Childhood cancer survivors are at risk for sexual dysfunction (SD) as a result of their cancer history and treatment [1-4]. Unfortunately, while SD affects between 20% and 50% of childhood cancer survivors, it is underrecognized and undertreated, leaving a gap in whole-person health care for these patients [1,2]. This unmet need exists despite national recommendations that sexual health, including sexual function, be addressed by oncology providers throughout the cancer care continuum [5-7]. Prior studies have identified numerous patient-level and provider-level barriers to SD communication, such as patient or provider discomfort and privacy concerns, which may be heightened in adolescent and young adult childhood cancer survivors [8-11]. Our prior research suggests that some of these barriers may be addressed through the use of standardized, patient-centered SD screening [9,11,12].

Unmet SD needs among adolescent and young adult childhood cancer survivors could be identified and addressed in pediatric oncology clinics, where patients are followed throughout and after treatment. Standardized, patient-centered SD screening could serve as the first step in ensuring appropriate sexual health care for this population. However, the routine use of patient-reported outcome measures is not common practice in pediatric oncology clinics, and barriers to implementation exist [13-15]. Additionally, SD in adolescent and young adult childhood cancer survivors can be a sensitive topic for both patients and clinicians unaccustomed to discussing this topic in oncology care [8,10,11]. These challenges are among several foreseeable barriers to the implementation of SD screening in this setting related to privacy, workflow, patient and provider buy-in, and more [9].

To develop an implementable patient-centered SD screening approach, we applied implementation science and user-centered design principles in an active, collaborative, and iterative approach to intervention design and implementation [16-18]. Through our prior and current work, this process leveraged considerations of context (guided by the CFIR [Consolidated Framework for Implementation Research] [19]) and engagement of oncology clinic partners to co-design an SD screening approach. We aimed to develop an approach that would fit within the contexts of pediatric oncology clinics, with future goals of achieving high Reach, Effectiveness, Adoption, Implementation, and Maintenance (RE-AIM) [20,21]. The RE-AIM and CFIR frameworks have been used successfully in many fields to systematically assess the robustness of interventions across settings and patient subgroups, with attention to the potential for scaling up to additional settings.

To support future dissemination and implementation of SD screening in pediatric oncology care, this study sought to develop a flexible intervention and implementation plan that can be tailored to a variety of contexts and practice settings. To do so, we adopted the implementation science–grounded concepts of “functions” and “forms” of an intervention: functions being the fundamental purpose or desired effect of an intervention or implementation strategy and forms being the intervention or strategy details, components, and activities selected to achieve specified functions. To achieve functions, forms may vary based on contextual factors influencing implementation [22,23].

Across this multiphase project, we followed the Wisconsin Interventions for Team Science Framework, beginning with a discover phase followed by an iterative design-build-test phase for the prototype SD screening approach [16]. In the discover phase, we first engaged patients and then providers through qualitative interviews, guided by the CFIR, to elicit perspectives on SD screening in the pediatric oncology setting, including potential barriers and facilitators that might impact implementation success as defined by RE-AIM outcomes [9,11,12]. These initial studies verified the need for a standardized SD screening intervention, identified the key functions of an SD screening approach, and highlighted potential forms for intervention and implementation design [9,11,12,24]. Both patient and provider participants found the National Institutes of Health Patient-Reported Outcomes Measurement Information System (PROMIS) Sexual Function and Satisfaction v2.0 Brief v2.0 (SexFS Brief) tool to be acceptable for this population and favored its use in facilitating SD conversations [9,11,12,25-27]. This tool is available in versions for male and female patients and is validated in adult cancer populations [25-27]. Subsequently, the design phase of the SD screening approach development process used collaborative, iterative co-design sessions with a group of pediatric oncology providers.

We frame this paper as a case study of the development of a consensus-driven intervention and implementation plan for an SD screening approach using the SexFS Brief in pediatric oncology clinics. The methods and results of this paper describe our process for engaging prospective users to determine the forms of the SD screening approach comprising the resulting prototype. More contentious or difficult decisions are highlighted, accompanied by descriptions of strategies for arriving at consensus in these scenarios. We also provide details about the specific design considerations and decisions for our SD screening approach, which can be used in future research investigating its effectiveness and potential for wide-scale dissemination and implementation.

