Cardiovascular disease prevention by personalized health promotion considering educational attainment

Abstract

To examine whether educational attainment modifies the effectiveness of personalized lifestyle recommendations for improving cardiovascular risk factors. This post-hoc analysis used data from a population-based randomized controlled trial in Girona, Spain, including adults aged 35–74 years without cardiovascular disease at baseline. Participants (n = 759; 48.7% men) were randomized to an intervention group (n = 380), which received personalized recommendations on diet and physical activity, or a control group (n = 379), which received a standard report of baseline results. Changes in systolic and diastolic blood pressure, LDL cholesterol, and physical activity energy expenditure from baseline to 12 months were analyzed. Multivariable linear regression models adjusted for age included an interaction term defined as group × educational attainment × time (1 year) to assess effect modification. Analyses were stratified by sex. Among women in the intervention group, significant interactions by educational attainment were observed. Compared with women with lower educational attainment, those with higher attainment showed more favorable changes in diastolic blood pressure [beta-coefficient (95% confidence interval): − 1.98 (− 4.23; 0.27) vs. 1.63 (− 0.21; 3.48)], LDL cholesterol [− 4.61 (− 11.40; 2.18) vs. 5.71 (0.25; 11.17)], and physical activity energy expenditure [0.52 (0.04; 1.23) vs. − 0.01 (− 0.36; 0.26)]. No significant interactions were found among men in the intervention group or among participants in the control group. Overall, within-group changes in these outcomes did not reach statistical significance. Personalized lifestyle interventions may provide greater benefits for women with higher educational attainment, although overall improvements were modest.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-36654-4.

Introduction

Cardiovascular diseases (CVDs) remain the leading cause of death worldwide, accounting for an estimated 17.9 million deaths annually¹. Evidence shows that adopting a healthy lifestyle can substantially reduce premature mortality from these conditions and extend life expectancy by up to 14 years for women and 12 years for men². Educational attainment plays a critical role in promoting health by enhancing self-awareness and improving access to healthcare services³. Cardiovascular risk factors—such as hypertension, elevated body mass index, and diabetes—partially mediate the relationship between education and CVD incidence^4,5. Despite widespread public health messages encouraging healthy behaviors, persistent barriers prevent certain population groups from fully benefiting from advances in information and communication technologies⁶. Limited educational attainment, in particular, acts as an invisible barrier to prevention and healthcare delivery, generating significant costs at both individual and population levels^7,9.

Several strategies have been proposed to implement personalized prevention effectively, including targeted educational programs, community awareness initiatives, population empowerment, and robust information systems¹⁰. An innovative self-screening method for cardiovascular risk—validated in a randomized crossover clinical trial—addresses these strategies by providing personalized recommendations on diet and physical activity to manage risk factors, showing promising results^11,13. Although scientific evidence consistently links higher educational attainment to better health outcomes¹⁴, its influence on the effectiveness of personalized health recommendations remains unclear and might vary by gender^15,17. We hypothesize that educational attainment modifies the impact of personalized lifestyle and cardiovascular risk management advice delivered through a validated self-screening system. Specifically, individuals with higher education levels are expected to achieve greater improvements in the control of cardiovascular risk factors compared to those with lower education. Furthermore, gender may influence how educational background affects individuals’ ability to benefit from such interventions, potentially leading to different outcomes in men and women. The objective of this study was to assess whether educational attainment influences the effectiveness of personalized recommendations for improving the management of cardiovascular risk factors, using sex-stratified analyses to explore potential gender differences.

Methods

Study design and participants

This study was a randomized, parallel-arm controlled trial with a 12-month follow-up. The detailed methodology has been described elsewhere^11,12. Participants were eligible if they were aged 35 to 74 years, resided in Girona or surrounding areas in northeastern Spain, and had no history of CVD at baseline. Eligible individuals were randomly selected from the population and randomly assigned to either the intervention or control group. Invitations were sent by postal mail, followed by a phone call from a field worker to confirm participation and schedule an appointment.

Participants were allocated to the intervention or control group using a computer-generated random sequence integrated into custom-designed data collection software. This sequence was concealed from the field workers responsible for enrollment. The intervention group received personalized preventive recommendations tailored to each individual’s cardiovascular risk profile, based on the latest scientific evidence. In contrast, the control group received only a standard letter reporting baseline examinations results, mailed within one month of data collection.

