Plant-based dietary patterns are increasingly accepted for ethical concerns and environmental sustainability(Reference Melina, Craig and Levin1), and when appropriately planned, can provide adequate nutrition across all stages of life(Reference Wickramasinghe and Breda2). While some vegetarian diets may include some animal-derived products such as dairy, eggs and honey, both vegetarian and vegan diets exclude the intake of animal products such as meat, poultry and fish. Vegan diets further exclude all animal-based foods and by-products(Reference Hargreaves, Rosenfeld and Moreira3). The prevalence of plant-based dietary patterns has been scientifically studied using health survey data. In Canada, a prevalence of 2·8 % was reported(Reference Valdes, Conklin and Veenstra4), and in the USA, a prevalence of 4·0 %(Reference Cramer, Kessler and Sundberg5). In the New Zealand health survey, a 2·04 % prevalence of vegetarians and a 0·74 % of vegans was reported(Reference Greenwell, Grant and Young6), and in Europe, studies reported prevalences <1 %(Reference Leahy, Lyons and Tol7,Reference Ponzio, Mazzarini and Gasperi8) . However, studies using representative population surveys have not applied survey-weighting factors, which allows accurate prevalence estimation to correct for sample bias, adjust for imbalances and allow for generalisable results in underrepresented groups, such as plant-based individuals(Reference Si, Lee and Heeringa9).
Individuals with plant-based diets have been reported to have healthier habits like consuming less alcohol or not smoking(Reference Acevedo Cantero, Ortega Santos and López-Ejeda10). Compared with omnivores, vegetarians and vegans often exhibit lower BMI, total cholesterol, LDL, TAG and blood glucose levels(Reference Crowe, Appleby and Travis11,Reference Orlich and Fraser12) . These reductions in risk factors are linked to a decreased likelihood of developing chronic diseases, including CVD, diabetes, obesity, chronic kidney failure and various cancers(Reference Huang, Huang and Hu13). Predominantly, plant-based diets improve metabolic control in individuals with diabetes(Reference Tonstad, Stewart and Oda14) and metabolic syndrome, reducing the risk of conditions such as ischemic heart disease(Reference Pettersen, Anousheh and Fan15), hypertension(Reference Tong, Appleby and Armstrong16), chronic kidney disease(Reference Wu, Tsai and Liu17), obesity(Reference Spencer, Appleby and Davey18), fatty liver disease(Reference Lv, Rong and Deng19) and cancer onset(Reference Tantamango-Bartley, Jaceldo-Siegl and Fan20,Reference Watling, Schmidt and Dunneram21) . However, plant-based diets can be associated with nutrient deficiencies if not adequately performed. A strict plant-based diet is associated with increased susceptibility to experiencing bone fractures, osteoporosis and haemorrhagic stroke(Reference Tong, Appleby and Armstrong16). Despite evidence suggesting health advantages associated with plant-based adherence, comprehensive studies remain lacking in integrating individuals’ health status and lifestyle factors perspective (e.g. physical activity, alcohol drinking habit or tobacco consumption habit). Moreover, few studies have compared these individuals to nonvegetarians with similar social determinants of health (e.g. educational level, household income or marital status).
Comparing individuals with an omnivore diet and a plant-based dietary pattern has explored the physiological outcomes associated with plant-based diets; however, these studies often failed to account for lifestyle habits and socio-demographic characteristics. Therefore, this study aimed to (i) determine the prevalence of vegetarian and vegan plant-based diets in the Spanish population and (ii) assess the impact of these dietary patterns on perceived quality of life, lifestyle, prevalence risk of common diseases and healthcare service utilisation in individuals with similar socio-demographic backgrounds.
Methods
Study design
We conducted a two-phase study: first, we assessed the prevalence of individuals adhering to a plant-based diet (vegetarians and vegans) in a representative sample of Spanish adults living in their households. We also examined the socio-demographic characteristics of the plant-based and omnivore groups. In the second phase, we conducted a matched cross-sectional analysis with all plant-based diet individuals matched to a sample of participants with similar socio-demographic characteristics and an omnivorous diet to study the association of plant-based diets with perceived quality of life, lifestyle factors, diseases and healthcare service utilisation in the last year.
