Correlation between health literacy and risk-taking behaviors among medical university students

Abstract

Background

Health literacy (HL) is a pivotal educational competency and a critical life skill that empowers individuals to make informed health decisions. Given the susceptibility of medical sciences students to various psychosocial stressors, elucidating the link between their HL capacities and behavioral choices is essential. This study aimed to investigate the association between health literacy and engagement in high-risk behaviors among students at Lorestan University of Medical Sciences.

Methods

This descriptive-analytical cross-sectional study was conducted in 2025 among 306 students selected via stratified random sampling. Data were collected using the Health Literacy for Iranian Adults (HELIA) and the Youth Risk Behavior Survey (YRBS) questionnaires. Statistical analyses included Pearson correlation and multiple linear regression. Model assumptions were verified, and the Variance Inflation Factor (VIF) was assessed to ensure the absence of multicollinearity.

Results

The mean scores for health literacy and high-risk behaviors were 132.96 ± 14.40 and 96.17 ± 10.74, respectively. A significant inverse correlation was observed between health literacy and high-risk behaviors (r=-.289, p<.001). Multiple regression analysis revealed that after controlling for demographic variables, health literacy was a significant inverse predictor of high-risk behaviors (β=-.334, p<.001), accounting for 12.5% of the variance (R²=.125). Dimensional analysis further indicated that the ‘Access to health information’ dimension was the most significant predictor of risk reduction (β=-0.23, p=0.002).

Conclusion

Higher health literacy is significantly associated with a reduction in high-risk behaviors among medical students. Specifically, the ability to effectively access and navigate health resources emerges as the most critical predictor of this association. These findings underscore the necessity of integrating functional health literacy training—focusing on information access and critical appraisal—into medical curricula to support student well-being and professional competency.

Introduction

Health literacy (HL) is increasingly recognized as a critical determinant of public health and a fundamental educational competency [1, 2]. It encompasses an individual’s capacity to obtain, process, and understand basic health information and services necessary to make informed and appropriate health decisions [2]. Beyond basic literacy, HL is a multifaceted life skill that enables individuals to navigate complex healthcare systems and engage in sustainable health-promoting behaviors [3]. The impact of HL extends across diverse populations and clinical settings; for instance, it has been strongly linked to the promotion of healthy behaviors among women during the COVID-19 pandemic [4] and is significantly associated with quality of life in both older adults [5] and nursing professionals in teaching hospitals [6]. In the context of higher education, HL is viewed not only as a personal asset but also as a vital outcome of the educational process, empowering students to manage their own well-being effectively [7, 8].

Medical sciences students, as future healthcare providers, occupy a unique position within the health system. Their level of HL is of paramount importance, as it governs not only their personal health-related choices but also their future professional capacity to communicate complex health information to patients and serve as effective health educators [9, 10]. Despite their advanced academic training, medical students often encounter significant psychosocial challenges, including intense academic pressure, chronic time constraints, and exposure to high-stress clinical environments, which may paradoxically hinder their ability to maintain healthy lifestyles [11]. Evidence suggests that gaps in functional HL among this population may lead to suboptimal health outcomes and a diminished capacity to act as role models for their patients [12]. Furthermore, data from other healthcare professionals, such as nurses, indicate that HL status is closely tied to professional well-being and quality of life [6], suggesting that theoretical academic knowledge is insufficient without the functional ability to apply that knowledge to one’s own health.

The theoretical mechanism through which HL mitigates risk-taking behavior is rooted in the cognitive process of health decision-making. Higher HL enables individuals to critically appraise health information, accurately perceive the magnitude of risks associated with specific behaviors, and anticipate long-term consequences. By enhancing the capacity for behavioral evaluation and informed decision-making—competencies that are pivotal in clinical shared decision-making processes [13]—HL acts as a cognitive buffer that transforms passive health knowledge into active behavioral change, thereby reducing the propensity to engage in maladaptive or high-risk activities.

Despite the established role of health literacy in general populations, the theoretical mechanisms through which it modulates risk-taking behaviors—particularly the transition from health knowledge to actual behavioral change—remain poorly understood. According to the Social Cognitive Theory, health literacy does not merely provide information but enhances an individual’s self-efficacy and cognitive appraisal of risks, which are critical in resisting peer pressure and managing stressful environments. However, previous studies have yielded inconsistent results, often failing to distinguish between theoretical knowledge and the functional ability to apply that knowledge in real-world scenarios—a phenomenon known as the ‘knowledge-practice gap.’ This inconsistency is particularly evident among medical students, who possess high clinical knowledge but may lack the functional health literacy required to navigate their own health-related decisions. Furthermore, there is a critical dearth of comprehensive data linking these distinct dimensions of health literacy to a broad spectrum of high-risk behaviors within the specific sociocultural context of western Iran, including the Lorestan region [14, 15]. Given the unique institutional dynamics and the evolving educational landscape of medical universities in this region, clarifying these associations is essential to inform targeted public health interventions.