The primary analysis of the randomized controlled trial on which this study is based demonstrated intervention efficacy among men, leading to reductions in systolic and diastolic blood pressure and improved control of total and LDL cholesterol levels. In contrast, women did not experience significant changes in any of the assessed cardiovascular risk factors¹². A non-stratified post-hoc analysis indicated promising effects on lifestyle behaviors—particularly improved adherence to the Mediterranean diet and smoking cessation—highlighting the potential of this personalized, evidence-based intervention as an innovative tool for cardiovascular prevention¹³.

All participants were fully informed about the study and provided written informed consent prior to enrollment. All methods were performed in accordance with the relevant guidelines and regulations. The clinical trial protocol was registered at ClinicalTrials.gov (#NCT02373319) and approved by the Institutional Research Board [Clinical Research Ethics Committee of Parc de Salut Mar (CEIC-PSMAR, #2014/5815/I)].

Variables collected

Examinations were conducted by a team of trained nurses using standardized questionnaires and measurement methods at both baseline and 12-month follow-up. Data on sex, age, and socio-demographic information were reported via self-administered standardized questionnaires. Participants were asked to indicate their educational attainment, categorized as follows: (1) lower educational attainment (no formal education, primary, or secondary education) and (2) higher educational attainment (university education).

A precision scale with easy calibration was used to measure weight, with participants wearing only their underwear. Height was measured with a standard measuring rod, with participants standing barefoot. Body mass index was calculated as weight divided by height squared (kg/m²). Blood pressure was measured using an automatic monitor with a cuff appropriately sized (young, adult, obese) for each participant’s upper arm circumference). After a 5-min rest, two measurements were taken at least 2 min apart, and the lower value was recorded for the study.

Adherence to the Mediterranean diet was assessed using the 14-item Mediterranean Diet Adherence Score (MEDAS) questionnaire, validated for the Spanish population¹⁸. The questionnaire comprises 12 questions on the frequency of food consumption and 2 questions on food intake habits characteristic of the Spanish Mediterranean diet. Each question was scored 0 or 1, with the final score ranging from 0 to 14. To measure energy expenditure from leisure-time physical activity (EEPA), the REGICOR questionnaire includes six two-part questions that gather information on the four dimensions of physical activity: type, frequency, duration, and intensity. EEPA was estimated by multiplying the intensity assigned to each activity by the number of days it was performed in a month and the average number of minutes per day. The metabolic equivalent of task values for the six activities were: walking (4), brisk walking (5), gardening (5), walking trails (6), climbing stairs (8), and any sport activity (10)¹⁹.

To determine lipid profile and glycaemia, blood samples were drawn within 60 s after a 10–14 h fast. Serum aliquots were stored at −80 °C. Total cholesterol and high-density lipoprotein (HDL) cholesterol concentrations were measured using enzymatic and direct methods, respectively (ABXHoriba, Montpellier, France). When triglycerides were < 300 mg/dL, low-density lipoprotein (LDL) cholesterol was calculated using the Friedewald formula. All analyses were conducted in a central laboratory.

Intervention with personalized preventive recommendations

The personalization process began by creating an individual profile for each participant based on collected data, including: (a) prevalence of cardiovascular risk factors, (b) tobacco use, and (c) physical activity level. Recommendations for the intervention group were based on the most up-to-date scientific evidence^20,22. Participants received a packet containing the results of baseline examinations (blood pressure, lipid profile, diabetes status, and smoking) along with their estimated cardiovascular risk, calculated using the Framingham–REGICOR risk function validated for the Spanish population²³. A trained nurse provided a detailed explanation of personalized recommendations during a 30-min consultation, focusing on strategies to improve cardiovascular health through weight control, smoking cessation, adherence to the Mediterranean diet, and increased physical activity tailored to individual performance levels (sedentary, moderate, or vigorous). Supplementary Materials include the personalized recommendations based on a Mediterranean dietary pattern and physical activity provided to each participant in the intervention group according to their individual profile.

Outcomes

At the end of the 12-month follow-up, a re-examination was conducted to measure various biomarkers. Outcomes were defined as the change in values between the baseline examination and the 12-month assessment. These outcomes consisted of continuous changes in the levels of various cardiovascular risk factors: (1) systolic blood pressure, (2) diastolic blood pressure, (3) LDL cholesterol, (4) HDL cholesterol, (5) glycaemia, (6) body mass index, (7) adherence to the Mediterranean diet, and (8) EEPA.

