Abstract
Background
Online health information-seeking behavior (OHISB) is important for nursing students, yet its heterogeneity and associations with digital health literacy and social support remain unclear.
Objective
To identify OHISB profiles among nursing students and examine the hypothesized indirect association between digital health literacy and profile membership through perceived social support.
Methods
This cross-sectional study included 772 nursing students. Data were collected using a sociodemographic questionnaire, the eHealth Literacy Scale, the Online Health Information-Seeking Behavior Scale, and the Perceived Social Support Scale. Latent profile analysis identified OHISB profiles. Multinomial logistic regression examined predictors of profile membership, with three-step analyses accounting for classification uncertainty. Structural equation modeling assessed indirect associations through perceived social support using bootstrap confidence intervals.
Results
Among the nursing students, latent profile analysis identified three distinct online health information-seeking behavior profiles: high (n = 286, 37%), moderate (n = 328, 42.5%), and low (n = 158, 20.5%) OHISB profiles. Place of origin, family structure, household income, education level, grade, and social support were associated with profile membership (p < 0.05). In adjusted three-step analyses, higher perceived social support was associated with lower odds of high- (OR = 0.956) and moderate-OHISB (OR = 0.959) membership relative to low-OHISB membership. Digital health literacy was positively associated with perceived social support (B = 0.293, p < 0.001). Perceived social support was negatively associated with high- (B = −0.311, OR = 0.733, p < 0.001) and moderate-OHISB membership (B = −0.215, OR = 0.807, p < 0.001) relative to low-OHISB membership. Direct effects of digital health literacy were not significant (p = 0.637 and 0.511). Significant indirect effects were observed for high versus low (a × b = −0.091, 95% bootstrap CI: −0.136 to −0.060) and moderate versus low (a × b = −0.063, 95% bootstrap CI: −0.102 to −0.036) contrasts.
Conclusion
Online health information-seeking behavior among nursing students was heterogeneous and characterized by three profiles. Perceived social support showed significant indirect associations between digital health literacy and OHISB profile membership. These findings highlight the potential importance of social support, while the cross-sectional design precludes causal inference.
Keywords: digital health literacy, latent profile analysis, nursing students, online health information-seeking behavior, social support, structural equation modeling
1. Introduction
The rapid growth of digital healthcare has revolutionized the dissemination and accessibility of health information. An individual’s capacity to effectively utilize digital tools to locate, interpret, and apply health information, known as digital health literacy, has become a pivotal determinant of personal health outcomes. Online health information-seeking behavior (OHISB), defined as the active pursuit, evaluation, and selection of health-related information online to satisfy knowledge needs and reduce uncertainty about one’s health status (1), is integral to evidence-based nursing practice and directly influences patient care quality (2). This competency is especially critical for nursing students, who must make informed decisions about their own health and accurately communicate health information to patients, thereby shaping health behaviors.
The Extended Model of Health Information Seeking Behavior conceptualizes individuals as both active seekers and passive recipients of information, shaped by environmental and personal factors (3). Empirical studies reveal significant associations among social support, digital health literacy, which encompasses the skills to search for, understand, critically evaluate, and integrate electronic health information from online sources to make informed health decisions, and OHISB (4–6). The relationship between digital health literacy and OHISB is complex and varies across populations (7, 8). Theoretically, digital health literacy may precede perceived social support because greater ability to access, evaluate, and use digital health resources may facilitate engagement with online health communities and interpersonal resources, thereby expanding individuals’ perceived access to social support. Social support may serve as a critical mediator, modulating the translation of digital health literacy into effective OHISB. However, the association may also be mutually beneficial, as greater social support may provide informational and interpersonal resources that help individuals improve their digital health literacy.
Research on OHISB has predominantly focused on college students (9, 10), older adults (11, 12), and patients with chronic conditions (13), while studies specifically targeting nursing students remain limited. Moreover, prevailing variable-centered approaches often overlook heterogeneity in individuals’ OHISB behaviors. Latent profile analysis (LPA), a person-centered statistical method, addresses this gap by identifying distinct OHISB subgroups based on response patterns, enabling nuanced characterization of group-specific traits (14).
This study employs LPA to delineate OHISB profiles among nursing students and use structural equation modeling (SEM) analysis to examines how digital health literacy influences perceived social support, which in turn affects the OHISB profiles. The findings aim to inform tailored, tiered health education interventions. We hypothesize that:
H1: Nursing students can be classified into multiple distinct OHISB profiles.
H2: Perceived social support is hypothesized to mediate the association between digital health literacy and OHISB profiles.
H3: The indirect association between digital health literacy and OHISB profile membership through perceived social support differs across OHISB profile contrasts.
2. Materials and methods
2.1. Study design and participants
A cross-sectional survey was conducted among nursing students from three medical colleges in Guangxi, China, between January and February 2026. Participants were recruited via a WeChat-based online questionnaire using convenience sampling. To prevent duplicate submissions, each IP address was restricted to a single response.
Nursing students currently enrolled in undergraduate or junior college programs, regardless of academic year, were eligible. Students with severe mental illness that could impede questionnaire completion were excluded. Consistent with a priori data-quality criteria, responses completed in fewer than 90 s or showing straightlining (identical responses across all items) were excluded. Finally no questionnaires were excluded by these criteria (minimum completion time = 101 s).
Sample size adequacy was evaluated in relation to the complexity of the planned analyses, including a three-indicator, three-class LPA, multinomial regression of profile membership, and a single-mediator SEM with a three-category latent outcome. Previous simulation studies have shown that LPA performance depends on sample size, the number and quality of indicators, the number of latent classes, and the separation between classes, rather than on a single fixed minimum sample size. In the present study, the three latent profiles comprised 286 (37.0%), 328 (42.5%), and 158 (20.5%) participants, respectively, with the smallest class containing 158 participants. The selected three-class solution also demonstrated good classification quality (entropy = 0.865). Thus, the final sample of 772 provided adequate representation of each latent class and sufficient observations for the subsequent multinomial regression and mediation analyses.
