Introduction

Despite substantial advances in oncology, up to 1 in 5 adolescents or young adults (AYA) will die from cancer [1]. AYAs with advanced cancer face multiple treatment decisions throughout their illness course; a study of hospitalized AYAs with cancer found 43% of patients received 4 or more treatment regimens prior to death [2]. These patients and their parents or trusted persons (PTPs) make treatment decisions often without identification of preferences and priorities for care. The advent of new therapies has further increased the number and type of options available [3]. Given the experimental nature of many of these new treatment options, the potential benefits and burdens may be unknown, increasing the uncertainty faced by patients and families. Each treatment has a unique risk and benefit profile, which can lead to the need to trade one desirable outcome for another. Adding to the complexity of this decision, stakeholders (AYAs, PTPs, and providers) likely characterize and value the risk/benefit ratios differently.

Two-thirds of AYAs desire an active role in decision-making, yet they report not feeling well informed about their options [4]. The evolving paradigm of oncology treatment and the multiple, complex medical decisions faced by patients therefore require the development and use of novel tools to amplify the AYA’s voice and role in decision-making. We propose that these tools must allow for explicit consideration of the various treatment factors, identification of the most important factors, support for patients as they uncover their own views about relative risk/benefit trade-offs and provision of tools to communicate these preferences with others. To our knowledge, there are no available clinical decision-making tools for use by AYAs with cancer and their families.

To address this need, we developed MyPref, an adaptive conjoint analysis (ACA)-based decision-making and communication tool. The conjoint analysis methodology assumes that each respondent chooses the option with the combination of attributes that gives him/her the highest utility. [5] ACA can be used to measure preferences in the face of multiple tradeoffs and provide a summary of preferences in real time [6]. Conjoint analysis-based interventions in adults have been shown to improve patients’ knowledge of the risks and benefits of different treatment options, leading to improved shared medical decision-making [7, 8]. However, this methodology has not previously been used in AYAs or in patients with advanced cancer.

We have previously described the iterative, step-wise process for developing MyPref [9]. The tool asks participants to consider a situation in which their current treatment is no longer working. By allowing AYAs to consider their preferences using a hypothetical situation at a time of clinical stability, we hypothesized that AYAs, even those with a deep sense of immortality, may be willing to explore their preferences for future treatments. By outlining factors involved in treatment decision-making and using ACA methodology to consider trade-offs involved in treatment options, we further hypothesized that AYAs would have improved insight into their preferences and be better able to communicate those preferences and make medical decisions. MyPref is novel in three ways: it is designed specifically for AYAs, it provides immediate information about patient preferences, and it can be used to aid in communication between patients and family members and to help with decision-making around further cancer treatment.

We conducted a single-arm, pilot study of MyPref in AYAs with advanced cancer, their PTPs, and oncology providers and assessed the utility of MyPref with a mixed methods design. A separate paper describes the preferred role in decision-making and preferences and priorities of AYAs and their PTPs identified using MyPref [10]. In this analysis, we sought to examine the clinical usefulness of MyPref, exploring if MyPref can help prepare AYAs to make medical decisions, if summary reports are understandable and useful in augmenting communication and decision-making, and the optimal timing of completion of MyPref during the treatment course.

Methods

Study recruitment and enrollment

Study methods have been described in full elsewhere [10]. Eligible participants included English-speaking patients aged 15–30 with a diagnosis of cancer that had progressed despite initial therapy. Patients were approached at a time of clinical stability, in which they had been on their current treatment regimen for at least 2 weeks without plans for routine radiographical evaluation in the upcoming 2 weeks. Eligible AYAs identified one PTP that may be helpful in decision-making, or they could participate alone. Informed consent was obtained from each participant. AYAs under 18 provided assent for enrollment and their legal guardians provided consent. The MyPref tool was initially piloted with 5 AYAs and their PTPs with resultant changes to the survey layout and design. Thereafter, we enrolled 10 additional AYAs and their associated PTPs.

Oncology providers were invited to participate if their patient participated in MyPref; additional oncologists were invited based on their clinical or research expertise. Eligible providers were sent an email inviting them to participate and verbal consent was obtained prior to the recorded interview.

Study procedures

Participants independently completed the MyPref on a tablet computer; an RA was present to answer questions. Sawtooth™ software (www.sawtoothsoftware.com) was used to administer the survey and results. Participants received gift cards for participation.

