The global prevalence of unhealthy dietary patterns and their impact on public health have received considerable attention in recent years(Reference Morze, Danielewicz and Hoffmann1,Reference Sotos-Prieto, Bhupathiraju and Mattei2) . Poor diet quality has been consistently linked to chronic diseases, including CVD, type 2 diabetes and certain cancers, underscoring the importance of understanding how dietary adherence influences overall mortality risk(Reference Nazar, Díaz-Toro and Petermann-Rocha3–Reference Ahmad, Moorthy and Lee5). Moreover, a robust body of evidence demonstrates that higher adherence to defined healthy dietary patterns – characterised by balanced intake of nutrient-dense foods and beverage groups – correlates with a reduced risk of mortality(Reference Sotos-Prieto, Bhupathiraju and Mattei2,Reference Shan, Wang and Li6,Reference English, Ard and Bailey7) .
However, dietary patterns and their health impacts vary by region and culture(Reference Jayedi, Soltani and Abdolshahi8,Reference Wang, Foster and Yi9) , and findings from developed countries may not be fully generalisable to other contexts(Reference Harmon, Boushey and Shvetsov10). Socio-economic factors also shape dietary behaviours, as they directly affect access to healthy foods and the ability to meet dietary recommendations(Reference Mackenbach, Nelissen and Dijkstra11).
In Latin America, where rapid urbanisation and economic disparities coexist, these factors often mediate the relationship between dietary patterns and health outcomes. Understanding this context is key to evaluating how adherence to healthy dietary patterns influences mortality, especially in populations where socio-economic inequalities may amplify barriers to achieving optimal dietary practices(Reference Mujica-Coopman, Navarro-Rosenblatt and López-Arana12). For example, a recent study conducted among elderly individuals in Costa Rica found a lower all-cause mortality associated with a traditional rural dietary pattern, where a major component was beans(Reference Yundan, Monica and Ana13).
Despite the global focus on dietary patterns, most studies have focused on regions outside of Chile, limiting the applicability of these findings to its population(Reference Schwingshackl, Bogensberger and Hoffmann14,Reference Petermann-Rocha, Diaz-Toro and Troncoso-Pantoja15) . Chile represents a unique context characterised by distinct health determinants, dietary practices and health profiles(Reference Martínez Arroyo, Corvalán Aguilar and Palma Molina16). Notably, 74 % of adolescents and adults over the age of 15 are classified as overweight or obese, highlighting the significant public health challenge posed by diet-related chronic conditions(Reference Lanuza, Morales and Hidalgo-Rasmussen17,18) . Although Chile’s geographical diversity provides access to a wide variety of foods, only 5 % of Chileans adhere to national dietary guidelines(Reference Lanuza, Zamora-Ros and Petermann-Rocha19).
While the associations between high diet quality and lower mortality risk are well established, most studies have been conducted in high-income countries and rely on dietary indices that may not reflect local or regional dietary patterns. In addition, conventional indices typically assign equal weight to all food groups, even though their impact on mortality risk differs considerably(Reference Burggraf, Teuber and Brosig20).
Therefore, the aim of this study was to evaluate the association between adherence to a healthy eating score (both unweighted and weighted) and all-cause mortality risk in a nationally representative sample. This applied approach offers a more nuanced and context-specific understanding of diet quality and its public health implications.
Methods
Study design and participants
This prospective study included participants aged ≥ 15 years, who underwent baseline assessments during the Chilean National Health Survey 2016–2017 (CNHS 2016–2017). The CNHS 2016–2017 was a large cross-sectional, nationally representative population-based study comprising 6233 participants(18). They were selected through a stratified multistage sampling of non-institutionalised individuals from urban and rural areas. Although the CNHS was originally designed as a cross-sectional survey, we conducted a prospective analysis by linking baseline data with mortality follow-up records from the Chilean Civil Registry and Identification. Trained interviewers collected data in two home visits, in which individuals were administered questionnaires (e.g. lifestyles), and measurements were taken, including anthropometric and physiological measures, as well as biological samples. Trained nurses conducted all clinical measures. From the original sample size (6233 participants), and after removing individuals with missing data on the exposure and covariates (n 897), the final analytical sample comprised 5336 participants (Fig. 1). Non-significant differences were observed between included and excluded groups regarding sex, geographical zone and healthy eating score (P > 0·05).
