Assessment of Population-Level Disadvantage Indices to Inform Equitable Health Policy

Excess Deaths in California During the COVID-19 Pandemic, by Healthy Places Index Quartile, February 2020-April 2022

OBJECTIVES

Place-based disadvantage indices have been used to assess health disparities and allocate funding and health resources. We assessed excess mortality in California during the COVID-19 pandemic by Healthy Places Index (HPI) quartile, a disadvantage index used by the California Department of Public Health to structure COVID-19 response efforts.

METHODS

We estimated expected deaths from all causes during the COVID-19 pandemic by fitting a quasi-Poisson regression model to actual deaths that occurred from 2014 through 2019. We estimated ranges of excess deaths by calculating differences between actual deaths and (1) the average expected number of deaths and (2) the upper bound of the 95% prediction interval. The percentage of excess deaths equaled the number of excess deaths divided by the corresponding threshold. We reported estimates overall and across demographic groups, stratified by HPI quartile; quartile 4 indicated communities with the most advantaged social and environmental conditions.

RESULTS

From February 2020 through April 2022, the number of excess deaths in California ranged from 81 245 to 107 806, with 93 309 deaths attributed to COVID-19. The number of excess deaths decreased across quartiles, from 27 924 to 35 615 (20.5%-28.0%) in HPI quartile 1 to 7757 to 14 477 (4.6%-9.2%) in HPI quartile 4. The Hispanic or Latine population had a disproportionate percentage of excess deaths across all quartiles, whereas American Indian or Alaska Native, Asian, Black or African American, Native Hawaiian or Pacific Islander, and White (all non-Hispanic) populations had percentage excess death estimates in quartile 2 that were similar to or higher than in quartile 1.

CONCLUSIONS

Health policies should supplement the use of place-based disparity measures with other measures that support groups at high risk for adverse health outcomes residing in more socially and environmentally advantaged communities.

Rethinking Vulnerability: Using Factor Analysis to Assess Census Tract-Level Vulnerability

OBJECTIVES

The Centers for Disease Control and Prevention/Agency for Toxic Substances and Disease Registry (CDC/ATSDR) Social Vulnerability Index (SVI) is used to guide policy making and resource allocation for emergency responses. However, limited research has explored the extent to which each variable contributes to the overall calculation of the SVI. We used a factor analysis approach to determine whether specific drivers of vulnerability exist at the state and national levels.

METHODS

We used the 2020 CDC/ATSDR SVI dataset to perform factor analysis separately for each state and nationally. We determined factor weights and scores and conducted a comparative analysis with CDC/ATSDR SVI. The final SVI for each census tract ranged from 0 to 1, with higher values indicating greater vulnerability.

RESULTS

At the national level, our factor analysis approach identified 4 primary variables that affected vulnerability the most: the percentage of the population below 150% of the federal poverty level (weight, 0.262), with housing cost burden (ie, households that spend >30% of their income on housing-related costs; weight, 0.226), in a racial and ethnic minority group (weight, 0.232), and without a high school diploma (weight, 0.138). However, at the state level, some analyses assigned low weights to the primary national-level drivers.

CONCLUSIONS

Our study highlights the need to consider context-specific vulnerability measures when characterizing community social vulnerability. The factor analysis SVI provides nuanced insight into vulnerability drivers at the national and state levels, laying the groundwork for more precise disaster response planning, resource allocation, and community resilience initiatives.

