Using Intervention Mapping to develop a novel pain self-management intervention for people with cirrhosis

Abstract

Background:

Chronic pain is common among patients with cirrhosis and is challenging to treat. While promising, pain self-management (PSM) interventions have not been tailored to this population’s needs.

Aims:

To design a PSM intervention for patients with cirrhosis.

Methods:

Semi-structured interviews with 17 patients with cirrhosis, 12 hepatologists, and 6 administrators from two medical centers were conducted to inform a rigorous, structured Intervention Mapping (IM) process. Qualitative content analysis was guided by Social Cognitive Theory (SCT) and the Consolidated Framework for Implementation Research (CFIR) and incorporated into intervention development. A planning group met regularly throughout the intervention, to reach consensus about how to use data and theory to develop the intervention through IM.

Results:

Participants described barriers to PSM behaviors, including the absence of simple, evidence-based interventions for pain for patients with cirrhosis, inadequate provider knowledge, time, and training, and lack of champions, funding, and communication. Patients described high motivation to treat pain using behavioral methods including meditation, prayer, and exercise. The intervention was designed to address barriers to PSM behaviors for patients with cirrhosis, using behavior change methods that address knowledge, self-efficacy, and outcome expectations. The LEAP (Liver Education About Pain) intervention is a 12-week, modular intervention delivered by phone via individual and group sessions with a health coach.

Conclusions:

People with cirrhosis, hepatologists, and administrators informed this theory-driven, tailored PSM intervention, which was designed to be implementable in the real world.

INTRODUCTION

Cirrhosis affects approximately four million people in the US[1–4] and is a leading cause of death and disability worldwide.[5] Chronic pain, defined as pain lasting more than three months, is reported by 14% of the general adult US population[6] but impacts 79–86% of patients with cirrhosis.[7, 8] Common causes of chronic pain in this population include liver-related pain (e.g., peritoneal stretch due to ascites), diffuse, inflammatory-type pain, and chronic pain of other etiologies (e.g., osteoarthritis), which may be exacerbated by the pro-inflammatory milieu that is present in people with cirrhosis.[9–12] Pain in this population is associated with mental health symptoms, pain-related disability, and increased healthcare utilization.[8, 13]

Pharmacologic chronic pain management in general populations typically includes a stepwise approach, starting with over-the-counter medications. However, safety concerns with common analgesics (e.g., high-dose acetaminophen and non-steroidal anti-inflammatory drugs) complicate pain management for patients with cirrhosis.[12, 14] Consequently, up to half of patients with cirrhosis are prescribed opioids each year,[7] leading to opioid-related complications,[15, 16] hepatic decompensation,[17] hospitalization,[13, 18, 19] and mortality,[7, 20] without addressing underlying pain.[8] Thus, there is a critical need to develop non-pharmacological approaches for pain management for this population.

Pain self-management (PSM) interventions are evidence-based behavioral interventions that are efficacious for managing chronic pain in general populations.[21–24] PSM interventions are also attractive because they are manualized and designed to be delivered by a variety of health care workers (e.g., social workers), making them scalable to real-world contexts.[25] While PSM interventions are promising in general, such interventions work best when they are tailored to the needs of the target population. Intervention Mapping (IM) is a stepwise approach to systematically developing complex, theory-driven, behavioral interventions.[25–31] Therefore, we aimed to develop a PSM intervention to meet the unique needs of patients with cirrhosis using IM.

METHODS

The intervention development process followed the six steps of Intervention Mapping (IM), as outlined by Eldridge et al. (Table 1). The investigator team included scientists experienced with and trained in IM, implementation scientists, intervention developers, qualitative methods experts, and a person with cirrhosis and chronic pain. This study was approved by both the VA Pittsburgh Healthcare System (VAPHS) and the University of Pittsburgh IRBs as a minimal risk study that included semi-structured interviews with administrators, hepatology providers, and patients. Throughout the study, the investigative team met weekly to review results. Per IM guidance, a planning group was formed to provide further feedback to the investigative team. This planning group included administrators, patients, providers, and researchers and met every 1–3 months over the course of 2 years, during the time when the data were collected and analyzed. Live notes were taken during each meeting and reviewed by the team. Consensus was reached by majority in general, with final decision at the discretion of the PI (SR) in the case of differences that could not be reconciled by consensus.

Table 1.

Steps of Intervention Mapping and tasks within each step as described by Eldridge et al.^*

IM Steps	Tasks within each step
Step 1: Develop a logic model of the problem	Work with a planning group Conduct a needs assessment (semi-structured interviews) Describe the context State program goals
Step 2: Identify program outcomes and objectives & logic model of change	Specify expected outcomes and performance objectives Select determinants for behavioral and environmental outcomes Construct matrices of change objectives Create a logic model of change
Step 3: Program Design	Generate program theme, sequence, scope, and components Choose theory and evidence-based change methods Select or design practical applications to deliver change methods
Step 4: Program Production	Refine program structure and organization Prepare plans for program materials Draft materials, protocols Pretest, refine, and produce materials
Step 5: Implementation Plan	Identify potential program users State outcomes and performance objectives for program use Construct matrices of change objectives for program use Design implementation interventions
Step 6: Evaluation Plan	Write process evaluation questions Develop assessment measures Specify evaluation design Complete evaluation plan

^*Key definitions:

Behavioral or environmental factors: contributors to the problem (behaviors, interpersonal, organizational, and social factors are included)

Determinants: barriers that need to change to control the behavioral and environmental factors contributing to the problem

Performance objectives: sub-behaviors that are desired (who has to do what)

Change objectives: theory and evidence-based changes expected to affect the behavioral and environmental outcomes; what will change the determinants to achieve the performance objectives?