Fig. 1. Participants included in the formal analysis. Chilean National Health Survey 2016–2017.
The CNHS 2016–2017 was funded by the Chilean Ministry of Health and led by the Department of Public Health, the Pontificia Universidad Católica de Chile. The CNHS 2016–2017 was conducted according to the guidelines laid down in the Declaration of Helsinki, and all procedures involving human subjects were approved by the Ethics Research Committee of the Faculty of Medicine at the Pontificia Universidad Católica of Chile and reviewed by the Chilean Ministry of Health. Written informed consent was obtained from all subjects (≥ 18 years) or from the caregivers of those younger than 18 years. Data are available on the Ministry of Chile webpage https://epi.minsal.cl/bases-de-datos.
Healthy eating score (unweighted)
The healthy eating score was developed to evaluate adherence to the national dietary guidelines. It assessed the frequency of consumption of six food groups: consumption of seafood, whole grains, dairy products, fruits, vegetables and legumes. These questions were designed by experts based on the food-based dietary guidelines for the Chilean population as described in the CNHS 2009–2010(21).
Frequency of consumption of seafood (How often do you consume fish and shell food?), whole grain (How often do you consume any whole-grain products like whole-grain bread, whole-grain cereal or any other food that contains whole-grain flour?), dairy products (How often do you consume milk, cheese, fresh cheese or yogurt?) and legumes (How often do you consume any type of legumes, such as beans, lentils, peas or chickpeas? The answers to these questions were scored according to the recommendations of the national dietary guidelines(Reference Olivares Cortés, Zacarías Hasbún and González22), from zero point for no compliance to two points for complete compliance, according to online Supplementary Table 1. For the intake of fruits and vegetables, it was possible to calculate an average intake/d according to the dietary report of the survey. An intake of 80 g/serving was considered as one standard portion size of fruits or vegetables according to the survey guidelines of use(18). Participants were scored according to tertiles of intake as g/d. Consumption of fruits and vegetables was estimated through the questions: ‘Typically, how many days a week do you eat fruits?’, ‘Typically, how many days a week do you eat vegetables or vegetable salad? (do not include legumes or potatoes)’ and ‘How many servings of fruits/vegetables or vegetable salad do you eat in one of those days? (supported by a deck of cards with standardised serving sizes)’. The healthy eating score was created, ranging from 0 to 12 points, where higher points indicated greater adherence to national dietary guidelines. This scoring method was constructed in prior research(Reference Lanuza, Petermann-Rocha and Celis-Morales23). The score was categorised into quartiles using rank cases, with the highest quartile representing the healthiest group.
Healthy eating score (weighted)
A weighted version of the healthy eating score was developed to account for the differential impact of individual food groups on all-cause mortality(Reference Burggraf, Teuber and Brosig20). Each food group was scored using a binary system: 1 if the recommendation is met and 0 if it is not. Relative risks (RR) for each food group in relation to all-cause mortality were obtained from a systematic review and meta-analysis of prospective studies (online Supplementary Table 1)(Reference Schwingshackl, Schwedhelm and Hoffmann24).
To construct a weighted score that reflects the relative contribution of each food group to mortality risk, we applied a logarithmic transformation to normalise the RR (ln(RR)). This approach is justified by the multiplicative nature of hazard ratios (HR): a given change in RR corresponds to a proportional (rather than additive) change in risk. The logarithmic transformation linearises this relationship, making it more interpretable and suitable for constructing weights that accurately reflect the relative importance of each food group.
The weighted diet score is then calculated as a weighted sum, following the structure: Diet Score = (1 – Recommendation met) × Weight for each food group. Recommendation met is a binary variable (1 = met, 0 = not met). The individual scores for each food group are summed to derive the total score. The final score was then categorised in quartiles using rank cases, with quartile 4 representing the healthiest group. The weighted score provides a nuanced measure of dietary quality by incorporating the relative impact of different food groups on mortality risk, offering a refined tool for analysing dietary patterns.