Source of data and context
The European Health Interview Survey in Spain (EHIS), conducted by the Spanish Statistics Institute and the Ministry of Health, is a household survey conducted every 5 years that collects health-related information on the population residing in Spain aged ≥15 years old, using a standard European questionnaire(22,23) . The EHIS data are a source of information for countrywide statistics on health, morbidity, the extent of access to and use of healthcare services and the factors influencing health-related outcomes(22). The EHIS collects a countrywide sample. It uses the family dwelling as the basic unit, gathering information voluntarily and anonymously. The sampling of the EHIS used a multistage clustering method with proportional random selection and independent samples from different Spanish regions. The survey was conducted in participants’ homes using face-to-face, computer-assisted personal interviews with trained interviewers. The methodology allows the delivery of long questionnaires, and the interviewer can assist in explaining the questions and demonstrating examples of how to complete the items. Data collected through this methodology are self-reported by the participants. The latest available data from 2020 included responses from 22 072 individuals. Further methodological details are available elsewhere(24).
Participants and matching criteria
We selected all individuals from the EHIS in Spain in the last available year (2020) with a complete form of a weekly FFQ included in the survey module on the dietary habits of the EHIS. It is a practical and standardised instrument that has been compared against a 7-day food diary and a seventy-four-item FFQ with strong reliability(Reference Tolonen, Koponen and Al-Kerwi25). It asks for fourteen food groups with six answer categories (once or more times a day, 4–6 times a week, three times a week, once or twice a week, less than once a week and never). We considered vegetarian to those participants reporting ‘never’ consuming meat, fish or processed meat products (e.g. salami, sausages and hot dogs); vegans if answered ‘never’ on meat, fish, processed meat, dairy or eggs and omnivores if answered ‘never’ to meat, fish and processed meat groups. The control participants (omnivores) did not avoid meat or other animal-derived foods in their diet (omnivores), according to the FFQ.
Study variables
Data on socio-demographic variables were collected on age (years), sex, living situation (living with a spouse, with cohabitating partner or living alone), marital status, completion of higher academic studies, household income and social class, categorised with the six-group classification adopted by the Spanish Society of Epidemiology, which is based on the occupation of the household reference person(Reference Domingo-Salvany, Bacigalupe and Carrasco26) and health insurance coverage used.
The prevalence of chronic diseases was examined using the item ‘Has a physician diagnosed any of the following diseases?’. We used the Patient Health Questionnaire for depression symptoms. The Patient Health Questionnaire-8 is a valid instrument for monitoring depression and studying its prevalence(Reference Kroenke, Strine and Spitzer27). Self-perceived health was studied with the EHIS question ‘How is your health in general?’ reported as very good, good, fair, bad and very bad. The variables on lifestyle habits included tobacco consumption (non-smokers, smokers and ex-smokers and hazardous alcohol drinking, which were measured using the Standard Drink Units (SDU). In Spain, one fermented beverage (e.g. beer, wine) is set to one 1 SDU, and a distilled beverage (e.g. spirit, liquor) is set to two SDU(Reference Dauer, Solé and Llácer28). We studied the weekly hazardous drinking (weekly alcohol intake >28 SBU in men or >17 SBU in women)(Reference Colom and Segura-García29). Physical activity was measured according to the WHO enhancing physical activity recommendations (1): aerobic compliance (≥150 min of at least moderate-intensity aerobic physical activity per week), muscle-strengthening compliance (≥2 times per week), and aerobic and muscle-strengthening if both. Variables on the use of healthcare services included a reported consultation (yes/no) during the preceding year with a General Practitioner, a physician specialist (not a General Practitioner), a physiotherapist, a kinesiotherapist, a chiropractor or osteopath, a psychologist, psychotherapist or psychiatrist, a registered nurse or nurse practitioner and a complementary medicine practitioner (homeopathist, naturopath or acupuncturist). Consumption of medicines was studied using the following questions: (During the past two weeks, have you used any medicines prescribed by a doctor? (yes/no) and have you used any medicines, herbal medicines or vitamins for the past two weeks? (yes/no), we also examined the reported consumption of vitamins, minerals or tonics and naturopathic or homeopathic products.
