Health and Health Care Use Strongly Associated with Cumulative Burden of Social Determinants of Health

Abstract

Understanding health outcomes and patterns of health care utilization associated with patients' cumulative social determinant of health (SDOH) risk is essential to supporting better health care. This study compared mental and physical health outcomes and health care utilization by increasing number of social needs among a clinical adult population. Surveys were sent to 6000 patients with recent visits to 7 primary care clinics in Portland, Oregon in 2018. The final study sample included respondents who matched to medical claims data, N = 1748. The authors used a modified logistic regression model to estimate risk ratios for the relationship between cumulative SDOH factors and self-reported chronic conditions, and a 2-part model to estimate the effects of cumulative SDOH risk on health care utilization. Increased SDOH need was associated with increasing likelihood of worse self-reported health outcomes, especially mental health. Compared with those with no SDOH need, having 1–2 SDOH need(s) (adjusted risk ratio [aRR] 1.25; 95% confidence interval [CI]: 1.06–1.46) and 3 or more SDOH needs (aRR 1.45; 95% CI: 1.22–1.73) had a greater risk of reporting any behavioral health condition. However, the number of SDOH had a graded but inverse impact on use of mental health care services where fewer visits were observed among those using care. Having SDOH was associated with increased likelihood of having an emergency department visit and increased number of primary care visits. This study demonstrates the compounding impact of SDOH on health and health care use. This highlights the importance of collecting SDOH, including the total number of SDOH needs, when considering a patient's health and health care.

Introduction

Acknowledging both an ethical and fiscal responsibility to address social needs among patient populations,¹ health care organizations have started incorporating social determinants of health (SDOHs) as a key dimension of care provision.² SDOHs refer to the upstream social factors—such as income, education, housing security, and food security—that influence health and health disparities.³ These social conditions have been identified as fundamental causes of health inequity⁴ and a significant body of evidence connects SDOHs to poorer health outcomes⁵ and increased health care utilization.^6,7 SDOHs are so highly impactful on health outcomes that high-quality well-coordinated medical care alone cannot ensure optimal health.³

Decades of research have revealed a stepwise relationship between SDOH and health such that increased magnitude of a social need is associated with poorer health outcomes.³ However, these studies tend to employ unidimensional measures of SDOH need such as income, education, or occupational status.⁸ Similarly, unidimensional measures of SDOH have been used to show the association between increased health care expenditure alongside rising social needs.⁹ Although these studies provide strong evidence of the impact of SDOHs on health outcomes and health care utilization, their findings are limited as socially complex patients often experience multiple SDOH needs simultaneously.¹⁰ Increasingly, studies seek to understand cumulative SDOH risk and its relationship to health outcomes and health care utilization.^11
–13

Understanding health outcomes and patterns of health care utilization associated with patients' cumulative SDOH risk is essential to supporting the provision of better health care especially among socially complex patient populations.¹⁴ Increased knowledge of cumulative SDOH risk would allow for the prioritization of vulnerable patients into care coordination, better aligning outreach, and intervention to those who need it most.^15,16 In addition, deeper insight into cumulative SDOH risk could improve the development of risk prediction models used to estimate health care need and utilization among patient populations.¹⁷

Despite the need for increased understanding of cumulative SDOH risk on health outcomes and care utilization, little research has been done to comprehensively assess this relationship in a health care setting.^11,13,16 Prior examinations of cumulative SDOH risk used restricted samples of specific subpopulations such as college-aged students¹⁸ and older adults,¹⁹ or have focused on specific health outcomes such as diabetes.^11,16 Even so, these previous studies point to the relationship between compounding social need and poorer physical and behavioral health outcomes.^{12,13,16,18,19} Prior evidence also indicates that cumulative SDOH risk among certain populations is associated with health care utilization, finding increased emergency department use and decreased use of preventive medicine services.^13,19

This study expands on the prior studies by incorporating multiple domains of social need to explore cumulative social risk among a clinical adult population. Using both administrative and self-reported data, the study objectives were to compare both mental and physical health outcomes and health care utilization for those with and without an increasing number of social needs among a clinical adult population oversampled for low-income patients.

Methods

Study design and sample

The study uses data collected for an evaluation originally designed to assess the impact of a clinic-based community resource desk on short-term health outcomes, health care utilization, and costs. The evaluation employed a cross-sectional study survey among a sample of patients with recent visits to 7 primary care clinics in Portland, Oregon metro area.

Patients were sampled using pseudo-random sampling stratified by insurance type to create the study population and maximize the number of people surveyed who would have unmet social and medical need. Patients were divided by insurance type and randomly sampled according to a predefined ratio (75% Medicaid, 15% Medicare, and 10% Commercial). Medicaid patients were oversampled because they are more likely to experience SDOH factors.¹³ This study was reviewed and approved by the Providence Institutional Review Board (#2017000640).