Consequently, the present study was designed to examine the association between health literacy and high-risk behaviors among students at Lorestan University of Medical Sciences. We hypothesized that: [1] overall health literacy would exhibit a significant inverse correlation with the prevalence of risk-taking behaviors, and [2] specific dimensions of health literacy, particularly the ‘Access to health information’* dimension, would serve as more potent predictors of risk reduction compared to other dimensions. By elucidating these critical drivers, this study aims to provide an empirical foundation for developing integrated educational strategies that enhance students’ health decision-making capacities and foster a safer, healthier campus environment.

Methods

Study design and participants

This descriptive-analytical cross-sectional study was conducted between 2025 and 2026 among students at Lorestan University of Medical Sciences (LUMS), Khorramabad, Iran. The primary objective of the study was to investigate the association between health literacy and high-risk behaviors.

Inclusion and exclusion criteria

The inclusion criteria were: [1] being an active undergraduate or graduate student at LUMS during the study period, and [2] providing written informed consent to participate. Participants were excluded if they had: [1] incomplete submission of the questionnaires (defined as more than 10% missing data), or [2] a self-reported history of severe psychological disorders. The latter exclusion was specifically implemented to prevent potential confounding effects, as severe psychiatric conditions can independently influence both an individual’s health literacy levels and their propensity for risk-taking behaviors, which could otherwise mask the primary association under investigation.

Sampling and sample size

To ensure a representative sample across the university, a stratified random sampling method was employed. Each faculty was treated as a stratum, and participants were selected proportionally to the total number of students in each faculty using a random number table. The distribution of the sample across faculties was as follows: [Medicine: 115, Paramedical: 85, Nursing and Midwifery: 66, Dentistry: 19, Pharmacy: 21].

The sample size was estimated based on the Pearson correlation coefficient ($r = 0.29$) reported in previous literature [14]. With a type I error rate of 1% ($\alpha = 0.01$) and a statistical power of 99%, the minimum required sample size was calculated as 270. To account for a potential non-response rate of approximately 15%, a total of 340 students were invited. Ultimately, 306 participants completed the survey, resulting in a response rate of 90% (306/340). To assess the potential for non-response bias, a comparison of the baseline demographic characteristics between respondents and non-respondents was performed; no significant differences were observed, suggesting that the sample remains representative of the student population.

Data collection instruments

Data were collected using a structured demographic questionnaire and two standardized instruments:

1. Health Literacy for Iranian Adults (HELIA)

Health literacy was assessed using the HELIA tool, a 33-item instrument specifically developed and validated for the Iranian population by Montazeri et al. [16]. The tool evaluates HL across five distinct dimensions: Reading, Access, Comprehension, Assessment, and Decision-making. In the present study, the overall internal consistency (Cronbach’s alpha) was 0.93. To ensure the reliability of each domain, the Cronbach’s alpha values for the subscales were calculated as follows: Reading (α = 0.72), Access (α = 0.79), Comprehension (α = 0.86), Assessment (α = 0.77), and Decision-making (α = 0.89).

To enable comparability across domains, raw scores were transformed to a standardized scale ranging from 0 to 100 using the following formula: [Standardized Score = (Raw Score - Minimum Possible Score) / (Maximum Possible Score - Minimum Possible Score) × 100]

Based on the standardized scores, health literacy levels were categorized into four groups: inadequate (0–50), marginal (50.1–66), adequate (66.1–84), and excellent (84.1–100).

• 2.Youth Risk Behavior Survey (YRBS)

Engagement in high-risk behaviors was measured using the YRBS. To ensure cultural and linguistic appropriateness, the Persian version of this instrument, which was rigorously validated and adapted for Iranian students by Rahmati et al. [17]. through extensive face and content validity tests, was utilized. The internal consistency of the YRBS in the current study was 0.79. Each item was measured using a five-point Likert scale (always, often, sometimes, rarely, never), scored from 1 to 5. Depending on the nature of the items, reverse scoring was applied in some domains, such that higher scores consistently indicated lower engagement in risky behaviors.