Statistical analysis
All participants were classified as vegetarians, vegans or omnivores according to their FFQ answers. We examined the frequency of weekly intake in the fourteen food groups. The prevalence of plant-based diets was computed using the whole EHIS sample, and population estimates with its 95 % CI were calculated using weighting factors for the entire Spanish population and 1eighteen Spanish regions. The socio-demographic characteristics of participants with plant-based and omnivore diets were described using central and frequency statistics, and differences between groups across socio-demographic characteristics were examined using ANOVA or χ2 tests.
In the second phase, the matching process of participants with a plant-based diet to a sample of participants with an omnivore diet based on socio-demographic factors that showed statistically significant differences in the first phase (P value < 0·05) was carried out using a propensity score analysis with the nearest neighbor criteria (Caliper = 0·08, r = 1:4). The matching method was used to reduce confounding bias and dependence by balancing the group of participants with omnivore and vegetarian diets in selected covariates. After matching, the final data were expected to be closer to a block-randomised design, and we examined the balance with the standardised mean differences of the propensity scores(Reference Ho, Imai and King30,Reference Stuart31) .
The association of plant-based diets with the prevalence of chronic diseases, self-perceived general health (bad or very bad categories), lifestyle and risk factors and use of healthcare services and medicine consumption were calculated using logistic regression models in two steps: Initially, we computed logistic regression models with the plant-diet or omnivore diet as the response factor. Subsequently, we adjusted the association for the influence of residual differences and variance between groups by introducing the socio-demographic covariates from the first analysis into the models. The results are presented as OR and adjusted OR with their respective 95 % CI and P values. In all analyses, the level of statistical significance was set at α = 0·05. All analyses were performed using R version 4.2.2(32).
Results
We identified eighty-six individuals following a plant-based diet, with a population prevalence in Spain of 5·62‰ (95 % CI: 4·33, 7·28). Of these, eighteen reported consuming a vegan diet (prevalence 1·31‰, 95 % CI: 0·70, 2·20), and sixty-eight followed a vegetarian diet (prevalence 4·31‰, 95 % CI: 3·18, 5·69). Three regions had no vegetarians, and eight had no vegans. Due to the small sample sizes of the vegan and vegetarian groups, we combined both groups. The Iberian Peninsula’s northern and central Mediterranean regions had higher prevalences, ranging from 10 to 12·23, respectively, whereas the central and southern regions had lower rates, between 1·52 and 6·64. Figure 1 presents a map of prevalence distribution across Spanish regions, with detailed regional data in Tables A, B and C in the online appendix.
Population prevalence in Spanish regions.
Figure 2 presents a stacked bar chart illustrating weekly food intake, with color intensity indicating the intake frequency by food group. Darker shades represent higher weekly intake (e.g. daily consumption). Individuals adhering to a plant-based dietary pattern had higher proportions in the never category across all food groups, with differences in snacks (50 % v. 35 %), fast food (58 % v. 36 %) and sweets (33 % v. 14 %). Individuals following a plant-based dietary pattern reported higher intake frequencies of vegetables, salads and greens, with 82·6 % consuming them once or more times per day compared with 43·9 % of omnivores. Additionally, over two-thirds (66·2 %) of plant-based individuals consumed legumes three or more times per week, compared with 33·3 % of omnivores. In contrast, differences in the never consumption of staple plant-based foods – such as bread and cereals (6 % v. 1 %), pasta, rice or potatoes (6 % v. 0 %), and legumes (6 % v. 1 %) – were minimal, as both groups showed relatively high overall consumption levels.
Weekly food intake by food groups, as reported by plant-based and omnivore participants.