Area-Level socioeconomic disadvantage and access to primary care: A rapid review

Social drivers of health aggregate geographically, contributing to health inequities that primary care access may mitigate. Two area-level measures of social disadvantage are the Area Deprivation Index and Social Vulnerability Index. This rapid review aimed to assess the association between these measures and primary care access. We conducted a rapid review of US studies published through February 11, 2025. Included studies were categorized as defining primary care access by self-reported access to primary care, geographic accessibility, or utilization. We analyzed 31 studies, of which 68% (N = 9/13 Area Deprivation Index, N = 12/18 Social Vulnerability Index) found that greater area-level social disadvantage was consistently associated with reduced primary care access. This association was most consistently observed in studies measuring primary care access via self-report (N = 2/2), vaccine uptake (N = 5/7), and via a higher odds of using telephone vs audio-visual or in-person primary care in areas of high socioeconomic disadvantage (N = 5/5). These findings have implications for telemedicine payment policy and care redesign. The possible expiration of Medicare's expanded telemedicine reimbursement may disproportionately reduce access points to primary care for individuals living in high socioeconomic disadvantage areas. These findings also support the need for community-level interventions to increase access to primary care administered vaccines.

The Child Opportunity Index and Children's Health: A Meta-Analysis

CONTEXT

Quantifying the impact of place on pediatric health is difficult given the diverse methodologies used to measure place-based social determinants. However, the Child Opportunity Index (COI) is increasingly used to study these relationships.

OBJECTIVE

To synthesize associations between the COI and pediatric health.

DATA SOURCES

Fifteen databases, 4 gray literature sources, and diversitydatakids.org searched from 2014 to 2024.

STUDY SELECTION

US-based observational studies that evaluated children, the COI, and at least 1 pediatric health outcome.

DATA EXTRACTION

Protocol registered with PROSPERO (CRD42023418407). Random-effects models created pooled odds ratios (ORs) comparing very low/low COI to high/very high COI for mortality, emergency department (ED) use, and hospital use. Clinically relevant subgroups were explored.

RESULTS

Most studies (n = 61 of 85; 72%) reported inverse associations between the COI and an adverse outcome. Lower COI was associated with higher odds of mortality (OR, 1.50; 95% CI, 1.31-1.94; tau squared [τ2] = 0.045; 15 associations from 13 studies). Overall, ED visits were similar (OR, 1.38; 95% CI, 0.97-1.95; τ2 = 0.312; 10 associations from 6 studies), but the subgroup of all-cause ED visits were significantly higher among children with lower COI (OR, 1.66; 95% CI, 1.19-2.31; τ2 = 0.198; 7 associations from 5 studies). Select hospitalization subgroups (medical, surgical/trauma, and >30-day rehospitalizations) were significantly associated with COI, but not overall hospitalizations (OR, 1.15; 95% CI, 0.96-1.36; τ2 = 0.090; 12 studies).

LIMITATIONS

Meta-analyses were unadjusted.

CONCLUSIONS

Place is a risk factor for children's mortality and select measures of health care use. Shifting the focus from identifying place-based disparities to cocreating community-engaged strategies that mitigate disparities may effectively advance children's health equity.

The integration of health equity into policy to reduce disparities: Lessons from California during the COVID-19 pandemic

Racial and ethnic minoritized groups and socioeconomically disadvantaged communities experience longstanding health-related disparities in the United States and were disproportionately affected throughout the COVID-19 pandemic. How departments of public health can explicitly address these disparities and their underlying determinants remains less understood. To inform future public health responses, this paper details how California strategically placed health equity at the core of its COVID-19 reopening and response policy, known as the Blueprint for a Safer Economy. In effect from August 2020 to June 2021, "the Blueprint" employed the use of the California Healthy Places Index (HPI), a place-based summary measure of 25 determinants of health constructed at the census tract level, to guide activities. Using California's approach, we categorized the state population by HPI quartiles at the state and within-county levels (HPIQ1 representing the least advantaged, HPIQ4, the most advantaged) from HPI data available to demonstrate how the state monitored crude COVID-19 test, case, mortality, and vaccine rates and unadjusted rate ratios (RR) using equity metrics developed for the Blueprint. Notable patterns emerged. Testing disparities disappeared during the summer and winter surges but resurfaced between surges. Monthly case RR peaked in May 2020 for HPIQ1 compared to HPIQ4 (RR 6.61, 95%CI: 6.41-6.81), followed by mortality RR peaking in June 2020 (RR 5.06, 95% CI: 4.34-5.91). As the pandemic wore on, disparities in unadjusted case and mortality RRs between lower HPI quartiles relative to HPIQ4 reduced but remained. Utilizing a place-based index, such as HPI, enabled a data-driven approach that used a determinants of health lens to identify priority communities, allocate resources, and monitor outcomes based on need during a large-scale public health emergency.