IM is a six-step, community-engaged process for developing a theory-based intervention. Steps guide investigators to develop a 1) logic model of the problem, 2) performance outcomes and objectives—logic model of change, 3) program design, 4) program production, 5) program implementation plan, and 6) evaluation plan.[26, 31] Each step of IM, as well as the associated tasks, are outlined in Table 1. The planning group and research team collaborated to complete these steps, which were informed by behavior change theory, implementation theory, and qualitative interviews (see below).

Conceptual Frameworks

IM encourages intervention developers to use multiple behavior change theories and implementation frameworks. Our intervention development process, based on input from our planning group (see below), was framed by the Social Cognitive Theory (SCT) and the Consolidated Framework for Implementation Research (CFIR). SCT underlies many behavioral interventions for pain and posits that thoughts, feelings, and behaviors are integrally and reciprocally connected.[32, 33] SCT constructs that are typically addressed in PSM interventions include: (1) self-efficacy, a person’s belief in their own ability to change their behavior; (2) positive outcome expectations, a person’s belief that something good will come from participating in a particular intervention; and (3) self-regulation, the willingness to change a behavior now in order to benefit in the future.

The CFIR assesses barriers to implementation across 5 domains, each with subconstructs: the characteristics of the intervention (e.g., how simple vs. complex pain management is for patients with cirrhosis), the characteristics of the individuals delivering the intervention (e.g., provider knowledge), the implementation process (e.g., the presence of champions and the degree to which the setting has planned for implementation), and the contextual factors influencing implementation, including the inner setting (e.g., local leadership) and outer setting (e.g., national policy) factors.[34, 35] We used the CFIR to evaluate implementation barriers to PSM interventions for patients with cirrhosis and consider implementation early in the intervention process.

Advisory interviews

Because IM is iterative and requires advisory input, we formed both a standing planning group, which met every 1 to 3 months, and conducted interviews to inform intervention development. Interview guides were reviewed with study team and planning group input. Patient, provider, and administrator participants were recruited from the University of Pittsburgh Medical Center (UPMC) and the VAPHS, with the goal of increasing generalizability of the findings across multiple settings and to include a sufficient sample size to achieve overall thematic saturation. Following guidance about achieving thematic saturation for each group of advisors (patients and providers/administrators), or achieving the point where no new themes are generated, we aimed to recruit at least 12 patients, 12 providers, and 6 administrators.[36] Providers and administrators were approached by email and were consented by phone.

To identify potentially eligible patients, the research coordinator prescreened charts of consecutive patients with upcoming hepatology appointments and selected potentially eligible patients, who were called (UPMC) or met in the clinic (VAPHS). Participants were required to be ≥18 years of age and capable of providing informed consent with a hepatologist-confirmed diagnosis of cirrhosis and evidence of possible chronic pain in the medical record. Potentially eligible patients were screened for chronic pain (lasting at least three months).[37] We excluded prior transplant recipients and patients receiving hospice care, given the unique pain management needs of these populations. Agreeable participants provided informed consent and arranged time for a phone interview.

Data Collection and Analysis

A trained, experienced qualitative interviewer conducted 30- to 60-minute recorded phone interviews, which were transcribed using denaturalized transcription, deidentified and uploaded into NVivo. Two coders used content analysis to code the narrative data;[38] double coding 25% of the interviews and meeting to adjudicate differences. A priori codes organized around SCT and CFIR constructs were also included in the codebook. The study team members regularly met with coders to adjudicate differences, review the codebook, and evaluate whether new information was still being collected through additional interviews. The results below present each step and task in IM and the ways in which the qualitative data and planning group informed intervention development.

RESULTS

IM Step 1: Develop a logic model of the problem

IM Step 1, Task 1: Develop and work with a planning group

Members of our multidisciplinary study team met weekly to inform and guide each IM step. This team includes investigators with lived experience and expertise in mental health care (MC, SHT, WD), behavioral intervention development (JM and WD), implementation science (SR, MC, JL), qualitative inquiry (GK, JM, MC), primary care (JL, KK), and hepatology (SR, NJ). Our planning group included additional patients, providers, and administrators and met every 1–3 months to inform IM.

IM Step 1, Task 2: Conduct a needs assessment to develop a logic model of the problem

The planning group first defined the key health problem as “Chronic pain is common and inadequately managed in people with cirrhosis.” Because PSM interventions were rarely considered, people with cirrhosis often received opioid prescriptions, which can lead to poor functioning, increased healthcare utilization, and increased morbidity and mortality.[9–13, 15, 18, 39–45]

The next IM Step 1 subtask was to define behavioral factors contributing to inadequate pain management and the determinants of those behaviors. Information from 35 semi-structured interviews (Table 2) was organized by SCT and CFIR constructs to identify the determinants (Table 3). Below we highlight the priority determinants and factors, per planning group review, and denote respondent type using “Pt” for patient, “Pr” for provider, and “A” for administrator quotes.