The questionnaire covers multiple domains including injuries and violence-related behaviors, tobacco use, substance abuse, sexual risk behaviors, physical inactivity, unhealthy nutrition, and mental health. The total possible score for the risk behavior section was 175, distributed as follows: tobacco use (5 points), substance abuse (25 points), sexual risk behaviors (45 points), physical inactivity (30 points), unhealthy nutrition (15 points), and injury-related behaviors (55 points). Additionally, the mental health section had a maximum score of 30.

To facilitate interpretation and comparison, raw scores were linearly transformed to a 0–100 scale using the following formula: [Transformed Score = ((Raw Score - Minimum Score) / (Maximum Score - Minimum Score)) × 100]

In this scaling method, higher scores indicate healthier behaviors and lower levels of risk.

Demographic Questionnaire:

• A supplementary demographic form was used to collect data on age, gender, faculty, and socioeconomic status, including self-reported monthly family income.

Ethical considerations

The study protocol was approved by the Ethics Committee of Lorestan University of Medical Sciences. Written informed consent was obtained from all participants. To ensure ethical integrity, data anonymity and confidentiality were strictly maintained throughout the recruitment and analysis process.

Data analysis

Data were analyzed using IBM SPSS Statistics version 26.0 [18]. To ensure the accuracy of the multivariate analysis, all categorical variables were dummy-coded for regression analysis (e.g., Gender: 0 = Male, 1 = Female; Residence Place: 0 = Urban, 1 = Rural). Descriptive statistics, including means, standard deviations, and frequencies, were used to summarize the demographic characteristics of the participants.

To examine the associations between health literacy and risk-taking behaviors, Pearson correlation analysis was performed for both the total HELIA score and its five individual dimensions. Subsequently, multiple linear regression (MLR) analysis was conducted to determine the predictive power of the HL dimensions on risk-taking behaviors, while adjusting for relevant demographic covariates (e.g., age, gender, and socioeconomic status).

To ensure the validity of the regression model, all fundamental assumptions were rigorously verified. Multicollinearity was assessed using the Variance Inflation Factor (VIF), with a threshold of VIF < 2.5 considered indicative of the absence of significant multicollinearity [19]. The normality of residuals was examined using P-P plots and the Kolmogorov-Smirnov test, which indicated that the residuals were normally distributed. Additionally, the analysis of standardized residual plots confirmed homoscedasticity. To prevent the influence of outliers, Mahalanobis distance was calculated, and any extreme cases were identified and handled according to standard statistical protocols.

To avoid overfitting and ensure model stability, the sample size was evaluated relative to the number of predictors; the current sample of 306 participants exceeds the recommended ratio of 15 observations per predictor, ensuring sufficient statistical power. All regression coefficients were reported with their corresponding 95% confidence intervals (CIs) to provide a precise estimate of the effect sizes. Statistical significance was predefined as p < 0.05.

Results

Socio-demographic characteristics of participants

The descriptive analysis revealed that the participants were predominantly female (59.2%) and single (87.9%). Regarding their living conditions, a significant majority resided in urban areas (95.1%) and lived in university dormitories (65.7%). From a professional standpoint, the vast majority of students were unemployed (88.2%).

Within the academic domain, the highest concentration of students was enrolled in bachelor’s degree programs (46.7%), with those in their fourth semester forming the largest academic cohort (19.3%). Medicine emerged as the most frequent field of study, representing 31.0% of the participants.

Analysis of continuous variables showed that the participants’ age ranged from 19 to 56 years, with a mean of 25.36 (SD = 5.91). The wide age range (19–56 years) was primarily due to the inclusion of graduate and PhD students, as well as those returning to medical education after a hiatus, which is common in the diverse academic structure of the university. The mean monthly family income was 22,248,366.01 Tomans (SD = 8,630,317.16), and the average birth order was 2.02 (SD = 1.09). The comprehensive socio-demographic profile is detailed in Table 1.

Table 1.