All-cause mortality
The outcome of the current study was all-cause mortality. The date of death was obtained at follow-up from death certificates linked to the Chilean Civil Registry and Identification. Mortality data were available until the 31st of December 2021. Therefore, mortality status was censored on this date or the date of death if this occurred earlier.
Covariates
Sociodemographic data were collected at baseline and included age, sex, zone of residence (rural or urban), income level (low, middle and high), geographical zone (north: I–VI regions, centre: VII–IX, south: X–XV) and education level (elementary: <8 years (low), secondary: 8–12 years (middle) and higher education: ≥12 years (high)).
Health-related conditions were self-reported in response to the question: ‘Has a doctor, nurse or another health professional ever told you that you have had or currently have hypertension, high cholesterol, diabetes, peripheral artery disease or previous CVD events (e.g. myocardial infarction or stroke)?’ These long-term conditions included hypertension, hypercholesterolaemia, diabetes, peripheral artery disease and prior CVD events. They were then used to construct a multimorbidity score, categorised as follows: no long-term conditions, one long-term condition or two or more long-term conditions.
Lifestyle factors included alcohol consumption, tobacco use, sleep duration, physical activity and sedentary behaviour. Alcohol consumption was self-reported and collected using the ‘Alcohol Use Disorders Identification Test’ (AUDIT) questionnaire developed by the WHO and adapted for the Chilean population(Reference Alvarado, Garmendia and Acuña25). Tobacco status was classified as non-smoker, ex-smoker or current smoker, based on self-reported responses. Sleep duration (in h/d) was self-reported using nationally validated questionnaires. Physical activity levels, including moderate and vigorous intensities and transport-related physical activity, were determined using the Global Physical Activity Questionnaire version 2 (QPAQ v2)(26). Physical activity was categorised into inactive individuals (<600 MET/min/week) and active individuals (≥600 MET/min/week)(Reference Concha-Cisternas, Lanuza and Waddell27). Sedentary behaviour was derived using the following question: ‘How much time do you usually spend sitting or reclining on a typical day?’.
Lifestyle variables were further stratified based on recommended criteria, including low-risk alcohol consumption (AUDIT score < 8 points), never smoking, adequate sleep duration (7–9 h/d), sufficient physical activity (≥ 600 MET/min/week) and low sedentary time (< 4 h/d). These criteria have been described in detail elsewhere(Reference Petermann-Rocha, Diaz-Toro and Troncoso-Pantoja15).
Finally, BMI was calculated as weight/height (kg/m2) and classified using the WHO criteria for adults (normal: 18·5–24·9 kg/m2; overweight: 25·0–29·9 kg/m2; obese: ≥ 30·0 kg/m2)(28) and the Pan American Health Organization criteria for older adults (normal: 23·0–27·9 kg/m2; overweight: 28·0–31·9 kg/m2; obese: ≥ 32·0 kg/m2)(29). Participants who were underweight were excluded due to the potential for reverse causality (n 197).
Statistical analyses
Descriptive characteristics by healthy eating score quartiles are presented as means with standard deviations (sd) for continuous variables and as frequencies and percentages for categorical variables.
Crude Kaplan–Meier curves were constructed to estimate 5·1-year survival for categories of healthy eating score (quartiles). Kaplan–Meier curves were selected as they visually represent survival probabilities across quartiles, allowing for an intuitive comparison of trends over the follow-up period. In addition, sensitivity analyses were conducted in the full adjusted model, using a 2-year landmark that excluded all participants who died within the first 2 years of follow-up (n 79).
Associations between healthy eating score quartiles and all-cause mortality were investigated using Cox proportional hazard models. Associations between healthy eating score quartiles and all-cause mortality were investigated using Cox proportional hazard models. This method was chosen for its suitability in analysing time-to-event data while accounting for potential confounders, providing robust estimates of relative HR for all-cause mortality. Individuals in the quartile 4 (healthiest eating score) were used as reference. The results are reported as HR with their 95 % confidence intervals (95 % CI). Duration of follow-up was used as the time variable.