The intervention and implementation development process used a series of co-design sessions [28-32] with a multidisciplinary group of pediatric cancer providers (physicians, advanced practice providers, and nurses) to develop an SD screening approach using the SexFS Brief in pediatric oncology clinics. Co-design session content was informed by the prior discover phase, in which providers and patients shared their preferences and recommendations for SD screening in qualitative interviews [11,12]. With the goal of developing a prototype SD screening approach through collaborative decision-making, as opposed to idea generation or brainstorming and reflecting, we engaged the same group of provider participants throughout the iterative design phase, following user-centered design principles [17]. The results of this process (ie, functions and forms of the SD screening approach as well as themes defining the decision points during the development process) will be used in the build and test phases, aiming to refine and pilot test the implementation and effectiveness of the SD screening approach.

All study procedures were reviewed and approved by the Colorado Multiple Institutional Review Board (20‐1679). Reporting follows the StaRI (Standards for Reporting Implementation Studies) guidelines, with the co-design process considered the implementation strategy [33]. All participants provided consent for their participation in this research at the beginning of the first co-design session. Upon completion of the 5 planned sessions, each participant received a US $100 gift card.

This study was conducted at a National Cancer Institute-designated cancer center within a quaternary academic children’s hospital in the western United States. The children’s hospital is a referral center with approximately 27 oncology physicians treating over 300 patients with cancer each year, drawing from an 8-state catchment region. The patient population is largely non-Hispanic White, with approximately 45% insured by Medicaid.

Participants for the design phase were purposively recruited from provider interviews conducted in the discover phase (N=25), exploring perspectives on the use of the SexFS Brief in pediatric oncology clinics [9]. Eligible individuals were the 25 discover phase pediatric oncology medical provider participants serving in any of the following roles: faculty physician, fellow physician, advanced practice provider, registered nurse serving as a nurse care coordinator, or medical assistant (MA). There were no exclusion criteria. Co-design sessions require substantial participant engagement and rely on thoughtful in-depth discussion; additionally, our design phase participants were expected to attend several sequential co-design sessions and ultimately arrive at consensus decisions. Therefore, consistent with other studies involving medical providers in the co-design of implementation strategies [32,34-36], we sought to engage a smaller total sample size of 6 to 8 highly engaged participants. Recruitment was completed on a rolling basis with targeted outreach to achieve variation by provider role and clinical team (ie, leukemia/lymphoma, central nervous system tumors, or solid tumor). In addition to the desired variation in clinical roles, the research team considered the depth of discover phase interview responses in determining the order in which to recruit eligible participants, aiming to balance variation in roles with demonstrated capacity to share perspectives. A total of 10 participants were approached, 9 expressed interest in participation, and 6 were selected for the final sample based on the alignment of schedules and ability to commit to all or nearly all co-design sessions. The final group of participants (n=6) who committed to attending planned co-design sessions included all provider roles except for MA and represented each clinical team.

Demographic and professional characteristics were collected from participating providers through a brief background questionnaire during the discover phase. Audio recordings and field notes for each session served as qualitative data for analysis.

This study used rapid qualitative analysis with a deductive-inductive approach, identifying key themes emerging from co-design sessions. The research team members who were engaged in the co-design sessions and qualitative analyses (JD and JR) practiced reflexivity throughout, including reflective discussions during post-session debriefs and collaborative analysis meetings. As biases of research team members were identified and noted, their potential effects on co-design facilitation and qualitative analyses were discussed with a team member (CS) who was not present at co-design sessions, was less immersed in data collection and analysis, and did not have clinical or research relationships with participants. Through this approach, alternative interpretations of the process and data through an implementation science lens were included to minimize the effects of bias on the research process. In addition, when the research team agreed internally about the potential feasibility of co-design decisions made by participants, strategies such as gathering and objectively presenting feasibility data were used to revisit decisions and ensure consideration from multiple perspectives (moderator strategies highlighted in the Results section).

