Skip to main content
  • Research article
  • Open access
  • Published:

Social risks and social needs in a health insurance exchange sample: a longitudinal evaluation of utilization



Health systems are increasingly attempting to intervene on social adversity as a strategy to improve health care outcomes. To inform health system efforts to screen for social adversity, we sought to explore the stability of social risk and interest in assistance over time and to evaluate whether the social risk was associated with subsequent healthcare utilization.


We surveyed Kaiser Permanente members receiving subsidies from the healthcare exchange in Southern California to assess their social risk and desire for assistance using the Accountable Health Communities instrument. A subset of initial respondents was randomized to be re-surveyed at either three or six months later.


A total of 228 participants completed the survey at both time points. Social risks were moderate to strongly stable across three and six months (Kappa range = .59-.89); however, social adversity profiles that included participants’ desire for assistance were more labile (3-month Kappa = .52; 95% CI = .41-.64 & 6-month Kappa = .48; 95% CI = .36-.6). Only housing-related social risks were associated with an increase in acute care (emergency, urgent care) six months after initial screening; no other associations between social risk and utilization were observed.


This study suggests that screening for social risk may be appropriate at intervals of six months, or perhaps longer, but that assessing desire for assistance may need to occur more frequently. Housing risks were associated with increases in acute care. Health systems may need to engage in screening and referral to resources to improve overall care and ultimately patient total health.

Peer Review reports


Social determinants of health (SDoH) are defined as factors in a person’s life that affect their health and ability to engage in healthcare [1, 2]. Mounting research suggests SDoH are more accurate predictors of premature mortality than genetics or healthcare access [3,4,5,6,7,8,9,10]. As a result of the compelling evidence that social adversity significantly influences health and costs of care [11,12,13,14,15,16], health systems are increasingly attempting to intervene on social adversity as a strategy to improve health care outcomes. Health system activities often begin with screening patients for social risks (e.g., housing insecurity, food insecurity, financial strain). However, insufficient research exists to inform how frequently screening must occur to accurately capture peoples’ experience of social adversity. Few longitudinal studies of social risks exist in the absence of an intervention to identify patterns over time.

As social adversity screening has increased in health care settings, several studies have documented that patients who endorse social risk, such as the experience of financial strain, may in parallel decline related assistance from their healthcare system [17]. The reasons for this are likely varied and may include peoples’ priorities and competing demands, prior experience with social services, and prior experience with their healthcare system. It is unclear whether this phenomenon is unique to a subgroup of people, severity, or type of social risk domain, or if it reflects a transitional period wherein people may go on to endorse interest in assistance if the social risk persists. Ideally, health systems would help people address a given social risk before it has adverse effects on their health and wellness.

Screening results are often used to identify people who may benefit from additional services or referrals to community-based organizations [17]. However, these interventions can be difficult to staff without a better understanding of the stability of the social adversity experience. Moreover, a recent review of health system social risk interventions found that while most interventions improved relevant outcomes, few studies had even explored the impact on healthcare utilization [17]. Evidence from a recent randomized clinical trial of a 3-month social risk intervention did not reveal reductions in healthcare utilization [18]. One potential reason for the lack of impact on healthcare utilization could be that both social risk and patterns of healthcare use, are dynamic [19].

To address these gaps in the literature, this manuscript reports on a longitudinal survey of social risk in a subpopulation of US residents insured via the subsidized exchange in an integrated healthcare system. This manuscript was guided by three aims: (1) to examine the stability of social risk by key domains (e.g., transportation, housing, food security) over time; (2) to explore how people’s social risk and interest in assistance with social risk change in relation to one another over time; and (3) to evaluate if and how endorsement of social risk is associated with subsequent healthcare utilization. We anticipate that findings from this study could help inform national social care-related initiatives by shedding light on the ideal timing of repeated social health screening efforts, the types of social health interventions in which people may be most interested, and the business case to justify additional staffing needs to bolster these efforts.



The study sample was recruited from 134,355 Kaiser Permanente (KP) Southern California members who obtained insurance coverage through the subsidized exchanges as of September 1, 2018. This population is an important, yet understudied group, who have income levels that preclude their eligibility from many governments and/or community financial assistance programs but are still low enough to experience financial strain and related social risks. The subsidized exchanges provided health insurance to people who met specific eligibility criteria, including income (between 100 and 400% of the Federal Poverty Level), ineligibility for Medicaid or other public health care programs, US residency status, and availability of health insurance coverage through their employer—a program that was created under the Affordable Care Act. Potential participants (N = 1,008) were randomly selected across age (18–26, 27–44, 45–61, 62 and over), gender, and language (English vs. Spanish) strata.