The study was approved by the institutional review board of the Science and Technology Ethics Committee of Guangxi Medical University Cancer Hospital (approval no. KY2026788), and informed consent was obtained from all participants.
2.2. Measures
2.2.1. Demographic characteristics
A self-designed questionnaire was used to collect demographic characteristics, including age, sex, student origin (urban/rural), family structure (nuclear/extended), family monthly income (<3,000, 3,000–5,999, and >6,000 CNY), education level (junior college/undergraduate), academic year (freshman to senior), and self-rated health (good/fair/poor).
2.2.2. Online health-seeking behavior scale (OHSB)
The OHISB was assessed using the scale, which comprises three subscales: information seeking (8 items), offline supplementation or replacement of professional care (11 items), and online social support seeking (18 items) (15). Items are rated on a 5-point Likert scale (5 = very frequent, 4 = frequent, 3 = sometimes, 2 = occasionally, and 1 = never), yielding a total score of 37–185, with higher scores indicating greater engagement in OHISB. The Cronbach’s α coefficient of the scale was 0.981, indicating very high internal consistency and substantial homogeneity among the items. Subscale sum scores were used as indicators in the LPA. A confirmatory factor analysis (CFA) supported the three-factor structure (RMSEA = 0.055, CFI = 0.893, SRMR = 0.056), justifying the use of the three subscale scores.
2.2.3. Digital health literacy
The eHealth Literacy Scale (eHEALS) was used to assess digital health literacy. It was originally developed by Norman et al. (7) and translated and validated in Chinese by Guo et al. (16). It consists of eight items, and items are rated on a 5-point Likert scale ranging from 1 (“strongly disagree”) to 5 (“strongly agree”). The total score ranges from 8 to 40, with higher scores indicating greater digital health literacy. The Cronbach’s α coefficient of the scale was 0.972. CFA supported a unidimensional structure (after allowing three residual correlations between conceptually similar items: CFI = 0.943, TLI = 0.906, SRMR = 0.041, RMSEA = 0.087), with all standardized loadings ≥ 0.58. Measurement invariance of the eHEALS across sex was supported (ΔCFI = 0.008 < 0.01 for configural vs. metric models, with invariant intercepts).
2.2.4. Perceived social support
Perceived social support was measured using the 12-item Perceived Social Support Scale (PSSS), originally developed by Zimet et al. (17) and translated and validated in Chinese by Jiang et al. (18). It comprises three dimensions: family support, friend support, and other support, with a total of 12 items. Items are rated on a 7-point Likert scale ranging from 1 (“strongly disagree”) to 7 (“strongly agree”). The total score ranges from 12 to 84, with higher scores reflecting greater perceived social support. The Cronbach’s α coefficient of the scale was 0.922. CFA supported the three-factor structure (RMSEA = 0.034, CFI = 0.987, SRMR = 0.021).
2.3. Statistical analysis
2.3.1. Descriptive statistics
Descriptive statistics were used to summarize participant characteristics by OHISB profile. Categorical variables were presented as frequencies and percentages, and continuous variables were summarized as means and standard deviations, as appropriate. Group differences in categorical variables were assessed using χ2 tests, whereas differences in digital health literacy and perceived social support across latent profiles were examined using the BCH three-step method to account for classification error.
2.3.2. Latent profile analysis
The LPA was conducted on the three OHISB subscale sum scores using Mplus 8.3. Models with 1–5 classes were estimated and compared on the AIC, BIC, sample-size-adjusted BIC, entropy, the Lo–Mendell–Rubin likelihood ratio test (LMR), and the parametric bootstrap likelihood ratio test (BLRT). The final model was selected based on statistical criteria, interpretability, and class sizes. Sensitivity analyses included a four-class solution and a re-analysis using mean subscale scores (subscale sums divided by item counts) as indicators, which reproduced the three-class solution exactly (n = 286, 158, 328; entropy = 0.865).
2.3.3. Three-step regression
To examine sociodemographic factors and social support as predictors of profile membership while accounting for classification uncertainty, a multinomial logistic regression was estimated using the manual three-step approach, with the low-OHISB profile as the reference. Results are reported as unstandardized logit coefficients (B), standard errors, Wald χ2, odds ratios (OR) with 95% confidence intervals (CI), and p-values.
2.3.4. Mediation analysis (SEM)
A structural equation model tested whether perceived social support mediated the association between digital health literacy and OHISB profile membership. Digital health literacy predicted social support (path a); social support predicted profile membership (paths b₁, b₂, multinomial logit); and digital health literacy directly predicted profile membership (paths c′₁, c′₂). Profiles were modeled as a latent categorical variable (three classes) to avoid treating class membership as observed. The model was estimated with the ML estimator and ALGORITHM = INTEGRATION, using a multinomial logit link for the categorical outcome. Indirect effects (a × b) and their 95% CIs were computed via MODEL CONSTRAINT and 1,000 bootstrap draws. A sensitivity analysis re-estimated the model with perceived social support as a latent variable indicated by its three subscales.
Analyses were conducted in Mplus 8.3; two-tailed p < 0.05 was considered statistically significant.
3. Results
3.1. Profile differences in demographic characteristics, digital health literacy, and social support
Participants’ ages ranged from 18 to 37 years, with a mean of (20.75 ± 1.79) years.