Analyses were adjusted for confounders based on previous literature(Reference Lanuza30) using the following two models: model 1 was adjusted by age, sex, zone of residency and educational level, while model 2 was additionally adjusted for lifestyle variables (alcohol consumption, tobacco status, sleep duration, physical activity and sitting time), BMI and multimorbidity.
Finally, we also investigate whether the association between the healthy eating score (unweighted) categories and all-cause mortality differed by subgroups. We tested for interactions, and all of them were found to be non-significant in the fully adjusted model. Nevertheless, we stratified the analyses based on well-established insights from previous studies(Reference English, Ard and Bailey7), considering factors such as age (≥ and < 60 years), sex (men and women), zone of residence (urban and rural), geographical zone (regions) and BMI categories (online Supplementary Table 2).
The significance level was defined as P < 0·05. IBM SPSS 29.0 was used for statistical analyses. This study followed the STROBE reporting guidelines for cohort studies.
Results
Over a median follow-up of 5·1 years (interquartile range: 5·0 to 5·2 years), 276 participants (5·2 %) died. Baseline characteristics according to healthy eating score quartiles are shown in Table 1. Of the total sample, 22·3 % of participants were in quartile 1 (least healthy), 31·6 % in quartile 2, 14·9 % in quartile 3 and 31·3 % in quartile 4 (most healthy) of the healthy eating score. The mean age of participants was 48·9 (19·1) years, with minimal variation across quartiles.
Table 1. General characteristics of the study population by quartiles of healthy eating score. Chilean National Health Survey 2016–2017
AUDIT, Alcohol Use Disorders Identification Test; AMI, acute myocardial infarction; n, number; IQR, interquartile range.
Notable differences were observed in education level, with a higher percentage of individuals with low education in the lowest quartile of the healthy eating score, compared with those in the highest quartile. Additionally, participants in the lower quartile were more likely to reside in rural areas and exhibit less adherence to healthy lifestyle factors, such as physical activity. Conversely, participants in the highest quartile were more likely to be women and to reside in urban areas.
Crude Kaplan–Meier survival estimates by healthy eating score quartiles are shown in Fig. 2. Participants in the lowest quartile of the healthy eating score had lower survival rates, followed by those in the middle quartiles (second and third). Associations between the healthy eating score (unweighted and weighted) and all-cause mortality are presented in Tables 2 and 3, respectively. In model 2, the most adjusted model, participants in the lowest quartile of the healthy eating score had a 1·61 (95 % CI: 1·14, 2·27), while those in quartile 2 and 3 had 1·44 (95 % CI: 1·04, 1·99) and 1·47 (95 % CI: 1·00, 2·16) times higher risk of mortality due to any cause, respectively, compared with those in the highest quartile of the healthy eating score. This trend persisted across all models, including the 2-year landmark analysis.
Fig. 2. Crude Kaplan–Meier curve to estimate 5·1-year survival for healthy eating score. Chile, Chilean National Health Survey 2016–2017. Healthy eating score (Q4: highest and Q1: lowest); error bars (95 % CI).
Table 2. Associations between unweighted healthy eating score and all-cause mortality in Chilean adults. Chilean National Health Survey 2016–2017
HR, hazard ratio. Analyses are presented as HR and their 95 % CI. Individuals in the quartile 4 were used as the referent. Model 1: adjusted by age, sex, zone of residency and educational level; model 2: as per model 1 but additionally for lifestyle variables (alcohol consumption, tobacco status, sleep duration, physical activity and sitting time), BMI and multimorbidity.
A 2-year landmark was carried out as a sensitivity analysis, excluding people who died during the first 2 years of follow-up. *Using covariates from model 2.
Table 3. Associations between weighted healthy eating score and all-cause mortality in Chilean adults. Chilean National Health Survey 2016–2017
HR, hazard ratio. Analyses are presented as HR and their 95 % CI. Individuals in the quartile 4 were used as the referent. Model 1: adjusted by age, sex, zone of residency and educational level; model 2: as per model 1 but additionally for lifestyle variables (alcohol consumption, tobacco status, sleep duration, physical activity and sitting time), BMI and multimorbidity.
