Addressing Socioeconomic Inequities in Children’s Cardiovascular Health via Positive Experiences

83

Socioeconomic disadvantage leads to poor cardiovascular health and this 84 relationship may be mediated by positive childhood experiences. This study aimed to 85 estimate the extent to which promoting positive experiences could reduce socioeconomic 86 inequities in children’s cardiovascular health. 87 88

Data source: The Longitudinal Study of Australian Children Child Health 89 CheckPoint (N=1874). Exposure: Maternal education (low/medium/high) as a key indicator 90 of family socioeconomic position during pregnancy. Outcome: Cardiovascular health (11-12 91 years) (poor/good) quantified by four health behaviors and four health factors. Mediator: 92 Multiple positive experiences ( ≥ 2/<2) indicated by positive parenting, supportive 93 relationships, environments, and high social engagement (2-11 years). We conducted a causal 94 mediation analysis using an interventional effects approach, adjusting for childhood adversity 95 and other potential confounders. 96 97

Children with low (risk difference=4.9%, 95% CI=-3.2%, 13.0%) or medium (risk 98 difference=5.6%, 95% CI=-1.2%, 12.5%) maternal education had a higher risk of poor 99 cardiovascular health compared to those with high maternal education. Causal mediation 100 analysis estimated that increasing the levels of positive experiences in children with low or 101 medium maternal education to be like their high maternal education peers could reduce these 102 risk differences by 1.0% (95% CI= -0.8%,1.5%) and 0.5% (95% CI=-0.5%, 1.5%) respectively, 103 reducing cardiovascular inequities by 20.4% and 8.9%. 104 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 16, 2025. ; https://doi.org/10.1101/2025.07.15.25331608doi: medRxiv preprint 6 105

Targeted policy interventions that promote positive experiences are potential 106 opportunities to reduce socioeconomic inequities in children’s cardiovascular health. 107 However, such interventions should be considered within a broader and multipronged 108 approach that includes addressing socioeconomic disadvantage itself and other socially 109 distributed drivers of cardiovascular diseases to achieve the maximum impact. 110 111

maternal education, positive experiences, health inequities, cardiovascular health, 112 longitudinal, children, interventional effects113 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 16, 2025. ; https://doi.org/10.1101/2025.07.15.25331608doi: medRxiv preprint 7

114 Cardiovascular disease (CVD) is the leading cause of mortality globally, accounting for 32% 115 of all deaths in 2020. 1 The economic burden of CVD is substantial, with global costs 116 estimated at US$1 trillion in 2030. 2 Socioeconomic disadvantage is a well-established 117 determinant of CVD, contributing to differences in the incidence and mortality of CVD. 3 118 Addressing socioeconomic inequities in CVD is a priority of governments worldwide. 119 Evidence suggests that more than 80% of CVD can be prevented or modified in early life by 120 following healthy lifestyles and addressing risk factors such as high blood pressure and 121 diabetes.4 122 123 The American Heart Association (AHA) introduced the concept of ideal cardiovascular health 124 (CVH) in 2010, 5 which refers to not merely the absence of CVD but the presence of 125 favorable health behaviors (e.g., no smoking, healthy diet, regular physical activity) and 126 health factors (e.g., normal body mass index, healthy blood pressure and lipid levels). These 127 components were updated in 2022 to reflect the Life’s Essential 8 (LE8). 6 This paradigm 128 represents a shift in cardiovascular research from a deficit-focus approach to a 129 strengths-based approach.5 Monitoring CVH at the population level over the life course is 130 essential for identifying CVH disparities and informing targeted interventions. 131 132 Children’s CVH is shaped by the social environments where they live and develop across the 133 life span.6,7 Socioeconomic inequities in CVH emerge as early as childhood. 7 Data from the 134 2013-2018 US National Health and Nutrition Examination Survey indicate that family 135 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 16, 2025. ; https://doi.org/10.1101/2025.07.15.25331608doi: medRxiv preprint 8 income is associated with differences in CVH indicators such as nicotine exposure, body 136 mass index, and diet among children aged 2 to 19 years. 8 These disparities are driven by an 137 unequal distribution of material resources as well as structural barriers that disproportionately 138 affect children from socioeconomically disadvantaged families. 