Table 2.

Demographic characteristics of participants in needs assessment

Characteristic	Patients N=17	Providers N=12	Administrators N=6
Recruitment site, n (%)
UPMC	8 (47)	8 (67)	3 (50)
VAPHS	9 (53)	4 (33)	3 (50)
Age (mean ± sd)	65.5 ± 9.0	45.9 ± 11.4	52.2 ± 11.1
Male gender, n (%)	11 (65)	7 (58)	4 (67)
Non-LatinX white, n (%)	13 (76)	9 (75)	4 (67)
Non-physicians, n (%)	n/a	2 (16)	2 (33)

Table 3.

Determinants of pain self-management for patients with cirrhosis, organized by Consolidated Framework for Implementation Research and Social Cognitive Theory constructs

Domain/Construct	Illustrative Quote(s)
Characteristics of intervention
Complexity	I felt that those were my two main kind of challenges. One is providers undertreating pain that probably needs to be treated. And, secondly, you know, are cirrhotics experiencing pain that we can’t really explain very well or manage very well. (Pr007) [It’s] very difficult to treat these patients and a challenge, clinically, for all of us. (Pr011)
Trialability	[Y]ou just have to get the right buy-in from the right people and kind of make little changes at once. (Pr012) [W]e would have to see how it could be implemented in the clinic and then it would probably, you know, you’d have to run a few patients through that specific, you know, intervention and see how it’s affecting, you know, the workflow and things like that. (A202)
Strength of evidence	There’s no studies and then there’s little evidence-base…We play by ear. We don’t have good evidence to support our decisions. This is one of the problems of this field. Homemade medical decisions, no evidence-base. (Pr011) I think we have a long way to go in terms of having robust support for non-pharmacologic therapy. (Pr003) There’s very few outcome studies specifically in this group but the little data that we have suggest that they don’t do very well on opioids. (A201)
Relative advantage	I think you’d have to show, at the end of the day, that there’s a benefit to doing it. So, you’d have to show that outcomes are going to be different. (Pr006)
Characteristics of individuals: providers
Self-efficacy	I don’t feel very confident, frankly. (Pr010)
Knowledge	I also am concerned that there may not be as much awareness as there should be about the potential for liver injury resulting from chronic high dose use of non-steroidal anti-inflammatories. (Pr003) Okay, I, so I think, I think you have to have something for providers. I think, you know, there should be…an educational component that simply states the, you know, some of the issues, some of the drugs to be, drug classes to be avoided, some of the concerns with the other drug classes and, you know, then we’d have some, you know, some strategies that a physician could employ pharmacologically. (A302) I try to prioritize Tramadol, since there are no drug-to-drug interactions. (Pr008) I will prescribe Tramadol, which I believe is technically considered an opioid, I’m not sure. (Pr007) It’s hard ‘cause opioids are effective in controlling pain. (Pr006)
Stigma	I think the other reality that you have to recognize is that patients with cirrhosis can be a very challenging group and they may not be the most motivated people to, you know, do a self-directed intervention. (A302) So, the problem is I feel like a lot of the time people are kind of dismissive of pain. And, then they say, ‘Oh, okay, we’ll do this non-pharmacologic based therapy and then don’t think about the other comorbid issues or not willing to recheck in with the patient and kind of see where they are. Because I feel like a lot of the time patients need, they just need closer monitoring. (Pr001)
Communication	I think that’s what’s most frustrating to patients when they suffer from chronic pain is that they feel like they’re not being listened to. And I think if you get into a shared decision-making process, it’s a lot easier. (Pr001)
Characteristics of individuals: Patients/SCT-mediators of personal behavior change
Self-regulation	[Other] people have different motivations. Me, it’s personally, hey, I want to be around for these grandkids of mine and all that stuff, you know, so that’s my motivation. (Pt310) …Taking them to that place where they can say, “Okay. I’ve been on these medications for a long period of time. What are my goals? How am I trying to improve my quality of life? Am I where I want to be?” (Pr001)
Self-efficacy	I just keep on saying to myself that I’m stronger than the pain or that I will like think of a blank piece of paper. I’ll say that to myself. (Pt208)
Outcome expectations	I think that it has to do with the person wanting to go in with an open mind and believing that it could happen that they don’t need pills for the pain. (Pt208) I think mental health is super important for kind of helping to teach some coping skills. You know, we have these people who tend to get super sensitive to pain and have a very low pain tolerance, and everything hurts them. If you touch them funny, I swear sometimes if you look at them the wrong way, it hurts them. I think, I think pain, I think psychology and mental health are super important in talking to them and helping them to, you know, really define what’s an acceptable pain level. (Pr012)
Knowledge	I used to be from the old school. “No pain, no gain.” What I’m finding out later in life, pain tells you, “Hey, there’s something wrong, you got to figure it out.” …Maybe these people figure that…whatever [physical activity] they’re going to do is going to cause them more pain, cause more injury. I guess if they’re taught how to do it the right way, they may have a different view on it. (Pt310) Teach [patients with cirrhosis] when you talk with them that that work, you know, that that really help, so they can do that kind of thing. So they can help themselves…Everybody don’t have the strength you know…or knowledge or whatever, education. (Pt203)
Inner Setting
Complexity of the medical system	And then, the other thing is, too, that I feel like the more hands you get in the pot, the more complex pain management gets for these people. You know, and he’s one that has frequent hospital admissions so, you know, you get every inpatient provider every time he’s in so it’s, you know, at least a resident and an attending on the medicine service. And then, it’s whatever specialty services. And then, there’s hepatology. And then, there’s transplant. And then, there’s you know, all of these other people. So, people like him are far more complex. And, I would say that I don’t think I particularly do a great job of figuring out even how to facilitate managing his pain. (Pr012)
Leadership support	We would have to have buy-in from like leadership from above, right, to kind of make more of the, the whole timing of the disciples. And, then you’d have to have, you know, obviously the players themselves, you know, the psychiatrist, primary care, hepatology, you know…the heads of the departments certainly would be helpful… (Pr004) And, you really have to have administrative support to that. And, that’s where we sometimes have hit a wall with some of our pursuits for pain related programs or whole health programs, that maybe the administrative support is not as strong as it should be especially if it’s not something absolutely mandated, I hate to say that, but I think that’s the case. (A301)
Relative priority	So, it’s easy as a provider to dismiss the subjective complaints of a liver patient because you have so many other objective abnormalities to deal with. (A202) Of all the things providers have on their plates, dealing with patients, chronic pain is probably the least of what they’re willing to be involved with, in all honesty. (Pr006) Well, the issue is that it needs to be a high priority issue for the gastroenterologist…And you know, they’re facing all kinds of massive pain problems… (A201)
Staffing and time	Time is tough. Staffing is tough, depending on who all you would want to have there. … Space can be tough… I think physician availability is tough. (Pr012) You know, I think the barriers are all is the same and that is staff time and space. (A301) In the past, a lot of people keep embedding things in primary care but there’s so many embedded different things in primary care that now they’re overwhelmed. (A201)
Infrastructure	So, I think we could be doing a lot better in terms of providing some non-pharmacologic infrastructure. (Pr003)
Outer Setting
Funding, insurance coverage	How can I count them up? The first nine [barriers] I would say would be, out of ten, would be resources. (Pr003) You have to actually have some kind of a reimbursement structure so that all these costs get covered. You know, if this is gonna actually be a sustainable intervention. (A201)
Patient needs and resources	[N]obody wants to drive in for a CT scan and then drive out, and then drive back for some blood work, and then drive back, and then drive in to see psychology and then drive back, especially because parking is terrible. You’re in pain. (Pr012) But my guess is that you will find a disproportionate number of patients who are homeless, a disproportionate number of patients who have ongoing substance use disorder issues. And, just those two things alone are huge challenges. (A302) [I]t’s difficult for me to get around now even, you know, to see doctors and stuff. (Pt202) If you have small children, you don’t have time. (Pt204)
Implementation
Planning	At the end of the day, is the patient and the physician willing to do it? And, are there plans to do it? Is there time for the physician to do? Because that’s the main issue here. (Pr005)
Champions	[E]ven someone, at least some identifiable champion within those departments, for lack of a better phrase. (Pr004) You know, if you had, if in the different systems, systems of care, if there were champions, you know, that would really be, that would really be great, I think. Anytime you have a champion for whatever the intervention is, that always is really terrific. (A302)