Statistical analysis

Data were analyzed using a per-protocol approach including participants who completed the 12-month follow-up (80% of the total sample). Normality of continuous variables was visually assessed using Q–Q plots. All results were stratified by sex and educational attainment. Baseline and 12-month values for systolic and diastolic blood pressure, LDL and HDL cholesterol, body mass index, glycemia and adherence to the Mediterranean diet were summarized using means, while EEPA was summarized using medians. These values were plotted by educational attainment (low vs. high) and intervention/control groups, separately for women and men. The percentage change at 12 months was calculated by subtracting the baseline biomarker value from the follow-up value, dividing the result by the baseline value, and multiplying by 100.

To assess effect modification, we tested whether educational attainment influenced the intervention effect on changes in cardiovascular risk factors from baseline to 12 months. Multivariable linear regression models adjusted for age included an interaction term defined as group × educational attainment × time (1 year), which was the main analysis. Analyses were stratified by sex. Interaction significance was evaluated using the − 2 log-likelihood test comparing nested models with and without the interaction term. For outcomes showing significant interactions, adjusted beta coefficients and 95% confidence intervals were estimated to illustrate the magnitude of change within the intervention group by educational attainment. EEPA values were log-transformed prior to inclusion in the models due to non-normal distribution. Model assumptions—linearity, homoscedasticity, normality of residuals, and absence of multicollinearity—were verified using standard diagnostic tools (residuals vs. fitted plots, Q–Q plots, and variance inflation factors).

All statistical analyses were conducted using the R Statistical Package (R Foundation for Statistical Computing, Vienna, Austria; V.4.2.2)²⁴.

Results

The study included 759 individuals (48.7% men), having high educational attainment 140 (36.0%) women and 134 (36.2%) men.

Tables 1 and 2 present the baseline characteristics of women and men, respectively, stratified by educational attainment. Participants with lower educational attainment exhibited a worse cardiovascular profile. Supplementary Table 1 shows the distribution of cardiovascular risk factors across the full sample.

Table 1.

Characteristics at baseline and 12-month change in women by educational attainment.

Low educational attainment	Control (N = 125)	12- month percent change	Intervention (N = 124)	12- month percent change
Age (years)*	53 (10)	–	55 (11)	–
Systolic blood pressure (mmHg)*	107 (16)	1.0 [− 6.9; 8.3]	110 (19)	0.9 [− 6.8; 8.4]
Diastolic blood pressure (mmHg)*	72 (10)	− 1.4 [− 8.7; 5.5]	74 (12)	0.0 [− 6.1; 5.7]
LDL cholesterol (mg/dL)*	127 (31)	0.0 [− 8.4; 15.4]	142 (33)	2.4 [− 4.0; 11.7]
HDL cholesterol (mg/dL)*	59 (13)	− 1.7 [− 8.2; 7.4]	60 (13)	− 0.9 [− 8.9; 7.5]
Glycaemia (mg/dL)*	92 (16)	1.2 [− 3.3; 6.3]	93 (13)	2.2 [− 3.8; 7.1]
Body mass index (kg/m2)*	26.1 (5.0)	0.0 [− 0.5; 0.5]	27.0 (4.7)	0.1 [− 0.6; 0.8]
Adherence to Mediterranean diet*	7.1 (2.0)	0.0 [− 12.5; 22.2]	7.1 (2.0)	0.0 [− 11.1; 20.0]
EEPA (kcal/day)†	1790 [958; 2769]	− 12.1 [− 47.9; 79.9]	1856 [1096; 2947]	− 16.6 [− 52.5; 32.8]
High educational attainment	Control (N = 71)	12−  month percent change	Intervention (N = 69)	12- month percent change
Age (years)*	46 (8)	–	47 (9)	–
Systolic blood pressure (mmHg)*	102 (13)	0.8 [− 6.0; 7.7]	101 (11.8)	− 0.9 [− 6.4; 6.9]
Diastolic blood pressure (mmHg)*	69 (10)	1.4 [− 5.4; 7.8]	69 (9)	− 1.8 [− 8.4; 5.0]
LDL cholesterol (mg/dL)*	127 (27)	6.2 [− 7.8; 15.3]	125 (30)	3.2 [− 2.7; 10.1]
HDL cholesterol (mg/dL)*	60 (13)	4.0 [− 6.4; 8.9]	61 (13)	− 4.0 [− 8.2; 6.9]
Glycaemia (mg/dL)*	88 (9)	− 1.1 [− 3.8; 5.8]	88 (9)	1.3 [− 5.1; 5.8]
Body mass index (kg/m2)*	25.0 (4.4)	0.0 [− 0.5; 0.5]	24.4 (4.0)	0.0 [− 0.6; 0.5]
Adherence to Mediterranean diet*	7.2 (1.9)	14.3 [− 4.5; 28.6]	7.5 (1.7)	0.0 [− 12.5; 16.7]
EEPA (kcal/day)†	1497 [946; 2352]	1.3 [− 34.0; 66.1]	1911 [751; 2532]	24.9 [− 13.1; 104.6]

