The relationship between digital healthy diet literacy, physical activity, and health literacy: a community-based study

Abstract

Background

Health literacy and digital healthy diet literacy are increasingly recognized as important determinants of health behaviors. However, the relationship between these literacy domains and physical activity remains complex and not fully understood. This study aimed to examine the interrelationships between health literacy, digital healthy diet literacy, and physical activity levels in a community-based adult population.

Methods

This cross-sectional, community-based study included 1,047 adults aged 18–65 years recruited from three districts with different socioeconomic characteristics in Istanbul, Türkiye. Health literacy was assessed using the Health Literacy Scale-Short Form (HLS-SF), digital healthy diet literacy using the Digital Healthy Diet Literacy Scale (DHDLS), and physical activity levels using the International Physical Activity Questionnaire-Short Form (IPAQ-SF). Correlation analyses, group comparisons, and binary logistic regression were conducted to examine associations between variables.

Results

The median HLS-SF score was 20.83 (range: 0–50), indicating limited health literacy, while the median DHDLS score was 29.17 (range: 0–50). Nearly half of the participants (46.6%) had low physical activity levels. A moderate positive correlation was found between health literacy and digital healthy diet literacy (r = 0.488, p < 0.001). Health literacy and digital healthy diet literacy scores were positively correlated with age and body mass index. However, higher health literacy was associated with a lower likelihood of engaging in moderate-to-high physical activity (OR = 0.970, p < 0.001). Male gender, district of residence, family history of obesity, and prior physical activity counseling were positively associated with achieving moderate-to-high physical activity levels.

Conclusion

This study demonstrates that while health literacy and digital healthy diet literacy are correlated, higher health literacy does not necessarily translate into higher physical activity levels. Physical activity appears to be more influenced by environmental, social, and structural factors than by individual knowledge alone. Public health interventions should move beyond information-based approaches and adopt multicomponent strategies that address contextual barriers to behavior change.

Introduction

Health is defined by the World Health Organization (WHO) as “a state of complete physical, social and mental well-being and not merely the absence of disease or infirmity” [1]. Health literacy, on the other hand, is defined by the WHO as an individual’s ability to access, understand, and use health-related information in order to protect and maintain health [2]. Health literacy is considered one of the key determinants of health [2]. Unfortunately, low levels of health literacy have been reported among children, adolescents, and adults even in economically developed countries [3]. Increasing health literacy not only improves individual health outcomes but also enables individuals to participate more effectively in health-related collective and societal actions [2].

According to the WHO, chronic diseases are among the most significant health problems worldwide and are considered a global health threat [4]. The leading risk factors for chronic diseases include unhealthy diet, physical inactivity, tobacco use, and excessive alcohol consumption [5]. These risk factors are behavioral in nature and are therefore modifiable [6]. Several studies in the literature have reported that individuals with higher levels of health literacy are more likely to adopt healthy lifestyle behaviors such as regular physical activity and healthy eating habits [7–9]. For this reason, increasing health literacy is of critical importance for the prevention of chronic diseases and the promotion of health at the population level.

Physical activity is defined by the World Health Organization as any bodily movement produced by skeletal muscles that requires energy expenditure [10]. Being physically active plays an important role in the prevention of diseases such as cardiovascular diseases, diabetes, and cancer, as well as in the reduction of symptoms of depression and anxiety [10]. According to the WHO, adults aged 18 years and older are recommended to engage in at least 150 min of moderate-intensity physical activity or at least 75 min of vigorous-intensity physical activity per week [11].

Similar to physical activity, healthy nutrition has numerous positive effects on overall health and well-being. A healthy diet not only protects against all forms of malnutrition but also has a protective effect against non-communicable diseases such as diabetes, cardiovascular diseases, stroke, and cancer [12]. In this context, the way individuals access nutritional information has changed in the digitalized world, and the concept of digital healthy diet literacy has gained importance. Digital Healthy Diet Literacy is the ability to access, understand, appraise, and apply digital information to adopt healthier eating behaviors. As an extension of health literacy, it enables individuals to use online sources for dietary management [13]. Given the proliferation of nutrition-related content on social media and health websites, individuals’ ability to critically appraise digital dietary information has direct implications for their dietary behavior and, consequently, for the prevention of non-communicable diseases.

Although smartphone and internet access are widespread (over 85% nationwide), significant variation remains in individuals’ ability to critically evaluate online health and nutrition information [14]. Districts with lower socioeconomic development levels are more likely to have lower digital literacy, which may influence both access to and appraisal of digital health content [15]. These contextual factors should be considered when interpreting digital healthy diet literacy in community-based studies, as structural inequalities may shape digital health behaviors independently of individual motivation.