Survey overview and fielding procedures

The survey development is described in more detail elsewhere [20], as are the fielding procedures [21]. To address stability in social risk, the survey was administered once and then repeated at either three or six month (participants were randomly assigned to follow up time frame) to explore differential change over two-time intervals that may reflect practical screening windows. Participants received an advance letter along with a $2 pre-incentive and an online web link to the survey. Participants received follow-up emails, phone calls, text messages, and if requested, a paper survey in the mail. Those who completed at least 50% of the survey received a $20 cash incentive. The study was part of ongoing quality improvement efforts. As such, it was deemed not human subjects research after administrative review by the KP International Review Board (IRB) (Human Subjects Assurance Number/Federal Wide Assurance Number: 00002344); informed consent was not required.


Social Risks

Social risks were measured using items from the Accountable Health Communities (AHC) instrument developed by the Centers for Medicare and Medicaid Innovation [22]. The AHC social risk screening tool has six items assessing housing security, housing quality (e.g. mold, pests, smoke detectors not working), transportation difficulties, food insecurity (2 items), and trouble paying for utilities. This relatively newer tool has been utilized in a variety of settings and populations [22]. Only one study has explored evidence of psychometric strength [23] in which it demonstrated convergent validity with a similar tool used widely in KP [21].

Interest in Assistance with Social Risks

One survey item asked if participants desired assistance with the following social risks, an endorsement of which was considered a social need: housing, food access, transportation (medical and non-medical), and utilities. Participants could check multiple risk factors for which they wanted assistance, or they could check ‘none.’

Healthcare Utilization Outcomes

The use of healthcare was captured from electronic health records (EHR) and claims. Data were pulled for the 12 months before and the six months after the initial survey. Healthcare utilization outcomes included: primary care, specialty care, urgent care, and emergency department visits.

Analytic strategy

Stability of Social Risk

For Aim 1 analyses, participants were coded as either having the risk or not having the risk at the initial and follow-up survey. These risks were then characterized as “stable,” “improving,” or “worsening” between these two survey points. We used the bias- and prevalence-adjusted Kappa statistic (PABAK) to assess the stability of reported social risk over time within each domain [24, 25]. Because roughly half of the sample completed each follow-up (3 months and 6 months), Kappa statistics were produced within each subsample (i.e., initial survey to 3 months and initial survey to 6 months). Kappa values can be interpreted as follows: No agreement = 0-0.20; Minimal = 0.21-0.39; Weak = 0.40-0.59; Moderate = 0.60-0.79; Strong = 0.80-0.90; Almost Perfect = above 0.90 [26].

Social Risk and Needs

For Aim 2 analyses, participants were categorized as having at least one social risk, if they endorsed any of the risk factors in any of the domains on the AHC screening tool. Participants were categorized as having a social need, if they wanted help with at least one social risk; otherwise, they were categorized as not having a need. Participants were categorized into four groups at both the initial survey time point and at follow-up: no social risk/need; at least one social risk/no need; no social risk/need assistance with at least one domain; and social risk/need. We used Cohen’s Kappa statistic to assess the stability of the social risk/need groups over time.

Utilization Outcomes

To assess which outcomes were prospectively associated with specific social risks, we conducted regression analyses with each healthcare utilization outcome as a dependent variable in separate models. Poisson regression models controlled for age, gender, race/ethnicity, the Neighborhood Deprivation Index [6], and language, as well as the value for the respective outcome in the 12 months before the survey. Each of the five social risks from the initial survey timepoint were entered as a dichotomous predictor of healthcare utilization in the six months following the survey. We also considered results with a Type I error correction. We used a Bonferroni correction across all 20 significance tests for the healthcare utilization outcomes resulting in an alpha level of 0.0025 for each test with the Type I error correction.


Out of 1008 subsidized exchange members invited to complete the survey, 442 completed at least part of the survey (45% response rate) for the initial timepoint, of which 355 were eligible for randomization to complete a follow-up survey at either 3 months or 6 months later. Two hundred and twenty-eight participants completed a follow-up survey; 190 surveys were fielded at 3 months for a total of 124 completed surveys (response rate = 65.26%), and 161 were fielded at 6 months for a total of 110 completed surveys (response rate = 68.32%). Sample characteristics are reported in Table 1. Consistent with the stratified design, the initial sample was roughly equally distributed in terms of age and gender, with diversity in terms of race and ethnicity.