EEPA, Energy expenditure in physical activity. HDL, High-density lipoprotein. IQR, interquartile range. LDL, Low-density lipoprotein. SD, standard deviation.

*Mean (standard deviation). ^†Median [Interquartile range].

Table 2.

Characteristics at baseline and 12-month change in men by educational attainment.

Low educational attainment	Control (N = 123)	12- month percent change	Intervention (N = 113)	12- month percent change
Age (years)*	50 (10)	–	51 (10)	–
Systolic blood pressure (mmHg)*	118 (14)	− 1.7 [− 6.0; 8.2]	120 (15)	− 0.9 [− 6.7; 5.2]
Diastolic blood pressure (mmHg)*	76 (9)	1.2 [− 6.6; 7.1]	79 (9)	− 1.9 [− 8.8; 6.9]
LDL cholesterol (mg/dL)*	140 (34)	1.3 [− 5.6; 11.0]	142 (38)	− 0.9 [− 9.5; 13.3]
HDL cholesterol (mg/dL)*	50 (10)	− 0.2 [− 6.7; 6.2]	50 (11)	− 1.1 [− 6.8; 7.4]
Glycaemia (mg/dL)*	95 (12)	0.5 [− 4.5; 6.8]	100 (25)	2.1 [− 3.5; 7.0]
Body mass index (kg/m2)*	26.7 (3.8)	0.1 [− 0.4; 0.7]	27.5 (3.8)	0.1 [− 0.5; 0.8]
Adherence to Mediterranean diet*	6.7 (1.9)	0.0 [− 12.5; 25.0]	6.8 (2.2)	0.0 [− 14.3; 20.0]
EEPA (kcal/day)†	2308 [1087; 4192]	− 15.8 [− 46.3; 49.8]	2543 [979; 3916]	− 10.3 [− 50.0; 43.3]
High educational attainment	Control (N = 63)	12- month percent change	Intervention (N = 71)	12- month percent change
Age (years)*	50 (10)	–	49 (11)	–
Systolic blood pressure (mmHg)*	117 (15)	0.0 [− 5.1; 5.2]	117 (17)	− 1.5 [− 8.8; 4.1]
Diastolic blood pressure (mmHg)*	77 (10)	1.3 [− 5.1; 8.3]	77 (10)	− 1.4 [− 9.4; 6.1]
LDL cholesterol (mg/dL)*	134 (35)	3.5 [− 8.5; 14.4]	128 (27)	0.0 [− 6.6; 13.0]
HDL cholesterol (mg/dL)*	50 (12)	1.3 [− 7.7; 11.1]	50 (9)	0.2 [− 6.6; 6.6]
Glycaemia (mg/dL)*	98 (26)	1.1 [− 4.5; 5.7]	98 (34)	1.5 [− 4.5; 4.9]
Body mass index (kg/m2)*	26.8 (3.4)	0.1 [− 0.4; 0.7]	26.8 (4.7)	− 0.1 [− 0.4; 0.5]
Adherence to Mediterranean diet*	7.1 (2.0)	10.0 [− 5.0; 22.5]	7.5 (2.2)	0.0 [− 10.6; 26.8]
EEPA (kcal/day)†	2014 [1070; 3360]	− 15.1 [− 42.3; 27.0]	2023 [1136; 3399]	10.7 [− 35.2; 79.3]

EEPA, Energy expenditure in physical activity. HDL, High-density lipoprotein. IQR, interquartile range. LDL, Low-density lipoprotein. SD, standard deviation.

*Mean (standard deviation). ^†Median [Interquartile range].