A growing body of evidence supports the association between health literacy and health-promoting behaviours. These reviews found that individuals with higher health literacy are more likely to engage in regular physical activity, adhere to dietary recommendations, and utilise preventive health services [9, 16].

Despite growing research interest in both health literacy and digital health information behaviours, studies simultaneously examining general health literacy, digital healthy diet literacy, and physical activity within a single community-based sample remain scarce. It should be noted that the present study focused on digital healthy diet literacy as a specific subdomain of digital health literacy, rather than digital health literacy in its broader sense. In particular, it is still unclear whether digital nutrition literacy shares common determinants with general health literacy or whether it constitutes a distinct capacity with different behavioural correlates. This study aims to address these gaps by examining the interrelationships among these three constructs in a diverse urban community sample in Istanbul, Türkiye. In this regard, digital health diet literacy, physical activity, and health literacy levels were evaluated among individuals aged 18–65 years.

Methods

Study type, population and design

This cross-sectional study was conducted between May 6, 2024, and August 13, 2024, in the Sultanbeyli, Fatih, and Eyüpsultan districts of Istanbul, Türkiye. These districts were selected to represent different socioeconomic development levels, a classification strategy previously utilized in regional health literacy research previously employed by the authors [17]. The study population consisted of individuals aged 18–65 years registered with family physicians in these districts.

The target sample size was determined to be 384 participants per district (totaling 1,152), calculated based on a 95% confidence level and a 5% margin of error; and an assumed prevalence of 50% (p = 0.50), representing the most conservative estimate for single-proportion calculations. The sampling frame was constructed by identifying eligible individuals from the electronic family physician registry. Participants were selected using a computer-generated simple randomization algorithm. To minimize non-response bias, if a participant did not answer the initial call, a second attempt was made on a different day; individuals who did not respond to two consecutive calls were replaced by the next randomly selected individual from the list.

A total of 1,064 participants agreed to participate and completed the telephone interview (initial response rate: 89.8%). After excluding 17 respondents who fell outside the targeted 18–65 age range during the verification step, a final sample of 1,047 participants was included in the analysis (overall effective response rate: 88.4%). Data were entered directly into Google Forms by the interviewers; since all survey items were set as mandatory fields, the final dataset contained no missing values.

Questionnaire and scales

The first section of the questionnaire included sociodemographic questions and items related to participants’ health history. The second section comprised the Digital Healthy Diet Literacy Scale and the Health Literacy Scale. The final section included the 7-item International Physical Activity Questionnaire-Short Form.

Health literacy scale-short form (HLS-SF)

The scale developed by Duong et al. (2019) was adapted into Turkish by Karahan Yılmaz and Eskici (2021) [18, 19]. The scale consists of 12 items and employs a 4-point Likert-type response format ranging from 1 (very difficult) to 4 (very easy). For scoring, the formula “(Mean − 1) × 50 / 3” is used. The index value calculated using this formula ranges from 0 to 50, with higher scores indicating better health literacy. In the Turkish adaptation study, Cronbach’s alpha coefficient was reported as 0.856, while in the present study it was calculated as 0.866.

Digital healthy diet literacy scale (DHDLS)

The scale developed by Duong et al. (2020) was adapted into Turkish by Karahan Yılmaz and Eskici [13, 19]. For scoring, the formula “Index = (Mean − 1) × 50 / 3” is used. The index value calculated using this formula ranges from 0 to 50, with higher scores indicating better healthy diet literacy. The scale consists of 4 items and employs a 4-point Likert-type response format ranging from 1 (very difficult) to 4 (very easy). The Cronbach’s alpha coefficient was reported as 0.785 in the Turkish adaptation study and was calculated as 0.849 in the present study.

International physical activity questionnaire-short form (IPAQ-SF)

The International Physical Activity Questionnaire-Short Form (IPAQ-SF) is used to assess physical activity levels and Sağlam et al. conducted the Turkish validity and reliability study of the questionnaire [20]. The questionnaire evaluates individuals’ physical activity levels over the past week in four categories: vigorous activity, moderate activity, walking, and sitting. The MET values (multiples of resting oxygen consumption) assigned to each activity (8 METs for vigorous activity, 4 METs for moderate activity, and 3.3 METs for walking) are multiplied by the duration and frequency (number of days) of each activity to calculate the total weekly MET-minutes score. The total physical activity score (MET-min/week) is calculated by summing the scores of walking, moderate-intensity activity, and vigorous-intensity activity. Physical activity levels are then classified as low (< 600 MET-min/week), moderate (600–3000 MET-min/week), and high (> 3000 MET-min/week) [20, 21].