The socio-demographic characteristics in Table 1 show that individuals with a plant-based diet were more likely to be women than those with an omnivore diet (69·08 % v. 52·80 %, P < 0·001), they were younger (mean age 48·42 v. 54·60 years, P < 0·001) and more often not living with a partner (70·93 % v. 49·19 %, P < 0·001), with a majority being single (60·35 % v. 27·04 %, P < 0·001). Additionally, individuals on a plant-based diet had higher educational attainment (university degree or higher: 41·86 % v. 19·29 %, P < 0·001). They also reported higher household incomes and social classes than those with an omnivorous diet (P < 0·05) and a higher proportion of healthy weight than omnivores (42·76 % v. 67·47 %, P < 0·001). The matching process of eighty-six individuals following a plant-based diet to 342 matched omnivores (1:4 ratio) reduced to almost 0 the standardised mean differences in age (deciles), sex, household living situation, marital status, educational attainment, household income, BMI and social class (Figure A in the online appendix). Additionally, we conducted a sensitivity analysis to assess whether BMI differed between the omnivorous and vegetarian groups when it was not included in the matching process, and we did not find a statistically significant effect. This result is presented in Figure B in the online appendix.
Socio-demographic characteristics of individuals with plant-based diets and omnivores
The study on lifestyle factors, healthcare utilisation and medication use showed that individuals following plant-based diets were had a higher likelihood of visiting a mental health professional (adjusted OR: 3·21, 95 % CI: 1·38, 7·43) and were more likely to consult complementary medicine practitioners, such as acupuncturists, naturopaths or homeopaths (Adjusted OR: 3·00, 95 % CI: 1·06, 8·48). No significant differences in tobacco use, physical activity, weekly hazardous drinking or consultations with other healthcare professionals (general practitioners, registered nurses, physician specialists or physiotherapists) were observed; all these results are shown in Table 2.
Plant-based diets association with lifestyle factors, health care utilisation and medication
* : P value < 0.05.
† Adjusted by: sex, region, age, studies, household income and social class.
Plant-based diets were associated with a higher proportion of depressive symptoms according to the PHQ8 questionnaire (Adjusted OR 2·58, 95 % CI: 1·00, 6·65) and an increased risk of stroke (adjusted OR 7·08, 95 % CI: 1·27, 39·46); however, this result showed a small subgroup size and a wide CI. No associations were found between plant-based diets and other examined diseases. Table 3 presents the prevalence risks and unadjusted and adjusted logistic regression results.
Plant-Based diets association with risk of chronic disease
* : P value < 0.05.
† Adjusted by sex, region, age, studies, household income and social class. Diseases are self-reports form the survey item: ‘Has a physician diagnosed any of the following diseases?’ for depression symptoms, we used the Patient Health Questionnaire (PHQ-8).
Discussion
This study found a low prevalence of individuals following a plant-based diet in different regions. Those adhering to plant-based dietary patterns reported a high frequency of intake of bread and cereals, pasta, rice, potatoes, legumes, fruits and vegetables. They were generally younger, female, single, with higher education levels and household incomes. However, when comparing plant-based and omnivorous individuals with similar socio-demographic characteristics and BMI, lifestyle factors did not differ.
The proportion of individuals adhering to plant-based diets was 5·62 per 1000 inhabitants. This is consistent with a previous study that combined data from national and European surveys covering the period from 2001 to 2011 and reported a sample prevalence of 0·2 %(Reference Acevedo Cantero, Ortega Santos and López-Ejeda33). In neighboring France, a web-based study reported a 3·4 % prevalence of plant-based diets among its 93 823 participants(Reference Allès, Baudry and Méjean34); however, this result might be interpreted cautiously due to the absence of a random sampling. Comparatively, the proportion of vegetarians is estimated at 4 % in the USA and 9 % in Italy, Germany and the United Kingdom(Reference Leitzmann35). These figures, however, are derived from data provided by Western vegetarian societies and may be subject to bias, as individuals often overestimate their adherence to a plant-based diet(Reference Vinnari, Montonen and Härkänen36). Therefore, our study provides a reliable, population-based estimate of plant-based diet adherence, confirming that while this dietary pattern remains a minority, its prevalence is likely overestimated in self-reported data.
