Coastal Proximity and Visits are Associated with Better Health but May Not Buffer Health Inequalities

The study examines whether living closer to the coast and visiting it more frequently are associated with better self-reported general health, whether such coastal contact can mitigate (buffer) the adverse association between low household income and health (equigenesis hypothesis), and whether these relationships generalize across countries. Prior single-country studies in the UK, Belgium, and Spain indicated better health among people living nearer the coast, and one longitudinal study in England found better health in years when individuals lived within 5 km of the coast. Proposed mechanisms include reduced exposure to hazards (e.g., air pollution), increased physical activity, visual/indirect contact reducing psychological distress, and the fact that proximity is associated with more frequent coastal visits. Despite suggestive evidence, a 2017 systematic review found insufficient evidence linking blue-space exposure to health outcomes, with limited and heterogeneous studies. The equigenesis hypothesis posits that nature contact may weaken the link between socioeconomic disadvantage and poor health; evidence for green spaces is mixed, and two English studies suggested coastal proximity might mitigate income-related health disparities. This study preregistered hypotheses: H1 and H2 that proximity and visit frequency predict better health; H3 and H4 that proximity and visit frequency weaken the income–health association; and explored cross-country generalizability (RQ1–RQ2). It advances the field by considering both proximity and direct visits, testing moderation of income–health relationships, using representative samples from 14 European countries and Australia, and applying Bayesian multilevel modeling to quantify evidence strength and generalizability.

Past research shows mixed but generally positive associations between coastal (blue space) exposure and health. UK, Belgium, and Spain cross-sectional studies reported better self-reported health nearer the coast, and a longitudinal English study found better health when living within 5 km of the coast. Mechanisms proposed include reduced hazards, more physical activity (e.g., walking), social interactions, and psychological restoration. Coastal visit frequency tends to decline exponentially with distance, implying diminishing returns with greater distance. However, a 2017 systematic review judged evidence insufficient due to few, heterogeneous studies. Regarding modifiers, the equigenesis hypothesis suggests nature contact can buffer the negative health impact of low income; some studies and a systematic review support buffering effects of green space, others report null or even adverse effects in low-income suburban areas. Two English studies indicated coastal proximity might mitigate income-related inequalities in general and mental health. This study tests these ideas across 15 countries using Bayesian methods to evaluate support for or against effects and their generalizability.

Design and data: Cross-sectional, secondary analyses of SOPHIE (Europe) and SOPHIA (Australia) online surveys conducted March–April 2019 (14 European countries: Belgium, Bulgaria, Czechia, France, Germany, Greece, Ireland, Italy, Netherlands, Norway, Poland, Portugal, Spain, UK) and September 2019 (Australia). Samples per country were stratified by age, sex, and region; median completion time 18.2 minutes. Total N = 15,179 (about 1,000 per country), ages 18–99 (M = 46.20, SD = 15.81); 51.3% female; 48.3% university degree. Lower income quintiles were underrepresented using country-based 2019 income quintiles (lowest 14.2%; highest 21.9%). Ethical approvals were obtained; data were deidentified. Measures: Outcome was self-reported general health (SF1; 1 = very good to 5 = very bad). Predictors: self-reported home coastal proximity (1 = up to 1 km to 8 = more than 100 km) and coastal visit frequency over the last 12 months (1 = once a week or more to 6 = never). Household income was collected in country-adapted deciles and recoded to country-level relative income quintiles plus a missing category. Confounders included age and sex; household income was also included as a covariate where not an effect modifier. All key variables had “don’t know/prefer not to answer” options. Exclusions and weighting: Listwise exclusion for missing values on key variables in each model; age/sex had no missing. Czechia was excluded from proximity models (H1, H3). Analytical Ns ranged from 11,916 to 14,702. Sampling weights (age, sex, region) were rescaled to the analytic sample size. Analysis: Bayesian multilevel cumulative probit regressions with respondents nested within countries were fitted in R (brms/rstan). Proximity and visits were modeled both as categorical (reference: <1 km; once a week or more) and as monotonic ordinal predictors (allowing non-decreasing/non-increasing effects between adjacent categories). Due to multicollinearity between proximity and visits (Kendall’s τ = 0.60), they were modeled separately. Model sets for H1/H2: four specifications each—(1) categorical fixed slope; (2) categorical with random slopes across countries; (3) monotonic fixed slope; (4) monotonic with random slopes. For moderation (H3/H4), best-performing structures were extended to include fixed interaction with income, then random slope for income, and then random interactions (Models 1–6). Model comparison used LOOIC; evidence was quantified via Bayes factors and posterior distributions, including one-sided tests for directionality. Generalizability across countries was assessed via random effects and overlap of 90% credible intervals. Priors: weakly informative α, β ~ Normal(0,10); country-level variance ~ HalfCauchy(0,10); monotonic simplex parameters with uniform Dirichlet. Four MCMC chains, 4000 iterations (1000 warm-up); good convergence and posterior predictive checks reported. Sensitivity analyses: narrower priors (Normal(0,5)), inclusion of Czechia, exclusion of speeders, and additional covariates (education, work status, political orientation) yielded similar results. Preregistration on OSF; deviations documented.