Statistical analysis

Statistical analyses were performed using IBM SPSS Statistics (Version 24.0). Descriptive statistics were presented as medians and interquartile ranges (IQR) for continuous variables, and as frequencies (n) and percentages (%) for categorical data. The normality of continuous variables was assessed using visual (histograms and probability plots) and analytical methods (Kolmogorov-Smirnov and Shapiro-Wilk tests). Since the data were not normally distributed, the Mann-Whitney U test was employed for group comparisons, and Spearman’s correlation coefficient was used to evaluate relationships between numeric variables. The association between categorical variables was analyzed using the Pearson chi-square test.

District was purposively selected to represent three distinct socioeconomic development levels and was therefore included in all analyses as a theoretically meaningful contextual variable rather than merely a statistical covariate. Given that environmental determinants of physical activity were not directly measured, district functions as a composite proxy for these unmeasured contextual factors. To examine whether district moderated the association between health literacy and physical activity, interaction terms (HLS-SF×District) were added to the binary logistic regression model in an exploratory analysis.

Factors associated with achieving at least a moderate level of physical activity (categorized as ‘low’ [reference] vs. ‘moderate/high’) were identified using binary logistic regression (Enter method). The multivariable model included variables significant in univariate analyses (p < 0.05) and age as a potential confounder. The final model included health literacy total score (HLS-SF), gender, age, district, education level, BMI, smoking status, alcohol consumption, prior physical activity counseling, and family history of obesity. Additionally, the month of data collection (May–August) was found to have no significant confounding effect on physical activity levels in sensitivity analyses (p > 0.05). Odds ratios (OR) and 95% confidence intervals (CI) were calculated, and the level of statistical significance was set at p < 0.05.

Ethical approval

The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008. Ethics committee approval was obtained from Istanbul Medipol University (Date: 18 April 2024, Decision number: 405). Informed consent to participate was obtained from all of the participants.

Results

A total of 1,047 participants from three districts were included in the study. Of the participants, 55.1% (n = 577) were female, and 38.7% (n = 405) were primary school graduates. The median age was 42.0 years (range: 18.0–65.0). Other sociodemographic characteristics of the participants are presented in Table 1. The median body mass index (BMI) of the participants was 26.7 kg/m² (range: 16.5–50.6).

Table 1.

Sociodemographics of the participants (n = 1,047)

		N	%
District	Eyüpsultan	336	32.1
	Fatih	318	30.4
	Sultanbeyli	393	37.5
Gender	Female	577	55.1
	Male	470	44.9
Marital status	Married	770	73.5
	Single	212	20.2
	Divorced	40	3.8
	Partner dead	25	2.4
Educational level	Illiterate	15	1.4
	Literate	15	1.4
	Primary school	405	38.7
	High school	283	27.0
	University and above	329	31.4
Working status	No	515	49.2
	Yes	532	50.8
Income*	Income < Expenses	383	36.6
	Income ≈ Expenses	513	49.0
	Income > Expenses	151	14.4

*Income was assessed subjectively based on participants’ perception of whether their monthly income covered their monthly expenses

Health-related characteristics and health service utilization of the participants were evaluated (Table 2). The proportion of participants with a family history of obesity was 24.4% (n = 255). The prevalence of chronic disease and regular medication use were 38.6% (n = 404) and 37.1% (n = 388), respectively. The proportions of participants who reported smoking and alcohol consumption were 38.2% (n = 400) and 11.9% (n = 125), respectively. While 28.5% of participants (n = 298) had previously consulted a dietitian, 15.1% (n = 158) reported paying for this service. The proportion of participants who had previously received physical activity counseling was 13.0% (n = 136). In addition, 5.7% of participants (n = 60) had applied to a Healthy Life Center.

Table 2.

Descriptive characteristics related to health, behavior, and service use (n = 1,047)

		n	%
Obesity in family	No	792	75.6
	Yes	255	24.4
Chronic disease	No	643	61.4
	Yes	404	38.6
Medication use	No	659	62.9
	Yes	388	37.1
Smoking status	No	598	57.1
	Yes	400	38.2
	Quit	49	4.7
Alcohol use	No	922	88.1
	Yes	125	11.9
Dietitian visit	No	749	71.5
	Yes	298	28.5
Fee payment for dietitian	No	889	84.9
	Yes	158	15.1
Previous physical activity counseling	No	911	87.0
	Yes	136	13.0
HLC application	No	987	94.3
	Yes	60	5.7

HLC Healthy Life Center

When the total scores of the scales used to assess health literacy, digital healthy diet literacy, and physical activity levels were examined, the median total score of the HLS-SF was 20.83, while the median total score of the DHDLS was 29.17. The median total physical activity score measured by the IPAQ-SF was 693.0 MET-min/week. According to physical activity level classification, 46.6% of participants had a low level of physical activity, 45.3% had a moderate level, and 8.1% had a high level of physical activity (Table 3).