Data collection

Participants were selected from electronic health records between August 2018 and February 2019 in 4 waves of 1500 for a total population of 6000. Patients were required to be at least 18 years old and have had a recent (within a 30-day window) visit to 1 of the 7 study clinics. Each patient was assigned a unique identifier, mailed a paper survey with a $5 cash incentive, and simultaneously sent an e-mail to complete an online survey option. A reminder letter, e-mail, and phone call follow-up was conducted for nonrespondents. In total, 2380 individuals completed the survey (39.7% response rate).

Insurance identifiers from survey respondents' electronic health records were sent to the health plan of the health care system and matched, where possible, to the health claims data. Claims data were requested for the period from January 2018 to December 2019, so respondents could have at least 6 months of health care utilization either side of their qualifying clinic visit. In total, 1899 individuals (79.8% of respondents) were matched to health claims data. Respondents with <6 months of claims eligibility before their visit were excluded from analysis giving a final claims sample of 1748 (73.4% of respondents).

Study measures

The survey instrument was designed in partnership with the community health division of the nonprofit health care system. The survey included self-reported items on chronic physical and behavioral health conditions, a checklist of SDOH and medical-related need (based on services offered by the clinic-based community resource desk), and a set of demographic questions (Supplementary Appendix SA).

The predictor for the study was constructed from a list of 8 SDOH needs. Participants were asked if they had recently needed assistance with food, utility costs, transportation, clothing, housing/rent, services for children, jobs or employment, or education classes. Responses were summed (1 = needed assistance, 0 = no need) and categorized into “No SDOH need,” “1–2 SDOH needs,” and “3 or more SDOH needs.”

Survey outcomes comprised the set of items on self-reported chronic conditions; these conditions include diabetes, asthma, chronic obstructive pulmonary disease (COPD), high blood pressure, heart disease, depression, anxiety, post-traumatic stress disorder (PTSD), bipolar disorder, and addiction issues. Behavioral and physical conditions were also combined to create a measure of any reported behavioral condition, any reported physical condition, and any reported condition.

Health care outcomes for the 12-month period were constructed, using modified HEDIS metrics, to divide health care utilization into several unique categories: inpatient stays, emergency department visits, outpatient mental health visits, primary care visits, and specialty care visits. All utilization data are expressed as per member per year (PMPY).

Statistical analyses

Descriptive statistics were used to analyze the demographic makeup of each of the 3 SDOH risk groups. With only a few exceptions, all demographic information came from survey responses; age and insurance status were identified in the electronic health records. The authors also descriptively explored how differently SDOH need was reported in the groups with 1–2 need and 3 or more needs.

A generalized linear model with a modified Poisson distribution was used to estimate risk ratios for the relationship between self-reported chronic conditions and cumulative SDOH risk. The authors controlled for demographic differences in the sample by adjusting for age, gender, race, highest education level, hours worked, marital status, insurance status, health insurance need, and other medical need. To maintain power in the analysis, age was used as a continuous variable and categories for gender and race were collapsed to binary variables.

Owing to the unusually large number of legitimate zero values in health care claims data, a 2-part model was used to estimate the effects of cumulative SDOH risk on health care utilization. The first part of the model estimated the probability of utilization with the same generalized linear model used to estimate chronic conditions. The second part of the model estimated the frequency of utilization among those who had utilization with a simple linear model. Both parts of the model adjusted for age, gender, race, highest education level, hours worked, marital status, insurance status, health insurance need, and other medical need. All analyses were conducted in R 4.0.3.²⁰

Results

The majority of the sample was >50 years of age, White, female, and married. The sample also largely did not work and were Medicaid insured. The demographics were mostly similar across the SDOH groups population (Table 1). Those without SDOH need were older (56.3 years) than those with 1–2 SDOH needs (53.4 years) and 3 or more SDOH needs (50.7 years). As SDOH risk increased respondents were more likely to identify as female and were less likely to be White. Respondents were also less likely to work and be married with increasing SDOH need. The Medicaid sample was much larger in the respondents with 3 or more SDOH need (94.9%), compared with 1–2 SDOH needs (79.5%) and no SDOH need (50.7%).

Table 1.

