A multicenter cross-sectional study.

Between June and September 2024, a multistage stratified cluster sampling method was used to recruit nursing interns from universities in Hubei Province. During the first stage, Wuhan, Yichang, and Jingzhou were selected as sample regions based on their high rankings in gross domestic product and population density rankings. During the second stage, between two and four higher education institutions were selected per city according to the population size of their nursing programs. During the third stage, cluster sampling was conducted by selecting internship classes from institutions that met the inclusion and exclusion criteria. Ultimately, nine institutions (five undergraduate and four vocational colleges) were included. The inclusion criteria were (1) internship duration of 3 months or less, (2) full-time associate degree or undergraduate nursing interns, and (3) voluntary signing of informed consent. Exclusion criteria were (1) absence from duty for more than 1 month due to illness or personal leave, and (2) a non-continuous internship at the same hospital.

The sample size was estimated using Kendall’s formula, which recommends that the sample size be 10–15 times the number of independent variables (23). With 13 independent variables in this study, and considering a 20% attrition rate, the calculated sample size ranged between 156 and 234. In addition, studies indicate that LPA typically requires a sample size exceeding 500 to ensure robust model estimation (24). Ultimately, 1,308 participants were included in this study.

The survey instruments included a sociodemographic characteristics questionnaire, the Feedback-Seeking Behavior Scale (FSBS), and the Career Adapt-Abilities Scale (CAAS).

Data were collected online via the Wenjuanxing platform.¹ Upon obtaining authorization from internship coordinators at the participating universities, survey links were forwarded via WeChat to class advisors, who distributed them directly to the class groups of eligible early-stage nursing interns. To encourage participation and minimize non-response bias, a small incentive (handmade soap) was provided to each participant upon completion. The study purpose was outlined on the survey homepage, where electronic informed consent was obtained. To ensure data integrity, the system restricted submissions to one per IP address, and all items were mandatory. Based on a pilot survey of 30 interns, a benchmark for focused completion time (120–261 s) was established, and attention check items (i.e., trap questions) were embedded in the main questionnaire. Following data collection, two researchers conducted a rigorous manual review. Responses were excluded if they failed attention checks, fell outside the focused time range (<120 s or >261 s), or exhibited obvious response patterns, thereby ensuring the quality of the final dataset. Strict confidentiality was maintained throughout the process to protect participant privacy.

The questionnaire data were entered and analyzed after being verified by two individuals. Mplus 8.3 was used to conduct LPA on the early-stage nursing interns’ FSB, with the optimal number of profiles selected based on the following criteria: (1) the Akaike information criterion (AIC), Bayesian information criterion (BIC), and sample-size-adjusted BIC (aBIC), with lower values indicating a better model fit; (2) entropy, where values closer to 1 denote higher classification accuracy (32); and (3) the Lo–Mendell–Rubin (LMR) test and bootstrap likelihood ratio test (BLRT), where p < 0.05 indicates that the K-class model is superior to the K − 1-class model. These metrics guided model selection; however, the interpretability of each profile was also considered (33). Statistical analyses were performed using SPSS version 23.0. Categorical data were expressed as frequencies and percentages (%), and continuous data as means ± standard deviations (x ± s). Group comparisons were performed using t-tests and analysis of variance. Multivariate logistic regression was used to assess factors influencing different profiles, with a significance level of 0.05.

This research was conducted in accordance with the Declaration of Helsinki and was approved by the Medical Ethics Committee of Tianyou Hospital affiliated with Wuhan University of Science and Technology (Approval No.: LL2025-04-01-01). All participants were voluntarily recruited and provided electronic informed consent before their participation. The survey was administered anonymously, with all data kept confidential, and participants retained the right to withdraw from the study at any time.

A common method bias test was performed using Harman’s single-factor test. The results showed that seven factors had eigenvalues greater than 1, and the first factor accounted for 29.057% of the variance, which is below the 40% threshold. Therefore, this study does not suffer from serious common method bias (Table 1).

A total of 1,370 questionnaires were collected, of which 1,308 were valid, with an effective recovery rate of 95.47%. The sample comprised 155 male (11.9%) and 1,153 female (88.1%). The demographic details are presented in Table 2.

The mean score of the FSBS was 5.06 ± 1.08, and the specific scores are presented in Table 3.

This study used the FSB dimension scores as manifest indicators to fit models with one to five latent profiles. The fit indices for these models are presented in Table 4. First, as the number of profiles increased, the AIC, BIC, and aBIC values decreased progressively; the three-profile model represented a turning point at which the downward trend of each value slowed. Second, the LMR tests for Models 4 and 5 were not statistically significant (p > 0.05). Third, the three-profile model had an entropy exceeding 0.80, and both the LMRT and BLRT were statistically significant (p < 0.05). In addition, the average probability for each profile exceeded 90%, indicating that the classification results of Model 3 were highly reliable. These model comparisons are detailed in Table 5. Collectively, the results indicated that Model 3 was the optimal fit.