Naming and Framing: Six Principles for Embedding Health Equity Language in Policy Research, Writing, and Practice

Policy Points Science communication and health policy language often fail to adequately define and contextualize systemic barriers-like structural racism and wealth inequity-that contribute to disparities in health outcomes. Health policy practitioners should understand best practices for communicating research and policy findings to various audiences and understand how to disseminate messages that are culturally and linguistically responsive to different community needs. As no perfect term exists, adopting health equity language principles can help health policy practitioners avoid dehumanizing and exclusionary language as well as ill-suited terminology that perpetuates racist systems and leads to inequities in population health.

CONTEXT

Language specificity in research, advocacy, and writing is an important tool to ensure more equitable health policies. All health policy practitioners working at the intersection of health care, health policy, and health equity have a role in upholding ethical standards that promote the use of humanizing, inclusive, and antisupremacist language.

METHODS

We conducted an environmental scan and synthesized themes across commonly used and publicly available health equity language guides to provide specific guidance to health policy practitioners to inform their policy research, analysis, writing, and dissemination.

FINDINGS

We identify and describe six guiding principles to dismantle systems that work against the goals of health equity through policy-focused research, writing, and communications. These principles include avoiding blaming language, contextualizing health inequities, acknowledging that systems are not passive, understanding that one-size-fits-all terminology does not exist, seeking input from community members, and paying attention to omissions.

CONCLUSIONS

Applying these principles will better equip health policy practitioners to develop or inform equitable policies and meaningfully engage in dialogue with community members to advance equitable health policy.

Applying individual- and residence-based equity measures to characterize disparities in crash outcomes

INTRODUCTION

Transportation safety priorities emphasize the importance of incorporating equity into efforts to reduce deaths and injuries. Using integrated data, we investigated relationships between individual- and residence-based measures of equity and rates of crash involvement in New Jersey, 2016-2019.

METHODS

We used statewide integrated data that includes linked crash reports, hospital discharge data, and residence-based equity measures. We calculated crash rates among drivers involved in and injured in a crash by residential census tract. Using generalized Poisson regression, we estimated rate ratios and 95% confidence intervals (aRR, 95% CI) in separate models for race and ethnicity categories and for six previously developed, multi-dimensional equity measures, controlling for driver sex and age.

RESULTS

We identified 1,629,219 drivers involved in crashes of whom 8.3% were injured. Hispanic and non-Hispanic Black drivers had higher rates of crash involvement than non-Hispanic White drivers (aRR, 1.67 [95% CI, 1.65-1.68] and aRR, 1.78 [95% CI, 1.77-1.80], respectively). For community equity measures, drivers who resided in census tracts with poorest equity scores had higher crash rates than those living in census tracts with most favorable equity scores (e.g., Index of Concentration at the Extremes: aRR, 2.10 [95% CI, 2.07-2.12]). We observed similar results for injury crash rates. Model fit improved for both all crashes and injury crashes models after adding each equity measure to baseline.

CONCLUSIONS

Rates of all crashes and injury crashes were consistently higher among drivers of minoritized race and ethnicity groups and among those who lived in less equitable communities. Associations among crash rates and different equity measures provided similar evidence that disparities in traffic safety outcomes are related to inequity.

PRACTICAL APPLICATIONS

The usefulness of individual and residence-based equity measures lies in the opportunity to identify communities with higher crash risks for tailored intervention to improve traffic safety and to reduce disparities.

Neighborhood opportunity and residential instability: associations with mental health in middle childhood

BACKGROUND

Neighborhood quality may contribute to child mental health, but families with young children often move, and residential instability has also been tied to adverse mental health. This study's primary goal was to disentangle the effects of neighborhood quality from those of residential instability on mental health in middle childhood.