Participants completed demographic questions and the Control Preferences Scale (CPS) [11] prior to the ACA survey. To introduce MyPref, participants were asked to imagine a situation in which their cancer/child’s cancer is no longer responding to treatment and to consider the types of things that would be important. Participants were then provided a list and description of the nine treatment attributes identified during study development [12]. After initial consideration of attribute preferences, participants were presented with a series of 12 unique combinations of attributes and asked to identify their preferred treatment and strength of preference. Estimates of the respondent’s preferences were automatically updated after each question and used to generate a new combination of treatments to consider. Upon completion, participants received a printed version of their MyPref Summary Report, which included estimated calculated attribute importance, scaled from 0 to 100, with higher numbers indicating greater degree of preference. An example version of MyPref is available in Online Resource 1.

Approximately 1–2 weeks later, participants completed the Preparation for Decision-making (PrepDM) scale [13], the MyPref Experience Questionnaire (MPEQ), and participated in an audio-recorded semi structured interview. The interview guide further queried the domains assessed in the MPEQ and examined the participant’s experiences with MyPref, clarity of the MPEQ, and how and in what ways the tool may be used.

Oncologists that agreed to participate were sent an example version of the MyPref to review. They met separately with the study investigator (JS) and completed the provider version of the MPEQ and participated in a recorded semi-structured interview, with a focus on the clinical utility and applicability of MyPref. Providers of AYAs that completed MyPref were shown a copy of their patient’s MyPref Summary Report and asked to comment. Providers without paired AYA participants reviewed a generic Summary Report. This study was approved by the Dana-Farber Cancer Institute Institutional Review Board.

Study instruments/outcomes

MPEQ participant-specific versions were developed de novo to characterize the utility and effectiveness of MyPref in decision-making and communication based on the International Patient Decision Aid Standards (IPDAS) Collaboration [14]; additional questions were adapted from surveys used in the PediQUEST trial [15]. The MPEQs addressed ease of understanding, accuracy, usefulness of MyPref, communication of preferences between patient/parent and patient/parent/provider, and potential stress or benefit experienced from participation. Participants also completed the 10-item, validated PrepDM scale, a tool commonly used in decision aid trials, to assess the utility of MyPref in preparing the patient to communicate with their healthcare provider and to make a healthcare decision [13].

Analysis

MPEQ participant responses were measured on a 5-point Likert scale (1 = strongly disagree, 3 = neutral, 5 = strongly agree) and used as a continuous variable in analyses. PrepDM scores ranging from 1 to 5 (correlating with answers not at all, a little, somewhat, quite a bit, and a great deal) were summed with higher scores indicating higher perceived preparation for decision making. We used a weaving approach to combine quantitative results with qualitative findings [16], focused on key domains of interest, perceived utility, and benefit.

Thematic analysis of interview transcripts was conducted utilizing both inductive and deductive approaches [17, 18]. An interdisciplinary team (GH, JS, AR) collaboratively and iteratively developed a comprehensive codebook. Initial coding was inductive, starting with open-coding, combined with prefigured codes from the interview guide. Two unique codebooks were developed to analyze participant interviews, one for the AYAs/PTPs and a second for oncology providers. The AYA/PTP codebook had 8 main domains, and the provider codebook included the same domains and included perspectives on integration of MyPref into practice. Each transcript was independently coded by two team members (GH, JS); the coders engaged in reflective discussion to resolve differences and refine codes until consensus was reached. Thematic analysis was also a collaborative team-based process that identified key impressions, contexts, and patterns across participants in relation to MyPref. An audit trail including coding dictionary and codebook adjustments was maintained to allow for scrutiny of the analytic process. Ethnographic software (NVivo 12) was used for data management and to facilitate analysis.

Results

Participant demographics

Thirty-four participants were enrolled, 15 AYAs, 7 PTPs, and 12 providers; 31 participants completed all study requirements. Thirty potentially eligible AYAs were identified. Table 1 includes demographic data for enrolled participants by group. Thirteen of the 15 AYAs (87%) that completed MyPref completed follow-up surveys, 1 died and 1 was lost to follow-up. No patients experienced further disease progression between MyPref and completion of follow-up study items. The median AYA age was 20 years (range 15–28). Most AYAs were male (n = 12, 80%), with a solid tumor (n = 9, 60%). AYAs had been on treatment 17 months on average. Nine AYAs identified a PTP; two PTPs were not available to participate. Of the 7 enrolled PTPs most were mothers (86%), white (71%), with some higher education. All PTPs completed follow-up items. Of the 15 providers invited to participate, 12 (80%) enrolled and completed all study items. Oncology provider participants were primarily female (80%), under 50 years old, had been in practice for 10 years or less, and had variable areas of practice in oncology.