6,8-10 Addressing CVH 139 inequities in children is likely to yield greater cost-effective benefits than interventions later 140 in life, given the cumulative impact of early life exposures on long-term health outcomes.8,11 141 142 The mechanisms linking socioeconomic disadvantage to CVH are complex, 12 including both 143 adverse and positive experiences. While childhood adversity (e.g., family violence, child 144 abuse) has been well-established as a risk factor of CVH, 13,14 positive childhood experiences 145 warrant specific focus because they are valued by families and communities, and efforts to 146 promote positive experiences are considered highly acceptable, avoiding stigma and aligning 147 with strengths-based practices and policies.15,16 Positive experiences refer to a range of events, 148 activities, or situations that foster flourishing and better health outcomes.17 Although variably 149 defined, emerging evidence suggests that positive experiences are associated with better 150 CVH,14,16,18-20 with possible pathways such as enhanced self-esteem and lower rates of 151 substance use.21 152 153 Compared to CVH in adulthood, very few studies have explored CVH in childhood from a 154 life course perspective.8,22 While there is increasing evidence showing the benefits of positive 155 experiences, the extent to which promoting positive experiences would reduce socioeconomic 156 inequities in CVH remains unknown. To inform intervention opportunities and policy actions 157 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 16, 2025. ; https://doi.org/10.1101/2025.07.15.25331608doi: medRxiv preprint 9 on CVH improvement at the population level, we estimated the extent to which promoting 158 positive experiences could reduce socioeconomic inequities in children’s CVH. 159 160

161 Data source 162 We drew on a subset of data from the birth cohort (B-cohort) of the Longitudinal Study of 163 Australian Children (LSAC), which commenced in 2004 when children were aged 0-1 year 164 (n=5107). A two-stage clustered design was employed to select a sample that was broadly 165 representative of the Australian child population except those living in remote areas. 23 166 Children were followed up every two years. We drew data when children were aged 0-1 years 167 (Wave 1; n=5107), 2-3 years (Wave 2; n=4606), 4-5 years (Wave 3; n=4386), 6-7 years 168 (Wave 4; n=4242), 8-9 years (Wave 5; n=4085), 10-11 years (Wave 6; n=3764) and 11-12 169 years (CheckPoint wave; n=1874). The CheckPoint wave was a one-off national-wide 170 cross-sectional physical health and biomarker module, nested between LSAC Waves 6 and 7.24 171 Multiple information sources were utilized, including parent interviews, parent-report and 172 child-report questionnaires. 173 174 Despite the requirement for children to attend multi-hour, in-person clinic assessments in the 175 CheckPoint wave, over 1,800 families participated, demonstrating a strong commitment and 176 willingness to invest time and travel resources. We found that children who had lower 177 maternal education, came from Aboriginal or ethnic minority backgrounds, and lived in low 178 socioeconomic status neighborhoods were likely to be missed out (see Supplementary file 1). 179 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 16, 2025. ; https://doi.org/10.1101/2025.07.15.25331608doi: medRxiv preprint 10 180 Measures 181 Our conceptual model (Figure 1) shows the hypothesized causal pathway from maternal 182 education (during pregnancy) to children’s CVH (11-12 years), via positive experiences (2-11 183 years) as an intervention target of interest, informed by current knowledge (see 184 Supplementary file 2). Figure 1 was used to guide the selection of measures and inform the 185 analytic approach. 186 187 Exposure (during pregnancy) 188 Maternal education at Wave 1 was used as a key indicator of socioeconomic resources in the 189 family environment during pregnancy, assuming maternal education did not change 190 significantly from pregnancy to just after birth. 25 In keeping with previous studies, 25 we 191 categorized it into three groups: low (Year 12 or below); medium (Certificate I/II/III/IV or 192 Advanced Diploma); and high (Bachelor’s degree or above). 193 194 Mediator (2-11 years) 195 Informed by the Health Outcomes from Positive Experiences (HOPE) framework 26 and 196 previous validation work, 17 we quantified overall positive experiences using 17 indicators, 197 each mapping to one of the four domains of positive experiences prospectively collected from 198 2 to 11 years: (1) positive parenting practice; (2) trusting and supportive relationships; (3) 199 supportive neighborhood and home learning environments; and (4) social engagement and 200 enjoyment (see Supplementary file 3). Each indicator was first dichotomized (yes/no) using 201 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 16, 2025. ; https://doi.org/10.1101/2025.07.15.25331608doi: medRxiv preprint 11 the top quartile to indicate the presence of exposure to a positive experience at each wave. 17 202 Next, we summed the number of positive experiences at each wave for each domain and 203 dichotomized this count at ‘two or more’ to indicate multiple positive experiences in each 204 domain at each wave. To measure multiple positive experiences across four domains at each 205 wave, we summed the number of positive experiences at each wave and dichotomized this 206 count at ‘two or more’ at each wave. Finally, to measure multiple positive experiences over 207 the follow-up period (2-11 years), we summed the number of positive experiences across 208 these waves and dichotomized this count at ‘two or more’, given that a cluster of positive 209 experiences is likely to have a cumulative benefit on health.27 210 211 Outcome (11-12 years) 212 We quantified CVH at 11-12 years using the LE8 