The three profiles differed significantly in all demographic variables (all P ≤ 0.05). Digital health literacy and perceived social support showed an inverse gradient across profiles: both were highest in the low-OHISB profile (32.12 ± 5.94 and 67.75 ± 11.48, respectively) and lowest in the high-OHISB profile (30.36 ± 5.73 and 58.62 ± 10.89, respectively), with the moderate-OHISB profile showing intermediate values (31.23 ± 5.55 and 62.37 ± 11.53, respectively). BCH-adjusted comparisons confirmed significant differences across profiles in digital health literacy (p = 0.008) and perceived social support (p < 0.001) (Table 1).
Table 1.
Univariate analysis of factors influencing latent profiles of OHISB among nursing students.
| Variable | Total (n = 772) | Low OHISB (n = 158) | Moderate (n = 328) | High (n = 286) | χ 2 | P |
|---|---|---|---|---|---|---|
| Sex | 37.45 | <0.001 | ||||
| Male | 126 (16.3) | 51 (32.3) | 43 (13.1) | 32 (11.2) | ||
| Female | 646 (83.7) | 107 (67.7) | 285 (86.9) | 254 (88.8) | ||
| Student origin | 36.51 | <0.001 | ||||
| Urban | 261 (33.8) | 42 (26.6) | 150 (45.7) | 69 (24.1) | ||
| Rural | 511 (66.2) | 116 (73.4) | 178 (54.3) | 217 (75.9) | ||
| Family structure | 75.71 | <0.001 | ||||
| Nuclear | 196 (25.4) | 80 (50.6) | 78 (23.8) | 38 (13.3) | ||
| Extended | 576 (74.6) | 78 (49.4) | 250 (76.2) | 248 (86.7) | ||
| Family income (CNY) | 83.41 | <0.001 | ||||
| <3,000 | 128 (16.6) | 14 (8.9) | 40 (12.2) | 74 (25.9) | ||
| 3,000–5,999 | 265 (34.3) | 30 (19.0) | 109 (33.2) | 126 (44.1) | ||
| >6,000 | 379 (49.1) | 114 (72.2) | 179 (54.6) | 86 (30.1) | ||
| Education | 81.87 | <0.001 | ||||
| Junior college | 194 (25.1) | 17 (10.8) | 53 (16.2) | 124 (43.4) | ||
| Undergraduate | 578 (74.9) | 141 (89.2) | 275 (83.8) | 162 (56.6) | ||
| Grade | 60.11 | <0.001 | ||||
| Freshman | 104 (13.5) | 49 (31.0) | 31 (9.5) | 24 (8.4) | ||
| Sophomore | 304 (39.4) | 58 (36.7) | 123 (37.5) | 123 (43.0) | ||
| Junior | 217 (28.1) | 34 (21.5) | 108 (32.9) | 75 (26.2) | ||
| Senior | 147 (19.0) | 17 (10.8) | 66 (20.1) | 64 (22.4) | ||
| Self-rated health | 14.30 | 0.006 | ||||
| Good | 425 (55.1) | 101 (63.9) | 183 (55.8) | 141 (49.3) | ||
| Fair | 313 (40.5) | 56 (35.4) | 131 (39.9) | 126 (44.1) | ||
| Poor | 34 (4.4) | 1 (0.6) | 14 (4.3) | 19 (6.6) | ||
| Digital health literacy | 31.09 ± 5.73 | 32.12 ± 5.94 | 31.23 ± 5.55 | 30.36 ± 5.73 | – | 0.008† |
| Social support | 62.08 ± 11.75 | 67.75 ± 11.48 | 62.37 ± 11.53 | 58.62 ± 10.89 | – | <0.001† |
Data are n (%) or mean ± SD.
†BCH-adjusted test across profiles (37), accounting for classification error.
3.2. Identification of OHISB profiles
Latent profile analysis on the three OHISB subscale sum scores identified a three-profile solution as the most parsimonious and interpretable model (Table 2). Although information criteria (AIC, BIC, aBIC) continued to decrease with additional classes, the LMR test was non-significant for four- and five-class solutions (p = 0.058 and 0.059, respectively), and the five-class solution contained a very small class (n = 6, 0.8%) lacking practical meaning. The three-class solution showed the highest entropy (0.865), all average latent class probabilities exceeded 0.90 (range 0.929–0.957), and a sensitivity analysis using mean subscale scores reproduced the three-class solution exactly (n = 286, 158, and 328; entropy = 0.865). The three-profile solution and the corresponding profile characteristics are presented in Figure 1.
Table 2.
Model fit indices for the LPA solutions (k = 1–5).
| Model | Log-likelihood | Free params | AIC | BIC | aBIC | Entropy | LMR (P) | BLRT (P) | Class sizes, n (%) |
|---|---|---|---|---|---|---|---|---|---|
| 1-class | −8580.546 | 6 | 17173.09 | 17200.99 | 17181.93 | – | – | – | 772 (100.0) |
| 2-class | −8162.656 | 10 | 16345.31 | 16391.80 | 16360.05 | 0.852 | <0.001 | <0.001 | 445 (57.6), 327 (42.4) |
| 3-class | −8007.463 | 14 | 16042.93 | 16108.01 | 16063.56 | 0.865 | <0.001 | <0.001 | 286 (37.0), 158 (20.5), 328 (42.5) |
| 4-class | −7963.627 | 18 | 15963.25 | 16046.94 | 15989.78 | 0.843 | 0.058 | <0.001 | 159 (20.6), 218 (28.2), 315 (40.8), 80 (10.4) |
| 5-class | −7925.548 | 22 | 15895.10 | 15997.37 | 15927.51 | 0.866 | 0.059 | <0.001 | 6 (0.8), 307 (39.8), 159 (20.6), 219 (28.4), 81 (10.5) |
Bold indicates the chosen model. LMR = Lo–Mendell–Rubin likelihood ratio test; BLRT = parametric bootstrap likelihood ratio test.