One key behavioral factor was low adoption of evidence-based PSM approaches. One patient stated, “I was raised old school. You just, you kind of toughed it out” (Pt309), and others described taking drugs and alcohol to cope with pain. Behavioral determinants of these ineffective approaches included lack of knowledge of PSM interventions, low confidence in the ability to manage pain, and unclear outcome expectations (Table 3).

Key environmental factors were provider non-referral to PSM interventions and opioid overprescribing. The determinants of these factors aligned with the CFIR and included intervention-level barriers such as providers’ perception that pain management in cirrhosis is complex and lacks evidence, resulting in a perceived lacked skills and low self-efficacy around pain management. One provider stated, “I don’t feel very confident, frankly.” (Pr010) Table 3 highlights providers’ knowledge gaps and other environmental determinants/implementation barriers, including lack of staffing, time, infrastructure, reimbursement, relative priority, and complexity of navigating the medical system. Outer setting determinants included stigma, transportation, and childcare.

Stigma was evident from provider perspectives, as highlighted here:

“80% of the patients come to your visit saying, ‘I have horrible pain.’ Because they know they’re likely to get opioids…everybody has pain because everybody knows that just claiming that without tests, without any diagnosis check to see why you have that pain, you will get the drug.”

(Pr011)

And also by patients, who perceived this attitude:

“Other than the doctors like actually listening, when somebody tells you that, when you’re a doctor and a patient tells you they’re in pain, right now the assumption is everybody is looking to get opioids and get high. And I believe that there’s a lot of people in legitimate pain.”

(Pt309)

These data, behavior change theory, and planning group input were used to develop the logic model of the problem (Figure 1).

Figure 1.