metrics, including four health behaviors 213 (diet, physical activity, cigarette smoking, and sleep) and four health factors (body mass 214 index, non-high-density lipoprotein, blood pressure, and blood glucose). For each child, each 215 of the LE8 metrics was scored on a scale of 0 to 100 (see Supplementary file 4). We then 216 calculated an overall CVH score, using the average value across all eight metrics. According 217 to the AHA recommendation, 6 we dichotomized the overall CVH score using “0 to 79” to 218 indicate children with poor CVH. 219 220 Confounders 221 Baseline confounders (at birth). We posited four baseline confounders: child’s sex 222 (male/female), child’s ethnicity (Anglo or European/minoritized ethnic group/Indigenous), 223 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 16, 2025. ; https://doi.org/10.1101/2025.07.15.25331608doi: medRxiv preprint 12 child’s disability status (yes/no), and neighborhood socioeconomic status (top 75% - not 224 disadvantaged/bottom 25% - disadvantaged) assessed by the Socioeconomic Indexes for 225 Areas of Relative Socioeconomic Advantage and Disadvantage.28 Due to the possible overlap 226 between maternal education and neighbourhood socioeconomic status, we conducted 227 sensitivity analyses by removing neighborhood socioeconomic status as a baseline 228 confounder . 229 230 Intermediate confounder (0-1 year). We posited four intermediate confounders: gestational 231 age in weeks (<37 weeks/ ≥ 37 weeks), maternal age at childbirth (<27 years/ ≥ 27 years), 232 family composition (single parent/two parents), and multiple childhood adversities (<2/ ≥ 2) 233 measured by parent legal problems, family violence, household member mental illness, 234 household member substance abuse, harsh parenting, parental separation, unsafe 235 neighborhood, and family member death (see Supplementary file 3).27 236 237 Statistical analysis 238 The analytic sample consisted of all children who attended CheckPoint (N=1874). Participant 239 characteristics were summarized overall and by maternal education, using descriptive 240 statistics. Preliminary analyses were first conducted to confirm whether data were consistent 241 with the expected associations depicted in Figure 1. Specifically, generalized linear models 242 with a log-Poisson link were used to examine unadjusted and confounder-adjusted 243 associations between exposure, mediator and outcome. Descriptive analyses and preliminary 244 analyses were conducted using Stata 18.0. 245 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 16, 2025. ; https://doi.org/10.1101/2025.07.15.25331608doi: medRxiv preprint 13 246 In all analyses, we ignored the clustering due to postcodes as the correlations between the 247 outcome measures within the postcodes were negligible (intra-cluster correlation= 0.0005). 248 We also did not incorporate the CheckPoint sampling weights in the analyses, 29 as the 249 incorporation of sampling weights appropriately in the causal mediation approach used below 250 is still an ongoing area of research. 251 252 Causal mediation analysis 253 We then conducted a casual mediation analysis using an interventional effects approach to 254 answer the causal question of interest:30,31 what would be the reduction in risk of poor CVH if 255 we could offer effective interventions that promote positive childhood experiences among 256 children with low or medium maternal education? As well-defined interventions that can 257 collectively address the composite measure of positive experiences are not available in the 258 community, we examined the question by conceptualizing hypothetical interventions that 259 map to a ‘target trial’.32 260 261 We first estimated the confounder-adjusted absolute difference in the risk of poor CVH in 262 children with low or medium maternal education compared to their high maternal education 263 peers, separately, using g-computation.30,31 These adjusted differences provided estimates of 264 the overall CVH inequities that we sought to reduce. 265 266 Next we evaluated the reduction in risk of poor CVH that would be achieved by a 267 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 16, 2025. ; https://doi.org/10.1101/2025.07.15.25331608doi: medRxiv preprint 14 hypothetical intervention that would shift the distribution of the positive experiences in 268 children with low or medium maternal education, to be similar to that in children with high 269 maternal education, using an extended g-computation estimation procedure (see 270 Supplementary file 5).32 This provided estimates of the absolute risk differences achieved if 271 we could offer an intervention that promotes positive experiences among children with low or 272 medium maternal education. 273 274 The difference between the initial overall CVH inequities and the reduction in risk achieved 275 by the hypothetical intervention provides an estimate of CVH inequities that would remain 276 after the hypothetical intervention. We also report the relative reductions in the gap achieved 277 for children with low or moderate maternal education using these estimates (i.e. reduction in 278 risk of poor CVH achieved by the hypothetical intervention as a ratio of existing 279 socioeconomic inequities in CVH). Standard error estimates were computed using a bootstrap 280 procedure. All mediation analyses were implemented using R Statistical Software 4.3.1 using 281 R package medRCT.33,34 282 283 Missing data 284 In the analytic sample, the percentage of missing data across any of the study variables was 285 59%. We used multiple imputation by chained equations to reduce bias due to incomplete 286 records, under the missing at random assumption. 