The event series applied transdisciplinary pedagogy by intentionally convening experts and learners from diverse fields, such as public health, health care, environmental science, sociology, policy, and others, to co-create knowledge. This approach aligns with pedagogical models that promote integration across disciplinary boundaries to foster innovative, systems-level solutions to complex issues like climate change and human and environmental health (9).

Drawing on the Interprofessional Education Collaborative (IPEC) Core Competencies (10), of Values and Ethics, Roles and Responsibilities, Communication, and Teams and Teamwork, the events provided experiential learning opportunities for health professions students and faculty to engage with peers in health and non-health disciplines. The collaboration with CAIEP ensured alignment with national IPE standards.

The structure and content of the series reflected key competencies identified by the Planetary Health Alliance and the Planetary Health Competency Framework: interconnection within Nature, the Anthropocene and health, equity and social justice, movement building and systems change, and systems thinking and complexity (11, 12). These competencies informed the thematic design of each event.

The educational design of these events also aligns with the Core Competencies for Public Health Professionals developed by the Council on Linkages Between Academia and Public Health Practice (13). These competencies emphasize interdisciplinary collaboration, systems thinking, and the integration of evidence-based decision-making into public health practice. The UNE Planetary Health series reflected several domains of this framework by creating opportunities for participants to engage with complex, real-world health and environmental challenges through transdisciplinary dialogue.

Participant feedback was collected through post-event surveys developed and administered by CAIEP. At the conclusion of each event, participants were invited to complete the survey via a QR code displayed on the final slide and through a survey link shared in the Zoom chat; these brought participants to the survey in Google Forms, where the surveys were completed voluntarily.

The survey included both quantitative and qualitative items. The quantitative questions asked respondents to identify their primary academic major or professional discipline. Demographic data was not collected about the participants. The qualitative component consisted of two open-ended prompts: “Please let us know what was successful about today’s event” and “Please let us know what you would change about today’s event.”

Survey data were provided by CAIEP for analysis in an Excel worksheet with timestamps of each survey submission. The data was de-identified, protecting confidentiality. All data is preserved in a UNE Google Drive with access restricted to key CAIEP staff only. UNE’s Office of Research Integrity determined that this project did not require Institutional Review Board review and approval because it does not involve research with human subjects as defined by 45 CFR 46.102.

De-identified, open-ended responses from post-event surveys were analyzed using an inductive thematic analysis approach guided by established qualitative research methodologies (14–16). The primary author (AC) first reviewed all responses to the prompts across all events to gain familiarity with the data and identify recurring patterns and concepts. This process resulted in a preliminary set of thematic categories reflecting common dimensions of participant feedback.

To refine and validate the preliminary coding framework, the primary author conducted a pilot coding exercise using responses from a single event (PFAS, Part 2). These responses were manually coded relative to the initial theme categories, allowing for clarification of theme boundaries and ensuring that the categories were simultaneously distinct and inclusive.

After this manual process, the same dataset was analyzed using a large language model (LLM), ChatGPT-5, to independently identify emergent themes. The themes generated by the model were compared with the primary author’s preliminary categories and were found to be conceptually consistent, supporting the clarity and coherence of the initial thematic framework.

Following this validation step, open-ended responses from each event were entered sequentially into the LLM, and the model was prompted to identify the themes represented within each event’s response set. The primary author then compared the themes generated across events and determined that the same core categories were consistently present in all datasets. Based on this cross-event comparison, final theme definitions were refined and confirmed by the primary author.

Each individual response was then coded manually by the primary author using the finalized theme categories. Responses were color-coded by event and by theme to support visual inspection. Responses were not constrained to a single category: short or focused comments (e.g., “Powerpoints were smooth”) were assigned one theme, when appropriate, while more complex responses (e.g., “Variety of speakers, timely, like the open discussion opportunity”) were coded across multiple themes when they reflected several distinct ideas (e.g., comments referencing speaker diversity, topic relevance, and event format). All responses were categorized within the established thematic framework. No comments were excluded due to lack of thematic fit; rather, variability in responses was captured through single- or multi-theme coding.

Following manual coding, the primary author conducted a descriptive tally of the frequency with which each theme appeared within and across event datasets. These counts reflected the number of responses referencing a given theme, recognizing that individual responses could contribute to multiple thematic categories. Frequency summaries were used to support comparative interpretation across events and to contextualize patterns in participant feedback. All frequency determinations were derived from the primary author’s manual coding decisions and analytic judgment.

This analytic process aligns with widely accepted guidelines for thematic analysis. Consistent with Braun and Clarke’s six-phase approach (14), the primary author engaged in repeated familiarization with the data, generated initial codes through iterative review and pilot coding, developed and reviewed themes across event datasets with AI-assisted pattern checking, defined and refined final thematic categories through cross-event comparison, and produced the final analytic narrative through manual coding and synthesis of representative responses.

Artificial intelligence was used as a supportive analytic tool to assist with pattern recognition and theme confirmation. The LLM was employed after initial theme development to corroborate emergent categories, assess consistency of themes across event datasets, and enhance analytic efficiency, given the volume of responses. All thematic decisions, final category definitions, and coding assignments were determined by the primary author through critical review.