Table 3.

Physical activity levels and questionnaire scores (n = 1,047)

HLS-SF Total score, median (IQR)		20.83 (15.28–26.39)
DHDLS Total score, median (IQR)		29.17 (16.67–37.50)
IPAQ-SF Total score, median (IQR)		693 (330–1386)
Physical activity level, n (%)	Low	488 (46.6)
	Moderate	474 (45.3)
	High	85 (8.1)

HLS-SF Health Literacy Scale-Short Form, DHDLS Digital Healthy Diet Literacy Scale, IPAQ-SF International Physical Activity Questionnaire-Short Form

Descriptive characteristics stratified by district revealed differences. Participants in Sultanbeyli had a median HLS-SF score of 20.83 (IQR: 13.89–30.56) and the highest proportion of low physical activity (58.5%), whereas Eyüpsultan and Fatih showed comparatively higher physical activity levels (63.4% and 57.6% moderate-to-high, respectively), with median HLS-SF scores of 19.44 (IQR: 12.50–25.00) and 20.83 (IQR: 16.67–26.39), respectively. Regarding digital healthy diet literacy, DHDLS scores were broadly similar across Eyüpsultan (median: 29.17, IQR: 17.71–33.33) and Fatih (median: 29.17, IQR: 16.67–37.50), whereas Sultanbeyli showed a lower median DHDLS score (median: 25.00, IQR: 16.67–41.67).

A moderate, positive, and statistically significant correlation was found between the total DHDLS score and the total HLS-SF score (r = 0.488, p < 0.001). In addition, low but statistically significant positive correlations were observed between the total DHDLS score and BMI (r = 0.086, p = 0.005) and age (r = 0.207, p < 0.001). Similarly, the total HLS-SF score was positively and significantly associated with BMI (r = 0.176, p < 0.001) and age (r = 0.276, p < 0.001) (Table 4).

Table 4.

Spearman correlations between DHDLS, HLS-SF, BMI, and age variables (n = 1,047)

		DHDLS Total score	HLS-SF Total score	BMI	Age
DHDLS Total score	r	1.000	0.488	0.086	0.207
	p	.	< 0.001	0.005	< 0.001
HLS-SF Total score	r		1.000	0.176	0.276
	p		.	< 0.001	< 0.001

HLS-SF Health Literacy Scale- Short Form, DHDLS Digital Healthy Diet Literacy Scale, BMI Body Mass Index

Characteristics of participants with low physical activity levels were compared with those exhibiting moderate-to-high activity levels (Table 5). The median HLS-SF score and BMI were significantly higher among participants with low physical activity levels (p < 0.001 and p = 0.005, respectively). When analyzed by district, participants residing in Sultanbeyli had significantly lower physical activity levels compared to those in Fatih and Eyüpsultan (p < 0.001). Furthermore, moderate-to-high physical activity levels were significantly more prevalent among men, individuals with higher educational levels, alcohol consumers, nicotine users, those without a family history of obesity, and participants who had previously received physical activity counseling (p < 0.05).

Table 5.

Varibles associated with physical activity levels (n = 1,047)

Variables		Physical activity level		p-value
		Low (n = 488)	Moderate/High (n = 559)
HLS-SF Total Score (Median, IQR)		22.22 (16.67–29.17)	19.44 (12.5–25.00)	< 0.001 a
DHDLS Total Score (Median, IQR)		29.17 (16.67–37.50)	29.17 (16.67–33.33)	0.418a
Age (Median, IQR)		43 (34–51)	42 (32–51)	0.116a
BMI (Median, IQR)		27.32 (23.93–30.84)	26.30 (23.26–29.39)	0.005 a
District, n (%)	Eyüpsultan	123 (36.6)	213 (63.4)	< 0.001 b
	Fatih	135 (42.5)	183 (57.5)
	Sultanbeyli	230 (58.5)	163 (41.5)
Gender, n (%)	Female	297 (51.5)	280 (48.5)	< 0.001 b
	Male	191 (40.6)	279 (59.4)
Education, n (%)	Below high school	231 (53.1)	204 (46.9)	< 0.001 b
	High school and above	257 (42.0)	355 (58.0)
Income, n (%)	Less than expenses	183 (47.8)	200 (52.2)	0.564b
	Equals expenses/more than expenses	305 (45.9)	359 (54.1)
Marital status, n (%)	Married	372 (48.3)	398 (51.7)	0.066b
	Single/divorced/partner dead	116 (41.9)	161 (58.1)
Chronic Disease, n (%)	No	294 (45.7)	349 (54.3)	0.468b
	Yes	194 (48.0)	210 (52.0)
Alcohol, n (%)	No	443 (48.0)	479 (52.0)	0.011 b
	Yes	45 (36.0)	80 (64.0)
Smoking status, n (%)	No/Quit	329 (50.9)	318 (49.1)	< 0.001 b
	Yes	159 (39.8)	241 (60.3)
Obesity in Family, n (%)	No	354 (44.7)	438 (55.3)	0.029 b
	Yes	134 (52.5)	121 (47.5)
Dietitian Visit, n (%)	No	344 (45.9)	405 (54.1)	0.483b
	Yes	144 (48.3)	154 (51.7)
Previous PA Counseling, n (%)	No	447 (49.1)	464 (50.9)	< 0.001 b
	Yes	41 (30.1)	95 (69.9)
HLC Application, n (%)	No	455 (46.1)	532 (53.9)	0.180b
	Yes	33 (55.0)	27 (45.0)