METHODS

1,946 children from 1,652 families in the Upstate KIDS cohort from New York state, US, were followed prospectively from birth to age 10. Residential addresses were linked at the census tract level to the Child Opportunity Index 2.0, a multidimensional indicator of neighborhood quality. The number of different addresses reported from birth to age 10 was counted to indicate residential instability, and the change in COI quintile indicated social mobility. Parents completed three assessments of attention-deficit/hyperactivity disorder, problematic behavior, and internalizing psychopathology symptoms at ages 7, 8, and 10. Child and family covariates were selected a priori to adjust sample characteristics, increase estimate precision, and account for potential confounding.

RESULTS

In unadjusted models, higher neighborhood quality at birth was associated with fewer psychopathology symptoms in middle childhood, but associations were largely mediated by residential instability. In adjusted models, residential instability was associated with more psychopathology symptoms, even accounting for social mobility. Neighborhood quality at birth had indirect effects on child mental health symptoms through residential instability.

CONCLUSIONS

Children born into lower-quality neighborhoods moved more, and moving more was associated with higher psychopathology symptoms. Findings were similar across different timings of residential moves, for girls and boys, and for children who did not experience a major life event. Additional research is needed to better understand which aspects of moving are most disruptive to young children.

Neighborhood Social Determinants of Health are Associated With Metabolic Dysfunction-associated Steatotic Liver Disease Outcomes

BACKGROUND & AIMS

Significant health disparities exist in metabolic dysfunction-associated steatotic liver disease (MASLD), driven by social determinants of health (SDOH). Few studies have explored neighborhood-level SDOH in MASLD.

METHODS

This is a retrospective cohort study of patients with MASLD at a multi-state healthcare institution. Primary outcomes were MASLD burden, mortality, and comorbidities by neighborhood SDOH, assessed using the Social Deprivation Index in cross-sectional and longitudinal analyses.

RESULTS

A total of 69,191 patients with MASLD were included, 45,003 of which had over 365 days of follow-up. Patients living in the most disadvantaged neighborhoods, as compared with the least, had higher odds of cirrhosis (adjusted odds ratio [aOR], 1.42; P < .001), any cardiovascular disease (aOR, 1.20; P < .001), coronary artery disease (aOR, 1.17; P < .001), congestive heart failure (aOR, 1.43; P < .001), cerebrovascular accident (aOR, 1.19; P = .001), diabetes mellitus (aOR, 1.57; P < .001), and hypertension (aOR, 1.38; P < .001). They also had increased incidence of death (adjusted hazard ratio [aHR], 1.47; P < .001), liver-related events (aHR, 1.31; P = .012), diabetes mellitus (aHR, 1.47; P < .001), and major adverse cardiovascular events (aHR, 1.24; P < .001). Patients in the most disadvantaged neighborhoods compared to the least were disproportionately Hispanic, Black, and Native American/Alaska Native, more often spoke Spanish as their primary language, and were more often uninsured or had Medicaid. Even after adjustment for Social Deprivation Index, Native American/Alaska Native patients had higher incidence of death, cirrhosis, diabetes, and major adverse cardiovascular events compared with non-Hispanic White patients.

CONCLUSION

Neighborhood-level SDOH are associated with MASLD burden, comorbidities, and mortality and should be considered in clinical care, quality improvement, and further research.

Comparing Deprivation vs Vulnerability Index Performance Using Medicare Beneficiary Surgical Outcomes

Importance

Health care researchers, professionals, payers, and policymakers are increasingly relying on publicly available composite indices of area-level socioeconomic deprivation to address health equity. Implications of index selection, however, are not well understood.

Objective

To compare the performance of 2 frequently used deprivation indices using policy-relevant outcomes among Medicare beneficiaries undergoing 3 common surgical procedures.

