Introduction

Design thinking (DT), also referred to as human-centered design, user-centered design, and design innovation, is a problem-solving framework suited to complex, ambiguous challenges. Grounded in creativity and cognition, DT emphasizes iterative cycles of inspiration, empathizing with stakeholders and defining problems; ideation, generating and refining ideas; and implementing, introducing and improving the derived solution [1]. As a structured yet flexible approach, DT supports collaboration, surfaces assumptions, and enables stakeholders to reframe entrenched problems [1]. Within health professions education, DT has been utilized to explore various curricular and organizational questions, such as redesigning clinical rotations, identifying barriers to rural experiential education, and exploring pathways to healthcare administration [2–4].

Student selection in health professions education represents one such complex institutional challenge. Assessing the suitability of candidates for our programs is one of the most important tasks undertaken by health professions schools. The issue of student selection – that is, predicting who is most likely to succeed in a curriculum and become a successful practitioner – serves as the foundation of our healthcare system. As such, schools must develop robust selection processes that comply with professional standards, align with institutional values, and adapt to the ongoing changes facing healthcare and higher education [5].

Over the past decade, selecting students for health professions programs has become increasingly complex. Student selection has been influenced by various social, legal, economic, and technological changes, including the US Supreme Court ruling on affirmative action, shifting enrollments and projected workforce shortages, procedural adjustments in response to the COVID-19 pandemic, controversy surrounding the use of standardized admissions tests, and the advent of generative artificial intelligence (GenAI) [6–9]. In response to these challenges, scholars have advocated for expanding, updating, and enhancing our selection strategies [6,7]. In this context, admissions processes must balance institutional values, professional standards, and uncertainty, often with limited guidance on how to adapt [5,10]. As noted by Conrad and colleagues, “Selection — who gains access to a medical education and to a career as a physician, researcher, and/or faculty member — is as much art as science.” (11 p. 1472) Innovating our admissions processes to address these contemporary challenges will require creative thinking rooted in localized context [10,11].

This paper presents an exploratory, practice-based case study illustrating how DT can be applied to student selection in a health professions program. Rather than evaluating the effectiveness of an intervention, the aim is to document DT as a structured, participatory approach to stakeholder engagement, problem reframing, and idea generation within a localized admissions context. Specifically, we describe the design and implementation of a DT workshop at the University of North Carolina at Chapel Hill (UNC) Eshelman School of Pharmacy and examine the types of ideas and perspectives that emerged through this process. Framed as generative, workshop-based inquiry, this study draws on artifacts of co-creation (e.g., brainstorming outputs, notes, and reflections) to provide a process-oriented insight into how DT can be used to explore complex institutional challenges and inform future phases of prototyping, testing, and evaluation.

Methods

Results

Given the exploratory aim, results are presented as illustrative outputs of the DT process rather than evidence of effectiveness. The findings reflect how participants engaged with the workshop to generate, refine, and prioritize ideas within a specific institutional context (e.g., admissions). Of note, the ideated solutions presented in the results are not meant to be an exhaustive list for others to utilize; rather, they demonstrate specific ideas contextualized to the unique experiences, processes, priorities, and resources at our School.

During the ideation phase, participants generated approximately 76 ideas of which 43 were selected across three focal areas for further consideration. Ideas varied widely in scope and specificity. Voting was used as a pragmatic prioritization tool within groups, with selected ideas serving as starting points for further development rather than validated solutions. Examples of the ideas generated included: “have students evaluate an AI response to a scenario” (n = 7 votes), “technology that blocks the internet during interviews” (n = 10 votes) “add writing activity to the end of candidates day” (n = 13 votes), “new interview and application questions” (n = 1 vote), and “use GenAI to develop objective rubrics” (n = 1 vote). Vote counts per ideas ranged from 0–13. Groups prioritized prototypes for further development based on higher vote totals. At the end of the workshop, each of the three groups presented 1–2 prototypes for consideration (4 total prototypes were presented due to limited time), incorporating their suggestions through the perspectives of cost, equity, and geographical factors. As an example, in response to the launching point of reliable comparison across candidates, one group recommended incorporating a task-based interview station, such as real-time written essays or math problems that would preclude the use of GenAI. As it related to perspectives of cost, equity and geographical factors, the group opined that the real-time component might allow for fair comparisons across candidates on constructs of interest (e.g., in absence of standardized tests), ensure independent work, and help identify candidates who may warrant additional academic support upon enrollment. However, some candidates may feel intimidated or struggle with on-the-spot tasks, which could overshadow their other qualifications. It could also add time, cost, and complexity related to the design and evaluation of the task.