35,36 Imputations of incomplete variables 287 were carried out at the composite level where applicable rather than at the item level due to 288 convergence not being achieved. The imputation model included all study variables and four 289 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 16, 2025. ; https://doi.org/10.1101/2025.07.15.25331608doi: medRxiv preprint 15 auxiliary variables (birth weight, GlycA, family income, and parents’ disability status) as well 290 as all two-way interactions amongst exposure, mediator, outcome, and confounders. 37 Based 291 on the percentage of missing data,36 we produced 60 imputed datasets and used Rubin’s rules 292 to obtain the final imputed estimates of interest. 38 Results using multiply imputed data are 293 shown for preliminary analyses and causal mediation analyses. 294 295

296 Sample characteristics 297 Participant characteristics are summarized in Table 1. At 11-12 years, around half (53.6%) of 298 children had poor CVH. A larger proportion of children with low or medium maternal 299 education had poor CVH compared with their high maternal education peers (low: 55.0%, 300 medium: 57.4%, high: 50.5%). At 2-11 years, a larger proportion of children with low or 301 medium maternal education had fewer positive experiences (low: 63.4%, medium: 55.3%), 302 compared with those with high maternal education (high: 51.3%). 303 304 Associations between socioeconomic disadvantage, positive experiences, and poor CVH 305 Children with low or medium maternal education had a higher risk of poor CVH and fewer 306 positive experiences than their high maternal education peers, after adjusting for baseline 307 confounders (Table 2). Children who had fewer positive experiences had a higher risk (risk 308 ratio (RR)=1.17; 95% CI=1.00, 1.36) of poor CVH than those who had two or more positive 309 experiences, after adjusting for all confounders and maternal education. These findings 310 confirm the hypothesized associations depicted in Figure 1. 311 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 16, 2025. ; https://doi.org/10.1101/2025.07.15.25331608doi: medRxiv preprint 16 312 The extent to which intervening on positive experiences could reduce socioeconomic 313 inequities in poor CVH 314 Using the interventional effects approach, we estimated an absolute difference of 4.9% (95% 315 CI: -3.2%, 13.0%) and 5.6% (95% CI: -1.2%, 12.5%) in the prevalence of poor CVH in 316 children with low and medium maternal education when compared with their high maternal 317 education peers. If we were able to intervene to effectively increase the levels of positive 318 experiences amongst children with low maternal education to be equivalent to their high 319 maternal education peers, we could potentially reduce this absolute risk difference by 1% (95% 320 CI: -0.8%, 1.5%). This translates to a relative reduction of 20.4% of socioeconomic inequities 321 (Table 3). Similarly, hypothetical interventions that promote positive experiences amongst 322 children with medium maternal education to be like their high maternal education peers, 323 could potentially reduce the absolute risk difference by 0.5% (95% CI: -0.5%, 1.5%), leading 324 to a relative reduction of 8.9% of socioeconomic inequities. After the hypothetical 325 interventions, 3.9% and 5.1% absolute socioeconomic difference in CVH would remain 326 respectively among children with low and medium maternal education. 327 328

from the sensitivity analysis omitting the neighborhood socioeconomic status showed 329 that promoting positive experiences in children with low and medium maternal education to 330 be like their high maternal education peers could reduce the absolute risk difference by 1.2% 331 and 0.6% respectively (see Supplementary file 6). 332 333 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 16, 2025. ; https://doi.org/10.1101/2025.07.15.25331608doi: medRxiv preprint 17

334 Using prospective data from a national birth cohort, we estimated the potential reduction of 335 socioeconomic inequities in children’s CVH if we could offer effective interventions to 336 promote positive experiences among children with low or medium maternal education to be 337 equivalent to their high maternal education peers. Our findings suggest a positive effect of 338 positive experiences to reduce socioeconomic inequities in CVH, especially in children with 339 low maternal education. 