Feasibility Pilot Trial of a Tailored Medication Adherence-Promotion Intervention for Adolescents and Young Adults with Cancer: Study Design and Protocol

Abstract

Background:

Medication non-adherence is common among adolescents and young adults (AYAs) with cancer and associated with poor health outcomes. AYAs with cancer endorse multiple barriers to adherence that differ across individuals, suggesting that tailoring intervention content to an AYA’s specific barriers may have the potential to improve adherence. The purpose of this manuscript is to report on ORBIT-guided Phase I design efforts to create the first tailored adherence-promotion intervention for AYAs with cancer and the study protocol for the ongoing Phase II pilot feasibility trial.

Methods:

Phase I design included qualitative interviews (n = 15 AYAs) to understand patient preferences for adherence-promotion care, development and refinement of a best-worst scaling exercise barriers tool (n = 5 AYAs), and development of intervention modules and a tailoring algorithm. In the ongoing Phase II pilot feasibility trial, AYAs (ages 15-24 years) with cancer currently taking oral chemotherapy or prophylactic medication will be recruited from three children’s hospitals. Feasibility, acceptability, and usability will be assessed and these outcomes along with data on medication adherence will be used to inform the next phases of intervention development and testing.

Conclusions:

If promising, this program of research ultimately has the potential to equip clinicians with additional strategies for supporting adherence among AYAs with cancer. NCT05706610

1. Introduction¹

Each year, more than 11,000 adolescents and young adults (AYAs) ages 15 to 24 years are diagnosed with cancer in the United States.¹ For many of these AYAs, treatment includes oral chemotherapy and/or prophylactic (e.g., antibiotic) medications. Unfortunately, 21-60% of AYAs are non-adherent,² or demonstrate medication-taking behavior that does not align with the treatment protocol,³ placing them at an increased risk for poor health outcomes. Specifically, among patients with acute lymphoblastic leukemia, oral chemotherapy non-adherence is associated with a 2.7-fold increased risk of relapse.⁴ In addition, non-adherence to antibiotic medications is associated with lower survival rates among AYAs with multiple cancer types.⁵

Because the thousands of AYAs⁶ currently demonstrating non-adherence are at risk for devastating consequences which could be prevented, adherence-promotion efforts have been recognized as a critical component of AYA cancer care.^7-9 AYA adherence-promotion interventions, however, are limited and only the Re-Mission videogame has demonstrated efficacy in increasing chemotherapy and prophylactic adherence among AYAs with multiple cancer types.^10-12 This means that clinicians currently have minimal guidance about how best to support their patients.

A promising target for novel intervention efforts is an individual’s unique adherence-related barriers.^13,14 AYAs experience barriers spanning multiple domains including capabilities (e.g., difficulty swallowing pills, forgetting), opportunities (e.g., not wanting to take medication in front of peers, difficulty picking up prescriptions), and motivation.^13-16 While many of these barriers are amenable to behavioral intervention, the specific components or behavior change techniques (BCTs) required to overcome each barrier differ.^17-20 For example, an AYA who forgets to take medication may benefit from prompts/cues while an AYA who is struggling with low motivation may benefit from rewards and/or motivational interviewing techniques.²⁰ Guided by these data and theory,^20-22 we hypothesize that, to improve their adherence, AYAs must receive BCTs matched to their barriers.

One challenge to creating an intervention in which AYAs receive BCTs matched to their barriers is that 79% of AYAs endorse a different set of barriers than any of their peers.¹³ This suggests that most AYAs will need a unique set of BCTs, or an intervention with tailored content, to overcome their barriers and improve their adherence. Our goal is to follow the ORBIT model²³ to design and evaluate a tailored intervention, SUMMIT (SUpporting Medication Management with Individualized Treatment), that meets this need. This manuscript reports on Phase I design efforts and the study protocol for the ongoing Phase II pilot feasibility trial.

2. Phase I Overview

Phase I included three components: 1) qualitative interviews to understand patient preferences for an adherence-promotion intervention; 2) development and refinement of a best-worst scaling exercise (BWSE) barriers tool to identify each AYA’s top barriers; and 3) development of intervention modules and a tailoring algorithm linking each barrier to a module. Study procedures were approved by each site’s IRB.