A 2-year landmark was carried out as a sensitivity analysis, excluding people who died during the first 2 years of follow-up. *Using covariates from model 2.
Associations between the weighted healthy eating score and all-cause mortality are presented in Table 3. In model 2, the most adjusted model, participants in the lowest quartile of the healthy eating score had a 1·52 times higher risk of all-cause mortality (95 % CI: 1·03, 2·23) compared with those in the highest quartile. However, for participants in quartiles 2 and 3, the association lost statistical significance. This trend persisted in the 2-year landmark analysis.
Finally, associations between quartiles of the healthy eating score and all-cause mortality by subgroups are presented in online Supplementary Table 2. Notably, despite the absence of significant interactions, the increased mortality risk associated with lower healthy eating scores was more pronounced among participants aged 60 years and older, with those in the lowest quartile having a 1·58 (95 % CI: 1·08, 2·30) times higher mortality risk compared with those in the highest quartile. Additionally, participants from the North region exhibited higher mortality risks in relation to the lowest quartile (2·01 (95 % CI: 1·14, 3·54)). Women also showed an elevated risk in the lowest quartile, with an HR of 1·72 (95 % CI: 1·04, 2·87). Conversely, participants with a BMI below 24·9 or 27·9 had a substantially higher mortality risk in the lowest quartile (HR: 2·82 (95 % CI: 1·57, 5·06)) compared with those with higher BMI.
Discussion
This study found that individuals with lower scores on the healthy eating score (unweighted and weighted) had a significantly higher risk of all-cause mortality compared with those with higher scores. These findings suggest that lower adherence to a healthy eating score, characterised by reduced consumption of seafood, whole grains, dairy products, fruits, vegetables and legumes, is associated with an increased risk of mortality in the Chilean population.
Previous studies often relied on dietary scores that did not account for the combined or weighted impact of individual food groups, limiting their ability to capture nuanced associations with mortality risk(Reference Lanuza, Petermann-Rocha and Celis-Morales23,Reference Tapsell, Neale and Satija31) . Our study adds novelty by applying a healthy eating score derived from the Chilean national dietary guidelines, tailored to regional dietary habits and public health concerns. Furthermore, we introduce a weighted version of the score, accounting for the differential health impact of food groups, which enhances the specificity and interpretability of the diet mortality associations.
The results of this study align with previous research demonstrating an inverse association between diet quality and mortality risk. For instance, studies conducted in various populations have reported that greater adherence to healthy dietary scores, such as the Healthy Eating Index (HEI) and the Mediterranean diet, is associated with a reduced risk of mortality(Reference Morze, Danielewicz and Hoffmann1,Reference Brlek and Gregorič32,Reference Mente, Dehghan and Rangarajan33) . In particular, HEI or alternative HEI include components that assess intake of nutrients of concern (e.g. fat, sodium, added sugars) and penalise consumption of certain foods. While this approach allows for detailed nutritional assessment, it assumes access to comprehensive dietary recall data. In contrast, our score is based on key food groups promoted by the Chilean dietary guidelines and does not include nutrients per se or penalize the intake of unhealthy foods, reflecting a pragmatic, food-based framework more feasible in a national survey.
It is important to acknowledge that scores such as HEI have undergone extensive validation across multiple international cohorts and are supported by strong predictive validity. By contrast, the healthy eating score proposed in our study – while grounded in national recommendations – has not yet been validated externally. Nonetheless, it responds to an urgent need to develop culturally adapted, context-sensitive tools for nutritional epidemiology in regions like Latin America, where the applicability of global scores may be limited not only by food availability but also by economic aspects and culinary practices, among others.