HLS-SF Health Literacy Scale- Short Form, DHDLS Digital Healthy Diet Literacy Scale, BMI Body Mass Index, PA Physical Activity, HLC Healthy Life Center

^a Mann-Whitney U test, ^b Pearson Chi-Square Test

Binary logistic regression was used to identify factors associated with moderate-to-high physical activity (Table 6). In the multivariable model, an increase in the HLS-SF score remained significantly associated with a decreased likelihood of moderate-to-high physical activity (aOR = 0.970; 95% CI: 0.957–0.984, p < 0.001). Conversely, being male (aOR = 1.599, p = 0.001), residing in Eyüpsultan (aOR = 2.033, p < 0.001) or Fatih (aOR = 2.060, p < 0.001), current smoking (aOR = 1.452, p = 0.008), and prior physical activity counseling (aOR = 1.748, p = 0.008) were significantly associated with higher odds of activity. Although a family history of obesity (p = 0.029), BMI (p = 0.005), and alcohol consumption (p = 0.012) showed significant associations in univariate analyses, their independent effects were not maintained in the multivariable model (p > 0.05). Similarly, age (p = 0.948) and education (p = 0.759) did not show significant independent associations with physical activity levels.

Table 6.

Univariate and multivariable logistic regression analysis of factors associated with moderate-to-high physical activity level (n = 1,047)

Variables	B	Univariate OR (95% CI)	p-value	B	Adjusted OR (95% CI)	p-value
Health Literacy (HLS-SF) Total Score	-0.035	0.966 (0.954–0.978)	< 0.001	-0.030	0.970 (0.957–0.984)	< 0.001
Age	-0.008	0.992 (0.982–1.002)	0.118	0.001	1.000 (0.987–1.012)	0.948
Gender (Ref: Female)	0.438	1.549 (1.211–1.982)	< 0.001	0.470	1.599 (1.221–2.096)	0.001
Education (Ref: Low/Middle)	0.447	1.564 (1.221–2.003)	< 0.001	-0.048	0.953 (0.699–1.299)	0.759
District (Ref: Sultanbeyli)			< 0.001			< 0.001
District(1) (Eyüpsultan)	0.893	2.444 (1.812–3.296)	< 0.001	0.710	2.033 (1.460–2.831)	< 0.001
District(2) (Fatih)	0.649	1.913 (1.418–2.581)	< 0.001	0.723	2.060 (1.471–2.884)	< 0.001
BMI	-0.035	0.966 (0.943–0.990)	0.005	-0.009	0.991 (0.963–1.020)	0.536
Smoking (Ref: No)	0.450	1.568 (1.218–2.019)	< 0.001	0.373	1.452 (1.101–1.915)	0.008
Alcohol Consumption (Ref: No)	0.497	1.644 (1.116–2.422)	0.012	-0.033	0.968 (0.634–1.478)	0.880
Prior PA Counseling (Ref: No)	0.803	2.232 (1.513–3.292)	< 0.001	0.558	1.748 (1.154–2.648)	0.008
Family History of Obesity (Ref: No)	-0.315	0.730 (0.550–0.968)	0.029	-0.269	0.764 (0.557–1.048)	0.095

The reference category for physical activity level is “Low physical activity”. Multivariable model adjusted for all variables listed in the table

OR Odds Ratio, CI Confidence Interval, HLS-SF Health Literacy Scale- Short Form, PA Physical Activity, BMI Body Mass Index, Ref Reference category

Multivariate Model Fit: Omnibus Test χ2 = 98.177, p < 0.001; Nagelkerke R² = 0.120; Hosmer-Lemeshow χ2 = 9.269, p = 0.170