Design, Setting, and Participants

This cross-sectional study examined outcomes among Medicare beneficiaries (65 to 99 years old) undergoing 1 of 3 common surgical procedures (hip replacement, knee replacement, or coronary artery bypass grafting) between 2016 and 2019. Index discriminative performance was compared for beneficiaries residing in tracts with high- and low-deprivation levels (deciles) according to each index. Analyses were conducted between December 2022 and August 2023.

Main Outcomes and Measures

Tract-level deprivation was operationalized using 2020 releases of the area deprivation index (ADI) and the social vulnerability index (SVI). Binary outcomes were unplanned surgery, 30-day readmissions, and 30-day mortality. Multivariable logistic regression models, stratified by each index, accounted for beneficiary and hospital characteristics.

Results

A total of 2 433 603 Medicare beneficiaries (mean [SD] age, 73.8 [6.1] years; 1 412 968 female beneficiaries [58.1%]; 24 165 Asian [1.0%], 158 582 Black [6.5%], and 2 182 052 White [89.7%]) were included in analyses. According to both indices, beneficiaries residing in high-deprivation tracts had significantly greater adjusted odds of all outcomes for all procedures when compared with beneficiaries living in low-deprivation tracts. However, compared to ADI, SVI resulted in higher adjusted odds ratios (adjusted odds ratios, 1.17-1.31 for SVI vs 1.09-1.23 for ADI), significantly larger outcome rate differences (outcome rate difference, 0.07%-5.17% for SVI vs outcome rate difference, 0.05%-2.44% for ADI; 95% CIs excluded 0), and greater effect sizes (Cohen d, 0.076-0.546 for SVI vs 0.044-0.304 for ADI) for beneficiaries residing in high- vs low-deprivation tracts.

Conclusions and Relevance

In this cross-sectional study of Medicare beneficiaries, SVI had significantly better discriminative performance-stratifying surgical outcomes over a wider range-than ADI for identifying and distinguishing between high- and low-deprivation tracts, as indexed by outcomes for common surgical procedures. Index selection requires careful consideration of index differences, index performance, and contextual factors surrounding use, especially when informing resource allocation and health care payment adjustment models to address health equity.

Comparing Social Disadvantage Indices in Pediatric Populations

BACKGROUND AND OBJECTIVES

Place-based social disadvantage indices are increasingly used to promote health equity, but vary in design. We compared associations between 3 commonly used indices (Social Vulnerability Index [SVI], Area Deprivation Index [ADI], and Child Opportunity Index [COI]) and infant well-child check (WCC) attendance and adolescent obesity. We hypothesized that the COI would have the strongest association with child health outcomes.

METHODS

We conducted a cross-sectional analysis of 2014-2019 Duke University Health System electronic health record data. Eligible participants were ≤18 years old, had outpatient encounters during the study period, and resided in Durham County, North Carolina. We aggregated indices into deciles; higher deciles represented greater disadvantage. Multivariable logistic regression models quantified the association between each index and infant WCC attendance (ages 0-15 months) and adolescent obesity (11-17 years).

RESULTS

There were 10 175 and 14 961 children in the WCC and obesity cohorts, respectively. All 3 indices were similarly associated with WCCs (SVI odds ratio [OR] 1.10, 95% confidence interval [CI] 1.08-1.12; ADI OR 1.10, 95% CI 1.08-1.12; COI OR 1.12, 95% CI 1.10-1.14) and obesity (SVI OR 1.06, 95% CI 1.04-1.07; ADI OR 1.08, 95% CI 1.06-1.10; COI OR 1.07, 95% CI 1.05-1.08). ORs indicate the increase in the outcome odds for every 1-decile index score increase.

CONCLUSIONS

Higher disadvantage as defined by all 3 indices was similarly associated with adolescent obesity and decreased infant WCC attendance. The SVI, ADI, and COI may be equally suitable for pediatric research, but population and outcome characteristics should be considered when selecting an index.