2.1. Patient Population

The patient population for Phase I included AYAs (ages 15-24 years) with a diagnosis of cancer currently prescribed an oral chemotherapy or prophylactic medication as part of their cancer treatment regimen. An age range of 15-24 years was selected as AYAs endorse different barriers and plan, problem-solve, and prioritize decisions like medication adherence differently than children (under 15 years)^24,25 and older adults (≥ 25 years).^26-32 While the National Cancer Institute defines the AYA period as 15-39 years,³³ using a narrower definition allowed us to design an intervention to account for the developmental factors unique to ages 15-24 years.

AYAs participating in the qualitative interviews were recruited from two children’s hospitals in the United States and AYAs participating in the best-worst scaling exercise refinement were recruited from one Midwestern children’s hospital. Informed consent (AYAs 15-17 years: parent/guardian permission and AYA assent; AYAs 18-24 years: AYA consent) was obtained for all participants.

2.2. Phase I: Qualitative interviews

In the first component of Phase I, 15 AYAs participated in a semi-structured qualitative interview where they answered questions regarding their preferences for an adherence-promotion intervention. Demographic and clinical characteristics are presented in Table 1 and transcripts were analyzed using thematic analysis.³⁴ AYAs expressed a desire for personalized adherence-promotion care tailored to their barriers. AYAs reported that their reliance on caregivers/support persons varied over time and that they wanted the ability to choose if/how to involve a caregiver/support person in the intervention. AYAs also highlighted the importance of building trust and rapport with the interventionist. Finally, AYAs expressed a desire for regular contact with the interventionist that occurred no more than once a week. Findings were used to inform the selection of the intervention frequency and duration as well as intervention content (i.e., tailored BCTs, rapport-building exercises).

TABLE 1.

Demographic and Clinical Characteristics

	Interviews (n = 15)	Best-worst scaling (n = 5)
Characteristic	n (%)	n (%)
Age (in years) at Baseline, M(SD)	20.95 (2.21)	19.09 (3.72)
Sex Assigned at Birth
Female	5 (33%)	4 (80%)
Male	10 (67%)	1 (20%)
Race
American Indian or Alaska Native	0 (0%)	0 (0%)
Asian	0 (0%)	0 (0%)
Black or African American	1 (7%)	0 (0%)
White	13 (87%)	5 (100%)
Other Race	1 (7%)	0 (0%)
Ethnicity
Hispanic	2 (13%)	1 (20%)
Not Hispanic	13 (87%)	4 (80%)
Educational History
High school student	1 (7%)	3 (60%)
High school graduate	6 (40%)	0 (0%)
College student	6 (40%)	2 (40%)
College graduate	2 (13%)	0 (0%)
Household Composition
Resides with parents	9 (60%)	5 (100%)
Resides with other family members	1 (7%)	0 (0%)
Resides with friends/roommates	3 (20%)	0 (0%)
Resides with significant other	2 (13%)	0 (0%)
Cancer Diagnosis
Leukemia	7 (47%)	2 (40%)
Lymphoma	4 (27%)	2 (40%)
Sarcoma	4 (27%)	1 (20%)

