In this arm, participants received a behavioral intervention involving AI-supported education during a structured, classroom-based case-solving session. After informed consent and a pretest, students were divided into groups of six. Each group selected a representative with access to ChatGPT-4 Premium via researcher-provided credentials. These representatives interacted directly with the AI while others collaborated in real time to solve a pediatric surgical emergency case. The case included 10 structured questions and one open-ended item, based on the Bowtie model used in NCLEX. Each question was addressed in 5-minute intervals through team-based discussion. The intervention aimed to enhance clinical decision-making and case-specific knowledge. No drug, device, or clinical procedure was used; the AI-supported education was conducted entirely in an academic classroom setting using digital tools.

Participants in this group engaged in traditional, instructor-led case-solving sessions without access to AI tools. Students were divided into groups of six and analyzed the same pediatric surgical emergency case used in the intervention group. To ensure no use of AI platforms like ChatGPT, a classroom monitoring application was implemented. Instead, students used institutional academic databases and library resources. The activity's structure, including group size, timing, question sequence, and classroom setup, mirrored the AI-supported group's experience to ensure consistency. This arm served as the control condition to compare traditional education with AI-assisted learning in terms of clinical decision-making and knowledge development. No drug, device, or clinical procedure was used; this was a classroom-based educational activity only.

In the intervention group (C-CASE), after informed consent and pretest completion (including a sociodemographic form and CDMNS), the case scenario was introduced by the course instructor. Students were divided into small groups, and each group selected a representative who accessed ChatGPT-4.0 Premium via credentials provided by the research team. Using a collaborative problem-solving format, each group worked through a structured case scenario involving pediatric surgical emergencies. Questions were distributed sequentially, with 5-minute intervals allocated per question. Students used ChatGPT to support reasoning and clinical decision-making within their group. After each interval, responses were submitted, and the next question was handed out. Sessions were proctored by research assistants, and the full implementation, including discussion, lasted approximately two hours. The intervention aimed to foster decision-making, teamwork, and AI literacy in a clinical nursing education.

This outcome measures clinical decision-making using the Clinical Decision-Making in Nursing Scale (CDMNS), developed by Jenkins (1983) and validated in Turkish by Durmaz (2012). The 40-item scale is rated on a 5-point Likert scale ("always" to "never") and includes four subdimensions: Search for Alternatives or Options (SAO), Canvassing of Objectives and Values (COV), Evaluation and Re-evaluation of Consequences (ERC), and Search for Information and Unbiased Assimilation of New Information (SIUANI). Each subdimension includes 10 items. Total scores range from 40 to 200; higher scores reflect stronger decision-making. Of the 40 items, 22 are positively worded and 18 are negatively worded (reverse-scored). Minimum and maximum scores for subdimensions are not explicitly defined in the original scale; instead, changes in subdimension scores were analyzed based on increase or decrease. The scale was applied at pretest and posttest.

This measure evaluates nursing students' knowledge in pediatric surgical emergency management using a researcher-developed, scenario-based test aligned with the NCSBN "Bowtie" model. The test includes 10 structured items (one with two parts) and one open-ended question, totaling 11 questions. Items combine multiple-choice and short-answer formats, requiring students to prioritize, analyze, and justify decisions. Partial credit is awarded for correct responses and justifications; incorrect answers do not deduct points. Scores range from 0 to 100, with higher scores indicating greater case-specific knowledge. No cutoff was defined; mean scores were compared between groups at posttest. This is not a previously validated measurement tool (scale); however, content validation was conducted by five independent nurse educators with expertise in pediatric and surgical nursing. The test was developed by two PhD nurse educators and administered digitally via QR code.