustration of our analytical methods, we will first describe the pattern of missing data over time by group, paying particular attention to the possibility of differential drop-out. We will then analyze three outcome variables: postintervention minus baseline will address the mechanistic hypothesis pertaining to immediate effects of the intervention; 3 months minus baseline will address the (primary) hypothesis of short-to-medium term effects of the intervention; and 6 months minus baseline will address the hypothesis of longer-term effects of the intervention. Each hypothesis will be tested using the main effect of study group, adjusting for covariates. We anticipate relatively little missing data for the primary and mechanistic hypotheses – the issue of missing data will be addressed through sensitivity analyses using propensity scoring (which would essentially allow us to “compare likes with likes” even in the presence of differential drop-out).⁶²

Statistical Power for Aim 1.

For purposes of estimating power, we will define pain severity as the primary outcome variable. Using the methods of Cohen, and assuming an effective sample size of 72 per group (i.e., 90 per group minus 20% attrition), an unadjusted analysis will have 80% power to detect an effect size of approximately 0.48. This is a moderate effect size, which is a reasonable estimate of the minimum clinically important difference in patients with pain, and practically important and consistent with what will need to be shown for adoption of the intervention in clinical practice.^62–64 Further support for this effect size is gained from our pilot work that has been consistent (within the constraints imposed by study design and sample size) with an intervention having this level of efficacy.²⁰ In an unadjusted analysis, the effect size equals the difference between the group means divided by the pooled standard deviation. Such a power calculation is conservative, as the reduction in degrees of freedom associated with the covariates in an adjusted analysis should have less of an impact than the reduction in the mean squared error associated with those same covariates.

Analysis of Aim 2.

Our secondary Aim 2 will be to explore mediators (i.e., self-efficacy for pain control, pain catastrophizing) through which mPCST-Community may lead to reduction in pain severity, pain interference, fatigue, physical disability, and psychological distress. This hypothesis will be explored, for example, by first fitting a model with intervention group as the primary predictor and self-efficacy as an outcome, and then by fitting a second model with self-efficacy as the primary predictor and pain severity as an outcome.

Analysis of Aim 3.

We will use patient-level data to test for differences between mPCST-Community and the mHealth-Ed interventions conditions on medical resource use, medical costs, participant time costs, and productivity costs. We will use generalized linear models specified using negative binomial error distributions and log links for resource use and gamma error distributions and log links for costs. We will apply mixed-effects linear regression models to test for interactions between study interventions and time on EQ-5D-derived preference weights across the 6-month follow-up period. Baseline resource use and EQ-5D weights will be included as covariates in regression models.

For each patient, we will compute quality-adjusted life-years (QALY) using the trapezoidal rule applied to EQ-5D weights at baseline, postintervention, and 3- and 6-month follow-ups. Because the mHealth-Ed control was designed to control for patient contact time and does not represent standard care, accounting for the costs of providing this condition and patient time spent participating would inappropriately offset the costs associated with the mPCST-Community intervention. Therefore, we will modify our calculation of the numerator of incremental cost-effectiveness ratio (ICER), in which we will calculate the difference in mean medical costs and productivity costs between the mPCST-Community and mHealth-Ed control group, then add direct costs of resources required to provide mPCST-Community as well as participant time costs. To compute the difference in QALYs, we will subtract mean QALYs in the mHealth-Ed condition from mean QALYs in the mPCST-Community condition. Next, we will divide the estimated difference in mean costs by the difference in mean QALYs and generate a 95% confidence interval constructed using nonparametric bootstrapping. Then, we will calculate net health benefits (QALY*λ – Cost, where λ represents willingness to pay per QALY threshold of $100,000) for each patient, and apply linear regression analysis to evaluate whether intervention and patient baseline characteristics are associated with improved cost-effectiveness.

In sensitivity analyses, we will evaluate the impact of: 1) applying various durations of sustained benefit beyond the 6-month follow-up period; 2) excluding productivity gains/costs; and 3) assuming participants could engage in videoconferencing at home. To evaluate practicality and sustainability of PCST-Community, we will assess accrual, participant retention, and intervention protocol adherence.

Discussion

Pain in women with breast cancer is strongly related to negative outcomes, including worse physical symptoms, physical functioning, and psychological distress.^65–68 In medically underserved areas, breast cancer patients with pain are especially vulnerable to persistent pain and disability, making the reduction of pain particularly relevant for this group. Despite their efficacy,^68,69 non-pharmacological behavioral pain interventions are rarely available in medically underserved areas due largely to a lack of pain therapists with specific expertise in behavioral skills training protocols (e.g., Pain Coping Skills Training); furthermore, existing protocols are not adapted for low literacy populations. mHealth technologies have the potential to overcome access barriers (e.g., transportation, cost, travel time, burden) to better resourced major medical centers where state of the art behavioral pain interventions might be available. Still, there is limited research exploring the efficacy of mHealth interventions carefully adapted for the unique needs of breast cancer patients with pain in medically underserved areas.

Innovative Features of mPCST-Community Intervention

Our mPCST-Community protocol innovatively addresses these gaps in the following ways. First, the mPCST-Community protocol capitalizes on advantages of mHealth technologies that are likely to enhance the intervention’s efficacy. We developed the mPCST-Community protocol by focusing on specific factors of the Social Cognitive Theory¹¹ (i.e., mastery, vicarious learning, verbal encouragement, physiological/affective responses; Figure 1) that increase self-efficacy and create a potent mHealth intervention. We hypothesize that mPCST-Community will enhance patients’ self-efficacy for pain control by extending skill acquisition, practice, and mastery to their home environment. This delivery modality eliminates the assumption of full generalization from a medical center setting to home. A review of mHealth interventions published in Translational Behavioral Medicine (TBM)¹³ concluded that behavioral interventions using mHealth technologies could benefit from increased health behavioral theory application. This study will be one of the first to begin to explore how theoretical constructs relevant to pain control (i.e., self-efficacy for pain management and pain coping) explain the benefits of an mHealth behavioral intervention.

