A Surrogate Measure for Social Determinants of Health? The Utility of an Area Deprivation Index for Social Needs Screening

Background: Social determinants of health (SDoH) drive 40-60% of the majority of health outcomes, and clinicians indicate uniform agreement on the importance of social needs assessment during clinical encounters. However, universal screening has been shown to overwhelm referral processes. One solution may be to identify subpopulations of patients that would most benefit from comprehensive assessment, including non-English speakers. The Brokamp Area Deprivation Index (ADI) is a validated measure of environmental deprivation that that has been used to risk-stratify patients but has not yet been applied for screening. The measure can be derived based on one's residential address extracted from the electronic health record. Therefore, our objectives are to investigate the association of the Brokamp ADI with patients at higher risk for adverse SDoH and to examine differences between English and non-English speaking patients.

Methods: This prospective study enrolled 300 pediatric patient-families between December 2020 and April 2021 at a large private academic institution. Participants completed the Health Leads Social Needs survey over the phone or electronically in English, Spanish, or Arabic. A composite score was calculated based on the sum of “yes” responses for specific SDoH domains. The Brokamp ADI was derived using a principal component analysis that included six census tract level variables. Statistical analyses included Spearman’s correlations between ADI and composite Health Leads score for the sample and for each language subgroup, t-tests to compare the mean ADIs of “yes” and “no” responders for each Health Leads subitem, and a comparison of the lowest and highest quartile of ADI using a t-test of mean composite Health Leads scores and Pearson’s χ2 tests of each Health Leads subitem.

Results: Of 300 patient-families, 150 (50.0%) were English-speaking, 128 (42.7%) were Spanish-speaking, and 22 (7.3%) were Arabic-speaking. The mean ADI across the sample was 0.407, s.d. 0.120. Spearman’s correlations between Brokamp ADI and composite Health Leads score were as follows: total (rs=0.12; p=0.04), English (rs=0.10; p=0.2), Spanish (rs=0.04; p=0.6), and Arabic (rs=0.44; p=0.04). Mean ADIs of “yes” and “no” responders were significantly different for childcare (p=0.027) and transportation challenges (p=0.016). Analyzing the lowest (Q1) and highest (Q4) quartiles of ADI, mean composite Health Leads scores were significantly lower for Q1 than for Q4 (1.3 vs. 2.0, respectively; p=0.04; n=149). Pearson’s χ2 test demonstrated a significant difference for transportation challenges (p=0.016).

Conclusion: In this prospective study, the Brokamp ADI has a weak but statistically significant association for specific adverse SDoH domains, including childcare and transportation. Further investigation with additional tests to re-examine non-normal associations and a larger sample of Arabic-speaking patients is required to determine the clinical utility of the Brokamp ADI as a screen to identify those patients who may benefit from more comprehensive SDoH assessment.