Table 1 Participant demographics for pilot study of MyPref
Full size table

Experience with MyPref and timing of use

All AYAs and PTPs completed MyPref in one sitting with minimal assistance, with an average time of 28 min. All participants reported that the MyPref calculated preferences were an accurate reflection of their beliefs. Figure 1 shows participants’ responses to select MPEQ questions with representative quotes from follow-up interviews. While almost all AYAs and PTPs agreed that the survey was easy to complete, a few stated that certain phrases were difficult to understand or that it was helpful to have a study team member present to answer questions. All but two participants stated that this tool would be helpful to repeat again later (2 strongly disagreed/disagreed, 5 neutral, 13 agreed/strongly agreed). However, the interviews revealed that participants differed in opinion on when and how often MyPref should be repeated. Most emphasized the importance of repeating MyPref if a new treatment decision needed to be made, while several others thought there was benefit to repeating regularly regardless of clinical changes.

Fig. 1
figure 1

Participant responses to select questions from the MyPref Experience Questionnaire with example quotations from follow-up interviews. Includes all pilot participants that completed follow-up items, N = 13 adolescents and young adults (AYAs), N = 7 parent or trusted person (PTP), N = 12 oncology providers

Full size image

Provider responses to select MPEQ questions are summarized in Fig. 2. Three providers reviewed their patient(s)’ MyPref summary reports during the interview; all agreed that the summaries matched their own impressions of their patient’s priorities, while also noting the summary reports provided new information based on their MPEQ responses. Overall, all participants (13 AYAs, 6 PTPs, 12 providers) rated MyPref as helpful, with 23 (74%) characterizing it as ‘quite a bit’ or ‘very helpful.’

Fig. 2
figure 2

Oncology providers’ responses (N = 12) to select items on the MyPref Experience Questionnaire

Full size image

Benefit, clinical utility, and integration of MyPref into practice

Almost all AYAs and PTPs reported experiencing benefit from study participation (3 neutral, 17 agreed/strongly agreed). Interviews indicated participants found benefit primarily through three mechanisms: being more aware of, learning new or meaningful information about, and starting conversations with others regarding their preferences. Only one participant experienced stress due to participation. All participants agreed that MyPref would be useful for other AYAs.

AYA and PTP MPEQ responses indicated that participants believed MyPref was useful in communicating with others around preferences for future treatments. Seventeen of the AYAs and PTPs had spoken with their parent/child about MyPref at some point between survey completion and follow-up; 13 (76%) agreed/strongly agreed that the information from MyPref was useful in this communication. Five AYAs had spoken with their medical team after completion of MyPref, 4 of these (80%) agreed/strongly agreed that information from the survey was useful in this conversation. Thirteen AYAs/PTPs (65%) agreed/strongly agreed that MyPref results could be useful if they needed to consider a new treatment in the future; 6 participants indicated a neutral response.

All 12 (100%) providers agreed that MyPref made them think more about their patient’s preferences, the information from MyPref could be useful if they needed to consider future treatment options with their patient, and that other patients may find benefit from MyPref (Fig. 2). Nine of the 12 providers (75%) noted they planned to change the way they approach discussions about preferences for future treatments because of MyPref (6 agreed, 3 strongly agreed).

AYA/PTP participant’s responses to PrepDM are summarized in Fig. 3. All 20 participants (13 AYAs and 7 PTPs) thought MyPref prepared them to make a better decision and helped them organize their thoughts, think about how involved they wanted to be in decision-making, identify questions to ask, and prepare for future conversations with their doctor. Most also thought MyPref helped them realize a decision needs to be made, think about pros/cons of each option and helped them know that the decision depends on what matters most to them.

Fig. 3
figure 3

Adolescent and young adult (N = 13) and parent or trusted person (N = 7) participant responses for the Preparation for Decision Making (PrepDM) scale

Full size image

The eight codes and example quotations from interviews describing how MyPref could be used clinically by AYAs and PTPs are presented in Table 2. Generally, participants believed MyPref facilitated consideration, organization, and visualization of preferences and improved understanding of both the factors involved and treatment options. Most AYAs and PTPs thought there was benefit in completing MyPref as a dyad, allowing for comparison of preferences. Finally, AYAs and PTPs noted that completion of MyPref may help prepare AYAs to follow-up with their providers and be more engaged.

Table 2 Codes and example quotations related to how MyPref could be used clinically by adolescents and young adults and their parents or other trusted persons
Full size table

For example, a PTP noted, “[laughs] [cries] – It opened up a conversation that needed to be opened. To not think about the what-ifs before they happen is not really a wise thing to do because then you’re just in such an emotional state that it’s hard to make a true informed decision because your emotions are taking over. So, having it down on paper and having a conversation with your family about it, you’ve already broached the subject and it’s going to be easier in the long run.”

All participant groups identified that MyPref may help providers to better understand or confirm previous knowledge of the patient’s preferences. Moreover, MyPref provided a way to initiate and structure conversations about priorities, engage AYAs in longitudinal discussions about preferences, and prepare patients for the possibility of further disease progression.