Demographics by Social Determinant of Health Need

Count	N	No SDOH need		1–2 SDOH needs		3 or more SDOH needs
		1571		474		335
		%	95% CI	%	95% CI	%	95% CI
Age group
18–29	252	11.01	9.56–12.66	11.18	8.64–14.35	7.76	5.34–11.15
30–39	331	12.86	11.29–14.61	14.77	11.85–18.25	17.61	13.9–22.07
40–49	306	10.82	9.38–12.46	15.40	12.42–18.94	18.81	14.97–23.35
50–59	404	14.26	12.62–16.08	19.20	15.9–22.99	26.57	22.11–31.56
60–65	344	14.32	12.68–16.14	14.77	11.85–18.25	14.63	11.23–18.83
>65	743	36.73	34.38–39.14	24.68	21.01–28.77	14.63	11.23–18.83
Mean (SD)	2380	56.30	18.84	53.41	18.15	50.73	15.08
Gender
Male	657	31.30	29.06–33.64	26.67	22.88–30.83	21.47	17.4–26.19
Female	1579	67.61	65.26–69.88	72.04	67.83–75.9	74.85	69.92–79.2
Transgender or gender nonconforming	10	0.34	0.14–0.79	0.43	0.11–1.68	0.92	0.3–2.77
I prefer to self-describe	24	0.74	0.42–1.31	0.86	0.33–2.25	2.76	1.46–5.18
Race
White	1638	74.18	71.96–76.29	70.11	65.83–74.06	68.32	63.15–73.08
Hispanic	132	5.98	4.91–7.26	5.59	3.85–8.06	5.59	3.58–8.62
Black	100	3.33	2.55–4.34	7.31	5.29–10.03	5.28	3.34–8.25
Asian	149	7.68	6.46–9.1	5.59	3.85–8.06	3.11	1.7–5.61
Other	240	8.83	7.53–10.34	11.40	8.83–14.59	17.70	13.98–22.16
Highest education level
Less than high school	193	7.57	6.36–8.98	9.33	7.02–12.3	12.23	9.13–16.19
High school diploma	634	28.08	25.92–30.36	30.59	26.6–34.88	25.39	21.02–30.33
GED or a high school equivalent	212	7.63	6.42–9.06	11.71	9.11–14.93	14.42	11.05–18.6
Vocational training or 2-year degree	548	23.52	21.49–25.68	26.25	22.48–30.4	25.71	21.31–30.66
A 4-year college degree or more	660	33.20	30.91–35.57	22.13	18.61–26.09	22.26	18.12–27.03
Hours worked
I don't currently work	1309	53.83	51.35–56.28	63.58	59.15–67.79	69.54	64.4–74.23
<20 hours per week	176	7.10	5.93–8.48	8.62	6.41–11.51	9.85	7.08–13.53
20–39 hours per week	305	14.41	12.76–16.24	12.28	9.62–15.56	11.38%	8.4–15.25
≥40 hours per week	463	24.66	22.59–26.85	15.52	12.53–19.06	9.23	6.56–12.83
Marital status
Married or domestic partnership	975	51.59	49.12–54.06	31.59	27.56–35.92	21.36	17.3–26.08
Single, never married	559	20.95	19.01–23.03	29.85	25.9–34.12	34.98	30.06–40.25
Widowed/divorced/separated	672	26.44	24.32–28.68	34.64	30.49–39.04	38.08	33.03–43.4
Something else	51	1.02	0.62–1.65	3.92	2.5–6.09	5.57	3.57–8.6
Language
English	2034	89.45	87.83–90.87	91.94	89.12–94.07	90.06	86.37–92.84
Other	225	10.55	9.13–12.17	8.06	5.93–10.88	9.94	7.16–13.63
Insurance status
Other (incl Medicare/Commercial)	888	49.27	46.8–51.74	20.46	17.07–24.34	5.07	3.18–8.01
Medicaid	1492	50.73	48.26–53.2	79.54	75.66–82.93	94.93	91.99–96.82
Health insurance need
No	1907	90.92	89.4–92.25	74.30	70.18–78.04	66.46	61.23–71.32
Yes	364	9.08	7.75–10.6	25.70	21.96–29.82	33.54	28.68–38.77
Other medical need
No	1193	67.34	64.98–69.62	32.55	28.48–36.9	14.33	10.97–18.5
Yes	1078	32.66	30.38–35.02	67.45	63.1–71.52	85.67	81.5–89.03
Type of medical need
Quitting tobacco	145	2.64	1.95–3.56	9.21	6.91–12.16	19.21	15.34–23.78
Counseling	419	10.43	9.02–12.04	26.12	22.37–30.27	43.60	38.38–48.96
Alcohol or drug recovery	71	0.88	0.52–1.49	4.07	2.62–6.27	11.89	8.84–15.82
Eye care	569	15.72	14–17.6	34.69	30.54–39.09	53.35	47.99–58.64
Dental care	643	16.94	15.16–18.87	40.69	36.35–45.17	61.89	56.57–66.94

CI, confidence interval; SD, standard deviation; SDOH, social determinant of health.

Respondents with more SDOH need also reported more medical need, most notably with health insurance (No SDOH need; 9.1%, 1–2 SDOH needs; 25.7%, 3 or more SDOH needs; 33.5%), counseling (No SDOH need; 10.4%, 1–2 SDOH needs; 26.1%, 3 or more SDOH needs; 43.6%), and dental care (No SDOH need; 16.9%, 1–2 SDOH needs; 40.7%, 3 or more SDOH needs; 61.9%).

Those reporting 3 or more SDOH needs had higher proportions of individual needs for every SDOH domain measured (Table 2). With the exception of education, these proportions were at least 2-fold higher in the 3 or more SDOH needs group. The biggest differences were with utility costs (1–2 needs; 20.9%, 3 or more needs; 79.7%), housing/rent (1–2 needs; 13.5%, 3 or more needs; 73.1%), and clothing (1–2 needs; 3.2%, 3 or more needs; 43.6%).

Table 2.