Logic Model of the Problem

IM Step 1, Task 3: Describe the context for the intervention

This task aims to define social, information, policy/practice, and physical environment assets that could support the intervention. One example that the team identified was high patient motivation. One patient said, “I’d be willing to do anything to relieve this pain and be able to get even like 6 hours of straight sleep, I would give half my kingdom, not that I have much of a kingdom…” (Pt309) Another patient said, “I’ll try anything just to get rid of the pain…I’m open for any suggestion. You name it, I’ll take it.” (Pt306) The team additionally identified informational assets including the electronic medical record system and patient access to telephones. Policy assets included significant system-level efforts to limit opioids and the fact that patients with cirrhosis are generally regularly seen in hepatology clinic. Such facilitators can be leveraged to make the intervention work.

IM Step 1, Task 4: State program goal

The planning group reviewed the findings and reached consensus that the “program goal” should be: “Patients with cirrhosis and chronic pain who complete at least 80% of a 12-session PSM intervention will report adopting at least 2 new PSM behaviors and have a 30% reduction in pain severity.”

IM Step 2: Performance outcomes and objectives—logic model of change

IM Step 2, Task 1: State the expected outcome for the behavior and environment

To achieve the program goal, the planning group set the behavioral goal of patients increasing PSM behaviors and the environmental goals that included: “Providers adopt, implement, and maintain the PSM intervention” and that the “Healthcare system will fund PSM intervention.”

IM Step 2, Task 2: Specify the performance objectives

“Performance objectives” are the specific, observable, sub-behaviors that would lead to the expected outcome from Task 1. For any PSM behavior (e.g., exercising, meditating), patients must achieve the following performance objectives: make the decision to try the behavior, set a goal, practice the behavior, track and evaluate the behavior, report to the medical team about what they are doing, and maintain the PSM behavior. Interviews supported the inclusion of these performance objects. As one patient said: “Setting goals, it sounds like, has kept something in your mind. At least that there’s a plan, a plan of something trying to improve.” (Pt202)

IM Step 2, Task 3: Select determinants for the outcomes

The study team and planning group selected the final determinants of each of the performance objectives based on behavior change theory, the logic model of the problem, and the literature regarding effective PSM interventions. We specifically focused on SCT and self-regulation models and identified the key, changeable determinants of PSM behaviors as outcome expectations, self-efficacy, skills, and knowledge. Patient interviewees further confirmed that these determinants impacted their use of PSM approaches (Table 3).

IM Step 2, Task 4: Construct matrices of change objectives

Members of the study team then developed matrices of change objectives, or desired actions designed to address the determinants and lead to achievement of the performance objectives. were developed for each performance objective. For example, to address self-efficacy around practicing PSM behaviors, the goal would be to get participants to express their confidence about practicing that behavior (e.g., relaxation skills). Table 4 shows a matrix of change objectives that can apply to any PSM behavior (e.g., sleeping more; increasing physical activity).

Table 4.

Sample matrices of change objectives

Behavioral outcome: Patients with cirrhosis will increase their use of healthy PSM behaviors (IM STEP 2)
	Personal Determinants
Performance objectives	Knowledge	Skills and self-efficacy	Outcome Expectations
Make a decision to try a PSM behavior	State the consequences of non-PSM behaviors	Feel confident in being able to select a PSM	Recognize that deciding to use PSMs can decrease pain and improve function
Set a PSM behavioral goal	Acknowledge the need to change set a goal/make a plan	Express confidence in the ability to set a goal and make a plan	Recognize the advantages of making a plan/setting a goal
Practice PSM behavior	Define the importance of practicing	Express confidence in behavior practice	Recognize that PSMs will work better if they are practiced
Track and evaluate PSM behavior	Explain what is involved in tracking behaviors	Express confidence in ability to self-regulate and adjust behavior	Explain the importance of tracking and evaluating progress
Report about PSMs to medical team	State the importance of communicating about what works with the medical team	Demonstrate confidence in the ability to explain current behaviors to the medical team	Recognize that communicating with the medical team will improve pain
Maintain PSM behavior	Explain why it is important to maintain behaviors over time to address pain	Feel confident in ability to maintain behavior	Recognize that PSMs will work better if they are maintained
Environmental Outcome: Providers will implement PSM intervention (IM STEP 5)
	Personal Determinants
Performance Objectives	Knowledge	Skills and self-efficacy	Outcome Expectations
Providers will ask about pain/pain management	Acknowledge that asking about pain is important to patients	Feel confident in ability to address pain and symptoms	Recognize that talking about pain management will be well received by patients
Providers will refer to program	State that the program exists and what it does	Express confidence in ability to refer patients to program	Express that referral will be well received by patients

IM Step 2, Task 5: Create a logic model of change

The team and planning group finalized the logic model for change (Fig. 2).

Figure 2.

Logic Model of Change

IM Step 3: Design the intervention

IM Step 3, Task 1: Generate program themes, components, scope, and sequence

The study team met to review examples from the literature, data from qualitative interviews, and existing programs to outline the program themes and components. We determined that the program would be 12 weeks, based on other successful PSM interventions. The group brainstormed possible names for the program and determined that LEAP—Liver Education About Pain—would convey movement. A liver jumping over a hurdle was designated as the program “mascot” to enforce the idea that this program relates to the liver and to promote overcoming hurdles (Supplemental Figure 1).

In terms of scope and sequence, potential participants wanted the program to be individualized to their unique concerns and circumstances. One patient said, “I think managing pain is an individual’s thing…I guess you have to, sort of make it, you know, unique to each person.” (Pt310) One provider recommended that we “tailor the treatment for what, whatever the needs of the patient are.” (Pr008) In response to this request, we opted to develop a modular intervention with individual sessions that would allow patients to select relevant intervention components.