Regional differences in plant-based adherence in Spain, with the highest frequencies found in the northeast regions of the Mediterranean Sea (Catalonia and Valencia), the central northern regions (Basque Country and Navarre) and the Canary Islands. Conversely, the Southern and central regions had the lowest adherence to plant-based diets. These regional differences may reflect underlying cultural and economic disparities not associated with regional average meat consumption per person, where Navarra had the highest meat consumption. At the same time, Catalonia, Valencia and the Basque Country ranked 8th, 10th and 6th, respectively(37). Regarding Regional Gross Domestic Product per capita for 2020, the Basque Country, Navarra and Catalonia ranked 2nd, 3rd and 4th, while Valencia ranked 12th. The low adherence to plant-based diets and observed regional disparities highlight the need for more targeted studies to explore dietary patterns within the population in greater detail.
Individuals following a plant-based diet reported a higher frequency of ‘never eat’ responses than omnivores, particularly regarding sweets, sugar-sweetened beverages, fast foods and snacks. In contrast, the highest consumption frequencies among plant-based individuals were observed for bread and cereals, pasta, rice, potatoes, legumes, fruits and vegetables. A systematic review analysing the quality of plant-based diets reported that plant-based diets were of higher quality due to a greater intake of whole grains, total fruits, plant-based proteins and seafood protein, along with a lower intake of refined grains and overall protein foods(Reference Parker and Vadiveloo38). A study using the Modified Healthy Eating Index also reported that plant-based diet quality scored higher than an omnivore diet in food groups with high-energy density (fat), processed meat and plant-based meat alternatives(Reference Bruns, Mueller and Schneider39). In our opinion, the dietary intake data obtained from the Spanish EHIS could be used in future research to analyse the quality of the Spanish diet in specific population groups. This approach could allow for adjustments based on lifestyle factors, socio-demographic variables and other covariates that may influence dietary choices and composition.
As anticipated, we implemented a study design to compare individuals following a plant-based diet with omnivores of similar socio-demographic characteristics, and we did not find differences in lifestyle habits (tobacco use and physical activity and hazardous drinking), consultations with general practitioners, registered nurses, physician specialists or physiotherapists. Social determinants of health shape nutritional patterns and influence health status(Reference McCullough, Chantaprasopsuk and Islami40). However, studies on the impact of plant-based diets often overlook social determinants of health to assess the health outcomes of plant-based diets accurately. Our study’s results showed that lifestyle factors were similar between the matched samples of participants adhering to plant-based diets and omnivorous diets. However, adherence to a plant-based diet was associated with a higher prevalence risk of depressive symptoms and an increased frequency of visits to mental health professionals. A recent meta-analysis reported that meat consumption was associated with lower depression(Reference Dobersek, Teel and Altmeyer41). A study involving a population-based cohort of 90 380 participants in France reported an association between depression and a lacto-ovo-vegetarian diet, with an adjusted OR of 1·44 (95 % CI: 1·08, 1·92), and with a vegan diet, an OR of 1·18 (95 % CI: 0·56, 2·47)(Reference Matta, Czernichow and Kesse-Guyot42). These results are comparable to our findings, which showed an adjusted OR of 2·58 (95 % CI: 1·00, 6·65) for the association between depression and plant-based diets. Other studies also found that depression among individuals with plant-based diets, focusing on females or males, was associated with higher rates of depressive disorders(Reference Baines, Powers and Brown43,Reference Hibbeln, Northstone and Evans44) . While most studies, with few exceptions, suggest an association between plant-based diets and mental health issues, it is essential to note that causality cannot be established based on observational studies that do not account for participants’ baseline mental health status. Disentangling why mental health status is associated with adherence to a plant-based diet compared with individuals with similar socio-demographic characteristics would require a different research approach (i.e. a qualitative study), given that an experimental design is not feasible in practice in humans.