To further examine whether the association between health literacy and physical activity differed across districts, an exploratory interaction analysis was conducted by adding HLS-SF×District terms to the binary logistic regression model. The overall interaction was statistically significant (Wald χ²=9.067, df = 2, p = 0.011), indicating that the relationship between health literacy and physical activity was not uniform across the three districts. In Sultanbeyli (reference category), higher health literacy was associated with a decreased likelihood of moderate-to-high physical activity (B = − 0.043, p < 0.001). In Eyüpsultan, the direction and magnitude of this association were virtually identical to Sultanbeyli (interaction term: B = + 0.001, p = 0.970). In Fatih, however, the inverse association was substantially attenuated (interaction term: B = + 0.044, p = 0.005), resulting in a near-zero net effect of health literacy on physical activity (net B ≈ + 0.001).

Discussion

This community-based study examines the interrelationships between health literacy, digital healthy diet literacy, and physical activity levels. One of the key findings of this study was the presence of a moderate, positive correlation between digital healthy diet literacy (DHDLS) and general health literacy (HLS-SF) (r = 0.488; p < 0.001). This result is consistent with findings from studies conducted in Türkiye and similar middle-income country contexts. Karahan Yılmaz and Eskici (2021) reported a comparable positive correlation between health literacy and digital healthy diet literacy in a Turkish adult sample (r = 0.41, p < 0.001) [19]. Soylar and Demirel (2025) reported a moderate positive association between HL and digital diet health literacy (r = 0.420, p < 0.001) among adults in Türkiye [22]. Similarly, Duong et al., who developed both instruments used in the present study, reported moderate positive correlations between general health literacy and digital healthy diet literacy scores across Vietnamese and Taiwanese community samples, suggesting that this association may generalise across diverse cultural and health-system contexts [13, 18]. Consistent findings have also been reported among adults in South Korea, further supporting the robustness of this relationship [23].

The moderate positive correlation observed between DHDLS and HLS-SF scores (r = 0.488, p < 0.001) suggests that these two constructs share substantial common variance, indicating that individuals with stronger general health literacy tend to also report higher digital dietary literacy. This pattern is consistent with theoretical frameworks conceptualising digital health literacy as a multidimensional extension of foundational health literacy skills into digital environments [24, 25]. However, it should be noted that a moderate correlation does not, in itself, establish that digital healthy diet literacy is not an independent construct; confirmatory factor analyses or structural equation modelling approaches in future research would be needed to more rigorously test the boundaries of construct independence.

The general profile of the participants showed that the health literacy (HLS-SF) level (20.83 points out of 50) reflects limited abilities to access and use health-related information in the community; this finding is consistent with previous studies reporting limited or inadequate health literacy levels in the Turkish population [26]. In contrast, the level of digital healthy diet literacy (DHDLS) in the study group was relatively higher (29.17 points out of 50), and this value was similar to findings reported in another recent study conducted among adults in Türkiye (28.14 points) [22]. This finding suggests that the adult population has a certain level of perceived self-efficacy regarding digital nutrition-related information and that individuals may feel more competent in accessing and understanding nutrition information through digital platforms compared to navigating the general health system.

Correlation analyses demonstrated that as age and BMI increased, both general health literacy (HLS-SF) and digital healthy diet literacy (DHDLS) scores also increased. An increase in health literacy scores with advancing age has also been reported in similar samples [22]. This suggests that as individuals age and health risks increase, they tend to show greater interest and exposure to health and nutrition-related issues. However, a systematic review of health literacy in healthy adults indicated that both younger adults and those aged 65 years and older tend to exhibit lower health literacy levels compared to middle-aged groups, pointing to a life-course pattern rather than a linear age-related increase [27]. Additionally, evidence from a recent scoping review suggests that individuals with higher BMI may demonstrate higher general and nutrition-related health literacy, although this does not necessarily translate into lower body weight due to environmental, behavioral, and structural barriers [28].

Approximately half of the participants had low physical activity levels, and only a small proportion engaged in high levels of physical activity. These findings are consistent with the literature, which reports that most adults do not have regular physical activity habits [22]. Because self-reported physical activity may differ from objectively measured activity [29], the observed prevalence may underestimate the true magnitude of physical inactivity.

Although some studies have reported positive associations between both health literacy and digital health literacy and physical activity [30], international systematic reviews have shown that this relationship can be inconsistent and may vary across populations [9]. In our study, health literacy levels were significantly lower among individuals with moderate-to-high physical activity compared to those with low physical activity levels. Logistic regression analysis further demonstrated that an increase in the HLS-SF score was associated with a decreased likelihood of engaging in moderate or high levels of physical activity (aOR = 0.970, p < 0.001). Although the effect size of a one-unit increase in health literacy score on physical activity likelihood was marginal, this association represents a statistically significant trend in a large sample.