Group sessions allow participants to practice skills, set goals with the group, seek social support, and learn together. Most participants in interviews expressed interest in peer-based groups. For example, one patient said,

“I would do anything to be able to join with people, to talk amongst, amongst each other, what they done and to help them, what I give my comments for, to help me or the next person. And, all of us is there and we become a kind of a bond to help one another.”

(Pt211)

However, there was not consensus on this topic, with another patient stating, “I’m not gonna sit in a room with a bunch of people and tell them my problems.” (Pt304) The planning group debated the discordant opinions about group sessions and opted to pilot a hybrid group and individual format using alternating sessions, similar to another PSM.[46] The pilot will have flexibility to convert to only one-on-one if patients do not agree to participate in groups.

Table 5 outlines each intervention session. In the first individual session, patients will select relevant modules for future work. Potential modules were informed by PSM literature and interviews and finalized by the planning group (Table 5). Providers and administrators suggested virtual platforms, such as tech-based apps. However, patients wanted low-tech solutions and reported “I don’t like to text,” “I don’t have a cell phone,” and “I don’t do email.” To meet the needs of the patients, the decision was made to develop a low-literacy, written manual and to develop an intervention that could be delivered telephonically and without the use of complex computerized platforms.

Table 5.

Intervention sequence and content

Session Order, Type, and Content
Session	type	content
1	Individual session 1	Introduction to Chronic Pain and Cirrhosis
2	Group session 1	Relaxation exercises; goal setting, review challenges, model/practice skills
3	Individual session 2	Optional Module 1
4	Group session 2	Relaxation exercises; goal setting, review challenges; model/practice skills
5	Individual session 3	Optional Module 2
6	Group session 3	Relaxation exercises; goal setting, review challenges; model/practice skills
7	Individual session 4	Optional Module 3
8	Group session 4	Relaxation exercises; goal setting, review challenges; model/practice skills
9	Individual session 5	Optional Module 4
10	Group session 5	Relaxation exercises; goal setting, review challenges; model/practice skills
11	Individual session 6	Wrap-up and plan for maintenance
12	Group session 6	Wrap up and plan for maintenance
Optional Modules (participant selects 4 of 8 optional modules for individual sessions)
	Physical Activity and Your Pain	[I]t just makes you feel when you do exercise or you do anything like that you feel like you accomplished something. You feel better, you know. (Pt301) [W]e know that certain types of pain really do benefit from physical therapy and ongoing, you know, exercise… (A302)
	For people with metabolic liver disease: Losing Weight to Improve Your Pain	Based on the literature about weight loss and PSMs
	Stress Management and Your Pain	I meditate just about every day… I’m sure it helps [the pain], you know. (Pt309) [T]here’s a Mindfulness app that would probably be, you know, especially for patients…who have a problem kind of getting mentally past their pain. (Pr012)
	Sleeping Better to Help Your Pain	[I]t goes back to not being able to sleep more than 8 hours…If I can sleep, if I could sleep longer than 4 hours I would feel better. (Pt211)
	Thinking Differently About Your Pain	You gotta be mentally strong. I think you gotta be mentally strong and keep a positive attitude. Like when I was 18, an older guy who taught me how to play piano…he said, “Always stay positive and never think negative. What you can’t control throw out the back door.” And yeah, I live my life that way and it worked. It’s simple advice, but it works strongly. (Pt306) I talk to the Big Man upstairs. You know, he’s been pretty good…I’m always smiling and, you know, even my doctors, my shrink, he looks at me and goes, “Man, you got such a great attitude.” (Pt307) The brain…is the huge filter that determines our experience, in particular, our pain experience and how individuals cope… it can either be an amplifier or it can be a dampener, in terms of their pain response. (Pr009) I just keep on saying to myself that I’m stronger than the pain or that I will like think of a blank piece of paper. I’ll say that to myself. (Pt208)
	Talking with people about pain/navigating the medical system	I think it’s important that the patient is empowered to find things, empowered and educated well, to find things that are not just opiates…. I’ve always found that my patients want to feel like they have a sense of control. (Pr012)
	Taking Opioid Pain Medications safely	The patients are already on these opioids and you need to know how to manage them, which drugs, you know, are particularly risky in patients and should be avoided and which, which opioids might be used, but with appropriate caution. And, how to do it. (A302)
	Coping with Pain without Drugs or Alcohol	[W]hen you don’t have [opioids], you know, it’s just like before, when I had my back, man, I was taking left and right because I was getting rebound pain if I missed a pill. (Pt307) I knew from being a medic you shouldn’t take meds and drink, but I was sitting there bored to tears. So I’d have some scotch and water, I’d wash my Percocet down with that. And one time, I’m on the phone and after I was done talking my wife says, “Oh, who was that?” “That’s just—I don’t know.” I says, “Why you ask?” She says, “Whoever it was you had a hell of a good time talking to them.” I said, “I don’t—I don’t know.” But that day I remember that and I stay away, you know. Besides seeing other people that are, you know, having [a] bad time with [opioids]. (Pt304) Well years ago I used to drink. Until I found out I had a bad liver. That was back in ‘15, I believe. Since then it’s been clean and narrow. (Pt302)

In terms of interventionist, we asked participants where pain should be managed. There were discordant opinions. However, the clinicians indicated that specialists did not have time to adequately teach PSM behaviors, and administrators spoke about primary care being unable to serve in this role. Thus, we decided to use a health coach, since this would be scalable, did not extend already limited clinic resources, and was relatively inexpensive.