2.3. Phase I: Best-worst scaling exercise

Tailoring intervention content to each AYA’s individual barriers requires a tool capable of identifying an AYA’s most problematic barriers. Best-worst scaling is a survey methodology that quantifies the relative impact of multiple factors (e.g., barriers) on a behavior (e.g., adherence) and can be used to identify an AYA’s “top” barriers.^35-37 In a best-worst scaling exercise (BWSE), an individual is presented with several “choice tasks” (or items) comprised of different combinations of factors (barriers) and asked to pick the most (best) and least (worst) influential factor in each task (see Figure 1 for a sample choice task). As BWSEs have not yet been used to assess barriers to medication adherence,³⁸ a novel BWSE barriers tool was created. BWSE tool creation includes: 1) identifying a comprehensive list of factors (e.g., barriers) that may impact a behavior (e.g., adherence); 2) determining the number of choice tasks to be included; 3) determining the number of factors (barriers) to be included in each choice task; and 4) generating the specific combination of factors (barriers) to be included in each choice task.^35,36 Following best-practice guidelines,^35,36 results of our systematic literature review,² qualitative data from our semi-structured interviews (2.2 Phase I: Qualitative Interviews), and clinician feedback were synthesized to compile a list of 18 barriers to adherence faced by AYAs with cancer (Table 2, Column 1). Next, a balanced incomplete block design was used to determine the number of choice tasks and barriers per choice task, resulting in a BWSE tool with 34 choice tasks with 9 barriers each.³⁸ Finally, the Street and Burgess method was used to generate the specific combinations of barriers to be included in each choice task and choice tasks were programmed into Sawtooth software for adminsiration.^37,39 A sample choice task is included in Figure 1.

Figure 1.

Sample best-worst scaling exercise (BWSE) barriers tool choice task

TABLE 2.

Tailoring Algorithm Linking Barriers to Content (Modules of BCTs)

If this barrier	Deliver one or more of these BCTs
Knowledge of how/when to take it	• Instruction on how to perform behavior
Forget to take it	• Problem-solving • Prompts/cues • Associative learning
Don’t realize when I run out	• Problem-solving • Prompts/cues • Mental resources
Hard to time with meals	• Problem-solving • Instruction on how to perform behavior
Don’t like the side effects	• Instruction on how to perform behavior
Don’t like the taste	• Add objects to environment
Hard to swallow	• Graded tasks • Verbal persuasion
Too many pills to take	• Problem-solving
Don’t like taking in front of others	• Information about others’ approval • Restructure environment
Not home when doses are due	• Problem-solving • Prompts/cues • Restructure environment
Gets in the way of other activities	• Focus on past successes • Problem-solving
It’s inconvenient	• Focus on past successes • Problem-solving
Don’t get enough help	• Practical social support • Social reward
Hard to get refills	• Practical social support • Prompts/cues • Restructure environment
Can’t afford it	• Practical social support
Am tired of taking it	• Pros/cons • Goal setting • Incentives/rewards
Don’t feel like taking it	• Pros/cons • Goal setting • Incentives/rewards
Doesn’t matter if I take it	• Information about consequences • Pros/cons • Goal setting • Incentives/rewards

Five AYAs provided qualitative and quantitative feedback to assess BWSE barriers tool complexity, readability, and acceptability. Specifically, AYAs participated in a three-step test-interview⁴⁰ during which they completed the BWSE barriers tool and were asked to “think aloud” or verbalize their thought process. After the three-step test-interview, AYAs completed the System Usability Scale⁴¹ and System Usability Scale-Adjective Rating.⁴²

Demographics and clinical characteristics are included in Table 1. All 5 AYAs described the BWSE barriers tool as “easy to use,” with SUS scores exceeding recommended cut-points (M[SD] SUS=91.00[10.25]; M[SD] SUS Adjective Rating=6.00[0.71]). AYA feedback was used to refine the instructions and formatting and the final BWSE barriers tool is available upon request.

2.4. Phase I: Intervention modules and tailoring algorithm

Possible intervention targets include the 18 barriers comprising the BWSE barriers tool. The intervention includes 18 modules of BCTs - 1 for each barrier. Module development began by consulting the Theory and Techniques Tool,⁴³ a heat map created as part of Human Behaviour Change Project that triangulates¹⁹ results of a literature review¹⁷ and expert consensus¹⁸ to illustrate the strength of the relationship between each BCT and barrier. Using the Theory and Techniques Tool, BCTs with theoretical and/or empirical support for overcoming each barrier were identified. Feedback from oncologists and licensed pediatric psychologists providing adherence-promotion interventions to AYAs as part of their clinical practice was used to refine the list of BCTs for each module. Once each module was finalized, a tailoring algorithm linking each barrier to the appropriate module was created (Table 2). Interventionists follow this algorithm to identify appropriate BCTs for each barrier (see 3.5.1.). Finally, a treatment manual including a script for delivering each BCT was created with feedback from five pediatric psychologists whose practices are largely guided by cognitive-behavioral therapy (CBT) and acceptance and commitment therapy (ACT) principles. BCT scripts were informed by examples included in the BCT taxonomy¹⁷ and relevant clinical approaches (e.g., BCT: “Focus on past successes” – ACT; BCT: “Problem-solving” – CBT; BCT: “Pros/cons” – Motivational Interviewing) (Supplemental Table 1).