Figure 1.
Nursing students’ online health-seeking behavior profiles. The three profiles were characterized by low, moderate, and high scores across the three OHISB subscales: Information seeking (8 items, range 8–40), offline supplementation/replacement of professional care (11 items, range 11–55), and online social support seeking (18 items, range 18–90).
The three profiles were ordered by their OHISB subscale levels (Information Seeking, Offline Supplementation or Replacement of Professional Care and Online Social Support Seeking): 27.2/47.1/76.6 for the high-OHISB profile, 22.3/35.6/53.8 for the moderate-OHISB profile, and 17.9/24.7/37.3 for the low-OHISB profile, respectively. They were therefore labeled high-OHISB (n = 286, 37.0%), moderate-OHISB (n = 328, 42.5%), and low-OHISB (n = 158, 20.5%) profiles.
3.3. Multivariate analysis of factors influencing OHISB profiles among nursing students
Table 3 presents the three-step multinomial logistic regression with the low-OHISB profile as the reference. Perceived social support was negatively associated with profile membership: each one-point increase in social support was associated with lower odds of membership in the high-OHISB (OR = 0.956, 95% CI 0.925–0.988, p = 0.010) and moderate-OHISB (OR = 0.959, 95% CI 0.927–0.991, p = 0.015) profiles relative to the low-OHISB profile. Rural student origin was associated with lower odds of high-OHISB membership (OR = 0.404, p = 0.004), whereas extended family structure (OR = 3.051, p < 0.001), lower family income (<3,000 CNY: OR = 2.727, p = 0.022), junior college education (OR = 4.470, p < 0.001), and higher academic year (sophomore to senior: ORs 2.864–7.123, all p < 0.01) were associated with higher odds of high-OHISB membership.
Table 3.
Multivariate analysis of factors influencing OHISB profiles among nursing students (n = 772).
| Variable | B | SE | Wald χ2 | OR (95% CI) | P |
|---|---|---|---|---|---|
| High vs. Low OHISB | |||||
| Social support (per 1 point) | −0.045 | 0.017 | 6.61 | 0.956 (0.925–0.988) | 0.010 |
| Female (ref: Male) | 0.491 | 0.345 | 2.03 | 1.634 (0.83–3.21) | 0.154 |
| Rural (ref: Urban) | −0.906 | 0.318 | 8.15 | 0.404 (0.22–0.75) | 0.004 |
| Extended family (ref: Nuclear) | 1.115 | 0.311 | 12.90 | 3.051 (1.66–5.61) | <0.001 |
| Income <3,000 (ref: >6,000) | 1.003 | 0.439 | 5.24 | 2.727 (1.15–6.44) | 0.022 |
| Income 3,000–5,999 | 0.942 | 0.332 | 8.04 | 2.564 (1.34–4.92) | 0.005 |
| Junior college (ref: Undergraduate) | 1.497 | 0.399 | 14.12 | 4.470 (2.04–9.77) | <0.001 |
| Sophomore (ref: Freshman) | 1.052 | 0.376 | 7.81 | 2.864 (1.37–5.98) | 0.005 |
| Junior | 1.343 | 0.437 | 9.46 | 3.832 (1.63–9.02) | 0.002 |
| Senior | 1.963 | 0.504 | 15.20 | 7.123 (2.65–19.1) | <0.001 |
| Not-good health (ref: Good) | 0.069 | 0.269 | 0.07 | 1.071 (0.63–1.81) | 0.798 |
| Moderate vs. Low OHISB | |||||
| Social support (per 1 point) | −0.042 | 0.017 | 5.94 | 0.959 (0.927–0.991) | 0.015 |
| Female (ref: Male) | 0.525 | 0.330 | 2.52 | 1.690 (0.89–3.23) | 0.112 |
| Rural (ref: Urban) | −1.484 | 0.293 | 25.66 | 0.227 (0.13–0.40) | <0.001 |
| Extended family (ref: Nuclear) | 0.961 | 0.278 | 11.94 | 2.614 (1.52–4.51) | 0.001 |
| Income <3,000 (ref: >6,000) | 0.285 | 0.467 | 0.37 | 1.330 (0.53–3.32) | 0.541 |
| Income 3,000–5,999 | 0.588 | 0.323 | 3.32 | 1.800 (0.96–3.39) | 0.069 |
| Junior college (ref: Undergraduate) | 0.303 | 0.425 | 0.51 | 1.354 (0.59–3.11) | 0.475 |
| Sophomore (ref: Freshman) | 0.861 | 0.369 | 5.45 | 2.366 (1.15–4.88) | 0.020 |
| Junior | 1.208 | 0.402 | 9.03 | 3.347 (1.52–7.36) | 0.003 |
| Senior | 1.598 | 0.466 | 11.74 | 4.942 (1.98–12.3) | 0.001 |
| Not-good health (ref: Good) | 0.031 | 0.269 | 0.01 | 1.031 (0.61–1.75) | 0.909 |
3.4. Mediating effect of social support on the relationship between digital health literacy and OHISB
Table 4 and Figure 2 summarize the mediation model. Digital health literacy was positively associated with perceived social support (B = 0.293, p < 0.001; path a). Perceived social support was negatively associated with membership in the high-OHISB (B = −0.311, OR = 0.733, p < 0.001) and moderate-OHISB (B = −0.215, OR = 0.807, p < 0.001) profiles relative to the low-OHISB profile (paths b₁, b₂). The indirect effects of digital health literacy on profile membership through social support were significant and negative for both the high-vs.-low (a × b = −0.091, 95% bootstrap CI − 0.136 ~ −0.060) and moderate-vs.-low (a × b = −0.063, 95% bootstrap CI − 0.102 ~ −0.036) contrasts, whereas the direct effects of literacy were non-significant (p = 0.637 and 0.511). A sensitivity analysis treating social support as an observed total score yielded consistent results (indirect effects −0.078 and −0.054).