IM Step 3, Tasks 2&3: Choose theory- and evidence-based change methods and designing practical applications

The planning group selected theory-based change methods that had been used in other PSM interventions and were recommended by IM developers.[25–28] Table 6 provides an example of how the determinants of change objectives were addressed using behavior change methods and associated applications. For example, to address self-efficacy around PSM behaviors and skills, coaches will first model behaviors (e.g., relaxation exercises) and then engage participants in guided skill practice through subsequent sessions. Matrices were similarly created for each determinant and change objective.

Table 6.

Sample Methods and Applications that address change objectives at the behavioral/patient and environmental/provider levels.

Behavioral outcome of “practicing PSM behaviors”
Determinant and change objective	Method	Application
Self-efficacy in using PSM behaviors (e.g., relaxation skills)	Demonstration and Guided practice	Group sessions will start with guided relaxation exercises
Knowledge about PSMs behaviors (e.g., relaxation skills)	Information	Facilitator will start group session with a fun quiz about relaxation and PSM
Outcome expectations about engaging in PSM behaviors (e.g., relation skills)	Goal setting; anticipated success	Participants will set short an long-term goals with their coach and then imagine the outcome of successfully achieving these goals in their journals.
Environmental outcome of “clinicians implement PSM intervention”
Implementation determinants and change objectives	Strategy	Application
Clinicians recognize pain and refer appropriate patients to intervention (skills and self-efficacy)	Education/training	Presentation at faculty meeting using structured slide deck
Clinicians are aware of evidence supporting PSMs for cirrhosis pain (intervention evidence and relative advantage)	Persuasive communication Modeling	Pilot trial (next step) with dissemination of information to clinicians Collect and present testimonials of providers who have seen benefits
Clinicians expect that the implementation will lead to change (outcome expectations)	Persuasive communication	Collect and present patient testimonials Collect and present data at faculty meeting

IM Step 4: Program Production

IM Step 4, Task 1: Refine program structure and organization

Nuances of cirrhosis necessitated tailoring of the intervention. For example, encephalopathy-mediated cognitive impairment and organic sleep disturbances in this population render sleep hygiene interventions inadequate. Thus, the sleep module includes information about encephalopathy and cirrhosis-specific considerations. Similarly, weight is more complex in patients with cirrhosis than in general populations, related to issues of water retention, sarcopenia, and/or obesity. While many patients with cirrhosis often are underweight or sarcopenic, weight loss is central to managing non-alcoholic fatty liver disease (NAFLD). These nuances presented challenges that were addressed by tailoring the modules for subpopulations (e.g., weight loss was only recommended in the context of NAFLD) and allowing patients to select relevant intervention modules (Table 5).

IM Step 4, Tasks 2–4: Plan, draft, refine, and produce program materials

Program materials were prepared by the production team (SG, WD, SR, GH), with iterative input from the study team and planning group. Example manuals from other PSM interventions were reviewed and adapted based on the findings in Steps 1–3. The document was reviewed for readability by WD, who has expertise in low-literacy manual production. A sample group session outline and a sample individual module are presented in the Supplemental Materials. Materials were reviewed by the planning group and modifications were made iteratively based on feedback. For example, the language used to discuss addiction was modified based on input from an expert in opioid use disorder. A person with cirrhosis and hepatologist modified the liver-related aspects of the manual. Further testing and refinement will occur through the pilot intervention, which is the next step of the project.

IM Step 5: Program Implementation Plan

IM Step 5, Task 1: Identify potential program users

Program users are conceptualized as people who will adopt, use, and provide resources to sustain the program. People with cirrhosis reported that they received care from a variety of persons, including social workers, surgeons, primary care, mental health, and other providers. However, patients with cirrhosis have regular appointments with hepatology and primary care providers, who were deemed to be the primary targets of implementation at the environmental level. Likewise, administrators with the policy and financial assets to support the program were included in interviews and goals.

IM Step 5, Task 2: State outcomes and performance objectives for program use

The planning group identified the desired outcomes to include adoption, implementation, and maintenance by providers and administrators. Specific performance objectives to obtain these desired outcomes included discussing and referring eligible patients to the PSM intervention and identifying a sustainable financing plan (Figure 2).

IM Step 5, Task 3: Construct matrices of change objectives for program use

Implementation barriers included patient transportation and competing demands, provider knowledge, relative priority, and skills, and stigma, infrastructure, and funding. These environmental determinants were mapped to CFIR (Table 3) and included in the logic model for change (Figure 2). Change objectives for program use were developed by crossing determinants and performance objectives (Table 4).

IM Step 5, Task 4: Design implementation interventions

The implementation interventions, or strategies, we identified to address implementation barriers were selected based on literature and behavior change theory. To address provider knowledge, skills, and self-efficacy. we developed training and education materials for providers who would refer patients. Table 6 illustrates some sample applications and methods for implementation supports. The pilot trial (see below) will allow us to further modify and tailor the implementation supports before scaling the intervention.