Socio-demographic characteristics of student participants (N = 306)

Qualitative Variables	n	%
Gender
Male	125	40.80
Female	181	59.20
Type of Residence
Dormitory	201	65.70
Private residence	105	34.30
Place of Residence
Urban	291	95.10
Rural	15	4.90
Marital Status
Married	37	12.10
Single	269	87.90
Educational Level
Associate degree	6	2.00
Bachelor’s degree	143	46.70
Master’s degree	19	6.20
Clinical doctorate	135	44.10
PhD	3	1.00
Employment Status
Student (unemployed)	270	88.20
Employed	36	11.80
Academic Semester
1	21	6.90
2	55	18.00
3	16	5.20
4	59	19.30
5	12	3.90
6	39	12.70
7	13	4.20
8	39	12.70
9	10	3.30
10	15	4.90
11	5	1.60
12	8	2.60
13	7	2.30
14	7	2.30
Field of Study
Medicine	95	31.00
Paramedical sciences	60	19.60
Nutrition	8	2.60
Dentistry	19	6.20
Pharmacy	21	6.90
Public health	18	5.90
Nursing	66	21.60
Laboratory sciences	19	6.20

Quantitative Variables	Range (Min–Max)	Mean (SD)
Age (years)	19–56	25.36 (5.91)
Monthly income (Tomans)	8,000,000–50,000,000	22,248,366.01 (8,630,317.16)
Birth order	1–9	2.02 (1.09)

Health literacy and high-risk behavior scores

The descriptive assessment of the study variables indicated that the mean aggregate score for health literacy among the students was 132.96 (SD = 14.40). The overall mean score for high-risk behaviors was 96.17 (SD = 10.74). For a more granular perspective, the scores were decomposed into their constituent dimensions; health literacy was evaluated across reading, access, comprehension, appraisal, and decision-making dimensions, while high-risk behaviors were assessed across seven distinct domains, including mental health, nutrition, and substance abuse. The complete descriptive summary of these dimensions is detailed in Table 2.

Table 2.

Mean scores of health literacy and high-risk behaviors among students

Variables	Min	Max	Mean (SD)
Health Literacy (Total)	101	160	132.96 (14.40)
Reading literacy	8	20	15.91 (2.96)
Access	17	30	26.01 (3.68)
Comprehension	21	35	30.44 (3.72)
Appraisal	11	20	16.21 (2.23)
Decision-making and behavior	29	59	44.38 (6.92)
High-Risk Behaviors (Total)	70	129	96.17 (10.74)
Injuries	13	42	20.02 (4.79)
Tobacco use	1	5	1.91 (1.18)
Addiction and substance abuse	5	20	8.55 (3.16)
Sexual behaviors	14	41	31.98 (5.49)
Nutrition	3	12	7.18 (2.04)
Physical activity	10	23	17.37 (3.00)
Mental health	5	18	9.15 (2.55)

Bivariate analysis and correlations

The comparative analysis of health literacy (HL) scores across various demographic strata revealed that employment status was a highly significant factor associated with HL levels (p < 0.001); specifically, employed students exhibited significantly higher mean HL scores compared to their unemployed counterparts. In contrast, no statistically significant disparities were observed in HL scores relative to gender, marital status, or field of study (all p > 0.05).

Notably, a marginal trend toward significance was observed regarding the place of residence (p = 0.055) and educational level (p = 0.075). For the place of residence, the effect size was medium (Cohen’s d = 0.51), indicating a noticeable difference between urban and rural students. However, for the educational level, the effect size was small (η² = 0.026), suggesting that while a trend exists, the academic degree has a minimal impact on HL scores (Table 3).

Table 3.

Association of demographic variables with health literacy and high-risk behaviors among students

Variable	Health Literacy Mean (SD)	p-value	High-Risk Behavior Mean (SD)	p-value
Gender
Male	133.06 (14.81)	.920ᵃ	97.07 (11.34)	.226ᵃ
Female	132.89 (14.15)		95.55 (10.28)
Marital Status
Married	136.29 (11.60)	.133ᵃ	96.37 (11.38)	.903ᵃ
Single	132.50 (14.70)		96.14 (10.66)
Employment Status
Student (unemployed)	131.88 (14.44)	< .001ᵃ	96.44 (10.64)	.239ᵃ
Employed	141.05 (11.36)		94.19 (11.38)
Type of Residence
Dormitory	133.46 (14.58)	.403ᵃ	95.89 (10.75)	.528ᵃ
Private residence	132.01 (14.06)		96.71 (10.74)
Place of Residence
Urban	133.32 (14.33)	.055ᵃ	96.15 (10.90)	.876ᵃ
Rural	126.00 (14.34)		96.60 (6.92)
Educational Level
Bachelor’s	131.53 (14.24)	.075ᵇ	96.77 (10.82)	.524ᵇ
Master’s	141.05 (10.36)		92.57 (9.46)
Clinical doctorate	133.17 (14.42)		96.22 (10.90)
PhD	141.33 (18.14)		95.67 (11.59)
Associate	132.33 (21.11)		92.66 (8.73)
Field of Study
Medicine	133.62 (13.68)	.466ᵇ	96.48 (10.79)	.678ᵇ
Paramedical sciences	130.90 (16.10)		95.26 (10.67)
Nutrition	141.25 (11.49)		90.62 (10.66)
Dentistry	129.84 (18.01)		94.00 (9.23)
Pharmacy	134.19 (14.43)		97.04 (12.87)
Public health	134.05 (11.60)		95.33 (11.26)
Nursing	136.23 (9.71)		97.73 (10.18)
Laboratory sciences	136.57 (11.99)		97.73 (10.18)