Table 4.
Mediation of the association between digital health literacy and OHISB profile membership through perceived social support.
| Path | B (SE) | P | OR (95% CI) | Bootstrap 95% CI |
|---|---|---|---|---|
| a: Digital health literacy → Social support | 0.293 (0.028) | <0.001 | – | – |
| b₁: Social support → High vs. Low OHISB | −0.311 (0.061) | <0.001 | 0.733 (0.650–0.826) | – |
| b₂: Social support → Moderate vs. Low | −0.215 (0.056) | <0.001 | 0.807 (0.723–0.900) | – |
| c′₁: Literacy → High vs. Low (direct) | 0.013 (0.028) | 0.637 | 1.013 | – |
| c′₂: Literacy → Moderate vs. Low (direct) | 0.018 (0.028) | 0.511 | 1.018 | – |
| Indirect 1: Literacy → Support → High vs. Low (a × b₁) | −0.091 (0.019) | <0.001 | – | −0.136 ~ −0.060 |
| Indirect 2: Literacy → Support → Moderate vs. Low (a × b₂) | −0.063 (0.017) | <0.001 | – | −0.102 ~ −0.036 |
B = unstandardized coefficient (logit scale for profile paths); OR = odds ratio; CI = confidence interval. Indirect effects = a × b, computed via MODEL CONSTRAINT; bootstrap CIs exclude zero, indicating significant indirect effects.
Figure 2.
Path diagram of the hypothesized mediation model with perceived social support as a latent variable. *** < 0.001.
4. Discussion
4.1. Main findings and heterogeneity of OHIS
This study identified three distinct latent profiles of OHISB among nursing students: low (158 students, 20.5%), moderate (328 students, 42.5%), and high (286 students, 37.0%) profiles. This finding supports the first hypothesis and indicates substantial heterogeneity in OHISB among nursing students. The three profiles were characterized by relatively low, moderate, and high levels across the three dimensions of OHISB, including information seeking, offline supplementation or replacement of professional care, and online social support seeking. This finding is consistent with previous research suggesting that OHISB is multidimensional and shaped by multiple individual and contextual factors (8).
The observed heterogeneity may be explained by differences in individual characteristics, access to digital resources, and perceptions of online health information. Previous studies have shown that demographic characteristics, education level, and digital health literacy are associated with individuals’ ability and willingness to seek health information online (19–21). In addition, trust in online health information may influence how individuals search for and use digital health resources (22). Even among nursing students with professional training, difficulties in evaluating the credibility of online information may lead to inappropriate trust in low-quality sources or underuse of reliable information (23, 24).The findings show that students in the high-OHISB profile had the lowest levels of digital health literacy and perceived social support. However, students in the low-OHISB profile had the highest levels of both. This finding suggests that a higher frequency or intensity of OHISB should not necessarily be interpreted as more adaptive or higher-quality health information behavior. The OHISB scale includes not only general information seeking but also offline supplementation or replacement of professional care and online social support seeking. Thus, high scores may be related to intensive reliance on online information or online support rather than effective use of health information. Previous research has linked lower eHealth literacy with problematic online health behaviors, including cyberchondria and health anxiety, supporting the possibility that intensive online health information use may sometimes reflect uncertainty or dependence rather than greater competence (25).
Multinomial regression further showed that several sociodemographic characteristics were associated with profile membership. Rural origin was associated with lower odds of high- and moderate-OHISB membership, while extended family structure, lower household income, junior college education, and higher academic year were associated with higher odds of high-OHISB membership. These findings suggest that the patterns of online health information use may be related to socioeconomic resources and digital access. Previous research has emphasized the digital divide and sociodemographic characteristics can influence digital health literacy and Internet-based health information use (26–29). The finding that lower educational attainment was associated with higher odds of high-OHISB membership also suggests that intensive online health information seeking does not necessarily indicate greater digital competence.
Therefore, nursing education should focus not only on encouraging online health information seeking but also on strengthening students’ ability to critically evaluate and appropriately use digital health information.
4.2. Sociodemographic factors and profile differences
Multinomial logistic regression showed that rural student origin, extended family structure, lower family income, junior college education, and higher academic year were independently associated with OHISB profile membership. Specifically, rural origin was associated with lower odds of high- and moderate-OHISB membership, while extended family structure, lower household income, junior college education, and higher academic year were associated with higher odds of high-OHISB membership.
The observed rural–urban difference is consistent with research highlighting the impact of the digital divide and socioeconomic conditions on access to and use of digital resources (27, 30). Educational attainment and academic progression may also influence digital skills and online information-search strategies. However, the finding that junior college students had higher odds of high-OHISB membership suggests that intensive online health information seeking does not necessarily correspond to greater digital health literacy (21, 31). The person-centered studies among nursing students have similarly shown the differences in digital literacy profiles and associations with educational characteristic (32, 33).
Family structure was also associated with OHISB profiles, with students from extended families were more likely associated with moderate- and high-OHISB membership. This association showed differences in family communication, resource sharing, or access to interpersonal information sources.
These findings suggest that interventions should be tailored according to students’ educational and socioeconomic backgrounds rather than assuming that greater online information-seeking intensity represents greater competence. We can conduct more structured training in source evaluation, verification, and critical appraisal of online health information, which may really benefit students with high OHISB but relatively low digital health literacy.
4.3. Association between digital health literacy, social support, and OHISB
This study found a positive association between digital health literacy and perceived social support (B = 0.293). Digital health literacy was related to perceived social support, because people with greater ability to use digital health resources may take a more active part in online health communities and interpersonal resources, thereby getting more social support. This interpretation is consistent with previous research linking digital/eHealth literacy with the effective use of digital health resources and interpersonal support (34, 35).