340 341 Consistent with previous findings, 21,39-41 we found that children with more positive 342 experiences had lower risk of poor CVH, after controlling childhood adversity and other 343 confounders. We build on existing evidence by evaluating the potential benefit of a 344 hypothetical intervention on positive experiences to reduce socioeconomic inequities in 345 children’s CVH. Children with medium maternal education (5.6%) show a slightly greater 346 absolute risk difference with their high maternal education peers than the low maternal 347 education group (4.9%). However, promoting positive experiences appears to reduce a larger 348 proportion of inequity among children with low maternal education. The greater benefit 349 observed for children with low maternal education likely reflects the well-documented social 350 gradient in CVH, 7 where children from more disadvantaged families benefit more from 351 interventions. Our estimates show that promoting positive experiences could reduce CVH 352 inequities by up to 20.4%, highlighting the potential value of investing in positive 353 experiences as a key intervention target. These reductions in childhood are likely to 354 accumulate and translate to substantial social and health benefits in adulthood.25 355 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 16, 2025. ; https://doi.org/10.1101/2025.07.15.25331608doi: medRxiv preprint 18 356 Socioeconomic disadvantage can influence CVH through multiple interrelated pathways, 357 including positive experiences as examined here, and other mediators such as childhood 358 adversity, limited resources and healthcare access. 6,7 Given this complexity, no single 359 intervention will fully close the socioeconomic gap in children’s CVH. The residual 360 inequities that persist after the hypothetical intervention on positive experiences suggest that 361 a multi-faceted and stacked approach is needed to address both upstream and downstream 362 factors of CVH.42 363 364 Strengths and limitations 365 This study utilizes a national birth cohort that captured resourceful data on social, 366 environmental, and physical measures longitudinally. We also used the target trial framework 367 to provide clarity in the study design (e.g., eligibility criteria, treatment strategies). However, 368 some limitations should be noted: (1) Selection bias: Although we conducted multiple 369 imputation to reduce selection bias due to missing data in the sample, we were unable to 370 account for the CheckPoint sampling weights in the analysis, meaning that some selection 371 bias might remain and limit the generalisability of our findings; (2) Measurement bias: We 372 used maternal education as a single measure of family socioeconomic position, which may 373 underestimate the influence of socioeconomic disadvantage on the outcome. In addition, 374 measurement errors may exist for parent-report or self-report measures, particularly with 375 respect to the mediator and the outcome; and (3) Confounding bias: Despite adjusting for a 376 range of potential confounders, residual or unmeasured confounding (e.g., racism, cultural 377 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 16, 2025. ; https://doi.org/10.1101/2025.07.15.25331608doi: medRxiv preprint 19 norms, policy environment) remains a possibility. 378 379 Implications for future research and practice 380 Our findings suggest that enhancing positive experiences has the potential to reduce 381 socioeconomic inequities in children’s CVH. This study reinforces the importance of 382 strengths-based approaches in epidemiolocal research that examine the positive health assets 383 that allow populations to thrive, including in the face of adversity. 43,44 We focused on a 384 negative outcome in the present study; it is worthwhile to consider using a positive outcome 385 in future to check whether results are consistent. It would be also interesting to explore the 386 potential benefits of intervening on each type of positive experience to reduce socioeconomic 387 inequities in each CVH component. 21 Future work may also consider exploring the potential 388 benefits in other countries and populations such as First Nations children. 389 390 The Australian Government’s Early Years Strategy highlights the importance of a 391 strengths-based approach, leveraging positive resources to support children in reaching their 392 optimal health. 15 The hypothetical intervention in our study that would be capable of 393 achieving an increase from “fewer than two” to “two or more” positive experiences remains 394 undetermined (i.e., what the intervention is in practice and how to deliver it is not specified). 395 Currently, there is increasing attention to programs targeting positive experiences to improve 396 children’s health, such as the Healthy Communities Study in the US 45 and the Kids Building 397 Future Healthy Mission in Australia. 46 It is likely to achieve the maximum impact by 398 combining strategies that promote positive experiences with a multi-faceted and sustained 399 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 16, 2025. ; https://doi.org/10.1101/2025.07.15.25331608doi: medRxiv preprint 20 approach that considers other fundamental drivers of CVH inequities. 400 401

402 This study demonstrates that positive experiences partially mediate the relationship between 403 socioeconomic disadvantage and poor CVH among Australian children. While promoting 404 positive experiences has the potential to reduce socioeconomic inequities in CVH, addressing 405 socioeconomic disadvantage itself and other socially distributed drivers of CVD remain 406 imperative to achieve the maximum impact. 407 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 16, 2025. ; https://doi.org/10.1101/2025.07.15.25331608doi: medRxiv preprint 21