3. Phase II: Pilot Feasibility Trial

Phase II includes a pilot feasibility randomized controlled trial (RCT) of the tailored adherence-promotion intervention (SUMMIT) as compared to uniform standard of care. We hypothesize that the trial will meet enrollment, retention, fidelity, and assessment completion feasibility criteria (Hypothesis 1a) and that AYAs will rate the intervention as easy to use and acceptable (Hypothesis 1b). The secondary aim is to explore group differences in improvements in electronically-monitored medication adherence to inform the next phases of intervention development.^44,45 All study procedures were approved by a single IRB (IRB of record: Cincinnati Children’s Hospital Medical Center).

3.1. Participants

Participants will be recruited from three children’s hospitals in the United States. AYAs who are 15-24 years of age and currently taking an oral chemotherapy or prophylactic medication with an anticipated duration of at least four months as part of their cancer treatment regimen will be eligible. AYAs will be excluded if they are not fluent in English, evidence significant cognitive deficits that may interfere with intervention comprehension and/or participation, and/or are of a current medical status which precludes their participation.

3.2. Recruitment process, consent, and eligibility determination

Study staff will review patient lists to identify potentially eligible AYAs and verify eligibility via electronic medical record review. Eligible AYAs may be recruited in-person or remotely. AYAs recruited remotely may be mailed an opt-out letter and flyer describing the study procedures and then contacted via phone.

Informed consent will be obtained by trained research staff and may be obtained in-person or remotely. During the consent process, all pertinent aspects of consent/assent will be covered. In addition, the medication-taking regimen will be assessed to minimize the potential that the use of an electronic adherence monitoring device could adversely impact adherence (e.g., if a patient is currently using a pill box and would need to switch to a bottle) and ensure the patient feels their current role in medication-taking is appropriate for a patient-focused intervention. Patients will be informed that their care will not be affected by their participation decision. For participants under the age of 18, permission will be obtained from at least one parent or guardian. Assent will be obtained from all participants under 18 years of age.

AYAs who provide consent will participate in a 4-week run-in period during which they will store their eligible medication in an electronic adherence monitoring device. AYAs who demonstrate < 95% of adherence and use the electronic adherence monitoring device without difficulty will be asked to complete the pre-treatment assessment (Figure 2). AYAs who complete the pre-treatment assessment will then be eligible for randomization.

Figure 2.

Study procedures

3.3. Randomization and blinding

AYAs will be randomized 1:1 into the tailored program (SUMMIT intervention) or feedback program (uniform standard of care⁴⁶ control). Group assignment will occur according to a 3 x 2 randomization table stratified by site (3 sites) and age (younger AYAs: 15-17 years; older AYAs: 18 – 24 years) in which block sizes of 2 or 4 were randomly chosen within each stratum.

3.4. Assessment and outcome measures

AYAs will complete pre-treatment measures prior to randomization and post-treatment measures approximately 3-4 weeks after program completion (Figure 2, Table 3). AYAs will use the electronic adherence monitoring device for medication administration from study enrollment to post-treatment assessment completion.

TABLE 3.