Comparison of Social Determinant of Health Needs Between Those with 1–2 Needs and Those with 3 or More Needs

Count	1–2 SDOH needs		3 or more SDOH needs
	474		335
	%	95% CI	%	95% CI
Social needs
Food	34.60	30.45–39	91.64	88.16–94.17
Utility costs	20.89	17.46–24.78	79.70	75.05–83.67
Transportation	30.80	26.81–35.11	68.66	63.49–73.4
Clothing	3.16	1.92–5.18	43.58	38.36–48.95
Housing/rent	13.50	10.71–16.88	73.13	68.13–77.61
Services for children	5.70	3.93–8.18	15.22	11.76–19.48
Jobs or employment	12.24	9.58–15.51	31.94	27.17–37.13
Education^a	10.49	8.04–13.59	15.85	12.32–20.17

For example, health, GED, language.

After adjusting for potential confounding variables, only asthma and COPD remained strongly related to cumulative SDOH risk for self-reported physical health conditions (Table 3). Those with 3 or more SDOH needs had 1.45 (95% confidence interval [CI] 1.07–1.96) times the risk of asthma and 1.69 (95% CI 1.09–2.62) times the risk of COPD compared with those with no SDOH need. The difference was smaller when looking at the binary of having any self-report physical health conditions, where those with 1–2 SDOH needs and 3 or more SDOH needs had 1.17 (95% CI 1.00–1.35) and 1.20 (95% CI 1.00–1.44) times the risk, respectively.

Table 3.

Prevalence of Chronic Conditions by Social Determinant of Health Need

	Prevalence			Logistic regression^a
	N	%	95% CI	aRR	95% CI	P
Diabetes						0.3019
No SDOH need	1571	18.97	17.11–20.98	1.00	—
1–2 SDOH needs	474	25.95	22.2–30.09	1.21	0.95–1.54
3 or more SDOH needs	335	24.18	19.89–29.05	1.14	0.84–1.54
Asthma						0.0180
No SDOH need	1571	12.99	11.41–14.74	1.00	—
1–2 SDOH needs	474	24.05	20.42–28.1	1.38	1.06–1.79
3 or more SDOH needs	335	26.87	22.39–31.87	1.45	1.07–1.96
COPD						0.0349
No SDOH need	1571	6.05	4.97–7.34	1.00	—
1–2 SDOH needs	474	11.60	9.02–14.81	1.49	1.01–2.18
3 or more SDOH needs	335	13.43	10.18–17.52	1.69	1.09–2.62
High blood pressure						0.1385
No SDOH need	1571	38.70	36.32–41.14	1.00	—
1–2 SDOH needs	474	42.41	38.03–46.9	1.11	0.92–1.33
3 or more SDOH needs	335	42.09	36.91–47.45	1.25	1–1.56
Heart disease						0.6453
No SDOH need	1571	11.84	10.33–13.53	1.00	—
1–2 SDOH needs	474	13.29	10.52–16.66	1.12	0.8–1.56
3 or more SDOH needs	335	11.34	8.36–15.21	1.20	0.78–1.82
Depression						<0.0001
No SDOH need	1571	25.97	23.86–28.2	1.00	—
1–2 SDOH needs	474	46.20	41.76–50.71	1.27	1.06–1.53
3 or more SDOH needs	335	68.36	63.18–73.12	1.60	1.31–1.95
Anxiety						0.0008
No SDOH need	1571	28.33	26.15–30.61	1.00	—
1–2 SDOH needs	474	47.89	43.42–52.39	1.20	1–1.43
3 or more SDOH needs	335	66.57	61.34–71.42	1.46	1.2–1.77
PTSD						<0.0001
No SDOH need	1571	8.59	7.3–10.08	1.00	—
1–2 SDOH needs	474	22.57	19.03–26.56	1.57	1.18–2.08
3 or more SDOH needs	335	37.91	32.87–43.23	2.13	1.59–2.86
Bipolar disorder						0.0118
No SDOH need	1571	3.44	2.64–4.46	1.00	—
1–2 SDOH needs	474	6.33	4.46–8.91	0.82	0.49–1.35
3 or more SDOH needs	335	14.93	11.5–19.16	1.65	1.05–2.61
Addiction issues						0.0001
No SDOH need	1571	4.39	3.48–5.52	1.00	—
1–2 SDOH needs	474	11.60	9.02–14.81	1.43	0.96–2.11
3 or more SDOH needs	335	22.09	17.97–26.85	2.31	1.57–3.42
Any reported physical condition						0.0569
No SDOH need	1571	54.49	52.02–56.94	1.00	—
1–2 SDOH needs	474	67.09	62.73–71.17	1.17	1–1.35
3 or more SDOH needs	335	65.07	59.81–69.99	1.20	1–1.44
Any reported behavioral condition						0.0001
No SDOH need	1571	36.98	34.63–39.4	1.00	—
1–2 SDOH needs	474	61.18	56.71–65.47	1.25	1.06–1.46
3 or more SDOH needs	335	79.70	75.05–83.67	1.45	1.22–1.73
Any reported condition						0.0358
No SDOH need	1571	67.66	65.31–69.93	1.00	—
1–2 SDOH needs	474	85.02	81.52–87.96	1.14	1–1.3
3 or more SDOH needs	335	89.25	85.46–92.15	1.20	1.03–1.4

Adjusted for age (continuous), gender (male/other), race (White/other), highest education level, hours worked, marital status, insurance status, health insurance need, and other medical need.

aRR, adjusted risk ratio; COPD, chronic obstructive pulmonary disease; PTSD, post-traumatic stress disorder.