Comparison of Six Measures of Genetic Similarity of Interspecific Brassicaceae Hybrids F2 Generation and Their Parental Forms Estimated on the Basis of ISSR Markers

Genetic similarity determines the extent to which two genotypes share common genetic material. It can be measured in various ways, such as by comparing DNA sequences, proteins, or other genetic markers. The significance of genetic similarity is multifaceted and encompasses various fields, including evolutionary biology, medicine, forensic science, animal and plant breeding, and anthropology. Genetic similarity is an important concept with wide application across different scientific disciplines. The research material included 21 rapeseed genotypes (ten interspecific Brassicaceae hybrids of F2 generation and 11 of their parental forms) and 146 alleles obtained using 21 ISSR molecular markers. In the presented study, six measures for calculating genetic similarity were compared: Euclidean, Jaccard, Kulczyński, Sokal and Michener, Nei, and Rogers. Genetic similarity values were estimated between all pairs of examined genotypes using the six measures proposed above. For each genetic similarity measure, the average, minimum, maximum values, and coefficient of variation were calculated. Correlation coefficients between the genetic similarity values obtained from each measure were determined. The obtained genetic similarity coefficients were used for the hierarchical clustering of objects using the unweighted pair group method with an arithmetic mean. A multiple regression model was written for each method, where the independent variables were the remaining methods. For each model, the coefficient of multiple determination was calculated. Genetic similarity values ranged from 0.486 to 0.993 (for the Euclidean method), from 0.157 to 0.986 (for the Jaccard method), from 0.275 to 0.993 (for the Kulczyński method), from 0.272 to 0.993 (for the Nei method), from 0.801 to 1.000 (for the Rogers method) and from 0.486 to 0.993 (for the Sokal and Michener method). The results indicate that the research material was divided into two identical groups using any of the proposed methods despite differences in the values of genetic similarity coefficients. Two of the presented measures of genetic similarity (the Sokal and Michener method and the Euclidean method) were the same.

Aging, Race, and Health Disparities: Recommendations From the Research Centers Collaborative Network

Racial disparities in adverse health outcomes with aging have been well described. Yet, much of the research focuses on racial comparisons, with relatively less attention to the identification of underlying mechanisms. To address these gaps, the Research Centers Collaborative Network held a workshop on aging, race, and health disparities to identify research priorities and inform the investigation, implementation, and dissemination of strategies to mitigate disparities in healthy aging. This article provides a summary of the key recommendations and highlights the need for research that builds a strong evidence base with both clinical and policy implications. Successful execution of these recommendations will require a concerted effort to increase participation of underrepresented groups in research through community engagement and partnerships. In addition, resources to support and promote the training and development of health disparities researchers will be critical in making health equity a shared responsibility for all major stakeholders.

Place-Based Child Opportunity at Birth and Child Development from Infancy to Age 4

OBJECTIVE

The objective of this study was to evaluate whether the children's neighborhood quality, as a measure of place-based social determinants of health, is associated with the odds of developmental delay and developmental performance up to the age of 4 years.

STUDY DESIGN

Mothers of 5702 children from the Upstate KIDS Study, a longitudinal population-based cohort of children born from 2008 through 2010, provided questionnaire data and a subset of 573 children participated in a clinic visit. The Child Opportunity Index 2.0 was linked to home census tract at birth. Probable developmental delays were assessed by the Ages and Stages Questionnaire up to 7 times between 4 and 36 months, and developmental performance was assessed via the Battelle Developmental Inventory at the age of 4 years.

RESULTS

In unadjusted models, higher neighborhood opportunity was protective against developmental delays and was associated with slightly higher development scores at age 4. After adjusting for family-level confounding variables, 10-point higher Child Opportunity Index (on a 100-point scale) remained associated with a lower odds of any developmental delay (OR = .966, 95% CI = .940-.992), and specifically delays in the personal-social domain (OR = .921, 95% CI = .886-.958), as well as better development performance in motor (B = 0.79, 95% CI = 0.11-1.48), personal-social (B = 0.64, 95% CI = 0.003-1.28), and adaptive (B = 0.69, 95% CI = 0.04-1.34) domains at age 4.