ᵃ Independent samples t-test

ᵇ One-way ANOVA

Regarding high-risk behaviors (HRB), the analysis demonstrated a consistent lack of statistically significant association with any of the examined demographic variables, including gender, marital status, employment status, residence type, residential location, or educational level (all p > 0.05). These findings suggest that high-risk behaviors among the sampled students were distributed homogeneously across different socio-demographic groups, indicating that these behaviors may be driven by factors beyond basic demographic characteristics (Table 3).

Pearson’s correlation analysis

To further evaluate the associations between the study variables, Pearson’s correlation analysis was performed. The results revealed a statistically significant, moderate inverse correlation between aggregate health literacy and high-risk behaviors (r = -0.289, p < 0.001), indicating that higher levels of health literacy are associated with a notable reduction in risk-taking behaviors among students. Conversely, no statistically significant linear relationships were observed between high-risk behaviors and other continuous demographic variables, including age, academic semester, birth order, or income level (all p > 0.05), although income level showed a marginal trend toward significance (p = 0.061). These correlations are summarized in Table 4.

Table 4.

Correlations of health literacy, age, academic semester, income level, and birth order with high-risk behaviors among students

Outcome Variable	Statistic	Health literacy	Age	Academic semester	Income level	Birth order
High-risk behaviors	r	− 0.289	− 0.054	0.003	− 0.107	− 0.018
	p-value	< 0.001	0.348	0.962	0.061	0.752

r: Correlation coefficients were calculated using Pearson’s correlation test

Predictors of high-risk behaviors: general health literacy model

To determine the primary determinants of high-risk behaviors among the student population, a multiple linear regression analysis was performed using the Enter method. The model incorporated all socio-demographic characteristics alongside the aggregate health literacy score as independent predictors.

Prior to the primary analysis, the robustness and validity of the model were verified through several diagnostic tests. The independence of residuals was confirmed by a Durbin-Watson statistic of approximately 1.89. Furthermore, the absence of significant multicollinearity among the predictors was evidenced by the Variance Inflation Factor (VIF) and Tolerance values; as detailed in Table 5, VIF values ranged from 1.03 to 2.13 and Tolerance from 0.47 to 0.97, both falling well within established academic thresholds.

Table 5.

Multiple linear regression analysis of predictors of students’ high-risk behaviors

Variables	Unstandardized Coefficient (B)	95% CI (Lower, Upper)	Standard Error	β	t	p-value	VIF	Tolerance
Constant	147.78	128.0,167.0	9.99	-	14.796	< 0.001	-	-
Health Literacy	− 0.249	(-0.33,-0.17)	0.043	− 0.334	-5.729	< 0.001	1.153	0.867
Age	− 0.010	(-0.29,0.27)	0.143	− 0.005	− 0.068	0.946	2.136	0.468
Semester	− 0.130	(-0.51,0.25)	0.192	− 0.041	− 0.680	0.497	1.391	0.719
Income (Per 1 M Tomans)	− 0.179	(-0.33,-0.03)	0.075	− 0.118	-2.380	0.018	1.117	0.895
Birth Order	-1.241	(-2.36,-0.12)	0.572	− 0.127	-2.170	0.031	1.159	0.863
Gender	-2.485	(-4.94,-0.02)	1.250	− 0.114	-1.988	0.048	1.153	0.867
Marital Status	-2.174	(-6.22,1.87)	2.097	− 0.066	-1.036	0.301	1.376	0.727
Employment	− 0.993	(-5.78,3.81)	2.443	− 0.030	− 0.406	0.685	1.851	0.540
Residence Type	0.760	(-1.85,3.37)	1.320	0.034	0.576	0.565	1.162	0.861
Residence Place	-2.985	(-8.74,2.76)	2.908	− 0.060	-1.026	0.306	1.157	0.864
R2=0.125

Income was rescaled to 1,000,000 Tomans units to improve coefficient interpretability

The overall regression model was statistically significant and explained 12.5% of the total variance in high-risk behaviors (Adjusted R² = 0.095, F = 4.206, p < 0.001). After adjusting for all other covariates in the multivariate model, health literacy emerged as the most potent inverse predictor of high-risk behaviors (β= − 0.334, p < 0.001). Specifically, a one-unit increase in the health literacy score was associated with a 0.249-unit decrease in the mean high-risk behavior score.