In contrast, after adjustment all 5 reported behavioral health conditions remained strongly related with cumulative SDOH risk. Depression (adjusted risk ratio [aRR] 1.60; 95% CI 1.31–1.95), PTSD (aRR 2.13; 95% CI 1.59–2.86), and addiction issues (aRR 2.31; 95% CI 1.57–3.42) showed the biggest increases in risk for those with 3 or more SDOH needs when compared with those with no SDOH need. Those with 1–2 SDOH need (aRR 1.25; 95% CI 1.06–1.46) and 3 or more SDOH needs (aRR 1.45; 95% CI 1.22–1.73) had a greater risk of reporting any behavioral health condition when compared with those with no SDOH need.

For utilization of health care, the authors first examined the proportion with at least 1 visit for the measured domains of care in the past 12 months (Table 4). After adjusting for potential confounders, those with 3 or more SDOH needs had at a 1.61 (95% CI 1.18–2.20) times greater likelihood of utilizing the emergency department than those with no SDOH need (Table 4). SDOH need did not produce any differences in the proportion utilizing any other measured domains of health care, including primary care, specialty care, outpatient mental health care, and inpatient care.

Table 4.

Proportion of Population Utilizing Health Care by Social determinant of Health Need

	N	Proportion		Logistic regression^a
	N	%	95% CI	aRR	95% CI	P
Inpatient						0.1340
No SDOH need	1122	6.77	5.44–8.4	1.00	—
1–2 SDOH needs	355	9.30	6.68–12.79	1.34	0.81–2.16
3 or more SDOH needs	271	9.59	6.61–13.72	1.76	1–3.05
Emergency department						0.0083
No SDOH need	1122	19.52	17.3–21.94	1.00	—
1–2 SDOH needs	355	23.10	19.01–27.77	1.07	0.79–1.43
3 or more SDOH needs	271	33.58	28.21–39.41	1.61	1.18–2.2
Outpatient mental health						0.8335
No SDOH need	1122	6.15	4.89–7.71	1.00	—
1–2 SDOH needs	355	5.92	3.89–8.9	0.95	0.53–1.64
3 or more SDOH needs	271	7.01	4.52–10.73	1.16	0.61–2.14
Primary care physician						0.2414
No SDOH need	1122	64.71	61.86–67.45	1.00	—
1–2 SDOH needs	355	62.54	57.38–67.42	0.99	0.83–1.17
3 or more SDOH needs	271	71.59	65.93–76.64	1.17	0.95–1.42
Specialty care						0.3455
No SDOH need	1122	50.53	47.61–53.46	1.00	—
1–2 SDOH needs	355	47.89	42.73–53.09	1.01	0.83–1.23
3 or more SDOH needs	271	50.55	44.62–56.47	1.18	0.93–1.49

Adjusted for age (continuous), gender (male/other), race (White/other), highest education level, hours worked, marital status, insurance status, health insurance need, and other medical need

Next, among those with utilization in each domain, the authors examined the frequency of use defined as number of visits PMPY (Table 5). Those with 1–2 SDOH needs (8.99 visits PMPY) and 3 or more SDOH needs (7.10 visits PMPY) had lower frequency of outpatient mental health utilization than those with no SDOH need (11.13 visits PMPY). Those with 1–2 SDOH needs (5.19 visits PMPY) and 3 or more SDOH needs (5.29 visits PMPY) had more primary care visits than those with no SDOH needs (4.32 visits PMPY). A trend in increasing number of ED visits for those with greater SDOH needs was also observed, but there was greater variance in use and the result was not statistically significant. There were no differences in frequency of utilization by cumulative SDOH risk for inpatient and specialty care events.

Table 5.

Frequency of Health Care Utilization for Those Using Care by Social Determinant of Health Need

	N	PMPY		Linear regression^a
	N	Mean	SE	Estimate	SE	P
Inpatient						0.5121
No SDOH need	76	1.802	0.119	—	—
1–2 SDOH needs	33	1.698	0.165	0.230	0.254
3 or more SDOH needs	26	1.511	0.163	0.308	0.288
Emergency department						0.3235
No SDOH need	219	2.903	0.174	—	—
1–2 SDOH needs	82	4.482	1.171	1.304	0.868
3 or more SDOH needs	91	3.780	0.319	0.572	0.886
Outpatient mental health						0.0291
No SDOH need	69	11.131	1.372	—	—
1–2 SDOH needs	21	8.992	2.034	−3.350	3.299
3 or more SDOH needs	19	7.102	1.902	−9.505	3.578
Primary care physician						0.0160
No SDOH need	726	4.316	0.131	—	—
1–2 SDOH needs	222	5.186	0.291	0.840	0.324
3 or more SDOH needs	194	5.285	0.247	0.794	0.364
Specialty care						0.1244
No SDOH need	567	4.411	0.215	—	—
1–2 SDOH needs	170	4.846	0.457	1.003	0.533
3 or more SDOH needs	137	4.612	0.345	0.914	0.610

Adjusted for age (continuous), gender (male/other), race (White/other), highest education level, hours worked, marital status, insurance status, health insurance need, and other medical need.