While most of these studies have been conducted in high-income countries in Europe and America, Chile presents a unique context. According to the Global Burden of Disease(Reference Murray, Aravkin and Zheng4), Chile is officially classified as a high-income economy. Nonetheless, it is often considered an emerging or transition economy due to its rapid economic growth, ongoing industrialisation and reliance on natural resources like copper, along with significant social inequities and developmental challenges. This study provides valuable insights into how these associations may differ across socio-economic and cultural contexts.
Interestingly, our score is quite similar than the Prospective Urban Rural Epidemiology (PURE) healthy eating pattern (except for whole grains v. nuts), which, in a combined analysis from eighty countries, was associated with a lower risk of mortality when comparing the lowest v. highest quintile of the healthy diet score (HR = 0·70; 95 % CI: 0·63, 0·77)(Reference Mente, Dehghan and Rangarajan33). Undoubtedly, both unweighted and weighted dietary scores are necessary despite the potential complexity and similar directional results (as seen in our data or the PURE study)(Reference Mente, Dehghan and Rangarajan33). This is because we recognise, and the evidence supports, that different food groups have varying impacts on mortality risk(Reference Schwingshackl, Schwedhelm and Hoffmann24,Reference Eleftheriou, Benetou and Trichopoulou34,Reference Afshin, Sur and Fay35) .
From a different perspective, a recent large prospective cohort study in China found that while the total scores of three a priori dietary indices showed no significant associations with all-cause mortality, key components – such as greater dietary variety, a more balanced diet and better food adequacy – were associated with reduced all-cause mortality(Reference Zheng, Zhu and Li36). This suggests that, beyond cultural and regional differences in dietary patterns and food security issues, the methodological approach of evaluating overall diet indices or specific food groups and their components could present opposite, neutralised or combined effects on mortality. Additionally, varying weighting methodologies across dietary indices can influence results, as certain foods (e.g. nuts or olive oil in the Mediterranean diet or whole-grain cereals and fatty fish in the Nordic diet) are more prominent in high-income countries but may be less common in Latin America or Chile(Reference Eleftheriou, Benetou and Trichopoulou34). These variations, influenced by factors such as food security, food processing and culinary techniques, can affect both the strength and direction of the association between diet and mortality. Indeed, extreme geographical zones, such as the North (I–VI regions) and South (X–XV regions) of Chile, showed different associations with mortality risk compared with the central regions, highlighting the importance of context-specific dietary evaluations(Reference English, Ard and Bailey7,Reference Burggraf, Teuber and Brosig20) .
Several biological mechanisms may explain the increased mortality risk observed among participants with lower scores on the HEI. Foods included in this index, such as fruits, vegetables, legumes and whole grains, are rich in vitamins, minerals, fibre and bioactive compounds such as phytochemicals that have been shown to have protective effects against chronic diseases, including CVD, type 2 diabetes and certain types of cancer(Reference Lanuza, Zamora-Ros and Bondonno37–Reference Aune, Giovannucci and Boffetta39). In addition to traditional beneficial properties (e.g. antioxidant and antinflammatory, among others) linked with dietary fibre and (poly)phenols, these components have also been identified as key factors in direct microbe–host interactions and, therefore, in shaping the gut bacterial community(Reference Lanuza, Meroño and Zamora-Ros40). However, in recent years, plant-based dietary indexes have not consistently been associated with a lower risk of mortality(Reference Kim, Wilkens and Haiman41,Reference Wang, Liu and Han42) . Indeed, the quality of these diets (healthy v. unhealthy) has been suggested as crucial in determining their effects on cardiometabolic diseases and the risk of all-cause mortality(Reference Lim, Neelakantan and Lee43,Reference Delgado-Velandia, Maroto-Rodríguez and Ortolá44) .