CONCLUSIONS

Community-level opportunities are associated with some aspects of child development prior to school entry. Pediatric providers may find it helpful to use neighborhood quality as an indicator to inform targeted developmental screening.

Cultural context index: A geospatial measure of social determinants of health in the United States

Minority populations will continue to grow in the United States. Such pluralism necessitates iterative, geospatial measurements of cultural contexts. Our objective in this study was to create a measure of social determinants of health in geographic areas with varying ethnic, linguistic, and religious diversity in the United States. We extracted geographic information systems data based on community characteristics that have known associations with population health disparities from 2015 to 2019. We used principal component analysis to construct a Cultural Context Index (CCI). We created the CCI for 73,682 census tracts across 50 states and five inhabited territories. We identified hot and cold spots that are the highest and lowest CCI quintile, respectively. Hot spots census tracts were mostly located in metropolitan areas (84.8%), in the Southern census region (41.5%), and also had larger Black and Hispanic populations. The census tracts with the greatest need for culturally competent health care also had the sickest populations. Census tracts with a CCI rank of 5 ('greatest need') had higher prevalences of self-reported poor physical health (17.2%) and poor mental health (17.4%), compared to either the general population (13.9% and 14.5%) or to CCI rank of 1 ('lowest need') (11.9% and 10.8%). The CCI can pinpoint census tracts with a need for culturally competent health care and inform supply-side policy planning as healthcare and social service providers will inevitably come in contact with consumers from different backgrounds.

Quantifying Associations Between Child Health and Neighborhood Social Vulnerability: Does the Choice of Index Matter?

Importance

Policymakers have increasingly utilized place-based social disadvantage indices to quantify the impacts of place on health and inform equitable resource allocation. Indices vary in design, content, and purpose but are often used interchangeably, potentially resulting in differential assignments of relative disadvantage depending on index choice.

Objective

To compare associations between three commonly used disadvantage indices (Social Vulnerability Index (SVI), Area Deprivation Index (ADI), and Child Opportunity Index (COI)) and two epidemiologically distinct child health outcomes-infant well-child check (WCC) attendance and adolescent obesity.

Design

Cross-sectional analysis of Duke University Health System electronic health record (EHR) data from January 2014 to December 2019.

Participants

Children ≤18 years of age with outpatient encounters between January 2014 and December 2019, and who were Durham County residents were eligible. WCC attendance was assessed for infants ages 0-15 months; obesity was assessed for children ages 11-17 years.

Exposures

2014 Social Vulnerability Index (SVI), 2015 Area Deprivation Index (ADI), and 2015 Child Opportunity Index (COI) 2.0.

Main Outcomes

1) Infant WCC attendance: attending less than the minimum recommended six WCCs in the first 15 months of life, and 2) Adolescent obesity: BMI ≥ the 95th percentile at both the most recent encounter and an encounter within the prior 9-36 months.

Results

Of 10175 patients in the WCC cohort, 20% (n = 2073) had less than six WCCs. Of 14961 patients in the obesity cohort, 20% (n = 2933) had obesity. All three indices were associated with both WCCs (OR for SVI 1.10, 95% CI 1.08-1.12; OR for ADI 1.10, 95% CI 1.08-1.12; OR for COI 1.12, 95% CI 1.10-1.14) and obesity (OR for SVI 1.05, 95% CI 1.04-1.08; OR for ADI 1.08, 95% CI 1.06-1.10; OR for COI 1.07, 95% CI 1.05-1.08).

Conclusions and relevance

Higher social disadvantage as defined by all three indices was similarly associated with both adolescent obesity and decreased infant WCC attendance. While the COI incorporates a broader set of child-specific variables, the SVI and ADI may often be just as suitable for pediatric research. Users should consider population and outcome characteristics when selecting an index.