The inverse association between health literacy and physical activity observed in our study parallels what has been described as the “intention–behavior gap” in the health psychology literature, that is, the documented discrepancy between health-related knowledge or intention and actual behavior [31]. While our study did not directly measure behavioral intentions, this theoretical framework offers a plausible interpretive lens for understanding why individuals with higher health literacy did not demonstrate higher physical activity levels. Similarly, frameworks such as the Theory of Planned Behavior suggest that attitudes and perceived behavioral control (constructs not assessed in the present study) mediate the translation of knowledge into action [32]. These theoretical interpretations should therefore be regarded as potential explanatory hypotheses that warrant direct empirical testing in future longitudinal research, rather than conclusions supported by the present data.

When examining the relationship between physical activity levels and sociodemographic characteristics, univariate analyses showed that individuals with higher educational levels had significantly higher physical activity rates. However, when other variables such as district of residence, gender, and health literacy were included in the logistic regression model, the independent effect of educational level on physical activity was no longer significant. This finding suggests that formal education alone is insufficient to produce sustained behavioral change and that physical activity is more strongly influenced by environmental and structural factors than by individual knowledge levels.

Interestingly, factors such as age (p = 0.948), BMI (p = 0.536), and family history of obesity (p = 0.095) did not show independent associations with physical activity levels in the multivariable model. It should be noted, however, that statistical non-significance does not necessarily indicate absence of effect; these associations may reflect collinearity with other model variables or insufficient power to detect modest independent effects after adjustment.

In the univariate analysis, both alcohol consumption (p = 0.011) and smoking status (p < 0.001) were significantly associated with higher physical activity levels. However, in the multivariable model, only smoking remained a significant independent predictor (aOR = 1.452, p = 0.008), while the association with alcohol consumption disappeared. This suggests that the relationship between alcohol use and physical activity may be mediated by other lifestyle factors or demographic characteristics. The finding that smokers exhibited higher odds of moderate-to-high physical activity is consistent with some literature suggesting that individuals with certain health-risk behaviors might engage in compensatory physical activity, though this relationship remains complex. The positive association between smoking and physical activity should not be interpreted as a protective effect of smoking. Rather, this finding may reflect compensatory health behaviors or underlying lifestyle patterns, particularly among men, where risk-taking behaviors and physical activity coexist. Further studies are needed to clarify the direction and mechanisms of this relationship.

When sociodemographic and environmental determinants were evaluated together, the district of residence emerged as a key determinant of physical activity. Participants living in Sultanbeyli had significantly lower physical activity levels compared to those residing in Fatih and Eyüpsultan, supporting the notion that physical activity is shaped not only by individual preferences but also by urban resources and structural conditions. Drawing on the built environment literature, this pattern may reflect socio-ecological factors such as walkability, access to green spaces, and perceived safety, which are known to shape physical activity independent of individual motivation [33]. The regional disparities observed in our study are consistent with contemporary literature emphasizing that physical activity is shaped by the dynamic interaction between social and physical environments rather than solely by individual health literacy levels [33]. However, as these environmental determinants were not directly measured in the present study, such interpretations should be regarded as plausible hypotheses rather than conclusions directly supported by the present data.

Notably, an exploratory interaction analysis revealed that the inverse association between health literacy and physical activity was not consistent across the three districts (Wald χ²=9.067, df = 2, p = 0.011). While the negative relationship between health literacy and physical activity was evident in both Sultanbeyli and Eyüpsultan, it was markedly attenuated in Fatih, where health literacy showed virtually no independent association with physical activity levels. This pattern may reflect the moderating role of district-level contextual factors that shape how individual health knowledge translates into physical behavior. In districts with more favorable environmental conditions, physical activity may be less dependent on individual cognitive capacity and more accessible regardless of health literacy level, thereby diminishing the observable association between the two. Conversely, in districts with fewer structural resources, individuals with higher health literacy may be more sedentary due to occupational patterns, caregiving responsibilities, or other unmeasured lifestyle factors that co-vary with literacy in lower-resource settings. It should be noted, however, that this interaction analysis was exploratory in nature and was not part of the original study design. The mechanisms underlying this district-level moderation cannot be determined from the present cross-sectional data, and these findings should be interpreted with caution pending replication in studies that directly measure environmental determinants of physical activity.

Additionally, individual-level education lost statistical significance in the multivariable model, whereas district affiliation remained a strong independent predictor of physical activity. This pattern suggests that area-level socioeconomic conditions may exert influences on physical activity that are not fully captured by individual socioeconomic indicators alone, consistent with neighbourhood health research demonstrating independent contextual effects beyond individual characteristics [34].

