Social vulnerability index, poverty, and asthma in a prospective study of Puerto Rican youth

The Social Vulnerability Index (SVI) is a quantitative indicator of the socioeconomic and demographic factors that influence the resilience of communities¹. Although higher social vulnerability has been associated with worse asthma outcomes^2–4, no study has examined SVI and asthma in Puerto Rican youth, a group disproportionately affected with asthma⁵. We examined whether a higher SVI is associated with asthma and severe asthma exacerbations (SAEs) in a prospective study of Puerto Rican youth, and whether such association is modified by social determinants at the individual level.

Subject recruitment and procedures for the PROspective study of Puerto Rican youth and Asthma study (PROPRA) have been previously described⁶. In brief, 678 children ages 6 to 14 years with (n=351) and without (n=327) asthma were recruited from randomly selected households in San Juan and Caguas (PR) for a study conducted between 2009 and 2010 (PR-GOAL); and a second study (EVA-PR) was conducted between 2014 and 2017. Of the 543 participants in EVA-PR (269 with and 274 without asthma), 406 had previously participated in PR-GOAL and were included in PROPRA⁶. The median time between the two PROPRA visits was 5.2 (interquartile range=4.7–6.1) years. All studies were approved by the institutional review boards of the University of Puerto Rico and the University of Pittsburgh. Written parental consent and child assent were obtained for participants younger than 18 years, and written consent was obtained from participants 18 years and older. A total of 405 children with valid addresses were included in this analysis. Census track-level SVI from 2014 was retrieved from the CDC and constructed using 15 social factors⁷. Each tract receives a separate ranking for each of four themes (socioeconomic status, household composition and disability, minority status and language, and housing type and transportation) and an overall ranking. The overall SVI and the score for each theme ranges from 0 (least vulnerable) to 1 (most vulnerable).

All participants completed a protocol including administration of questionnaires and spirometry at the baseline and follow-up visits. One of the child’s caretakers (most often the mother) completed questionnaires about SES and the child’s respiratory health and household characteristics, and a semiquantitative food frequency questionnaire (FFQ). A dietary score, ranging from −2 (“unhealthiest diet”) to +2 (“healthiest diet”) was derived using FFQ data, and an unhealthy diet was defined as a nonpositive score⁸.

Body mass index (BMI) z-score was calculated using CDC growth charts. Spirometry was conducted with an EasyOne^® spirometer according to ATS/ERS recommendations modified for children. The annual average PM_2.5 level was estimated from Global PM₂.₅ satellite 12 months prior to the second visit. As in prior work, residential distance to a major road was used as a marker of traffic-related air pollution (TRAP) at the baseline visit⁹. Residential distance to a major road was classified as within vs. greater than 441 meters (the first to third quartiles vs the upper quartile)⁹.

Asthma was defined as physician-diagnosed asthma and ≥1 episode of wheeze in the year prior to the follow-up visit, and SAEs were defined as ≥1 emergency department (ED)/urgent care visit or ≥1 hospitalization for asthma in the year prior to the follow-up visit. Logistic regression was used for the multivariable analysis of the SVI at the baseline visit and asthma, which was adjusted for age, sex, parental asthma, second-hand smoke (SHS) at the baseline visit, persistent (i.e. at both visits) overweight or obesity, a persistently unhealthy diet, average PM_2.5 exposure in the year prior to the follow-up visit, and the time interval between visits. Models for SAEs in the year prior to the follow-up visit were adjusted for age, sex, persistent overweight/obesity, use of inhaled corticosteroids in the previous 6 months, average PM_2.5 exposure in the previous year, and the time between visits. All analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC).

At the baseline visit, children with asthma (cases), were significantly more likely to be male and to have parental history of asthma, an unhealthy diet, higher overall SVI, and higher BMI z-score but lower %predicted FEV₁ and FVC than children without asthma (controls)(data not shown).

In a multivariable analysis, participants with an overall SVI in the 3^rd and 4^th quartiles had similarly increased odds of asthma compared with those with an SVI in the 1^st quartile. Based on this result, we repeated the analysis after dichotomizing the overall SVI as at or above the median. In this analysis, participants with an overall SVI at/above the median had 1.65 times increased odds of asthma (95% confidence interval [CI]=1.02–2.68). This association was unchanged after additional adjustment for residential proximity to a road.

Next, we tested for interactions between the overall SVI and individual social determinants of health, including type of health insurance, annual household income, parental education, and perceived poverty. Given potential interactions between the overall SVI and household income or perceived poverty on asthma (P for interaction <0.10 in both instances), we stratified the analysis by household income or perceived poverty. In these analyses, an overall SVI at/above the median was significantly associated with asthma in those with an annual household income <$15,000 (odds ratio [OR]=2.73, 95% CI=1.39–5.38) but not in those with an income ≥$15,000 (the median household income in Puerto Rico in 2008–2009). Similarly, an overall SVI at/above the median was significantly associated with asthma in those perceiving themselves as almost poor or poor (OR=5.28, 95% CI=1.04–26.88) but not in others (Table 1). Moreover, the SVI was not significantly associated with SAEs.

Table 1.

Multivariable analysis of the social vulnerability index and asthma, stratified by annual household income or perceived poverty at the baseline visit.

	Overall social vulnerability index
	Q1-Q2	Q3-Q4
Social determinants at the individual level	OR (95% confidence interval), P-value
Annual household income
< $15,000	1.0	2.74 (1.39, 5.38), <0.01
≥ $15,000	1.0	0.97 (0.36, 2.63), 0.95
Perceived poverty
Almost poor/poor	1.0	5.28 (1.04, 26.88), 0.04
Live well/comfortably/check to check	1.0	1.45 (0.86, 2.44), 0.17

All models adjusted for age, sex, parental history of asthma, and exposure to second-hand smoke at the baseline visit, persistent (i.e., at both study visits) overweight or obesity, a persistently unhealthy diet, average exposure to PM_2.5 in the year before the follow-up visit, and the time interval between study visits.

Our study has substantial strengths, including prospective data in a high-risk underserved population and ability to control for potential confounders such as SHS exposure and PM_2.5 and proximity to a major road. We also recognize several limitations. First, our negative findings for SAEs may be due to limited statistical power to detect modest effects. Second, residential census tracts were used for SVI estimates, which may not capture neighborhoods where children spend most of their time. Third, the SVI did not account for some factors associated with asthma, such as housing characteristics¹⁰ or neighborhood violence. Lastly, our findings may not be generalizable to Puerto Ricans living in the US mainland.

In this study, a higher SVI (a composite measure of neighborhood social vulnerability) was associated with persistent or new-onset asthma in Puerto Rican youth followed for ~5 years. Further, this association was only significant in youth with low household income or greater perceived poverty, suggesting that youth who live in vulnerable communities and whose households are economically deprived are at highest asthma risk. Large longitudinal studies are needed to corroborate our findings and identify the role of individual components of SVI on asthma in Puerto Ricans and other underserved populations.

Abbreviations:

ATS

American Thoracic Society

BMI

body mass index

CDC

Center for Disease Control and Prevention

CI

confidence interval

ED

emergency department

ERS

European Respiratory Society

FFQ

food frequency questionnaire

OR

odds ratio

PM2.5

Particulate Matter less than 2.5 micrometers in diameter

PR

Puerto Rican

SAE

severe asthma exacerbations

SHS

second-hand smoke

SVI

Social vulnerability index

TRAP

traffic-related air pollution