IM Step 6: Evaluate the Intervention

IM Step 6, Task 1: Write process evaluation questions

Patient effectiveness outcomes will include pain, pain interference, and opioid dose. Implementation outcomes will include fidelity, acceptability, and appropriateness. Intervention dose, modules chosen, and changes to the program will also be evaluated.

IM Step 6, Task 2: Develop assessment measures

The planning group collaborated to select assessment measures. Pain and pain interference will be assessed with the PEG and PROMIS pain interference measure.[47, 48] Appropriateness and acceptability of each module will be evaluated using brief, validated measures.[49] Acceptability of the overall intervention will secondarily be operationalized as the approach:enrollment ratio and the number of sessions attended (out of 12). The health coach will track the fidelity and adaptations to the LEAP intervention using a checklist that we have developed for the intervention and the FRAME to track adaptations.[50]

IM Step 6, Tasks 3&4: Specify evaluation design and complete the evaluation plan

To evaluate the efficacy of the intervention, the study team will conduct a 40-patient randomized controlled pilot study in which patients will be block randomized in groups of 5 to intervention or usual care control.

DISCUSSION

This study employed Intervention Mapping to develop a novel pain self-management intervention for patients with cirrhosis. We identified several novel themes through this work, including areas of discordant opinions between patients and providers, novel intervention targets, and notable implementation barriers. Aiming to develop an intervention that was implementable in the real world, we proactively assessed potential implementation barriers using CFIR early in the intervention development process.

Short-term cognitive behavioral interventions have been previously associated with long-term benefits, [51, 52] and PSM interventions have a successful track record in other populations.[22–24, 53–56] As such, components of existing PSM interventions, such as using exercise and relaxation, were incorporated into the LEAP intervention. However, the unique needs of patients with cirrhosis necessitated adding novel components to the final intervention.

To address challenges regarding the complexities of cirrhosis care, we developed a novel intervention component targeting medical navigation. Health literacy predicts health outcomes but is challenging for patients with cirrhosis, given the complexity of their disease and multiple comorbidities.[57–61] Thus, developing a low-literacy, educational intervention could address multiple issues for patients with cirrhosis. There is significant stigma around addiction, pain management, mental health, and cirrhosis which are also potential barriers to seeking care.[62–68] Thus, this module also aims to empower patients to engage in self-advocacy, which also addresses a specified need identified in the interviews.

One novel facet of our approach was considering environmental determinants in developing a person-centered intervention. Most interventions, even evidence-based interventions, are never implemented in the real world.[69] This results in billions of wasted research dollars and inadequate access to evidence-based medical care for people with complex conditions.[69] As such, we aimed to understand implementation barriers early in the intervention development process, employing the CFIR in order to comprehensively evaluate barriers to and facilitators of providing pain management for patients with cirrhosis. We found that there was no consensus about how to potentially embed this in medical care, and most providers identified space, staffing, and training as barriers to implementing as a part of routine care. Therefore, we plan to pilot test an intervention that is outside of the clinic and telephonic to be flexible and meet the needs of both patients and the overwhelmed health care system. Choosing to implement the intervention outside of clinic also obviated the need to address the innumerable barriers we found to integrating programs into hepatology clinic.

These results highlight the importance of including diverse participant groups in program building processes. IM was selected to support intervention building because of its focus on engaging a diverse group of advisors to understand the problem from multiple perspectives. Illustrative of divergent perspectives, we heard non-patient advisory skepticism about perceived patient buy-in for behavioral interventions, despite high patient-reported motivation to participate. Similarly, providers reported patients would prefer an App or computer-based intervention, though patients reported the opposite preference. Several hepatologist participants spoke about the importance of listening to patient needs and aligning the intervention with their preferences, suggesting that these perceptions may be bridged through routine communication during medical encounters. A growing body of literature supports using shared decision to align medical care and education with patient preferences.[70, 71]

There were several limitations of this study. While we tried to address selection bias by including providers, administrators, and patients from two healthcare systems, there is always the potential for selection bias in qualitative studies with small sample sizes. The use of theory can help to mitigate this limitation, and it is reassuring that the themes identified through the interviews aligned closely with social cognitive theory and a widely accepted implementation framework. While the described barriers were similar in both sites, this may reflect the small sample size of providers and administrators available at each site, rather than universality of barriers. Further research is needed to evaluate the external validity of these findings. Another limitation was that this study included participants from only two medical centers in a single city, which may limit the external validity of these findings. Moreover, though we attempted to recruit caregivers, only two patients referred caregivers, such that these data could not be analyzed. Finally, there is inherently subjectivity of IM, where theories are selectively applied. We tried to address this by involving diverse advisors and using the CFIR, which is a meta-framework. However, there may be other theories or frameworks that could add to these findings or alter the interpretation of these data. Despite these limitations, this intervention was designed to meet the needs of a unique population using a rigorous IM process.

In conclusion, we developed a theory-based, advisory informed intervention to address chronic pain in patients with cirrhosis. In doing so, we identified divergent perspectives across advisory groups, illustrating the importance of including diverse voices in intervention development. We also identified specific unmet needs of patients with cirrhosis that were addressed through IM.