This approach is consistent with emerging literature that supports the responsible use of generative AI in qualitative research when applied transparently and under human oversight (17, 18). Prior studies have demonstrated that AI-assisted thematic analysis can enhance rigor and scalability while maintaining analytic credibility when researchers retain control over interpretation, reflexivity, and final decision-making (19, 20). In this study, AI functioned as an analytic aid rather than an interpretive authority, ensuring that thematic conclusions remained grounded in researcher judgment and contextual understanding.

Across the five PHC events, a total of 767 participants engaged meaningfully in active formats, including Zoom and in-person attendance. Learners included primarily undergraduate and graduate students from the health professions, environmental and social sciences, as well as faculty, professional staff, and occasional community participants. Collectively, 502 post-event surveys were gathered (a 65.4% response rate), representing participants from over 23 distinct academic programs and professional disciplines across UNE. One survey from the first event, four surveys from the second event, and one survey from the fourth event were excluded from the dataset because their timestamps indicated completion in the days and months following the live session, revealing they were submitted by passive viewers, rather than by active participants. Active participation, as well as passive viewing, program/discipline representation, and number of surveys collected are summarized in Table 3. Attendance varied by delivery format, with the highest participation observed in fully virtual events, and the lowest in the sole in-person-only session.

Of the 502 post-event surveys collected across the five PHC events, 312 (62.2% of survey respondents) participants provided open-ended responses to the prompt “Please let us know what was successful about today’s event,” and 235 (46.8% of survey respondents) participants responded to “Please let us know what you would change about today’s event.” These qualitative data formed the basis of the thematic investigation.

Across the five PHC events, survey completion patterns evolved over time. Increased response rates in later events may reflect participants’ growing familiarity with the event format and organizers’ more explicit emphasis on the importance of feedback for strengthening future programming. These improvements suggest that consistent messaging and integration of feedback culture into educational events can meaningfully enhance participation and the quality of evaluative data.

Attendance and participation patterns also offer insights for optimizing future delivery formats. While several respondents from virtual sessions expressed desire for in-person gatherings to encourage interpersonal connections, the attendance data suggest that virtual and hybrid formats were significantly more accessible and inclusive. The sole in-person-only event drew the smallest audience and generated limited survey data, offering insufficient evidence to support a transition to face-to-face programming alone. These findings highlight an important balance between accessibility and experiential connection, suggesting that hybrid approaches, which allow for both physical presence and remote participation, may best support broad engagement in planetary health education.

Interpretation of these findings should be considered in light of several important limitations. The voluntary nature of participation in the PHC event series introduces potential participation bias. Individuals who elect to attend may represent a group already interested in planetary health, interdisciplinary learning, or climate-related issues. A further source of participation bias stems from institutional requirements and elective incentives within the medical program. Beginning in 2022, first-year medical students were required to attend at least three CAIEP Knowledge Exchange events per academic year, and second-year students were required to attend at least two per academic year. In addition, many medical students voluntarily pursue the Interprofessional Honors Distinction, a nationally accredited digital credential, which requires attendance at four CAIEP events. These expectations likely contribute to higher participation rates among medical students relative to other disciplines and may partially explain the strong representation of this group in the survey data. While this engagement aligns with curricular goals, it also introduces an attendance pattern influenced by program requirements rather than solely by learner interest.

The descriptive nature of the study design further limits interpretation. Without demographic data, it is impossible to assess the representativeness of the event participants or survey respondents, leading to uncertain generalizability. Survey responses may also reflect response bias, in which participants with particularly favorable feedback (or conversely, particularly critical feedback) are more likely to respond. Additionally, feedback was not collected from passive viewers of the event recordings, resulting in a data gap from this segment of the audience.

Beyond these limitations of data collection, limitations related to analytical rigor warrant consideration. While the primary author reviewed, refined, and consolidated the resultant themes, the reliance on generative AI as an analytic support tool, as well as the lack of a second reviewing author, may limit the rigor of the inductive thematic analysis (21). Despite this, emerging evidence supports the cautious and transparent use of generative AI to assist with qualitative pattern recognition when paired with vital human oversight (22). Therefore, the findings should be interpreted as exploratory and descriptive, rather than definitive.

Despite the limitations, these findings offer meaningful insights into the design and delivery of planetary health education in higher education settings. The PHC-CAIEP partnership model demonstrates a scalable framework for convening diverse disciplinary perspectives around complex health-environment challenges in ways that are accessible, engaging, and responsive to learner feedback. The consistent appreciation for interdisciplinary dialogue, coupled with strong participation in virtual formats, suggests that similar models could be adapted across universities seeking to embed planetary health content into co-curricular settings.

Future research should build on this work by incorporating more systematic outcomes measures, including longitudinal assessment of learning, content application, and systems-thinking decision-making, with the addition of demographic data to assess representativeness and differential impacts across learner groups. Mixed-methods designs, multiple coders, and intentional alignment with a single guiding framework from the start may further strengthen analytical rigor. Together, these next steps can advance planetary health education from exploratory programs to evidence-based educational practice.