Self-Report Measure Administration Schedule

Timepoint
Domain	Measure	Pre-	Post-
		Usability	System Usability Scale (SUS)41		X
Usability	System Usability Scale Adjective Rating42		X
Acceptability	Treatment Acceptability		X
Acceptability	Open-Ended Acceptability Questions		X
Exploratory Outcome	Medical Adherence Measure47	X	X
Exploratory Outcome	Barriers Checklist	X	X
Potential Covariate	Demographic Characteristics	X
Potential Covariate	Clinical Characteristics	X
Potential Covariate	Allocation of Treatment Responsibility48	X
Potential Covariate	PROMIS Anxiety Short Form49,50	X
Potential Covariate	PROMIS Depression Short Form49,50	X

3.4.1. Feasibility.

The primary outcome is enrollment rate. Secondary feasibility outcomes include retention rate, fidelity, and assessment completion. Data to inform the calculation of enrollment, retention, and assessment completion rates will be tracked prospectively. For AYAs randomized to the tailored program (SUMMIT intervention), components of intervention fidelity will also be assessed (contact, length, duration, and content⁵¹) via a standardized fidelity checklist completed by a trained member of the study team following a review of session recordings.

3.4.2. Usability and acceptability.

Usability will be assessed via the System Usability Scale⁴¹ and System Usability Scale Adjective Rating.⁴² Acceptability will be assessed via a 7-item rating of psychological treatment acceptability developed by the authors to assess the domains of the theoretical framework of acceptability.^52,53 Open-ended questions selected based on guidelines for using qualitative data in a feasibility study^54-56 will also be asked to inform potential refinements based on patient (i.e., “What did you like best about the adherence program?”; “What could we change to make the program better?”) and interventionist (i.e., “Describe any feasibility, usability, or acceptability challenges”) feedback.

3.4.3. Adherence.

At enrollment, all AYAs will be given an eCAP^™ electronic adherence monitoring device and asked to use the eCAP^™ for eligible medication administration throughout the study. eCAP^™ devices include a computer chip that records the date and time of each device opening and demonstrate high accuracy.⁵⁷ Each AYA’s eCAP^™ data will be exported to Excel. Study staff will obtain the AYA’s prescribed medication regimen and non-monitored periods (e.g., hospitalizations) from the medical record and adjust the prescription to account for holds, non-monitored periods, and changes in the regimen. Adherence will be calculated by comparing eCAP^™ data to the prescribed regimen.

3.4.4. Covariates.

Demographic information (i.e., sex assigned at birth, gender, ethnicity, race, disability status, education, employment status, income, rurality,⁵⁸ family structure), clinical characteristics (i.e., diagnosis, date of diagnosis, medical regimen, additional adherence-promotion interventions delivered by medical team), anxiety symptoms (assessed via age-specific PROMIS short forms),^49,50 depressive symptoms (assessed via age-specific PROMIS short forms),^49,50 and treatment responsibility⁴⁸ will be collected and explored as potential covariates.

3.4.5. Additional outcomes.

Additional exploratory aim outcomes will include the Medical Adherence Measure,⁴⁷ a self-report measure of adherence, and an investigator-created barriers measure assessing the presence or absence of the 18 barriers targeted by the intervention.

3.5. Interventions

3.5.1. Intervention condition: Tailored program (SUMMIT).

AYAs in the intervention group will participate in 4 bi-weekly intervention sessions (1 every other week over 8 weeks) estimated to last 30-45 minutes each. On alternating weeks, AYAs will receive a text message check-in. The intervention dose and duration were selected to align with AYA preferences (2.2 Phase I: Qualitative Interviews) and the number of sessions is consistent with that of other pediatric adherence-promotion interventions.⁵⁹

Intervention content includes BCTs matched to each AYA’s top barriers. In Session 1, a 3-step process will be followed to tailor content. First, AYAs will complete the BWSE barriers tool and results will be analyzed to generate a rank-ordered list of barriers. Second, at the beginning of Session 1, the interventionist will review the AYA’s top three barriers and work with the AYA to determine which of these barriers the AYA would like to target. Third, the interventionist will follow the standardized tailoring algorithm (Table 2) and deliver the module of BCTs matched to the target barrier. At the beginning of the module, the interventionist will ask questions to understand the barrier (e.g., “Tell me more about times when you forget to take your medication”) and previous efforts to address the barrier (e.g., “Is there anything that you’ve tried to help you remember to take your medication? How did that go?”). Based on this information, the interventionist will deliver one or more BCTs from the module. Session 1 will conclude with the creation of an action plan. In Session 2, the interventionist will review adherence data and make modifications to the Session 1 action plan as necessary. As barriers may change over time, each AYA will complete the BWSE barriers tool a second time prior to Session 3 and the interventionist will partner with the AYA to choose a barrier to target in Sessions 3 and 4. AYAs may choose to target a new barrier in Sessions 3 and 4 or continue to work on the same barrier throughout the study. All four sessions include rapport building, psychoeducation, and a review of adherence data.