Educational attainment did not significantly interact with the 12-month modification of any cardiovascular risk factors in men. However, in women, it interacted with the association between the intervention and two biomarkers: diastolic blood pressure (p-value for interaction = 0.019) and LDL cholesterol (p-value for interaction = 0.029). Additionally, EEPA showed a marginally significant interaction (p-value for interaction = 0.057) in women (Figs. 1 and 2). Supplementary Table 2 presents the p-values for all interaction terms tested.

Fig. 1.

Baseline and 12-month values for systolic and diastolic blood pressure, LDL and HDL cholesterol, body mass index, glycaemia and adherence to the Mediterranean diet (summarized as means), as well as energy expenditure in physical activity (summarized as medians), stratified by educational attainment and intervention/control groups among women.

Fig. 2.

Baseline and 12-month values for systolic and diastolic blood pressure, LDL and HDL cholesterol, body mass index, glycaemia and adherence to the Mediterranean diet (summarized as means), as well as energy expenditure in physical activity (summarized as medians), stratified by educational attainment and intervention/control groups among men.

A stratified analysis by educational attainment was conducted to explore these significant interactions. The beta coefficients (95% confidence intervals [CI]) for diastolic blood pressure were − 1.98 (− 4.23; 0.27) for women with higher educational attainment and 1.63 (− 0.21; 3.48) for those with lower attainment. For LDL cholesterol, the beta coefficients were − 4.61 (− 11.40; 2.18) for women with higher educational attainment and 5.71 (0.25; 11.17) for those with lower attainment. In addition, log-linear regression models stratified by educational attainment were fitted to account for the non-normal distribution of EEPA in women. The beta coefficients (95% CI) were 0.52 (0.04; 1.23) for women with higher educational attainment and − 0.01 (− 0.36; 0.26) for those with lower attainment (Fig. 3).

Fig. 3.

Adjusted beta coefficients for absolute changes in diastolic blood pressure and LDL cholesterol, and log-linear beta coefficients for relative changes in energy expenditure in physical activity, among women by educational attainment, comparing intervention and control groups.

Discussion

Higher educational attainment was associated with better control of specific cardiovascular risk factors only in women. No statistically significant differences were observed in men. This population-based randomized controlled trial evaluated the efficacy of a personalized health promotion intervention and revealed significant interactions with educational attainment. After 12 months of follow-up, the influence of educational level was particularly evident in the control of diastolic blood pressure, LDL cholesterol, and increased EEPA among women. In contrast, the remaining risk factors showed changes of similar magnitude across educational levels, consistent with findings from previous studies^12,13. These results highlight the importance of considering socio-demographic variables, such as educational attainment, alongside cardiovascular risk profiles when designing personalized health promotion strategies^25,27.

Educational attainment is a modifiable risk factor

This population-based randomized controlled trial demonstrated that individuals with lower educational attainment exhibited a worse cardiovascular profile at baseline, consistent with findings from previous studies^28,29. In our sample, cardiovascular risk profiles significantly worsened with age. Additionally, a higher proportion of women had lower educational attainment compared to men. Consequently, all statistical analyses were stratified by sex, and multivariable models were adjusted for age—both variables known to contribute substantially to health inequalities^30,31. Clouston et al. have highlighted additional determinants such as adolescent cognition, non-cognitive skills, and the rate of cognitive decline as predictors of health literacy in later life, which are closely associated with sex and age³². These interrelated factors reinforce the notion that educational attainment is a modifiable determinant of health inequalities^33,34.

From a clinical perspective, these findings underscore the need to tailor cardiovascular prevention strategies to socio-educational contexts, particularly among women⁷. The observed associations between higher educational attainment and improved control of diastolic blood pressure, LDL cholesterol, and physical activity suggest that educational level may influence individuals’ ability to engage with and benefit from personalized health interventions. Although the intervention was uniformly delivered, its effectiveness varied by educational level, highlighting potential disparities in health literacy, self-management capacity, and access to supportive resources³⁵. Recognizing and addressing these differences is essential to ensure equitable health outcomes and to optimize the impact of preventive strategies in diverse populations³⁶.