In addition to health literacy, three demographic factors maintained significant predictive power within the model: birth order (β= − 0.127, p = 0.031), monthly income (β= − 0.118, p = 0.018), and gender (β= − 0.114, p = 0.048). Conversely, other variables—including age, academic semester, marital status, employment, and residential characteristics—did not contribute significantly to the model’s predictive capacity (p > 0.05). The comprehensive regression coefficients are summarized in Table 5.

Dimensional analysis of health literacy

To achieve a more granular understanding of the specific components of health literacy that mitigate high-risk behaviors, a secondary multiple linear regression analysis was conducted. In this model, the aggregate health literacy score was replaced by its five constituent dimensions—reading, access, comprehension, appraisal, and decision-making—while maintaining the same set of demographic control variables. The dimensional regression model demonstrated overall statistical significance, explaining approximately 12.2% of the variance in high-risk behaviors (Adjusted R² = 0.122, F = 4.035, p < 0.001). This approach aimed to isolate the primary drivers of behavioral change among students.

As evidenced by the results in Table 6, the decomposition of health literacy revealed that the ‘Access to health information’ dimension was the only significant negative predictor of high-risk behaviors (β= -0.23, p = 0.002). While the dimensions of comprehension (β= -0.18, p = 0.053) and decision-making (β= -0.12, p = 0.071) exhibited a marginal trend toward significance, they did not reach the conventional statistical threshold of p < 0.05. Other dimensions, including reading and appraisal, did not demonstrate a significant association with the outcome.

Table 6.

Multiple linear regression analysis predicting High-Risk Behaviors (HRB)

Variable	Unstandardized Coefficient (B)	95% CI (Lower, Upper)	β	t	p
(Constant)	137.48	(123.13,151.83)	-	18.79	< 0.001***
Reading Dimension	0.14	(-0.45,0.73)	0.04	0.47	0.639
Access Dimension	− 0.67	(-1.10,-0.24)	− 0.23	-3.07	0.002**
Comprehension Dimension	− 0.52	(-1.05,0.01)	− 0.18	-1.95	0.053
Appraisal Dimension	0.49	(-0.08,1.06)	0.10	1.67	0.096
Decision-making Dimension	− 0.19	(-0.40,0.02)	− 0.12	-1.81	0.071
Gender	-2.01	(-4.34,0.32)	− 0.09	-1.69	0.093
Birth Order	-1.40	(-2.50,-0.03)	− 0.14	-2.51	0.013*
Monthly Income (Million Tomans)	− 0.17	(-0.32,-0.02)	− 0.13	-2.27	0.024*

*p < 0.05, **p < 0.01, ***p < 0.001;Adjusted R² = 0.122, F = 4.035,p < 0.001

Notably, the predictive influence of birth order (β= -0.14, p = 0.013) and monthly income (β= -0.13, p = 0.024) remained remarkably consistent with the findings from the general model.

Collectively, these findings suggest that the overall protective effect of health literacy observed in the aggregate model is primarily driven by the students’ capacity to effectively access health resources. This indicates that the ability to locate and obtain health information is the most critical literacy skill in reducing high-risk behaviors within this population. This is further supported by the stability of the demographic predictors observed in Table 6.

Discussion

Health literacy (HL) serves as a pivotal determinant of both individual and societal well-being, encompassing the multifaceted capacity to access, comprehend, and apply health information to facilitate informed decision-making. Given the current socio-economic and cultural shifts, elucidating the determinants of HL—particularly among medical sciences students—is imperative for the development of evidence-based health professions curricula [20]. The present study identified a significant positive association between employment status and health literacy, with employed students exhibiting higher HL levels than their unemployed counterparts.

This association may be linked to several synergistic factors. Primarily, employment may provide experiential learning opportunities that bridge the gap between theoretical knowledge and the practical application of health concepts in real-world settings. This aligns with findings by Lipson et al. (2022), who noted that professional engagement is often associated with superior communication and problem-solving skills, which may facilitate the acquisition of functional health literacy [21]. Furthermore, the professional environment may grant students increased access to informational resources and expanded social networks, both of which are critical for HL enhancement [22]. However, it is essential to consider alternative explanations; for instance, students who secure employment may possess higher baseline levels of self-discipline or psychological maturity—traits that independently correlate with higher HL. Additionally, the socioeconomic advantages inherent in employment may be a contributing factor, as higher income often facilitates enhanced access to supplementary educational resources [23].