PMPY, per member per year; SE, standard error.

Discussion

In this study, the authors found that increasing number social determinant risk factors were associated with increasing likelihood of worse self-reported health outcomes, especially mental health. The number of SDOH had a graded but inverse impact on use of mental health care services, where more SDOH needs was associated with decreasing frequency of use of mental health care among those engaged in care. They also found that having high SDOH needs was associated with greater likelihood of having an emergency department visit and greater number of primary care visits; however, a compounding relationship was not present for these outcomes. These findings support the strong impact of SDOH on health and health care and demonstrates how cumulative social risk can worsen outcomes.

The authors found the strongest relationship between the number of SDOH needs and self-reported mental health burden. This is consistent with previous studies showing that SDOH such as economic insecurity is associated with negative mental health outcomes.^21,22

Social determinant impacts on mental health are thought to manifest through increased stress and allostatic load that impacts regulation of multiple biological systems.^23,24 Increasing number and persistence of stressors can lead to further maladaptive physiological responses and worsen health outcomes.²³ This study shows that increasing number of SDOH needs was associated with increasing likelihood of reporting poor mental health, including depression, PTSD, and addiction issues. Thus, the compounding impact of SDOH may manifest through increased dysregulation stemming from physiological responses to stress, but additional research is needed to further understanding in this area.

Research has also shown a 2-way relationship where having poor mental health can increase the likelihood of experiencing negative social determinants.²⁵ For example, research has shown that having mental health challenges in young adulthood can impact education, employment, and involvement with the criminal/legal system, and these experiences can have long-term impacts that impede socioeconomic improvement in the future.^26,27 Thus, it is possible that the population with high SDOH needs already had high mental health burden and that this contributes to their current SDOH risk factors.

Despite the strong relationship between SDOH and mental health need, the authors found that increasing SDOH was not associated with an increased proportion using mental health care. However, among those using mental health care, increased SDOH need was associated with decreased frequency of visits. This suggests that those with more SDOH have greater mental health needs but are not getting as much mental health care as those without SDOH needs. Research indicated that barriers to mental health care exist broadly and suggest that globally 70% of young people and adults with mental illness receive no mental health care.²⁸ Reasons for not accessing care include lack of knowledge to identify mental illness or treatment options and prejudice and stigma about people with mental illness.^28,29

The current data indicate that SDOH may also be a key reason for not accessing mental health care. This could stem from the stigma and poor health care experience associated with facing SDOH challenges³⁰ as well as from the complexity of accessing mental health care when managing experiences such as housing instability, financial strain, lack of transportation, and food insecurity. Thus, understanding the type and amount of SDOH challenges patients are facing when managing mental health issues should be a key consideration for treatment.

The current research also demonstrate an impact of SDOH on physical health outcomes and physical health care; however, the relationship was not as strong as with mental health and the authors did not observe a compounding relationship with the number of SDOH needs. The physical health diagnoses associated with SDOH were ones that are often connected to environmental conditions such as asthma.³¹ The authors did observe some impact of SDOH on use of physical health care where there was increased proportion of using of emergency care and increased frequency of outpatient care visits.

There was some evidence of increased number of emergency care visits, but the variance in the data suggests the need for further investigation. Overall the findings are consistent with previous studies focused on specific SDOH such as homelessness, where chronically homeless populations show increased use of emergency department care.³² Although not as profound as mental health, these data still suggest that SDOH impacts physical health and health care and should be considered in the context of care and treatment.

The current data clearly demonstrate the importance of capturing SDOH in the context of health care. Furthermore, not only does SDOH need to be captured, but also the number of SDOH needs to be considered as they have a compounding impact. This is similar to other known social risk factors such as adverse childhood experiences (ACEs), where increasing number of ACEs is associated with increased risk of poor health outcomes.³³ Many health care organizations are recognizing the importance of collecting SDOH information to better understand health risk in their population and to better provide whole-person care.^34
–36 Some health care organizations are also finding ways to address SDOH as evidence suggests that they can help improve outcomes.

For example, there is extensive evidence demonstrating the connection between stable affordable housing and health and health care use and cost.^37,38 As a result, several health care organizations have begun to invest in housing and housing supports.³⁹ Thus, collecting SDOH information and having a strategy to address SDOH is especially important as health care is increasingly held accountable for population outcomes through value-based care that focused payment on outcomes instead of number of services.