Dahlgren and Whitehead’s model of health determinants illustrates the interplay between individual lifestyles—shaped by social norms and networks—and broader socioeconomic and cultural conditions³⁷. Our findings partially reflect this model: after 12 months, disparities in cardiovascular risk profiles by educational attainment widened (e.g., prevalence of optimal blood pressure, systolic blood pressure). This trend may be explained by the sample’s average educational level and the intervention design, which offered personalized recommendations based on cardiovascular risk but did not account for educational differences. Education is closely linked to occupation and income, and multiple pathways connect it to health outcomes—primarily through health literacy, the ability to access, understand, and apply health information³⁸. Since health literacy is a modifiable risk factor for CVD, improving educational attainment could support both primary and secondary prevention^7,9,39. However, this strategy requires structural and policy-level efforts to reduce socioeconomic inequalities and ensure equitable access to education^3,40. Personalized health promotion programs should, therefore, integrate socioeconomic factors and literacy-sensitive communication strategies, rather than focusing solely on individual risk factors. As Geoffrey Rose compellingly argued, effective prevention of CVD requires not only targeting individuals at high risk but also addressing the broader social, structural, and environmental determinants that shape population-level risk distributions. His concept of the “cause of causes” underscores the importance of shifting the entire risk curve through upstream interventions—such as policies that reduce social inequalities, improve living and working conditions, and promote healthier environments. Although this population-based strategy may produce less visible short-term effects, it has the potential for a much greater overall impact. It also places responsibility on policymakers to create conditions that support cardiovascular health for all⁴¹.

The effect of educational attainment in women

Evidence-based health promotion, which considers individual lifestyles, social norms, networks and living and working conditions, is essential for maintaining and improving cardiovascular health. This holistic approach underscores the urgent need to implement effective healthcare practices that empower individuals to take greater control over their health and enhance overall well-being⁴². Importantly, this strategy must also address how gender shapes health outcomes, as women and men often encounter different barriers and opportunities influenced by education, social roles, and access to healthcare services^22,23. Womenface unique challenges in health literacy and empowerment, often rooted in gender-specific social and economic dynamics²⁴. Therefore, one of the central challenges in cardiovascular prevention is to promote health literacy and empower individuals—especially women—while minimizing the risk of exacerbating existing health inequalities^25,26. Addressing these disparities requires that interventions designed to improve cardiovascular health explicitly incorporate gender-sensitive approaches, ensuring that differences between men and women are carefully considered in both the design and implementation of these programs^29,43,44.

Limitations

This analysis included individuals who completed the 12-month follow-up of a randomized controlled trial (81%). Although the study sample was randomly selected from the population, participants with higher educational attainment were more likely to complete the follow-up. Consequently, the distribution of educational attainment was slightly altered: the percentage of individuals with university education increased from 35.1% to 36.1%, which may have introduced selection bias. Baseline differences in cardiovascular risk factors by education level were also observed, with individuals with lower educational attainment exhibited a worse cardiovascular risk profile. While changes in biomarkers were considered to mitigate this limitation, residual confounding cannot be ruled out.

In addition, information on medication use during follow-up was not collected. Although this fact introduces some uncertainty, the limitation likely biases toward the null hypothesis, as women with lower educational attainment had higher baseline LDL cholesterol and diastolic blood pressure values compared to those with higher educational attainment (142 mg/dL vs. 125 mg/dL and 74 mmHg vs. 69 mmHg, respectively), making them more likely to initiate treatment. Finally, the sample size did not allow for stratified analyses by level of higher education (e.g., bachelor’s, master’s, doctoral degrees) or by age. Future studies with larger samples and longer follow-up periods will be necessary to explore cohort and period effects in greater detail.

Conclusion

Personalized interventions tailored to control cardiovascular risk factors may offer greater benefits for women with higher educational attainment, particularly in improving diastolic blood pressure, LDL cholesterol levels, and EEPA. However, as within-group improvements in these outcomes did not reach statistical significance, the findings should be interpreted with caution. These results underscore the importance of incorporating literacy-sensitive communication strategies and addressing socioeconomic disparities when designing effective health promotion programs.

Author contributions

M.D.Z., D.A-J., M.G conceptualized and designed the study, interpretation of data and drafted the initial manuscript. C.P. and C.V. performed the statistical analyses and critically revised the final manuscript. A.D., D.T., N.S., C.P-F., A.T-R. and A.R. contributed to the final analyses and interpretation of data and critically revised the final manuscript.

Funding

This study was financed by Spain’s Ministry of Science through the Carlos III Health Institute FEDER (CM12/03287, PI17/00250, CPII17/00012, and FIS14/00449).

Data availability

Raw data is provided in https://zenodo.org/records/14565573

Declarations

Competing interests

The authors declare no competing interests.