Conversely, the professional burden may act as a double-edged sword. Keech et al. (2018) observed that work-related stress and professional pressures can be associated with the neglect of personal health, potentially undermining an individual’s health literacy [24]. Such divergent findings across studies may stem from variations in measurement instruments, sampling strategies, and sociocultural contexts. For instance, discrepancies often arise when assessment tools prioritize theoretical knowledge over practical, functional competencies [23, 24].

The multivariate analysis revealed a modest but statistically significant inverse association between HL and high-risk behaviors. The observed moderate negative correlation (r = -0.289) is consistent with the magnitude of associations reported in analogous populations, reinforcing the protective role of HL across diverse cohorts, from hospitalized women [4] to elderly patients [5] and nursing professionals [6]. Moreover, the capacity of HL to enhance shared clinical decision-making in cancer patients underscores the universal necessity of these competencies in optimizing health outcomes [13].

Notably, the regression model explained 12.5% of the total variance in high-risk behaviors (R² = 0.125). This relatively low explained variance suggests that while HL is a significant predictor, it operates within a complex interplay of other unmeasured determinants. It is likely that psychological factors—such as stress, anxiety, and burnout, which are prevalent among medical students—as well as social determinants, including peer pressure and personality traits like sensation-seeking, contribute substantially to risk-taking behaviors and were not captured in the current model [25]. Given the cross-sectional design of the present study, these associations should be interpreted as correlations rather than causal relationships.

Crucially, the dimensional analysis revealed that the ‘Access to health information’ subdomain was the primary significant driver of risk reduction (β = -0.23, p = 0.002). This implies that for medical students, the ability to effectively locate and obtain reliable health resources is more strongly linked to a reduction in risk behaviors than other literacy components [26, 27]. Interestingly, while it was initially hypothesized that the ‘Decision-making and Behavior Evaluation’ dimension would be the most potent predictor, our findings indicate that ‘Access’ plays a more critical role. This discrepancy highlights a paradoxical gap: the susceptibility to high-risk behaviors among students may not be primarily rooted in a deficit of decision-making capacity, but rather in the structural and functional challenges associated with accessing and navigating reliable health resources within a high-stress academic environment.

From a theoretical perspective, this observation is consistent with the tenets of Social Cognitive Theory, which posits that health literacy—particularly the competency to access and synthesize information—is often associated with higher levels of self-efficacy and an enhanced perception of risk. Such cognitive frameworks may be linked to a student’s capacity to regulate their behaviors and resist maladaptive coping strategies [25, 28]. In alignment with this framework, gender emerged as a significant predictor; the observed higher prevalence of high-risk behaviors among male students is consistent with the magnitude of effects reported in previous studies [29, 30]. Similarly, the association between birth order and risk-taking tendencies suggests that family dynamics and socialization processes may play a nuanced role in shaping these behavioral patterns [28].

These insights carry profound implications for the evolution of medical education. Given that students in the medical sciences are future healthcare providers, the cultivation of robust health literacy competencies is imperative. Integrating these skills—specifically the capacity to access and critically appraise health information—into the core curriculum could potentially optimize the personal health outcomes of the students while simultaneously enhancing their future clinical efficacy in communicating complex health information to patients [27, 31].

Despite these insights, several limitations must be acknowledged, which are elaborated upon in the subsequent section.

Limitations

Despite the significant insights provided by this study, several limitations warrant consideration. First, the inherent nature of the cross-sectional design precludes the establishment of definitive causal inferences. Consequently, the observed associations between health literacy and high-risk behaviors should be interpreted as correlational, and longitudinal research is required to elucidate the temporal relationship between these variables.

Second, the reliance on self-reported instruments for assessing high-risk behaviors may have introduced social desirability bias. Given that several analyzed behaviors—such as substance use or sexual activity—are sensitive and potentially stigmatized, participants may have under-reported these actions to align with perceived social norms. This likely resulted in an underestimation of the true prevalence of high-risk behaviors within this population.

Third, while multiple demographic covariates were controlled, the possibility of residual confounding cannot be entirely excluded. The relatively modest variance explained by the regression model (R²=0.125) underscores the likelihood that health literacy operates within a complex interplay of unmeasured determinants. It is probable that latent factors—such as psychological distress, personality traits, family history of addiction, or systemic environmental stressors—exerted an influence on risk-taking tendencies that was not captured in the current model.