408 1. Coronado F, Melvin SC, Bell RA, Zhao G. Peer Reviewed: Global Responses to Prevent, 409 Manage, and Control Cardiovascular Diseases. Preventing Chronic Disease. 2022;19 410 2. Reddy KS, Mathur MR. Global Burden of CVD. In: Kickbusch I, Ganten D, Moeti M, 411 eds. Handbook of Global Health. Springer International Publishing; 2021:423-437. 412 3. Bann D, Wright L, Hughes A, Chaturvedi N. Socioeconomic inequalities in 413 cardiovascular disease: a causal perspective. Nature Reviews Cardiology . 414 2024;21(4):238-249. 415 4. Ioachimescu OC. From Seven Sweethearts to Life Begins at Eight Thirty: A Journey 416 From Life's Simple 7 to Life's Essential 8 and Beyond. Am Heart Assoc; 2022. p. e027658. 417 5. Lloyd-Jones DM, Hong Y , Labarthe D, et al. Defining and setting national goals for 418 cardiovascular health promotion and disease reduction: the American Heart Association’s 419 strategic Impact Goal through 2020 and beyond. Circulation. 2010;121(4):586-613. 420 6. Lloyd-Jones DM, Allen NB, Anderson CAM, et al. Life’s essential 8: updating and 421 enhancing the American Heart Association’s construct of cardiovascular health: a presidential 422 advisory from the American Heart Association. Circulation. 2022;146(5):e18-e43. 423 7. Qureshi F, Bousquet-Santos K, Okuzono SS, et al. The social determinants of ideal 424 cardiovascular health: A global systematic review: Social determinants of ideal CVH. Annals 425 of Epidemiology. 2022; 426 8. Lloyd-Jones DM, Ning H, Labarthe D, et al. Status of cardiovascular health in US adults 427 and children using the American Heart Association’s New “Life’s Essential 8” Metrics: 428 prevalence estimates from the National Health and Nutrition Examination Survey (NHANES), 429 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 16, 2025. ; https://doi.org/10.1101/2025.07.15.25331608doi: medRxiv preprint 22 2013 Through 2018. Circulation. 2022;146(11):822-835. 430 9. Perng W, Francis EC, Schuldt C, Barbosa G, Dabelea D, Sauder KA. Pre- and Perinatal 431 Correlates of Ideal Cardiovascular Health during Early Childhood: A Prospective Analysis in 432 the Healthy Start Study. (1097-6833 (Electronic)) 433 10. Henriksson P, Henriksson H, Labayen I, et al. Correlates of ideal cardiovascular health in 434 European adolescents: the HELENA study. Nutrition, Metabolism and Cardiovascular 435 Diseases. 2018;28(2):187-194. 436 11. Arteaga SS, Gillman MW. Promoting ideal cardiovascular health through the life span. 437 Pediatrics. 2020;145(4) 438 12. Suglia SF, Campo RA, Brown AGM, et al. Social Determinants of Cardiovascular Health: 439 Early Life Adversity as a Contributor to Disparities in Cardiovascular Diseases. The Journal 440 of pediatrics. Feb 25 2020;doi:10.1016/j.jpeds.2019.12.063 441 13. Suglia SF, Koenen KC, Boynton-Jarrett R, et al. Childhood and adolescent adversity and 442 cardiometabolic outcomes: a scientific statement from the American Heart Association. 443 Circulation. 2018;137(5):e15-e28. doi:10.1161/CIR.0000000000000536 444 14. Ortiz R, Kershaw KN, Zhao S, et al. Evidence for the association between adverse 445 childhood family environment, child abuse, and caregiver warmth and cardiovascular health 446 across the lifespan: the Coronary Artery Risk Development in Young Adults (CARDIA) study. 447 Circulation: Cardiovascular Quality and Outcomes. 2024;17(2):e009794. 448 15. Australian Government. The Early Years Strategy: Discussion Paper . 2023. 449 https://engage.dss.gov.au/wp-content/uploads/2023/02/early-years-strategy-discussion-paper.450 pdf 451 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 16, 2025. ; https://doi.org/10.1101/2025.07.15.25331608doi: medRxiv preprint 23 16. Singh SS, Stranges S, Wilk P, Tang ASL, Frisbee SJ. Influence of the Social Environment 452 on Ideal Cardiovascular Health. Journal of the American Heart Association . 2023/02/21 453 2023;12(4):e026790. doi:10.1161/JAHA.122.026790 454 17. Guo S, O'Connor M, Mensah F, et al. Measuring Positive Childhood Experiences: 455 Testing the structural and predictive validity of the Health Outcomes from Positive 456 Experiences (HOPE) framework. Academic Pediatrics . 2021/11/18/ 2022;22(6):942-951. 457 doi:https://doi.org/10.1016/j.acap.2021.11.003 458 18. Slopen N, Chen Y , Guida JL, Albert MA, Williams DR. Positive childhood experiences 459 and ideal cardiovascular health in midlife: Associations and mediators. Preventive Medicine. 460 2017/04/01/ 2017;97:72-79. doi:http://dx.doi.org/10.1016/j.ypmed.2017.01.002 461 19. Appleton AA, Buka SL, Loucks EB, Rimm EB, Martin LT, Kubzansky LD. A 462 prospective study of positive early-life psychosocial factors and favorable cardiovascular risk 463 in adulthood. Circulation. 2013;127(8):905-912. 464 20. Deer LK, Han D, Maher M, et al. Positive childhood experiences and adult 465 cardiovascular health. Health Psychology. 2025;44(5):489. 466 21. La Charite J, Khan M, Dudovitz R, et al. Specific domains of positive childhood 467 experiences (PCEs) associated with improved adult health: A nationally representative study. 468 SSM-Population Health. 2023;24:101558. 