On alternating weeks, AYAs will receive a text message check-in including their action plan and adherence calendar (BCTs = self-monitoring of behavior and feedback on behavior). The AYA may respond to obtain additional support on how to implement the action plan if desired.

3.5.2. Uniform Standard of Care (USC) control condition: Feedback program.

Per purpose-guided trial design,⁶⁰ the pilot feasibility RCT control group should match that of the future efficacy RCT. As the goal of the subsequent RCT will be to evaluate how well the SUMMIT intervention works compared to a clinically-relevant alternative, a uniform standard of care⁶¹ control group was selected. In the uniform standard of care group, or feedback program, AYAs will receive a weekly text including a calendar depicting their eCAP^™ adherence data and instructions to contact a medical team member should they desire additional support (BCTs = self-monitoring of behavior, feedback on behavior, social support [unspecified]; Supplemental Figure 1). Per the NIH OBSSR Panel’s Pragmatic Model for Comparator Selection,⁴⁶ uniform standard of care is an appropriate control group as it includes the routine assessment of adherence recommended as a standard of care¹² and represents the highest level of care currently provided at any of our sites. A uniform standard of care control group also addresses the ethical concerns inherent in failing to provide any support.⁶¹ Text messages were selected to match the treatment modality to that of intervention check-ins. While the purpose of a uniform standard of care control differs from that of an attention control group,⁴⁶ the feedback group was designed to include an equivalent number of contacts (8 total) to the intervention.

3.5.3. Interventionists.

Interventionists are psychosocial providers with at least a master’s degree and are supervised by a licensed clinical psychologist with expertise in AYA oncology and adherence science. All interventionists are trained by the supervisor in AYA oncology, adherence science, and BCTs. Interventionists attend weekly supervision meetings during which questions, concerns, and deviations in treatment fidelity are reviewed.

3.6. Data safety monitoring

As the proposed trial is expected to be of minimal risk, a data and safety monitoring board (DSMB) was not convened. However, unanticipated problems related or possibly related to participation that place participants or others at a greater risk of harm than was previously known or recognized will be reported to an Independent Data Safety Monitor and the IRB.

3.7. Data management

All data will be managed by the main study site, Cincinnati Children’s Hospital Medical Center. Data from pre-treatment and post-treatment assessments will be directly entered by participants and stored in REDCap. Medication adherence data will be transmitted from the AYA’s eCAP^™ to the CCHMC study team via the CertiScan^® cloud and then saved in a secure location at the main study site. Medical record reviews and fidelity data will be entered in REDCap and double-checked for accuracy.

3.8. Power analyses and sample size considerations

As the primary aim is to evaluate feasibility, an enrollment target of 40 was set to evaluate our ability to achieve an enrollment rate (primary outcome) of 70% based on the anticipated eligible patient population. Assuming retention rates ≥ 80%, at least 32 AYAs (16 in each group) are expected to complete all study procedures. A sample size of 32 is consistent with sample size recommendations for pilot RCTs with the goal of informing future efficacy trials.⁶²

3.9. Analytic plan

A Statistical Analysis Plan (SAP) document has been developed for this trial.