High consumption of seafood and dairy products has also been associated with better cardiovascular health and reduced inflammation, which could contribute to a decreased mortality risk(Reference Krittanawong, Isath and Hahn45). The n-3 PUFA have been attributed with properties such as anti-arrhythmic and anti-inflammatory effects and improved vascular function(Reference Rimm, Appel and Chiuve46). However, it is worth mentioning that fried fish consumption is likely associated with an increased risk of CVD events, reinforcing the importance of considering both the type of food and its matrix, as observed in the Costa Rican study, where, by contrast, beans are the main source of nutrients(Reference Yundan, Monica and Ana13). Regarding dairy products, the evidence is more controversial, and the effects may depend on the amount and the type of food: low-fat or full-fat dairy products such as milk, cheese and yogurt(Reference Schwingshackl, Schwedhelm and Hoffmann24,Reference Mazidi, Mikhailidis and Sattar47,Reference Tutunchi, Naghshi and Naemi48) . For example, a systematic review of prospective cohort studies did not find an association between dairy product consumption and CVD mortality(Reference Bhandari, Liu and Lin49), while another systematic review found an inverse association between yogurt consumption and the risk of all-cause and CVD mortality(Reference Tutunchi, Naghshi and Naemi48).
The interplay between socio-economic factors and dietary patterns highlights the importance of addressing social and economic inequalities as part of broader public health policies and strategies. Limited access to affordable, nutrient-dense foods and structural barriers, such as food deserts in rural or underprivileged urban areas, may exacerbate disparities in adherence to healthy dietary patterns. These challenges underline the necessity of integrating targeted policies and interventions that not only promote dietary education but also improve the affordability and availability of healthy foods, particularly for vulnerable populations. Recently, the new Dietary Guidelines for Chile have been relaunched(50), which delivers ten new nutritional guidance messages, highlighting two of them: consumption of fresh food from fairs and established markets and sharing kitchen tasks, enjoying new and traditional preparations.
This study has several strengths, including using a large, nationally representative cohort, the application of both unweighted and weighted healthy eating scores and the adjustment for multiple confounding factors in the analyses. However, it also has some limitations. First, dietary information was self-reported, which may lead to recall bias or inaccurate reporting. Although several confounding factors were adjusted for, the presence of residual or unmeasured confounders cannot be entirely ruled out. Finally, the observational nature of the study limits the ability to establish causal relationships.
In addition, the relatively short follow-up period and limited number of deaths may reduce the statistical power to detect associations and increase the risk of reverse causality, despite the 2-year landmark sensitivity analysis conducted. Moreover, the absence of cause-specific mortality data limited our ability to explore diet-disease associations in greater detail.
Future research could focus on longitudinal and interventional studies with a longer duration to observe how changes in dietary patterns over time affect mortality risk. Also, it is important to consider the type and quality of the dietary pattern and its food groups or food items to avoid misunderstandings between studies because the complexities and relationships of the diet and health depend on the food processing, sources and preparations or culinary techniques, among others.
Additionally, qualitative studies could explore the barriers and facilitators to adhering to a healthy diet in different subgroups of the Chilean population. It would also be beneficial to investigate how other lifestyle behaviours, such as physical activity, alcohol consumption and smoking, interact with dietary patterns to influence health outcomes(Reference Petermann-Rocha, Zhou and Mathers51). In conclusion, the findings of this study reinforce the importance of promoting healthy eating habits as a key strategy to improve public health and reduce the risk of mortality in the Chilean population.
Supplementary material
For supplementary material/s referred to in this article, please visit https://doi.org/10.1017/S0007114525104212.
Acknowledgements
We thank all participants for their cooperation and the Chilean Health Ministry, the Sub-secretary of Public Health and the Pontificia Universidad Católica de Chile for designing and conducting the National Health Survey (2009–2010). All the results in the study are the responsibility of the authors, and the Chilean Health Ministry was not involved in the study.
F. L. receives support from the Chilean Government by the Agencia Nacional de Investigacion y Desarrollo (ANID) (FONDECYT Iniciación 11250095).
F. L., F. D-T. and C. C-M. conceived and designed the original research; F. L. analysed the data and wrote the first draft of the article. F. L., F. D-T., G. N., Y. C-C., M. M., S. P-S., N. L-L., T. M., F. P-R. and C. C-M. interpreted the results, and all authors critically revised the research article and approved submission of the final manuscript.
The authors declare no conflicts of interest.
Raw data and all questionnaires from the CNHS are freely available on the Ministry of Chile webpage https://epi.minsal.cl/bases-de-datos
Open access