However, the mediation analysis does not establish a causal sequence. Perceived social support was negatively associated with membership in the high-OHISB (B = −0.311, OR = 0.733) and moderate-OHISB (B = −0.215, OR = 0.807) profiles relative to the low-OHISB profile. The indirect effects of digital health literacy through perceived social support were significant and negative for both the high-vs.-low (a × b = −0.091, 95% bootstrap CI − 0.136 ~ −0.060) and moderate-vs.-low (a × b = −0.063, 95% bootstrap CI − 0.102 ~ −0.036) contrasts. Higher digital health literacy was statistically associated with higher perceived social support, which was in turn associated with low OHISB profiles. Several explanations are possible. First, differences in the operationalization of OHISB may be important: the present scale includes behaviors involving supplementation or replacement of professional care and online social support seeking. Higher scores may be associated with seeking a lot and even being dependent. Second, students with lower offline social support may compensate by seeking information and validation online. Third, demographic and socioeconomic factors may shape both digital competence and online information-use patterns. Therefore, high-frequency online health information seeking should not automatically be equated with high-quality or adaptive information behavior.
The absence of a statistically significant direct effect of digital health literacy should also not be interpreted as proof of “full mediation.” Rather, the results indicate a statistically significant indirect association through perceived social support, while the direct associations were not statistically significant. Given the cross-sectional design, these findings should be interpreted as an indirect association consistent with the proposed mediation model rather than evidence of a causal mechanism.
4.4. Directionality, alternative models, and implications
In this study, we placed digital health literacy before perceived social support in the model because students who are better at finding and evaluating online health information may also be better able to search and use support resources, whether online health communities or face-to-face networks. The positive association we observed (B = 0.293) is consistent with this idea. The reverse direction is just as plausible, however. Strong social networks may themselves build digital skills—through encouragement, informal guidance, and everyday opportunities to practice using health information online (36). With cross-sectional data we cannot tell which pathway comes first, and the two may well reinforce each other. We therefore regard the mediation model as one reasonable specification rather than evidence of temporal order, and there is no causal claim.
The practical implications follow from the profile pattern regardless of direction. Students in the high-OHISB profile sought the most health information online while reporting the lowest digital health literacy and perceived social support. For these students, it’s not useful to continuously seek online. We should provide more training in critical appraisal on judging sources, verifying claims, and recognizing unreliable content. In addition, real-world support from peers, faculty, and family could also help students get more information, instead of only depending on online health information seeking. We offer these suggestions cautiously, because a cross-sectional study cannot establish that improving literacy or support would actually reduce heavy online health information use.
5. Limitations
This study has several limitations. First, its cross-sectional design restricts the ability to establish causal or temporal relationships among variables; therefore, future research should employ longitudinal designs to validate these findings. Although digital health literacy was theoretically specified as an antecedent of perceived social support, the cross-sectional design cannot establish temporal ordering, and reciprocal or alternative pathways remain possible; therefore, the indirect effects should be interpreted as hypothesized associative relationships rather than causal mechanisms. Second, the sample was drawn exclusively from nursing colleges within a single region, which limits the generalizability of the results; thus, multi-center and cross-regional studies are necessary to enhance representativeness and applicability. Because institution identifiers were not collected, clustering by institution could not be explicitly modeled, and residual between-institution heterogeneity may have affected the precision of the estimates. Third, although LPA statistically identifies distinct classes, the labeling and interpretation of these categories remain somewhat subjective. Future research should incorporate qualitative methods to gain deeper insights into the underlying behavioral motivations and contextual factors influencing these profiles. The LPA was based on three OHISB subscale scores rather than item-level indicators; although CFA and sensitivity analyses supported the identified structure, alternative profile solutions cannot be completely excluded. Three-step and BCH-adjusted analyses were used to account for classification uncertainty, although some uncertainty in profile assignment may remain. Fourth, all variables were measured using self-reported questionnaires, which may introduce common method bias despite the use of validated instruments. Recall and social desirability biases may also have affected the responses; future studies should incorporate objective behavioral measures where possible.
6. Conclusion
This study classified nursing students’ OHISB into low, moderate, and high profiles. The three profiles were characterized as low (20.5%), moderate (42.5%), and high (37.0%) OHISB profiles. Student origin, family structure, income, education level, grade, and social support were significant predictors of these OHISB profiles. Higher perceived social support was associated with low OHISB profiles. Digital health literacy was positively associated with perceived social support, while higher perceived social support was negatively associated with membership in the high OHISB profiles. Mediation analysis indicated significant indirect associations of digital health literacy with OHISB profile membership through perceived social support. These findings highlight social support as a key bridge linking digital health literacy to OHISB. However, given the cross-sectional design, these findings should be interpreted as associative rather than causal, and the directionality of the digital health literacy–social support relationship remains uncertain. Interventions should focus on enhancing both digital skills and social support networks, adopting stratified, dynamic, and personalized educational strategies. Particular attention may be warranted for students with intensive online health-information seeking but lower digital health literacy and perceived social support.
Acknowledgments
We would like to thank the students who participated in this study for their tremendous cooperation.
Funding Statement
The author(s) declared that financial support was not received for this work and/or its publication.
Edited by: Shaozhuo Huang, Erasmus Medical Center, Netherlands
Reviewed by: Zhihui Liu, People's Hospital of Zhengzhou University, China
Harun Aslan, Kastamonu University, Türkiye
Abbreviations: OHISB, Online Health Information-Seeking Behavior; LPA, Latent Profile Analysis; SEM, Structural Equation Modeling; eHEALS, the eHealth Literacy Scale; OHSB, Online Health-seeking Behavior Scale; PSSS, Perceived Social Support Scale; AIC, Akaike Information Criterion; BIC, Bayesian Information Criterion; aBIC, adjusted Bayesian Information Criterion; LMRT, Lo–Mendell–Rubin adjusted likelihood ratio test; BLRT, Bootstrap Likelihood Ratio Test; OR, Odds Ratio; CI, Confidence Interval.
