Two institutions participated—University of California, San Francisco (UCSF), and Stanford University. From 2018 to 2019, SB—a UCSF oncologist—designed 12 scaffolds, using Google Slides covering the solid tumor chapters from the American Society of Clinical Oncology’s Self-Evaluation Program (ASCO-SEP) textbook [5]. Hematology, gynecologic oncology, and neuro-oncology were omitted for this pilot. Scaffolds included text and images synthesized from ASCO-SEP and National Comprehensive Cancer Center guidelines. For brevity, the slides instructed fellows to adhere to length limits when making edits.

We emailed scaffold links to all first- to third-year UCSF (n=21) and Stanford University (n=27) oncology fellows in July 2019 and July 2020. Use was optional, and fellows could access and update the scaffolds anonymously at any time. Updates were audited by SB.

In December 2021, to evaluate feasibility outcomes (fidelity: degree to which the innovation was implemented as intended; appropriateness: perceived fit of the innovation; self-efficacy: belief in the ability to execute the innovation’s goals) [6], we reviewed updates tracked in Google Slides and conducted 2 voluntary feedback focus groups (UCSF: facilitated by SB; Stanford University: facilitated by MS—a Stanford University oncology fellow) with 4 fellows each. Focus group size was determined by responses to recruitment emails. Consent and demographic information were obtained. Participants did not need to use the scaffolds, as we were also exploring barriers to use. Focus groups were recorded and professionally transcribed. SB and MS independently reviewed the transcripts and generated themes through iterative discussion [7].

The scaffolds were updated in 2023 by SB (available on Google Drive) [8].

UCSF and Stanford University institutional review boards granted exemption (#20-31645) and approval (#57766), respectively. Participants received an information sheet and verbally consented before each focus group. Transcripts omitted personal identifiers, and interviewers never revealed participant identities to the rest of the study team. Participants received a US $10 electronic gift card.

From July 2019 to December 2021, fellows made 60 updates (Table 1), ranging from new trials to changes in management; none were erroneous. SB made 9 edits for brevity.

Focus group participants (N=8) were women and included Asian (n=3, 37.5%), White (n=3, 37.5%), Black (n=1, 12.5%), mixed-race (n=2, 25%), first-year (n=5, 62.5%), second-year (n=2, 25%), and third-year (n=1, 12.5%) fellows. Most (n=7, 87.5%) used the scaffolds. Qualitative analysis (Table 2) revealed that fellows felt the scaffolds were accessible and succinct learning tools, addressed the dearth of similar resources, served as effective preparation materials for clinical work and examinations, provided structured information for rapid reviews, and made interactions with complex resources easier.

Qualitative analysis revealed barriers to updating the scaffolds—fellows’ lack of ownership over the scaffolds and low confidence regarding appropriate updates.

This pilot explored the feasibility of implementing constructivist scaffolds for oncology fellows. We found evidence of fidelity and appropriateness and delineated next steps to optimize self-efficacy. The scaffolds [8] can be downloaded and modified to avoid generating institution-specific scaffolds from scratch. To promote ownership and confidence, we recommend assigning fellows to update the scaffolds under faculty mentorship.

Despite demonstrating superior outcomes when compared to passive lectures, constructivist learning is rarely studied at the fellowship level [9-11]. We recommend evaluating constructivist learning modalities, such as scaffolds, in graduate medical education to enhance learning outcomes.

Though the focus groups suggested that multiple fellows used the scaffolds, Google Slides did not track how many fellows accessed or updated them. We did not incorporate multimedia components beyond images and tables (some needed to be removed before publication to respect copyright), nor did we include assessments in this pilot. We recommend that institutions consider incorporating multimedia content and assessments into the scaffolds. The number of focus group participants was small and not gender-diverse. Future studies should quantitatively evaluate usage patterns and user satisfaction to examine what factors drive utilization.

We piloted a novel constructivist approach to fellow learning and found evidence of feasibility. Oncology educators may use and modify the scaffolds [8] to jump-start constructivist education for fellows at their institutions. Educators in other fields may wish to apply this model to their specialties.