Data analysis followed reflexive thematic analysis approach [9,11] of individual brainstorming outputs, group-generated themes, creative reflection worksheets, and facilitator note identified three overarching areas of focus: interview format and delivery; AI integration and compliance; and admission criteria and evaluation. Within these, subthemes captured a range of perspectives including approaches to in-person and online interviews, supervised assessment, guidelines and disclosures, screening and awareness, detection and monitoring tools, and holistic evaluation strategies. In some cases, participants offered suggestions for strategies already used in the admissions process (noted by italics in Table 1); however, these suggestions often offered new perspectives or nuanced ideas for consideration.

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Key themes, sub-themes, and example brainstormed strategies identified through an admissions DT workshop.

https://doi.org/10.1371/journal.pone.0352192.t001

Post-workshop survey data (n = 15, 94% response rate) are presented as descriptive indicators of participant experience. Participants generally reported positive perceptions of the workshop. They strongly agreed that the workshop encouraged creative problem solving (M = 4.7, SD = 0.5), helped them to identify potential strategies (4.8 ± 0.4), and enhanced the quality of ideas generated (M = 4.9, SD = 0.4). Most strongly agreed that the workshop was a good use of their time (Mean = 4.4, SD = 0.5) and that the results will be useful to the School (M = 4.8, SD = 0.4). Open-text responses emphasized the value of engaging diverse perspectives and highlighted the workshop as a space for structured dialogue. One participant described the process as engaging and effective for generating new ideas and meaningful discussion. Another participant suggested expanding the use of DT for future initiatives and across other committees or working groups. Suggestions for improvement included allowing more time for idea development and in one case, rotating group membership to further broaden perspectives and increase opportunities for cross-group interaction. These findings should be interpreted cautiously given the small sample and the potential for response and social desirability bias.

Additional reflections emerged during member-checking discussions with admission committee members, who offered additional strategies and raised several critical points for consideration. For example, one committee member suggested generating an MMI response from GenAI and presenting it to a candidate for critique. In addition, committee members challenged the utility and value of changing the admissions process (e.g., Would we do anything different in our decision-making process?; How might we leverage admissions data in the curriculum once candidates enroll?; Can we obtain feedback from candidates?)

Overall, these results demonstrate the range and diversity of ideas and perspectives elicited through a DT approach, highlighting its potential utility as a structured method for collaborative exploration and problem reframing in complex institutional contexts. However, no strategies were implemented or evaluated at the time of this writing, and the findings do not address feasibility, effectiveness, or impact.

Discussion

This study illustrates how a DT approach can be used to explore complex and evolving challenges in student selection, including the advent of GenAI and sunsetting of standardized admissions tests. Specifically, we used a 3-phase DT approach to engage students, recruiters, faculty, and administrators in a process aimed at elucidating challenges and identifying potential solutions. The findings show how this process supported the articulation of key issues in contemporary admissions practices and enabled the development of a range of context-specific ideas, such as the use of GenAI, interview format (e.g., in-person/remote), and design of MMI stations.

Given the exploratory scope of this work, the findings do not demonstrate effectiveness or implementation. Instead, they provide insight into how stakeholders engaged with emerging issues. For example, the workshop highlighted differing perspectives on the role of GenAI in application materials and candidate evaluation. This reflects a broader shift in the literature, which points to increasing use of GenAI in admissions-related tasks. While the long-term impact of GenAI on student selection remains to be seen, scholars have quickly tested and critiqued its capabilities – with some demonstrating its ability to efficiently generate plausible college admissions essays [1,10]. Research suggests that applicants are increasingly likely to utilize GenAI, with one study citing a significant increase in its estimated use on medical school essays from 2021 to 2024 [2]. Further, Smith and colleagues reported that most medical residency applicants either used or anticipated using GenAI for assistance in developing their personal statement [12]. Through the DT workshop, our participants wrestled with the potential role of GenAI for applications and its implications for how we might evaluate candidates. We ideated potential strategies, such as removing pre-written components of the application and adding real-time writing assessments during the interview process, which may or may not be feasible for others, and we encourage contextualized exploration of localized challenges for further exploration of these issues.