Data availability statement
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.
Ethics statement
The studies involving humans were approved by The Science and Technology Ethics Committee of Guangxi Medical University Cancer Hospital. The studies were conducted in accordance with the local legislation and institutional requirements. The participants provided their written informed consent to participate in this study.
Author contributions
LY: Methodology, Conceptualization, Writing – review & editing, Writing – original draft, Formal analysis. LL: Investigation, Writing – review & editing, Data curation, Validation. HH: Investigation, Writing – review & editing, Validation, Data curation. KP: Investigation, Validation, Data curation, Writing – review & editing. QD: Validation, Writing – review & editing, Data curation, Investigation. XM: Writing – original draft, Validation, Writing – review & editing, Conceptualization, Project administration, Methodology, Formal analysis, Supervision.
Conflict of interest
The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Generative AI statement
The author(s) declared that Generative AI was used in the creation of this manuscript. During the preparation of this manuscript, the authors used ChatGPT (OpenAI) to assist with language refinement, clarity, and editorial suggestions. The tool was not used for study design, data collection, statistical analysis, interpretation of the findings, or the generation of research data. All AI-assisted content was critically reviewed, verified, and revised by the authors, who take full responsibility for the accuracy and integrity of the manuscript.
Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.
Publisher’s note
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
References
- 1.Moreno MA. Seeking health information online. JAMA Pediatr. (2017) 171:500. doi: 10.1001/jamapediatrics.2016.3109, [DOI] [PubMed] [Google Scholar]
- 2.Gilmour J, Strong A, Chan H, Hanna S, Huntington A. Primary health-care nurses and internet health information-seeking: access, barriers and quality checks. Int J Nurs Pract. (2016) 22:53–60. doi: 10.1111/ijn.12361, [DOI] [PubMed] [Google Scholar]
- 3.Longo DR, Schubert SL, Wright BA, LeMaster J, Williams CD, Clore JN. Health information seeking, receipt, and use in diabetes self-management. Ann Fam Med. (2010) 8:334–40. doi: 10.1370/afm.1115, [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Wong DK, Cheung MK. Online health information seeking and eHealth literacy among patients attending a primary care clinic in Hong Kong: a cross-sectional survey. J Med Internet Res. (2019) 21:e10831. doi: 10.2196/10831, [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Shi Z, Du X, Li J, Hou R, Sun J, Marohabutr T. Factors influencing digital health literacy among older adults: a scoping review. Front Public Health. (2024) 12:1447747. doi: 10.3389/fpubh.2024.1447747, [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Zhao BY, Huang L, Cheng X, Chen TT, Li SJ, Wang XJ, et al. Digital health literacy and associated factors among internet users from China: a cross-sectional study. BMC Public Health. (2024) 24:908. doi: 10.1186/s12889-024-18324-0, [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Norman CD, Skinner HA. eHEALS: the eHealth literacy scale. J Med Internet Res. (2006) 8:e27. doi: 10.2196/jmir.8.4.e27, [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Chen J, Yang Y, Xia H, Duan Y, Da C, Cai T, et al. Patterns and associated factors of online health information seeking behaviors among young women diagnosed with breast cancer in China. Asia Pac J Oncol Nurs. (2025) 12:100700. doi: 10.1016/j.apjon.2025.100700, [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Zhang D, Zhan W, Zheng C, Zhang J, Huang A, Hu S, et al. Online health information-seeking behaviors and skills of Chinese college students. BMC Public Health. (2021) 21:736. doi: 10.1186/s12889-021-10801-0, [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Chen J, Tian H. Associations and gender differences between OHI-seeking behaviors and eHealth literacy among Chinese university students. Rev Esc Enferm USP. (2024) 58:e20230340. doi: 10.1590/1980-220X-REEUSP-2023-0340en, [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Zhu X, Li C. Information support or emotional support? Social support in online health information seeking among Chinese older adults. Healthcare. (2024) 12:1790. doi: 10.3390/healthcare12171790, [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Ma X, Liu Y, Zhang P, Qi R, Meng F. Understanding online health information seeking behavior of older adults: a social cognitive perspective. Front Public Health. (2023) 11:1147789. doi: 10.3389/fpubh.2023.1147789, [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Zeng L, Zhang YP, Zhang X, Chen JF, Bai CQ, Lei XX, et al. Relationship between the latent profiles of health information-seeking behaviour and glycated haemoglobin levels in patients with type 2 diabetes: a cross-sectional survey in rural areas of China. BMJ Open. (2025) 15:e088891. doi: 10.1136/bmjopen-2024-088891, [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Muthén B, Muthén LK. Integrating person-centered and variable-centered analyses: growth mixture modeling with latent trajectory classes. Alcohol Clin Exp Res. (2000) 24:882–91. doi: 10.1111/j.1530-0277.2000.tb02070.x, [DOI] [PubMed] [Google Scholar]
- 15.Peng JL, Qu XY, Li SK, Liu CQ, Wang HY, Deng JX. Chinese translation and validation of the online health information-seeking behavior scale. J Nurs Sci. (2025) 40:104–7. [Google Scholar]
- 16.Guo SJ, Yu XM, Sun YY, Nie D, Li XM, Wang L. Chinese translation and applicability of the eHealth literacy scale (eHEALS). Chin J Health Educ. (2013) 29:106–8. doi: 10.16168/j.cnki.issn.1002-9982.2013.02.019 [DOI] [Google Scholar]