3.9.1. Hypothesis 1a: Feasibility.

The hypothesis (Hypothesis 1a) that the trial will meet enrollment, retention, fidelity, and assessment completion feasibility criteria will be supported if descriptive statistics indicate that pre-determined thresholds are met. The percent of eligible approached AYAs who enroll will be calculated to explore the hypothesis that the trial will meet the enrollment rate goal of 70%. The percent of AYAs completing all study procedures and with complete data will be calculated to evaluate the hypothesis that the retention and assessment completion rates will be “acceptable” as defined as ≥ 80%.⁶³ In addition, the percent of AYAs randomized to the intervention arm who receive the intervention contact, length, duration, and content as specified in the protocol will be calculated to evaluate the hypothesis that each component of intervention fidelity will be “high” as defined as ≥ 80%.⁵¹

3.9.2. Hypothesis 1b: Usability and acceptability.

The hypothesis (Hypothesis 1b) that AYAs will rate the trial conditions as easy to use and acceptable will be supported if descriptive statistics indicate that pre-determined usability and acceptability thresholds are met. Usability and acceptability scores for the intervention and control conditions will be calculated separately. The mean System Usability Scale and System Usability Scale Adjective Rating scores for each group will be calculated to explore the hypothesis that AYAs will rate the programs as easy to use as defined by a mean System Usability Scale score ≥ 68⁴¹ and a mean System Usability Scale Adjective Rating ≥ 4.⁴² Mean scores on the treatment acceptability measure for each group will be calculated to explore the hypothesis that AYAs will rate each program as acceptable as defined by a score ≥ 28.

3.9.3. Exploratory Aim 2: Adherence.

Consistent with guidelines for feasibility pilot trials,⁴⁵ “preliminary efficacy” will not be calculated as this trial is not adequately powered to evaluate impacts on adherence. However, longitudinal trajectories of adherence will be explored by group to inform the next phases of intervention development.

4. Discussion

While non-adherence to cancer treatment protocol medications is common among AYAs with cancer, evidence-based treatment options are limited.¹⁰ Multiple studies suggest that barriers may be a promising target for adherence-promotion efforts,^13,14 with individual-level differences in barriers highlighting the need for personalization. The goal of the research described here is to develop and test a novel adherence-promotion intervention for AYAs with cancer (SUMMIT) in which individuals receive BCTs matched to their unique barriers. As AYA adherence-promotion interventions to date have delivered the same BCTs to all AYAs, the current intervention represents a shift from “one-size-fits-all” to tailored and personalized care consistent with the broader precision medicine efforts within oncology.

Although tailored adherence-promotion interventions have been developed for other populations, this intervention includes two unique components with the potential to advance adherence science. First, we created a BWSE barriers tool to identify each AYA’s top barriers. Historically, trials have largely relied on clinical interviews or Likert-type scales to identify barriers,^59,64,65 but both methods have limitations. Specifically, clinical interviews do not produce the standardized results required for replication and Likert-type scales may result in AYAs rating multiple barriers as equally problematic or endorsing no barriers, making it difficult to discern which barrier(s) to prioritize. Best-worst scaling exercises overcome both limitations as they produce a rank-ordered list of barriers for each individual AYA. AYAs rated our BWSE barriers tool as easy to use during initial testing and if this tool is feasible to implement within the trial, offers an additional tailoring method for future adherence-promotion efforts. Second, we leveraged the extensive efforts of the Human Behaviour Change Project ^17-19,43 to create a module of BCTs for each barrier. Historically, many tailored interventions have allowed interventionists to select BCTs they believe will target the patient’s barrier,^59,64,65 but this strategy means that even AYAs with the same barrier could receive different BCTs.⁶⁶ While interventionists can still use their clinical judgement to select which BCT(s) from the module to deliver, our approach ensures that each BCT delivered in the intervention is supported by theory and/or evidence. If this intervention is promising, this approach could be used in other tailored interventions to balance clinical judgement with theory and evidence.

In sum, results of this pilot trial will provide information about the feasibility, usability, and acceptability of a novel adherence-promotion intervention for AYAs with cancer. Pending promising findings, the next step will be to conduct a large-scale efficacy trial to evaluate the impact of the intervention on medication adherence in a more diverse sample of participants whose demographic and clinical characteristics represent the larger population of AYAs with cancer. Long-term, if this program of research is successful, it has the potential to expand our options for supporting adherence among AYAs with cancer.