Fourth, the single-center framework, conducted within a specific medical university, may constrain the external validity and generalizability of the findings. The institutional culture and environmental factors unique to this setting may not be fully representative of the broader medical student population across different regions or diverse sociocultural contexts.

Fifth, the use of self-reported monthly income as a proxy for socioeconomic status may be susceptible to recall bias or reporting inaccuracies, particularly within volatile economic environments. This potentially limits the precision of the socioeconomic analysis.

Finally, although the HELIA instrument is a robust and validated tool for assessing general health literacy, it does not encompass the specificities of digital health literacy (eHealth literacy). In the contemporary era of telemedicine and ubiquitous online health information, the ability to navigate and critically evaluate digital resources is a crucial competency for medical students that remained unaddressed in the present study [28, 31]. Future research should integrate validated digital health literacy scales and employ longitudinal designs to more comprehensively map the trajectory of these behavioral determinants.

Future research directions

To build upon the findings of the present study and address the identified methodological constraints, future research should prioritize the implementation of longitudinal and interventional designs. Such frameworks would be instrumental in transitioning from observed correlations to the elucidation of definitive causal mechanisms and in evaluating the long-term efficacy of targeted educational interventions. Furthermore, transitioning from a single-center approach to multicentric studies encompassing diverse academic and sociocultural settings would significantly augment the external validity and generalizability of these results.

To mitigate the inherent limitations of self-report bias and social desirability, future inquiries should employ mixed-method approaches. The triangulation of quantitative data with qualitative insights—such as semi-structured interviews or focus groups—could provide a more nuanced understanding of the lived experiences of students and the systemic drivers of high-risk behaviors. Additionally, the integration of objective behavioral assessments would enhance the precision of the findings.

Given the critical role of the ‘Access to health information’ dimension identified in this study, specific attention should be directed toward developing and testing curriculum-based interventions that prioritize functional health literacy. Research should investigate how enhancing a student’s ability to navigate and critically appraise health resources—particularly within digital environments (eHealth literacy)—moderates risk-taking tendencies. Such empirical evidence is essential for the design of evidence-based health promotion strategies that not only improve the personal well-being of future healthcare providers but also optimize their clinical capacity to empower patients through improved health communication.

Conclusion

The present study provides critical evidence of a significant, albeit modest, inverse association between health literacy (HL) and the prevalence of high-risk behaviors among medical sciences students. While employment status emerged as a positive correlate of higher HL, the multivariate analysis highlighted that male gender, lower household income, and birth order are associated with increased risk-taking tendencies. Most notably, the dimensional decomposition of HL revealed that the ‘Access to health information’ subdomain is the most significant predictor of risk reduction. This finding underscores a pivotal insight: for this specialized population, the ability to effectively navigate and secure reliable health resources is a critical competency associated with the mitigation of high-risk behaviors.

From a pedagogical perspective, these results signal a need for a paradigm shift in medical education. The findings suggest that the traditional emphasis on theoretical clinical knowledge may be insufficient for fostering personal health resilience. Instead, there is a compelling need to integrate functional health literacy training—specifically targeting the capacity to access and critically appraise health information—into the core medical sciences curriculum. Prioritizing these practical competencies could empower students to proactively manage their own health and make evidence-based decisions throughout their rigorous academic journey.

Ultimately, such an educational evolution serves a dual purpose: it supports the personal well-being of students during their formative years and potentially enhances their future professional efficacy as health educators and patient advocates. Although the cross-sectional design of this study precludes the establishment of definitive causal inferences, the identified associations provide a robust empirical foundation for future longitudinal and interventional research. Such studies are essential to evaluate the efficacy of targeted HL programs in reducing high-risk behaviors and optimizing the health trajectories of future healthcare providers.

Authors’ contributions

E.F.: Conceptualization, Data curation, Writing – original draft. M.S.: Methodology, Formal analysis, Writing – review & editing. Z.K.: Formal analysis, Investigation, Writing – review & editing. R.M.: Data curation, Writing – review & editing. H.F.: Data curation, Investigation, Supervision, Writing – review & editing. All authors read and approved the final manuscript.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Data availability

The datasets generated during and/or analyzed during the current study are not publicly available due to keeping participants’ information confidential but are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

This study was approved by the Ethics Committee of Lorestan University of Medical Sciences (Ethics Code: IR.LUMS.REC.1403.065). The study was conducted in accordance with the Declaration of Helsinki.

Written informed consent was obtained from all individual participants included in the study. Participants were informed that they could withdraw at any time without consequence.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.