We observed an association between the vegetarian diet and self-reported prevalence risk of stroke, with vegetarians showing a significantly higher adjusted OR of 7·08 (95 % CI: 1·27, 39·46) compared with omnivores. This finding aligns with a large cohort study reporting a 17 % increased risk of stroke among vegetarians compared with meat eaters (hazard ratio: 1·17, 95 % CI: 1·00, 1·40), and an hazard ratio of 1·35 for vegans, although the latter was not statistically significant(Reference Key, Papier and Tong45). Conversely, other extensive cohort studies and a meta-analysis indicated that vegetarian diets might reduce stroke risk, reporting a hazard ratio of 0·86 (95 % CI: 0·67, 1·11) for stroke among vegetarians(Reference Chiu, Chang and Wang46–Reference Baden, Shan and Wang48). The risk of stroke is primarily driven by chronic diseases and modifiable factors, such as hypertension, smoking, diet, physical activity and alcohol consumption, which together account for >90 % of the population-attributable risk(Reference O’Donnell, Chin and Rangarajan49). However, we found no significant associations between plant-based diets and these diseases. Other expected impacts between omnivorous and vegetarian groups on cardiovascular risk factors, such as hypertension or cholesterol, might not have been observed because the sample of vegetarians was relatively young. Studying the impact of dietary patterns in older populations adhering to a vegetarian diet may better capture variations in these risk factors to examine with precision the effect of plant-based diets and increase our understanding of how they affect ischaemic stroke risk and other CVD.
Practical implications
It should not be presumed that individuals adhering to a vegetarian diet are inherently healthier or experience better mental health status than those following other dietary patterns. Such assumptions must be evaluated within specific health and psychological outcomes, as dietary patterns alone do not necessarily correlate with improved health or increased mental health status. Policymakers and health professionals should prioritise personalised, culturally sensitive, inclusive, and evidence-based dietary guidance rather than promoting specific dietary patterns such as vegetarian, vegan, or omnivorous diets. Furthermore, public health initiatives promoting healthy eating should be accompanied by education on balanced nutrition to prevent potential deficiencies and manage expectations regarding health outcomes.
Limitations
Participants’ recall biases may influence the study findings. However, because this bias likely affects all participants systematically, the reliability of the results remains intact. In addition, no cases were identified in small regions. Although this may raise concerns regarding prevalence estimates, the questionnaire did not directly ask participants whether they consistently exclude animal products but assessed consumption over a recent reference period. As such, it may not fully capture the true proportion of individuals who follow a strictly vegan or vegetarian diet. While the questioning method could be critiqued, the EHIS data are a valuable and reliable reference for the EU. We could not determine when participants began the plant-based diet or how long they maintained it. However, this limitation also affects the group of omnivores who could have followed plant-based diets in the past. Dietary classification was based on self-reported intake using six frequency categories ranging from ‘once or more times a day’ to ‘never’. Participants who usually consume animal products but abstained during the reference period might be misclassified as vegetarians or vegans. Nevertheless, the use of the ‘never’ category provides a relatively strong criterion, helping to minimise the inclusion of recent or inconsistent adopters in the vegetarian or vegan group.
Conclusions
The prevalence of plant-based diets remains relatively low. Individuals who adhered to plant-based diets were generally younger, female, single and had higher education levels and household incomes. Self-perceived general health and lifestyle factors, such as tobacco consumption, hazardous drinking and physical activity, were similar to those of individuals following omnivorous diets with similar socio-demographic profiles. These findings highlight the importance of prioritising healthy behaviours over specific dietary patterns when addressing patient concerns regarding dietary patterns and personal well-being.
Supplementary material
For supplementary material accompanying this paper visit https://doi.org/10.1017/S1368980025100669
Data availability
Data described in the manuscript, code book and analytic code will be made publicly and freely available without restriction at https://www.sanidad.gob.es/en/estadEstudios/estadisticas/EncuestaEuropea/Enc_Eur_Salud_en_Esp_2020.htm
Authorship
L.G.d.P. and N.E. designed, conducted the research and wrote the paper; L.G.d.P. analysed the data. All authors contributed to the writing of the first draft and read and approved the final manuscript.
Financial support
This study had not received any grant funding.
Competing interests
The authors declare no conflicts of interest.
Ethics of human subject participation
This study used information from public sources, including a public, anonymous dataset from an online repository. Therefore, under Spanish legislation and EU rules, the approval of the Ethics Committee was unnecessary.
Open access