Many efforts to standardize and leverage available tools to support systematic and consistent collection of SDOH in health care exist. Some health care organizations have created SDOH screening modules in their electronic health records systems to support data collection. Standard SDOH screening tools have been created such as Protocol for Responding to and Assessing Patients' Assets, Risks, and Experiences (PRAPARE) that consists of a set of actionable SDOH questions, including housing stability, income, and social integration.⁴⁰ Natural language processing and machine learning approaches are also being developed to mine chart notes for SDOH information.³⁶ Collectively, these efforts demonstrate the movement to better incorporate SDOH into health care and broader adoption across the nation could further the health care system's ability to improve effectiveness of care.

Limitations

This study has several limitations. The population is limited to clinics in the Portland-metro area in Oregon. Although a random sample was selected, oversampled for Medicaid, from this clinic population, the results may not be broadly generalizable. Moreover, there was not sufficient power in the analysis to stratify by race, gender, or age which limits the ability to make inferences about experiences and outcomes for these groups. The collection of SDOH data relied on survey response that is subject to response bias and social desirability bias.

Survey data were also used to capture health conditions, which could have led to underreporting due to lack of diagnosis or understanding of their health conditions. The authors did not assess the magnitude or duration of SDOH experiences and suggest that is considered in future studies. They also did not explore whether specific types of SDOH emerged as most salient to health outcomes and utilization and also recommend this for future investigation. Finally, further research is needed to assess causation as well as understanding optimal approaches for addressing SDOH to improve health and health care outcomes.

Conclusion

This study demonstrates the compounding impact of SDOH on health and health care including a gradient effect on mental health whereby increasing number of SDOH was associated with increasing self-reported mental health need. The current data suggest that it is not only important to collect SDOH when considering a patient's health and health care, but also the number of SDOH needs should be considered as part of their health care risk and care. As health care embraces whole-person care, connection with support for social needs will likely be a key ingredient in supporting better health outcomes.

Footnotes

Acknowledgments

The authors thank the research staff who helped with this study such as development, fielding surveys, tracking responses, including Bill Wright, Aisha Gilmore, Heather Polonsky, Lauren Broffman, and Sheetal Kulkarni.

Author Disclosure Statement

The authors declare that there are no conflicts of interest.

Funding Information

The authors wish to express appreciation to their funders: University of California, San Francisco (UCSF) Social Interventions Research & Evaluation Network (SIREN) Innovation Grants (73799) and Providence Community Health Division.

Supplementary Material

Supplementary Appendix SA

References

Backrach

, Pfister

, Wallis

, Lipson

. Addressing patients' social needs: an emerging business case for Provider

Investment

. 2014. https://www.commonwealthfund.org/publications/fund-reports/2014/may/addressing-patients-social-needs-emerging-business-case-provider Accessed April 30, 2021 .

Solomon

Health care steps up to social determinants of health: current context. Perm J, 2018; 22:18–139.

Braveman

, Egerter

, Williams

. The social determinants of health: coming of age. Annu Rev Public Health, 2011; 32:381–398.

Phelan

, Link

, Tehranifar

. Social conditions as fundamental causes of health inequalities: theory, evidence, and policy implications. J Health Soc Behav, 2010; 51(1_suppl):S28–S40.

Adler

, Glymour

, Fielding

. Addressing social determinants of health and health inequalities. JAMA, 2016; 316:1641.

Hatef

, Ma

, Rouhizadeh

, Singh

, Weiner

, Kharrazi

. Assessing the impact of social needs and social determinants of health on health care utilization: using patient- and community-level data. Popul Heal Manag, 2021; 24:222–230.

Mick

, Alcusky

, Li

N-C

, et al. Complex patients have more emergency visits: don't punish the systems that serve them. Med Care, 2021; 59:362–367.

Braveman

, Cubbin

, Egerter

, Williams

, Pamuk

. Socioeconomic disparities in health in the United States: what the patterns tell us. Am J Public Health, 2010; 100(Suppl. 1):S186–S196.

Tarasuk

, Cheng

, De Oliveira

, Dachner

, Gundersen

, Kurdyak

. Association between household food insecurity and annual health care costs. Can Med Assoc J, 2015; 187:E429–E436.

10.

Singh

, Daus

, Allender

, et al. Social determinants of health in the United States: addressing major health inequality trends for the nation, 1935–2016. Int J MCH AIDS, 2017; 6:139–164.

11.

Berkowitz

, Meigs

, DeWalt

, et al. Material need insecurities, control of diabetes mellitus, and use of health care resources: results of the measuring economic insecurity in Diabetes Study. JAMA Intern Med, 2015; 175:257–265.

12.

Bisgaier

, Rhodes

. Cumulative adverse financial circumstances: associations with patient health status and behaviors. Health Soc Work, 2011; 36:129–137.

13.

McQueen

, Li

, Herrick

, et al. Social needs, Chronic conditions, and health care utilization among Medicaid beneficiaries. Popul Health Manag. [Epub ahead of print];DOI:10.1089/POP.2021.0065

14.