469 22. Perng W, Aris IM, Slopen N, et al. Application of Life’s Essential 8 to assess 470 cardiovascular health during early childhood. Annals of epidemiology. 2023;80:16-24. 471 23. Soloff C, Lawrence D, Johnstone R. LSAC Technical paper No. 1. Sample design. 2005. 472 https://growingupinaustralia.gov.au/sites/default/files/tp1.pdf 473 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 16, 2025. ; https://doi.org/10.1101/2025.07.15.25331608doi: medRxiv preprint 24 24. Clifford SA, Davies S, Wake M. Child Health CheckPoint: Cohort summary and 474 methodology of a physical health and biospecimen module for the Longitudinal Study of 475 Australian Children. BMJ open. 2019;9(Suppl 3):3-22. 476 25. Priest N, Guo S, Gondek D, et al. The potential of intervening on childhood adversity to 477 reduce socioeconomic inequities in body mass index and inflammation among Australian and 478 UK children: A causal mediation analysis. Journal of Epidemiology and Community Health . 479 2023;77(10):632-640. doi:https://doi.org/10.1136/jech-2022-219617 480 26. Sege RD, Browne CH. Responding to ACEs with HOPE: Health outcomes from positive 481 experiences. Academic Pediatrics. 2017;17(7):S79-S85. 482 27. Priest N, Guo S, Gondek D, et al. The effect of adverse and positive experiences on 483 inflammatory markers in Australian and UK children. Brain, Behavior, & Immunity-Health . 484 2022;26:100550. 485 28. Australian Bureau of Statistics. Y ear Book Australia. Australian Bureau of Statistics; 486 2006. 487 29. Department of Social S, Australian Institute of Family S, Australian Bureau of S. 488 Growing Up in Australia: Longitudinal Study of Australian Children (LSAC) Release 8 489 (Waves 1-8). doi:doi/10.26193/VTCZFF http://dx.doi.org/10.26193/VTCZFF 490 30. Hernán MA, Robins JM. Causal inference: what if. Boca Raton: Chapman & Hall/CRC; 491 2020. 492 31. Vansteelandt S, Keiding N. Invited commentary: G-computation–lost in translation? 493 American journal of epidemiology. 2011;173(7):739-742. 494 32. Moreno-Betancur M, Moran P, Becker D, Patton G, Carlin J. Mediation effects that 495 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 16, 2025. ; https://doi.org/10.1101/2025.07.15.25331608doi: medRxiv preprint 25 emulate a target randomised trial: Simulation-based evaluation of ill-defined interventions on 496 multiple mediators. Statistical Methods in Medical Research. 2021;30(6):1395-1412. 497 doi:https://doi.org/10.1177/0962280221998409 498 33. Moreno-Betancur M, Moran P , Becker D, Patton GC, Carlin JB. Mediation effects that 499 emulate a target randomised trial: Simulation-based evaluation of ill-defined interventions on 500 multiple mediators. Statistical Methods in Medical Research . 2021;30(6):1395-412. 501 doi:10.1177/0962280221998409 502 34. medRCT: Causal Mediation Analysis Estimating Interventional Effects Mapped to a 503 Target Trial. 2024. https://t0ngchen.github.io/medRCT/ 504 35. Van Buuren S, Groothuis-Oudshoorn K. mice: Multivariate imputation by chained 505 equations in R. Journal of statistical software. 2011;45:1-67. 506 36. White IR, Royston P, Wood AM. Multiple imputation using chained equations: Issues 507 and guidance for practice. Statistics in Medicine . Feb 20 2011;30(4):377-399. 508 doi:https://doi.org/10.1002/sim.4067 509 37. Dashti SG, Lee KJ, Simpson JA, Carlin JB, Moreno-Betancur M. Handling multivariable 510 missing data in causal mediation analysis estimating interventional effects. Epidemiology. 511 2025;doi:10.1097/EDE.0000000000001866 512 38. Rubin DB. Multiple imputations in sample surveys-a phenomenological Bayesian 513 approach to nonresponse. Proceedings of the survey research methods section of the 514 American Statistical Association. American Statistical Association Alexandria; 1978:20-34. 515 39. Kemp L, Elcombe E, Blythe S, Grace R, Donohoe K, Sege R. The Impact of Positive and 516 Adverse Experiences in Adolescence on Health and Wellbeing Outcomes in Early Adulthood. 517 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 16, 2025. ; https://doi.org/10.1101/2025.07.15.25331608doi: medRxiv preprint 26 International Journal of Environmental Research and Public Health. 2024;21(9):1147. 518 40. Huang CX, Halfon N, Sastry N, Chung PJ, Schickedanz A. Positive childhood 519 experiences and adult health outcomes. Pediatrics. 2023;152(1):e2022060951. 520 41. Guo S, Wijesuriya R, O'Connor M, et al. The effects of adverse and positive experiences 521 on cardiovascular health in Australian children. International Journal of Cardiology . 522 2024:132262. 523 42. Goldfeld SR, O'Connor E, Pham C, Gray S, Changing Children's Chances Investigator G. 524 Beyond the silver bullet: closing the equity gap for children within a generation. Medical 525 Journal of Australia. 2024;221(10):508-511. doi:https://doi.org/10.5694/mja2.52493 526 43. VanderWeele TJ, Chen Y , Long K, Kim ES, Trudel-Fitzgerald C, Kubzansky LD. 527 Positive Epidemiology? Epidemiology. 2020;31(2) 528 44. O'Connor M, Olsson C, Lange K, et al. Progressing “Positive epidemiology”: A 529 cross-national analysis of adolescents’ positive mental health and outcomes during the 530 COVID-19 pandemic. Epidemiology . 2025;36(1):28-39. 531 doi:10.1097/EDE.0000000000001798 532 45. Arteaga SS, Loria CM, Crawford PB, et al. The Healthy Communities Study: Its 533 Rationale, Aims, and Approach. Am J Prev Med . Oct 2015;49(4):615-23. 