The association between social needs and chronic conditions in a large, urban primary care population

Abstract

There is consensus that social needs influence health outcomes, but less is known about the relationships between certain needs and chronic health conditions in large, diverse populations. This study sought to understand the association between social needs and specific chronic conditions using social needs screening and clinical data from Electronic Health Records. Between April 2018-December 2019, 33,550 adult (≥18y) patients completed a 10-item social needs screener during primary care visits in Bronx and Westchester counties, NY. Generalized linear models were used to estimate prevalence ratios for eight outcomes by number and type of needs with analyses completed in Summer 2020. There was a positive, cumulative association between social needs and each of the outcomes. The relationship was strongest for elevated PHQ-2, depression, alcohol/drug use disorder, and smoking. Those with ≥3 social needs were 3.90 times more likely to have an elevated PHQ-2 than those without needs (95% CI: 3.66, 4.16). Challenges with healthcare transportation was associated with each condition and was the most strongly associated need with half of conditions in the fully-adjusted models. For example, those with transportation needs were 84% more likely to have an alcohol/drug use disorder diagnosis (95% CI: 1.59, 2.13) and 41% more likely to smoke (95% CI: 1.25, 1.58). Specific social needs may influence clinical issues in distinct ways. These findings suggest that health systems need to develop strategies that address unmet social need in order to optimize health outcomes, particularly in communities with a dual burden of poverty and chronic disease.

Keywords: Chronic disease; Electronic health records; Mental health; Population health; Primary healthcare; Social determinants of health.

Figure 1.

Unadjusted and adjusted^a prevalence ratio of outcomes by number of unmet needs among screened Montefiore patients, April 2018-December 2019 Figure reflects the unadjusted and adjusted prevalence ratios of selected chronic conditions by number of needs (no needs (referent; white circle); 1 need (white square); 2 needs (gray square); ≥3 needs (black square). Tests for trend used number of needs as a grouped linear variable. ^aAdjusted for age, sex, race/ethnicity, preferred language, health insurance, block-group poverty and public housing status; ^bDefined as those reporting being a current smoker in latest smoking survey in 1 year prior to the screener; excludes 7.5% not surveyed; ^cDefined as those with a diagnosis for the following: F11.x–F16.x, F18.x, F19.x, Z71.5, Z72.2, F10, E52, G62.1, I42.6, K29.2, K70.0, K70.3, K70.9, T51.x, Z50.2, Z71.4, Z72.1 in two years prior to screener; ^dDefined as those with BMI≥30 kg/m², based on latest BMI value in 1 year prior to screener; excludes 2.1% without BMI measurement; ^eDefined as those with I10.x diagnosis in two years prior to screener; ^fDefined as those with E10.X, E11.X, E12.X, E13.X, E14.X diagnosis codes in two years prior to screener; ^gDefined as those with J45.X diagnosis codes in two years prior to screener; ^hDefined as those with F20.4, F31.3–F31.5, F32.x, F33.x, F34.1, F41.X, F43.2 diagnosis codes in two years prior to screener; ⁱDefined as those with a score ≥2, based on latest PHQ-2 score in 1 year prior to screener; excludes 7.5% of sample not screened

Figure 2:

Adjusted prevalence ratio of outcomes by strongest social needs among screened Montefiore patients, April 2018-December 2019 Figure identifies the specific unmet that has the strongest fully-adjusted association with the identified chronic condition. Symbols in the figure indicate need type (healthcare transportation (black diamond); healthcare cost (white square); housing quality (white triangle); getting along (gray circle). ^aAdjusted for age, sex, race/ethnicity, preferred language, health insurance, block-group poverty and public housing status; ^bDefined as any patient reporting being a current smoker in latest smoking survey in 1 year prior to the screener; excludes 7.5% not surveyed; ^cDefined as any patient with a diagnosis for the following: F11.x–F16.x, F18.x, F19.x, Z71.5, Z72.2, F10, E52, G62.1, I42.6, K29.2, K70.0, K70.3, K70.9, T51.x, Z50.2, Z71.4, Z72.1 in two years prior to screener; ^dDefined as BMI≥30 kg/m², based on latest BMI value in 1 year prior to screener; excludes 2.1% without BMI measurement; ^eDefined as any patient with I10.x diagnosis in two years prior to screener; ^fDefined as any patient with E10.X, E11.X, E12.X, E13.X, E14.X diagnosis codes in two years prior to screener; ^gDefined as any patient with J45.X diagnosis codes in two years prior to screener; ^hDefined as any patient with F20.4, F31.3–F31.5, F32.x, F33.x, F34.1, F41.X, F43.2 diagnosis codes in two years prior to screener; ⁱDefined as those with a score ≥2, based on latest PHQ-2 score in 1 year prior to screener; excludes 7.5% of sample not screened