- 17.Zimet GD, Dahlem NW, Zimet SG, Farley GK. The multidimensional scale of perceived social support. J Pers Assess. (1988) 52:30–41. doi: 10.1207/s15327752jpa52012 [DOI] [PubMed] [Google Scholar]
- 18.Jiang QJ. Perceived social support scale (PSSS). Chin Ment Health J. (2000) (1999 Suppl):131–133. [Google Scholar]
- 19.Kim S, Oh J. The relationship between E-health literacy and health-promoting Behaviors in nursing students: a multiple mediation model. Int J Environ Res Public Health. (2021) 18:5804. doi: 10.3390/ijerph18115804, [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Xiong Z, Zhang L, Li Z, Xu W, Zhang Y, Ye T. Frequency of online health information seeking and types of information sought among the general Chinese population: cross-sectional study. J Med Internet Res. (2021) 23:e30855. doi: 10.2196/30855, [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Wang X, Wang Y, Chen X, Lv S, Ding W, Ma S, et al. Subtypes of digital literacy and their association with innovative behavior among Chinese undergraduate nursing students: a latent profile analysis. Front Public Health. (2026) 14:1717234. doi: 10.3389/fpubh.2026.1717234, [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Lee SY, Choi J, Noh GY. Factors influencing health-related internet activities and their outcomes. J Health Commun. (2016) 21:1179–86. doi: 10.1080/10810730.2016.1236853, [DOI] [PubMed] [Google Scholar]
- 23.Theron M, Redmond A, Borycki EM. Baccalaureate nursing students' abilities in critically identifying and evaluating the quality of online health information. Stud Health Technol Inform. (2017) 234:321–7. [PubMed] [Google Scholar]
- 24.Eysenbach G, Köhler C. How do consumers search for and appraise health information on the world wide web? Qualitative study using focus groups, usability tests, and in-depth interviews. BMJ. (2002) 324:573–7. doi: 10.1136/bmj.324.7337.573 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Kim M, Kim B, Park S. Social support, eHealth literacy, and mHealth use in older adults with diabetes: moderated mediating effect of the perceived importance of app design. Comput Inform Nurs. (2024) 42:136–43. doi: 10.1097/CIN.0000000000001081, [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Rong H, Lu L, He M, Guo T, Li X, Tao Q, et al. Online health information-seeking Behaviors among the Chongqing population: cross-sectional questionnaire study. JMIR Form Res. (2025) 9:e56028. doi: 10.2196/56028, [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Wang X, Ge L, Yu Z, Qiao Q, Guo X, Liu S, et al. Decomposition and comparative analysis of urban-rural differences in eHealth literacy among nursing students: a multisite cross-sectional study. Nurse Educ Today. (2025) 146:106562. doi: 10.1016/j.nedt.2024.106562, [DOI] [PubMed] [Google Scholar]
- 28.Zhong Z, Hu D, Zheng F, Ding S, Luo A. Relationship between information-seeking behavior and innovative behavior in Chinese nursing students. Nurse Educ Today. (2018) 63:1–5. doi: 10.1016/j.nedt.2018.01.004, [DOI] [PubMed] [Google Scholar]
- 29.Ashkanani H, Asery R, Bokubar F, AlAli N, Mubarak S, Buabbas A, et al. Web-based health information seeking among students at Kuwait University: cross-sectional survey study. JMIR Form Res. (2019) 3:e14327. doi: 10.2196/14327, [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Shi X, Li J, Ji X, Wu Y, Zang S. The mediating effects of self-efficacy, family health, and perceived stress on the relationship between perceived social support and eHealth literacy in nursing students: a structural equation model. BMC Nurs. (2024) 23:868. doi: 10.1186/s12912-024-02546-z, [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Xu L, Li N, Ding X, Wang H, Yin P, Pan W. Electronic health literacy and related factors among nursing students: a latent profile analysis. Front Public Health. (2026) 14:1845664. doi: 10.3389/fpubh.2026.1845664 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.van Kessel R, Wong BLH, Rubinić I, O'Nuallain E, Czabanowska K. Is Europe prepared to go digital? Making the case for developing digital capacity: an exploratory analysis of Eurostat survey data. PLoS Digit Health. (2022) 2:e0000013. doi: 10.1371/journal.pdig.0000013 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Estacio EV, Whittle R, Protheroe J. The digital divide: examining socio-demographic factors associated with health literacy, access and use of internet to seek health information. J Health Psychol. (2019) 24:1668–75. doi: 10.1177/1359105317695429, [DOI] [PubMed] [Google Scholar]
- 34.Erdat Y, Sezer Ceren RE, Ozdemir L, Uslu-Sahan F, Bilgin A. Influence of technical, cognitive and socio-emotional factors on digital literacy in nursing students assessed using structural equation modeling. Nurse Educ Today. (2023) 130:105937. doi: 10.1016/j.nedt.2023.105937, [DOI] [PubMed] [Google Scholar]
- 35.Aslan H, Koca GŞ. Examining the relationships between e-health literacy, cyberchondria, health anxiety, and health perception among foreign university students. BMC Public Health. (2025) 25:4368. doi: 10.1186/s12889-025-25636-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Cohen S, Wills TA. Stress, social support, and the buffering hypothesis. Psychol Bull. (1985) 98:310–57. doi: 10.1037/0033-2909.98.2.310, [DOI] [PubMed] [Google Scholar]
- 37.Asparouhov T, Muthén B. Auxiliary Variables in Mixture Modeling: Using the BCH Method in Mplus to Estimate a distal Outcome model and an Arbitrary Secondary model. Mplus Web Notes. (2014). Available online at: https://www.statmodel.com/download/asparouhov_muthen_2014.pdf (Accessed August 20, 2026).
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.