The finding that men were significantly more physically active than women (aOR = 1.599) should be interpreted within the context of gender roles and the limitations of measurement tools. In Türkiye, women’s extensive unpaid domestic labor (housework, childcare, etc.) may not be fully captured within the “leisure-time physical activity” domain of standardized instruments such as the IPAQ-SF, potentially leading to an underestimation of women’s actual activity levels. Consistent with our findings, male gender has been widely reported to be more strongly associated with physical activity in the literature [35].

While family history of obesity appeared to be a significant barrier in univariate analysis, its independent effect was not sustained in the multivariable model (p = 0.095), suggesting that other environmental or behavioral factors in the model may exert a stronger influence.

Furthermore, participants who had previously received physical activity counseling were approximately 1.75 times (aOR = 1.748) more physically active, highlighting the importance of structured support. In this context, expanding accessible services such as Healthy Life Centers (HLCs) and integrating evidence-based mobile health interventions into physical activity promotion strategies represent promising directions that warrant evaluation in future controlled studies. Mobile health applications offer significant potential as low-cost and accessible tools that support professional counseling processes and promote physical activity [36].

From a public health and policy perspective, these findings suggest that approaches focusing solely on increasing knowledge levels may have limited effectiveness. Policies aimed at increasing physical activity should move beyond individual information-based strategies and adopt comprehensive approaches that incorporate environmental modifications, urban planning, and structured counseling services within primary healthcare. Addressing the individual, social, and structural barriers that hinder the translation of health literacy into behavior is critical for the development of effective and sustainable public health interventions.

Strengths and limitations

This study has several strengths and limitations. Key strengths include the use of a large, community-based sample (n = 1,047) and a comprehensive approach examining health literacy, digital healthy diet literacy, and physical activity simultaneously across three districts with distinct socioeconomic profiles. Furthermore, the use of multivariable logistic regression to control for established confounders (including age, BMI, and lifestyle habits) strengthens the validity of the observed associations and enhances the analytical depth of the study.

However, several limitations should be acknowledged. First, the cross-sectional design precludes drawing causal inferences. Second, physical activity, height, weight, and health behaviors were assessed using self-reported data. Specifically, while the IPAQ-SF is a widely accepted and standardized instrument, it is subject to recall and social desirability bias; participants may have overestimated their activity levels or struggled to accurately recall duration and intensity. Third, environmental determinants of physical activity, such as walkability and access to green spaces, were not measured. Fourth, data collection was limited to the summer months; although internal sensitivity analyses showed no significant monthly variation, the results may not fully capture seasonal fluctuations in activity. Finally, the use of telephone interviews, while effective for reaching a broad sample, may have influenced the quality of responses among individuals with lower digital or general literacy.

Additionally, while an exploratory interaction analysis suggested that the association between health literacy and physical activity may vary by district, the absence of directly measured environmental variables (e.g., walkability or safety) precluded a mechanistic interpretation of this finding. Consequently, these observed district-level differences should be interpreted as context-specific findings, and further research incorporating objective spatial data is needed to enhance generalisability to other urban settings.

Conclusion

This community-based study examines the interrelationships between health literacy, digital healthy diet literacy, and physical activity levels. The moderate positive association between digital healthy diet literacy and general health literacy indicates that digital nutrition-related knowledge is closely linked to overall health literacy capacity. In contrast, the finding that higher health literacy levels were associated with a lower likelihood of engaging in moderate-to-high physical activity suggests that knowledge does not always translate into healthy behavior and highlights the complex determinants of physical activity.

Moreover, the strong associations between physical activity levels and environmental context, gender, and professional counseling underscore the importance of environmental and structural factors beyond individual knowledge. These findings suggest that public health interventions should not be limited to increasing health literacy alone but should be supported by environmental modifications, structured counseling services in primary care, and multicomponent strategies that facilitate behavior change. Holistic approaches that aim to translate health literacy from knowledge into action hold significant potential for reducing physical inactivity and the burden of related chronic diseases.

Authors’ contributions

SÇ, ZMA, MAS, ŞH, and MNA conceptualized the study and designed the research framework. Data collection was conducted by MAS, SÇ, and MNA. Data analyses were performed by ZMA and SÇ. SÇ and ZMA led the drafting of the manuscript, with contributions from MAS, ŞH, and MNA. All authors critically reviewed and approved the final manuscript for publication.

Funding

This research received no specific grant from any funding agency, commercial entity, or not-for-profit organization.

Data availability

The dataset used and analyzed during the current study is available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008. Ethics committee approval was obtained from Istanbul Medipol University (Date: 18 April 2024, Decision number: 405). Informed consent to participate was obtained from all of the participants.

Competing interests

The authors declare no competing interests.