- H1 (proximity → health): Very strong evidence that living nearer the coast is associated with better self-reported health within countries, controlling for age, sex, and income (BF+ = 82.33; slope b = 0.02, SE = 0.01; 90% CrI [0.01, 0.03]). Because lower scores reflect better health and nearer distance, a positive slope indicates better health with closer proximity. Largest marginal gain occurs between <1 km and 1–2 km (37.9% of total improvement; 90% CrI [7.7%, 64.7%]). Other adjacent-category improvements ranged from 6.7% to 18.2%. Predicted probabilities: within 1 km vs >100 km—very good health 10.4% vs 8.5% (1.22×), good 45.7% vs 43.2% (1.06×); fair 34.3% vs 36.7% (inverse 1.07×), bad 7.7% vs 9.2% (1.19×), very bad 1.8% vs 2.4% (1.31×). Country-level evidence ranged from BF = 2.24 (Italy) to BF = 341.86 (Norway), consistently positive; magnitudes were similar across countries (overlapping CrIs). - H2 (visits → health): Extremely strong evidence that more frequent coastal visits predict better health (BF+ → ∞; b = 0.11, SE = 0.02; 90% CrI [0.08, 0.13]) controlling for age, sex, income. Of the total improvement with increasing visit frequency, 13.3% (90% CrI [4.4%, 22.5%]) occurs between “once a week or more” and “once every 2–3 weeks,” and 53.4% (90% CrI [44.5%, 61.8%]) between “once or twice a year” and “never.” Predicted probabilities: at least weekly vs never—very good 12.4% vs 4.8% (2.60×), good 47.8% vs 35.3% (1.36×); fair 32.0% vs 41.7% (1.30×), bad 6.4% vs 13.7% (2.13×), very bad 1.4% vs 4.5% (3.29×). Evidence was at least strong in all countries except Italy; magnitudes varied: strongest in Ireland (b = 0.18, 90% CrI [0.13, 0.24]) and second strongest in Greece (b = 0.16, 90% CrI [0.09, 0.23]). - Generalizability (RQ1): Proximity–health relationship generalizes across countries in presence and similar magnitude; visits–health relationship generalizes in presence but not magnitude. - H3 (moderation by proximity): Contrary to buffering, very strong evidence that the income–health association is slightly stronger when living nearer the coast (BF = 39.68; interaction b = 0.01, SE = 0.00; 90% CrI [0.00, 0.01]). Effect is very small and CrI includes zero, compatible with no moderation; best-fitting model assumed fixed interaction across countries. - H4 (moderation by visits): Insufficient evidence that more frequent visits weaken the income–health relationship (BF+ = 1.08; interaction b = 0.00, SE = 0.00; 90% CrI [-0.01, 0.01]). Exploratory two-sided test provided extremely strong evidence that lower income is associated with poorer health regardless of visit frequency (BF01 = 2,348.96). Fixed interaction model fit best, indicating cross-country generalizability of the null. - Robustness: Sensitivity analyses (narrower priors, inclusion of Czechia, exclusion of speeders, additional covariates) yielded similar results.

The findings address the research questions by showing consistent positive associations between living near and visiting the coast with better self-reported health across Europe and Australia. The largest proximity-related health gains occur within the first 1–2 km, suggesting the benefits are strongest at very close distances, potentially due to increased opportunities for physical activity, social engagement, and psychological restoration. Visit frequency shows a strong graded relationship with health; however, part of the large difference between “never” and rare visitors may reflect reverse causality (e.g., chronic illness limiting mobility). Contrary to the equigenesis hypothesis, coastal proximity did not buffer income-related health inequalities; if anything, the income–health gradient was marginally stronger nearer the coast, though the effect was very small and compatible with null. No buffering by visit frequency was detected; lower income predicted poorer health regardless of coastal visits. Possible explanations include variation in coastal environmental quality and accessibility, especially in lower-income areas with potentially poorer-quality coastal environments. These results suggest that while coastal access and engagement can promote population health, they may not, by themselves, reduce income-related health disparities without targeted interventions. Cross-country analyses indicate broad generalizability of the presence of effects, with visit-related effect sizes varying by national context (e.g., strongest in Ireland, weakest in Italy), possibly due to differences in distance distributions, tourism, access, and privatization.

Living nearer to the coast and visiting it more often are associated with better self-reported health across 15 countries in Europe and Australia. However, these relationships do not preferentially benefit lower-income groups; there is no reliable evidence that coastal contact buffers the adverse association between low income and poorer health. Policymakers may consider enhancing fair and equitable access to healthy coastal environments to promote public health, while recognizing that such access alone is unlikely to reduce existing health inequalities without targeted efforts toward low-income populations. Future research should incorporate objective health measures, assess coastal environmental quality and accessibility, and test these relationships in low-income countries and at sub-national scales.

- Cross-sectional design precludes causal inference and cannot rule out selective migration or reverse causality (e.g., healthier individuals may choose to live near or visit the coast; individuals with chronic illness may visit less). - Country coverage limited to middle- and high-income nations; findings may not generalize to low-income countries where coastal environments may pose greater health risks. - Online survey modality may under-sample lower-income individuals; samples representative nationally by age, sex, and region but not at sub-national levels; lack of data on ethnic background limits subgroup analyses. - Reliance on self-reported measures (health, distance, visits); potential recall and seasonality biases in 12-month visit reports; distance may not capture perceived accessibility or travel time. - No direct assessment of coastal environmental quality or access/privatization, which may moderate health impacts, especially in lower-income areas. - Small moderation effect estimates with credible intervals including zero suggest practical insignificance regarding buffering; income quintile underrepresentation may affect moderation analyses.