After each co-design session, the co-moderator (JR) synthesized the data from field notes and audio recordings to develop a triangulated summary of the session. Summaries detailed the SD screening approach decision points discussed, including barriers, facilitators, potential solutions, variations in opinions expressed, and any consensus decisions that were made. The summary also included an assessment of the quality of the discussion (eg, participant insight, participation level, rapport, problems, line of questioning). After each session summary was developed, the research team (JD, JR, CRS) met to assess the accuracy and thoroughness of the session summary, referring to field notes and the audio recording as needed. This deductive-inductive approach allowed for organization by co-design discussion topics or decision points and the identification of expected themes based on the CFIR and on interviews conducted in the earlier discovery phase (eg, clinician discomfort discussing SD, relative priority of SD in the pediatric oncology setting) as well as emergent themes. The products of these analytic meetings were revised session summaries for each co-design session, organized by design decision. Session summaries were used to identify decision points characterized by little or no consensus among multidisciplinary design team members and those characterized by high levels of consensus. The results of each session’s rapid qualitative analysis were integrated into the information provided during the next co-design session as a form of member checking, in addition to any clarifications or additional information requested by providers (eg, information about electronic health record [EHR] capability, details about privacy or confidentiality regulations).

Following the completion of the co-design sessions, the research team completed a modified function and form matrix [24], mapping (1) the key themes of discover phase patient and provider interviews, (2) the core functions of the co-designed SD screening approach and strategies to support its implementation, and (3) the forms (options for intervention content and delivery to meet the defined core functions of the screening approach) recommended in the co-design process. Implementation strategies were also mapped to the Expert Recommendations for Implementing Change compilation of 73 established implementation strategies [22,38]. The product of this part of the design phase was the prototype SD screening approach.

The 4 co-design sessions and the process mapping session took place between December 2021 and February 2022. The 6 co-design provider participants are described in Table 2.

Four co-design sessions were attended by five participants, and one session was attended by four participants. The moderator (JD) met individually with absent participants following missed sessions to update them on the discussion and elicit their input. The content of each session is shown in Table 1.

The findings were organized into 2 major overarching categories consistent with the goals of the co-design process: intervention design and implementation strategy selection. Within intervention design, discussions focused on 7 decision components: target population for screening, introduction of the SD topic and the need for screening of patients, screening modality, workflow, timing and frequency, patient-specific or population-specific tailoring, and clinical management of SD screening results (Table 3).

Within the implementation strategy selection, discussions addressed ways to increase clinician acceptability, buy-in, and adoption of the SD screening approach, including reminders, provider education, resources, and ongoing involvement in design, implementation, and adaptation (Table 4).

With myriad potential decisions and multiple forms available for each function of the screening approach and implementation strategies, it was important to understand how decisions would impact each other. In several instances, a cascade effect in decision-making was evident in which one decision influenced and often limited options for future decisions. Throughout the iterative process, there was a need for the co-design session moderators to remind participants of the overarching considerations guiding intervention design (eg, patient privacy and comfort) and the implications of decisions made in earlier sessions as the process progressed. Below, the decision points and process for selecting intervention components and implementation strategies are described, concluding with identified strategies used by the moderator and co-moderator in facilitating discussion and consensus-building in co-design sessions.

In addition to discussing the specifics of SD screening delivery and clinical management, the co-design participants considered implementation strategies aimed at optimizing provider buy-in, reach, adoption, scalability, and sustainability. The moderator and co-moderator introduced these co-design session topics by integrating prior qualitative interview results with information about established implementation strategies [39].

The primary focus of this session was a process mapping exercise outlining consensus-driven decisions from the intervention start (patient education about SD and screening approach) to completion (management of identified SD needs). In the preparation for this session, the research team compiled all decisions made in prior sessions into a process diagram (Figure 1), used to guide the discussion of every aspect of the SD screening approach.

This process helped to identify additional issues, such as potential variability in patient education at the time of diagnosis and the importance of delaying the first screening so as not to screen patients on their first visit to the oncology clinic (Table 3). Additional potentially helpful implementation strategies were identified, such as the integration of a notification system, for the clinical team upon screening completion and the development of note templates for uniform documentation of sexual health team discussions.