Second, our intervention is unique in that it tests a behavioral cancer pain intervention that uses a hybrid mobile health approach of videoconferencing and an adjunct mobile application in a medically underserved population. Little work has been done to understand how mHealth advances may benefit patients in medically underserved areas. Our proposed intervention has been carefully designed to use mHealth technologies to improve efficacy and decrease critical access barriers (i.e., delivery by a well-trained pain therapist at in-person sessions at a major medical center, travel that is expensive and physically burdensome, negotiating an unfamiliar medical center, and low literacy) to behavioral cancer pain interventions.^13–15 We propose to have patients use videoconferencing at their familiar, community-based clinic to receive four mPCST-Community sessions from a well-trained pain therapist located at a major medical center. This will be combined with a simple mobile application that extends the intervention into the patient’s home. Our hybrid model of videoconferencing plus mobile application is unique in that it retains a critical “human touch” component which is glaringly absent in many mHealth behavioral pain intervention attempts; while this is being done some in the commercial arena, the combination is largely absent in efficacy research.

Third, this work is significant because it addresses a central recommendation of the Institute of Medicine’s Pain in America Report by providing patients with pain self-management strategies.⁷⁰ This recommendation is particularly important to act on in underserved communities where resources on self-management education for health and disease (e.g., obesity, diabetes, arthritis) is markedly lower than in academic medical center settings;^11,22 thus, when diagnosed with cancer, patients in underserved communities are likely to have little disease or symptom self-management strategies that could be beneficial due to years of suboptimal care. This lack of resources is further complicated by lower overall literacy levels in underserved communities. mPCST-Community is likely to provide patients with breast cancer in underserved communities with their first health self-management program specifically designed to meet the many challenges they face and address a wide range of cancer pain issues related to advanced disease, post-surgical pain, neuropathy, endocrine therapy, and/or comorbid pain conditions.

Anticipated Outcomes

This is one of the first trials to assess the efficacy of a mobile health pain coping skills training protocol for reducing pain in patients with breast cancer who live in medically underserved areas. If the mPCST-Community intervention proves efficacious, it will underscore the utility of using videoconferencing and an adjunct mobile application to enable breast cancer patients with pain in medically underserved areas to receive an intervention that has been carefully adapted and refined to meet their specific needs.

Around 60 million people live in medically underserved areas of the US, where there is a shortage of even primary care providers and total specialist care is estimated at only 13 providers for every 100,000 individuals.^71,72 In a traditional care model, breast cancer patients with pain who live in medically underserved areas have little to no chance of receiving known efficacious behavioral pain intervention strategies that can help them to manage their pain and improve other pain-related outcomes (e.g., pain interference, physical disability, quality of life). Advances in healthcare technologies have not yet realized their potential to drastically increase patient access to care. Results from this trial may begin to challenge the status quo and represent an important step toward addressing health disparities in these communities.

Additionally, the assessment of mediators through which mPCST-Community may lead to benefits will reinforce our current understanding of mechanistic variables underlying behavioral pain interventions. Stanton et al.⁷³ called for focused attention on mediators of efficacious psychosocial interventions for cancer as a necessary step to develop maximally effective interventions. We have shown low levels of self-efficacy for pain control and high pain catastrophizing to be related to negative outcomes in pain patients.^74–76 A small number of studies in other samples with persistent pain (i.e., back pain,⁷⁷ jaw pain⁷⁸) have also shown treatment-related changes in pain self-efficacy and catastrophizing to be mediators of outcomes. Findings from our trial will help clarify whether such mediators are also relevant for cancer populations in medically underserved areas.

Finally, to assess cost and cost-effectiveness, four types of information will be collected: all-cause medical resource use, participant time, productivity, and health utilities. A recent Cochrane review of 32 RCTs of telephone-delivered symptom management interventions for cancer patients found that none of the studies reported on cost-effectiveness.⁷⁹ This gap in knowledge is surprising given that cost savings is one of the primary arguments for remotely-delivered interventions. Moreover, cost-effectiveness data are crucial for determining an intervention’s sustainability and potential for widespread implementation. We designed this trial with these key considerations in mind. The trial results will allow us to determine if mPCST-Community is cost-effective in terms of its incremental cost per quality-adjusted life-year from a societal perspective, inclusive of healthcare, intervention, and patient-time costs. As healthcare costs are increasingly scrutinized, robust evidence on the relative costs and effectiveness of symptom management interventions will be valuable to providers and policymakers aiming to efficiently allocate scarce resources, especially since there is no consistent model of payment for behavioral pain coping programs.

Conclusions

This trial is the first to test the efficacy of mPCST-Community, a novel mHealth behavioral cancer pain intervention specifically designed for patients with breast cancer in medically underserved areas. Before deploying mHealth technologies for delivery of behavioral cancer pain interventions, the efficacy of such approaches must be validated. Results from this trial are expected to demonstrate that our mPCST-Community protocol decreases cancer pain and disability for patients in medically underserved communities. Our findings may lead to increased implementation of a potent and accessible behavioral pain intervention for cancer patients. Broad implementation of an efficacious mHealth-based intervention could improve overall quality of life among patients who face barriers to accessing in-person interventions.^66,67,80,81