Based on the FSBS scores provided by the nursing interns during the initial stage of their internship, the three latent profiles were identified. A visualization of these profiles, with the x-axis representing the dimensions and the y-axis representing the average scores, is shown in Figure 1. Each latent profile was named based on its scores across various dimensions and its external characteristics. Regarding the C1 category, the scores in all dimensions were low; therefore, it was named the “Low Feedback-Seeking Group,” comprising 273 (20.87%) participants. Regarding the C2 category, the scores in all dimensions were at a medium level; therefore, this was designated as the “Medium Feedback-Seeking Group,” comprising 501 (38.3%) participants. The C3 category was marked by high scores across all dimensions; therefore, it was named the “High Feedback-Seeking Group,” comprising 534 (40.83%) participants.

The univariate analysis revealed statistically significant differences among the FSB categories of early-stage nursing interns concerning education level, student leadership experience, internship hospital type, employment intention, and career adaptability scale scores (p < 0.05; Table 2). Following this, variables in the univariate analysis with statistical significance were incorporated into a multivariate logistic regression model, with the low and high feedback-seeking groups as reference groups. The results indicated that education level, student leadership experience, internship hospital type, and career adaptability were predictors of FSBs among early-stage nursing interns (p < 0.05).

The results revealed that early-stage nursing interns who had experience as student leaders (OR = 1.852, CI = [1.324, 2.590]), were from tertiary Grade A hospitals (OR = 2.546, CI = [1.392, 4.655]), were from tertiary Grade B hospitals (OR = 3.37, CI = [1.698, 6.690]), exhibited career curiosity (OR = 2.621, CI = [1.771, 3.878]), and exhibited career confidence (OR = 1.496, CI = [1.045, 2.141]) were more likely to be in the high feedback-seeking group than the low group. Additionally, those with a junior college degree (OR = 0.577, CI = [0.403, 0.825]) had a higher probability of being in the low feedback-seeking group. Further details are presented in Table 6.

The mean FSBS score among early-stage nursing interns was 5.06 ± 1.08, which was lower than that reported by Zhang et al. (9), thus indicating a moderately low level. This difference may be attributed to the characteristics of their career development stage. Compared with new nurses who possess some clinical experience (10), early-stage interns, transitioning from “knowledge recipients” to “practice explorers” lack sufficient professional knowledge and struggle to make accurate judgments or respond effectively to complex clinical situations (6). Consequently, they exhibit conservative feedback-seeking tendencies in clinical practice. This emphasizes the need for nursing managers to establish systematic, individualized feedback mechanisms, with senior nurses providing accurate feedback and fostering effective learning environments and positive ward atmospheres to enhance interns’ willingness to actively seek feedback.

Moreover, this study categorized early-stage nursing interns, into three groups based on their FSB: low (20.87%), moderate (38.3%), and high (40.83%). The “Low Feedback-Seeking Group” (20.87%) was characterized by lower educational levels and career adaptability. This group is likely accustomed to passive, teacher-centered learning (34). They often lack self-improvement awareness and undervalue feedback for professional growth. Due to insufficient adaptability, they are sensitive to “transition shock” and tend to use defensive avoidance strategies. Wan et al. (35) found through weekly diaries that consistent feedback-seeking significantly boosts work engagement and performance. For this group, experienced mentors should conduct weekly one-on-one dialogues. Guiding clinical reflection and feedback journaling can help enhance their intrinsic motivation.

The “Moderate Feedback-Seeking Group” (38.3%) was characterized by bachelor’s degrees, tertiary hospital internships, and high scores in career control and curiosity. These interns received systematic theoretical education (36) and possess a solid professional foundation. Research by Sümen and Adibelli (37) showed that inquiry-based learning, such as leading case discussions, effectively satisfies curiosity and improves clinical decision-making. Therefore, we recommend granting moderate autonomy during rotations. For example, interns could lead brief case reports during ward rounds. Such tasks expose knowledge gaps, transforming latent curiosity into concrete feedback-seeking. This helps them shift from passive acceptance to active problem discovery.

The “High Feedback-Seeking Group” (40.83%) scored high across most dimensions, excluding “monitoring of instructor feedback.” Key traits include student leadership experience and high career confidence. These students typically understand organizational mechanisms and the role of feedback in achieving goals. They are also adept at identifying key feedback sources. Driven by strong confidence and goal orientation (38), they are willing to expose deficiencies and actively seek growth. As noted by Baldwin et al. (39), peer-assisted learning offers reciprocal benefits: it reinforces the mentor’s professional identity while reducing the peer’s anxiety. For this group, we recommend a “peer mentor” mechanism. Organizing experience-sharing sessions or participating in training allows them to model positive behaviors, fostering a proactive feedback atmosphere.