The codes describing how MyPref could be used clinically by oncology providers were grouped into three themes: mypref helped to “identify and understand preferences and priorities,” “augment communication with patients and families,” and “support decision-making.” The codes related to the first theme are presented in Table 2 given their overlap with concepts identified in AYA/PTP interviews. Table 3 includes the codes from the latter two themes. Providers saw MyPref as a tool to ascertain prognostic understanding and create time and space to reflect on the current treatment plan and if and how this aligns with identified preferences. Finally, providers noted that MyPref could be used to discuss postponing or halting further cancer-directed treatment for patients in which their preferences clearly did not align with available options.

Table 3 Themes, codes, and example quotations related to how MyPref could be used clinically by oncology providers
Full size table

Discussion

Participants found MyPref easy to complete and useful in helping them think about their preferences. Most AYA/PTPs believed MyPref could augment communication around preferences and that completion of MyPref by both individuals allowed for comparison of preferences. AYAs and PTPs also indicated that MyPref helped them prepare to make a decision and to follow-up with the doctor, even though they were not facing an actual decision. Notably, less than half of the AYAs identified a PTP to participate in MyPref. This may indicate that AYAs may be or feel more alone in their decisional approach than previously recognized.

Providers noted that MyPref may be useful in initiating early discussions about preferences with AYAs that can be readdressed over time and to broach the ‘what ifs’ to help prepare patients for the possibility of further disease progression [19]. By providing a time and space to contemplate, compare, and discuss their preferences when less emotionally stressed, AYAs may be better prepared to engage in decision-making later. MyPref was seen by providers as a way to structure conversations around treatment options, including more explicit and focused discussion around risks and benefits that matter most to the patient. Furthermore, AYAs, PTPs, and providers also noted MyPref could be used to assess the current treatment and discuss ways it does or does not align with identified preferences.

For patients whose preferences clearly do not align with available treatment options, MyPref may provide an opportunity to discuss discontinuing further cancer-directed treatment, perhaps earlier on than this would have otherwise been considered. Conversations linking MyPref identified preferences with larger goals for care may be useful in promoting advanced care planning in this population. Importantly, providers also indicated that MyPref may allow AYAs to identify their own preferences separate from the influence of their parents, other family members, and medical team and even empower them to decline treatment offered if the treatment does not align with their preferences.

Notably, participants identified several barriers to integration and use of MyPref in clinical settings. First, they did not agree on the ideal time to use MyPref. While many AYAs and PTPs indicated the best time would be around an actual decision, they also noted that timeframe is highly stressful. Some providers thought MyPref would be best completed at a time in which there is no treatment decision to allow patients and family members to think about preferences in a less emotional context. However, other providers were concerned about causing distress for patients and families by asking them to consider a time in which the cancer had progressed.

There may be differences in how and when to utilize MyPref based on the cancer type, given distinctions in treatment options and timelines. For hematologic malignancies, decisions around a new treatment are often made quickly after relapse. It may not therefore be feasible to use MyPref before making a treatment decision. Also, this study did not evaluate if preferences change over time (i.e., are preferences different in a hypothetical scenario compared with when facing a real treatment choice?) Based on clinical experience and the feedback from participants, we recommend that MyPref be introduced when there is not an upcoming decision, allowing consideration of their preferences at less stressful time. If the patient’s cancer progresses or relapses, AYAs/PTPs should be offered the chance to repeat MyPref to examine if preferences are stable over time; new summary reports can be used in discussion around treatment decisions. We are currently conducting feasibility study of this two-phased approach to MyPref in AYAs with advanced cancer.

One provider highlighted that translating the information from the summary report into a clinical decision may be difficult. For example, predicting the amount of time until cancer grows can be challenging as the efficacy of many treatments may be unknown. Additionally, AYAs and PTPs may express almost near equal preference for options that may be dichotomous (e.g., time until cancer grows and side effects).

Lastly, the findings of this study may be limited given the lack of racial and ethnic diversity in the participants as well as the single-institution pilot study design. Despite these limitations, we believe that MyPref provides and important insight into the preferences of AYAs and seems to be useful to patients, parents, and providers alike.

Conclusion

AYAs with advanced cancer require tools to help them engage and actively participate in their cancer care. The results from this pilot study indicate that MyPref may be useful in allowing AYAs time and space separate from their parents and medical team to contemplate future treatments and the relative importance of the factors under consideration. This tool not only allows AYAs and their PTPs to consider, organize, visualize and better understand their preferences but also may be helpful in initiating and conducting conversations with others around preferences and priorities. MyPref may be a useful way for providers to engage and empower AYAs during conversations over the disease course and help prepare them to make future medical decisions. Incorporating feedback from pilot participants, we began enrolling participants in a two-phased examination of MyPref including use around an actual treatment decision in July 2020.