Kreuter

, Thompson

, McQueen

, Garg

. Addressing social needs in health care settings: evidence, challenges, and opportunities for public health. Annu Rev Public Health, 2021; 42:329–344.

15.

Gottlieb

, Sandel

, Adler

. Collecting and applying data on social determinants of health in health care settings. JAMA Intern Med, 2013; 173:1017.

16.

Cottrell

, O'Malley

, Dambrun

, et al. The impact of social and clinical complexity on diabetes control measures. J Am Board Fam Med, 2020; 33:600–610.

17.

Chen

, Tan

, Padman

. Social determinants of health in electronic health records and their impact on analysis and risk prediction: a systematic review. J Am Med Informatics Assoc, 2020; 27:1764–1773.

18.

Johnson

, Brookover

, Borden

General health and depression outcomes related to cumulative indicators of unmet social determinant of health needs for college students. J College Stud Psychother, 2021:1–22. [Epub ahead of print]; DOI: 10.1080/87568225.2021.1881752.

19.

Rhee

, Marottoli

, Cooney Jr

, Fortinsky

. Associations of social and behavioral determinants of health index with self-rated health, functional limitations, and health services use in older adults. J Am Geriatr Soc, 2020; 68:1731–1738.

20.

R Core Team. A language and environment for statistical computing. 2017. http://www.r-project.org/ Accessed November 23, 2021 .

21.

Benoit

, McCarthy

, Jansson

. Occupational stigma and mental health: discrimination and depression among front-line service workers. Can Public Policy, 2015; 41(Supplement 2):S61–S69.

22.

Silva

, Loureiro

, Cardoso

. Social determinants of mental health: a review of the evidence. Eur J Psychiatry, 2016; 30:259–292.

23.

Schneiderman

, Ironson

, Siegel

. Stress and health: psychological, behavioral, and biological determinants. Annu Rev Clin Psychol, 2005; 1:607–628.

24.

Beckie

TM.

A systematic review of allostatic load, health, and health disparities. Biol Res Nurs, 2012; 14:311–346.

25.

Bährer-Kohler

Social determinants and mental health. Hauppauge, NY: Nova Science Publishers, 2011.

26.

Livingston

. Mental illness-related structural stigma: the downward spiral of systemic exclusion. 2013. https://www.mentalhealthcommission.ca/sites/default/files/MHCC_OpeningMinds_MentalIllness-RelatedSructuralStigmaReport_ENG_0_0.pdf Accessed November 23, 2021 .

27.

Hale

, Viner

. How adolescent health influences education and employment: investigating longitudinal associations and mechanisms. J Epidemiol Community Heal, 2018; 72:465–470.

28.

Henderson

, Evans-Lacko

, Thornicroft

. Mental illness stigma, help seeking, and public health programs. Am J Public Health, 2013; 103:777–780.

29.

Thornicroft

Stigma and discrimination limit access to mental health care. Epidemiol Psichiatr Soc, 2008; 17:14–19.

30.

Gray

, Jones

, Wright

. Patients with health-related social needs more likely to report poor clinic experiences. J Patient Exp, 2021; 8:23743735211008307.

31.

Williams

, Sternthal

, Wright

. Social determinants: taking the social context of asthma seriously. Pediatrics, 2009; 123(Supplement 3):S174–S184.

32.

Moore

, Rosenheck

. Comprehensive services delivery and emergency department Use among chronically homeless adults. Psychol Serv, 2017; 14:184–192.

33.

Felitti

, Anda

, Nordenberg

, et al. Relationship of childhood abuse and household dysfunction to many of the leading causes of death in adults. The Adverse Childhood Experiences (ACE) Study. Am J Prev Med, 1998; 14:245–258.

34.

Gold

, Cottrell

, Bunce

, et al. Developing Electronic Health Record (EHR) strategies related to health center patients' social determinants of health. J Am Board Fam Med, 2017; 30:428–447.

35.

Cantor

, Thorpe

. Integrating data on social determinants of health into electronic health records. Health Aff, 2018; 37:585–590.

36.

Deferio

, Breitinger

, Khullar

, Sheth

, Pathak

. Social determinants of health in mental health care and research: a case for greater inclusion. J Am Med Informatics Assoc, 2019; 26:895–899.

37.

Wright

, Vartanian

, Li

H-F

, Royal

, Matson

. Formerly homeless people had lower overall health care expenditures after moving into supportive housing. Health Aff, 2017; 35:20–27.

38.

Martinez

, Burt

. Impact of permanent supportive housing on the use of acute care health services by homeless adults. Psychiatr Serv, 2006; 57:992–999.

39.

Mercy

Housing

, Low Income Investment

Fund

. Innovative models in health and housing. 2017. https://www.mercyhousing.org/wp-content/uploads/2018/11/Mercy_Innovation-Models-in-Health-and-Housing-2017.pdf Accessed November 23, 2021 .

40.

National Association of Community Health Centers. PRAPARE—NACHC. 2020. https://www.nachc.org/research-and-data/prapare/ Accessed July 27, 2021 .

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.06 MB