534 doi:10.1016/j.amepre.2015.06.029 535 46. Victorian Health Promotion Foundation. Kids Building Future Healthy. VicHealth. 20 536 December 2024, Accessed 25 May, 2023. 537 https://www.vichealth.vic.gov.au/programs-and-projects/campaigns-initiatives/future-healthy-538 minecraft 539 540 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 16, 2025. ; https://doi.org/10.1101/2025.07.15.25331608doi: medRxiv preprint 27 541 Figure 1. Directed acyclic graph depicting the assumed causal model conceptualizing the 542 pathway from maternal education to children’s cardiovascular health via positive childhood 543 experiences. 544 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 16, 2025. ; https://doi.org/10.1101/2025.07.15.25331608doi: medRxiv preprint 28 Table 1. Descriptive information for all study variables in our sample (N=1874). Observed 545 data are shown. 546 Frequency (%) Maternal education Missing High Medium Low Exposure Maternal education 0 High 826 (44.1) - - - Medium 619 (33.0) - - - Low 429 (22.9) - - - Mediator Fewer positive experiences (<2) 464 (24.8) No 631 (44.8) 313 (48.7) 206 (44.7) 112 (36.6) Yes 779 (55.2) 330 (51.3) 255 (55.3) 194 (63.4) Outcome Cardiovascular health 815 (43.5) Good 491 (46.4) 251 (49.5) 146 (42.6) 94 (45.0) Poor 568 (53.6) 256 (50.5) 197 (57.4) 115 (55.0) Baseline confounders Child's sex assigned at birth 0 Female 919 (49.0) 412 (49.9) 295 (47.7) 212 (49.4) Male 955 (51.0) 414 (50.1) 324 (52.3) 217 (50.6) Child's ethnicity 0 Anglo or European 1616 (86.2) 696 (84.3) 551 (89.0) 369 (86.0) Minoritized ethnic group 221 (11.8) 126 (15.3) 53 (8.6) 42 (9.8) Indigenous 37 (2.0) 4 (0.5) 15 (2.4) 18 (4.2) Child's disability status 0 No 1785 (95.3) 792 (95.9) 582 (94.0) 411 (95.8) Yes 89 (4.7) 34 (4.1) 37 (6.0) 18 (4.2) Neighbourhood socioeconomic status 0 Not disadvantaged 1322 (70.5) 658 (79.7) 411 (66.4) 253 (59.0) Disadvantaged 552 (29.5) 168 (20.3) 208 (33.6) 176 (41.0) Intermediate confounder Gestation age in weeks 13 (0.7) ≥ 37 weeks 1746 (93.8) 782 (95.5) 568 (92.2) 396 (93.0) <37 weeks 115 (6.2) 37 (4.5) 48 (7.8) 30 (7.0) Maternal age at childbirth 0 ≥ 27 1626 (86.8) 782 (94.7) 519 (83.8) 325 (75.8) <27 248 (13.2) 44 (5.3) 100 (16.2) 104 (24.2) Family composition 0 Two parents 1781 (95.0) 810 (98.1) 574 (92.7) 397 (92.5) Single parent 93 (5.0) 16 (1.9) 45 (7.3) 32 (7.5) Multiple adversities (≥ 2) 170 (9.1) No 1585 (93.0) 718 (94.6) 517 (92.3) 350 (90.9) Yes 119 (7.0) 41 (5.4) 43 (7.7) 35 (9.1) . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 16, 2025. ; https://doi.org/10.1101/2025.07.15.25331608doi: medRxiv preprint 29 547 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 16, 2025. ; https://doi.org/10.1101/2025.07.15.25331608doi: medRxiv preprint 30 Table 2. Generalized linear models examining the associations between maternal education, 548 positive childhood experiences, and poor cardiovascular health, using imputed data 549 (N=1874). 550 Risk ratio (95% CI) Model 1: Unadjusted Model 2: Adjusted for baseline confounders Model 3: Adjusted for baseline and intermediate confounders Model 4: Adjusted for baseline, intermediate confounders and maternal education Association with poor cardiovascular health Low maternal education (ref=high) 1.11 (0.92, 1.34) 1.09 (0.91, 1.32) - - Medium maternal education (ref=high) 1.11 (0.94, 1.30) 1.11 (0.94, 1.31) - - Fewer positive experiences (ref=2 or more) 1.18 (1.01, 1.37) 1.17 (1.01, 1.36) 1.17 (1.00, 1.36) 1.17 (1.00, 1.36) Association with fewer positive experiences Low maternal education (ref=high) 1.23 (1.05, 1.45) 1.19 (1.01, 1.41) - - Medium maternal education (ref=high) 1.10 (0.95, 1.28) 1.08 (0.93, 1.26) - - Baseline confounders controlled for were child’s sex, child’s ethnicity, child’s disability, and neighborhood 551 socioeconomic status. Intermediate confounders controlled for were gestational age in weeks, maternal age at 552 childbirth, family composition, and childhood adversity. CI, confidence interval; -, not applicable. 553 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 16, 2025. ; https://doi.org/10.1101/2025.07.15.25331608doi: medRxiv preprint 31 Table 3. Results of evaluation of mediator interventions to close socioeconomic inequities in 554 children’s poor cardiovascular health (CVH) using the interventional effects approach, using 555 imputed data (N=1874). 556 Group comparison Estimate of absolute risk difference (%) 95% CI p value Proportion of the socioeconomic gap in poor CVH Low versus high maternal education Existing socioeconomic inequities in poor CVH (before intervening on positive childhood experiences) 4.9 (-3.2, 13.0) 0.23 100 Reduction in inequities from intervening on positive childhood experiences 1.0 (-0.8,2.8) 0.26 20.4 Remaining inequities 3.9 (-4.3,12.1) 0.35 79.6 Medium versus high maternal education Existing socioeconomic inequities in poor CVH (before intervening on positive childhood experiences) 5.6 (-1.2,12.5) 0.10 100 Reduction in inequities from intervening on positive childhood experiences 0.5 (-0.5, 1.5) 0.32 8.9 Remaining inequities 5.1 (-1.8,12.1) 0.14 91.1 557 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 16, 2025. ; https://doi.org/10.1101/2025.07.15.25331608doi: medRxiv preprint