Real-time interviews can provide critical insight into a candidate’s suitability for a health professions program; however, the structure of those interviews can vary widely. In-person interviews and remote interviews both provide benefits and drawbacks, requiring schools to determine an approach best suited to their needs. Remote interviews remove several barriers for candidates, including the cost of travel, reduction of time, and more flexibility [7]. This can increase the number of candidates who are able to interview at a program. Based on our experiences, many interviewers also prefer remote interviews because of the flexibility in location and reduced commitment time of participating in remote interviews versus in person. However, in person interviews remove the concern of candidates using GenAI during the interview, having others present during the interview process, or recording the interview questions, overall assuring the integrity of the process. To ensure candidates were aware that the use of GenAI was not permitted during the interview process, we developed usage guidelines to share with the candidates. We also reminded them of this prior to their interview in a group setting and had a short reminder added to all the MMI scenarios posted for candidates to consider.

Similarly, the design of individual MMI stations can vary significantly, affording schools the flexibility to focus on constructs best aligned with their values and priorities [3]. For example, an MMI station may be situational, asking candidates to consider and respond to a hypothetical scenario, or behavioral, asking candidates to interact with an actor who plays a specific role [4]. Stations may also be individual, involving a one-on-one interaction with the interviewer, or collaborative, involving work with other candidates or actors to complete a task or solve a problem [12]. Further, numerous constructs can be assessed in an MMI (e.g., empathy, ethical reasoning, critical thinking, self-awareness). This flexibility offers space for schools to think creatively about how to best design and leverage the MMI for decision-making in admissions. Our workshop resulted in suggestions for rethinking our MMI structure (e.g., adding a group task, making a station task-focused) and, as reflected in our member-checking process, schools should align the design and focus of their MMI with curricular priorities [4]. Further, all changes should be evaluated in the context of local policies and laws, such as the American Disabilities Act, with university resources consulted as needed (e.g., ADA university officers).

Beyond the specific ideas generated by the workshop, this study reflects the utility of DT for tackling complex challenges in a structured and efficient manner. This work contributes to a small, yet growing, body of literature in health professions education that leverages DT for problem solving, and the first known study to do so for student selection [13]. Our workshop was user-centered, focused on the needs and experiences of faculty, students and staff dedicated to the admissions process. It encouraged a deeper understanding of the problems faced by users and promoted brainstorming to generate a wide range of ideas. Our participants found the process useful and meaningful, suggesting that this approach may have utility for other types of educational challenges. By applying DT to the admissions process, we believe schools can create a user-centered, efficient, and effective system that meets the needs of prospective students and institutional constituents.

These examples illustrate how institutions might begin to examine similar challenges within their own contexts. In addition to the ideas generated, this study provides a practice-based example of DT as a structured approach to collaborative problem solving in health professions education. The workshop facilitated dialogue across stakeholder groups, brought forward diverse perspectives, and supported the generation of multiple potential strategies. Participants’ feedback indicated the process was engaging and useful for idea generation, however these perceptions should be interpreted with caution given the small sample size and reliance on self-reported data.

No interventions were implemented or evaluated as part of this study, and conclusions about feasibility, effectiveness, or scalability cannot be drawn. Instead, this work represents an early-stage exploration that may inform future cycles of prototyping and evaluation. Further research is needed to examine how ideas generated using DT processes can be translated into practice and assessed for their impact on admissions outcomes.

Conclusion

DT offers a systematic, iterative methodology for exploring the shifting complexities of admissions. By emphasizing collaboration and creative problem-solving, it enables teams to generate and refine innovative ideas through structured ideation and prototyping. This study illustrates how such an approach can be used to explore challenges, surface diverse perspectives, and generate context-specific strategies. While effectiveness was not assessed, this work highlights the potential of DT as a foundation for future development and evaluation of admissions practices.