Table 1 Sample characteristics at baseline

Aim 1

From initial survey to three months, housing insecurity was the most stable risk (Kappa = 0.86; 95% CI: 0.77, 0.95), followed by difficulty paying for utilities (Kappa = 0.79; 95% CI: 0.68, 0.90), transportation (Kappa = 0.74; 95% CI: 0.62, 0.86), food insecurity (Kappa = 0.69; 95% CI: 0.56, 0.82), and housing quality (Kappa = 0.64; 95% CI: 0.50, 0.78). Across domains, 4–11% of social risks improved whereas 2–8% worsened. From the initial survey to six months, risks were again moderately stable with some variability across domains with the order of prevalence in stable/consistent risk as follows: housing quality problem (Kappa = 0.59; 95% CI: 0.44, 0.74), food insecurity (Kappa = 0.63; 95% CI: 0.48, 0.78), transportation (Kappa = 0.80; 95% CI: 0.68, 0.91), housing insecurity (Kappa = 0.83; 95% CI: 0.73, 0.94), and paying for utilities (Kappa = 0.89; 95% CI: 0.80, 0.98). Across domains, 3–12% of social risks improved whereas 1–6% worsened. Figure 1a and b depict the percentage of respondents who improved, worsened, or remained consistent in their risk profile between the two surveys.

Fig. 1
figure 1

a Stability of Social Risk over 3 Months (N = 117). b Stability of Social Risk over 6 Months (N = 108)

Improving means the participant reported the social risk at initial survey but not at the follow-up survey. Worsening means the participant did not report the social risk at the initial survey but reported it on the follow-up survey. Stable means the participant consistently reported no risk or the risk at both surveys

Due to the small sample size, we estimated precision post-hoc for the lowest and highest kappas on both the 3 and 6 months surveys using the R package kappa size [27]. For 3 months, the analysis for housing quality (kappa = 0.64) had an expected lower bound of 0.49 and the analysis for housing insecurity (kappa = 0.86) had an expected lower bound of 0.68. For 6 months, the analysis for housing quality (kappa = 0.59) had an expected lower bound of 0.44 and the analysis for utilities (kappa = 0.89) had an expected lower bound of 0.66. Given that studies should be ideally powered to detect a value greater than 0 such as 0.40 [25], these post-hoc analyses suggest the study was adequately sized to accurately estimate the kappas.

Aim 2

From initial survey to three (Kappa = 0.52; 95% CI: 0.41, 0.64) and six months (Kappa = 0.48; 95% CI: 0.36, 0.60), the social adversity profiles showed weak agreement; see Fig. 2a and b, respectively. The Sankey diagram depicts movement between risk profiles and indicates the greatest movement is toward risk resolution, from risk but no need to no risk or need. Only 50% of those with a risk and need at the initial survey (n = 22) reported a risk and need by six months (n = 11), with the majority still endorsing risk, but no longer desiring assistance, and several reporting complete resolution of risk by six months.

Fig. 2
figure 2

a Movement between Social Risk. Need Status from 3 month (N = 117). Only those participants who had social risk and need data at both the initial survey time point and 3 months later were included. Overall Kappa = .52; 95% CI: 0.41, 0.64. b Movement between Social Risk/Need Status from 6 months (N = 108). Only those participants who had social risk and need data at both the initial survey time point and 3 months later were included. Overall Kappa = .48; 95% CI: 0.36, 0.60

Aim 3

Overall, results of Poisson regressions showed few social risks were associated with any utilization differences at 6 months (Table 2). Housing insecurity was associated with increases in ED visits (relative risk [RR] = 2.69, p = 0.01), and problems with housing quality was associated with more UC visits (RR = 1.61, p = 0.04), which do not hold when the Bonferroni correction is applied.

Table 2 Social risk and healthcare utilization 6 months later


This study examined the stability of social risks and needs over three or six months in a sample of people receiving subsidies from the healthcare exchange. Social risks were generally moderately to strongly stable over three months. Over six months, social risks were moderately to strongly stable except for housing quality which showed potentially weak to moderate stability. Social adversity profiles were showed weak stability over both three and six months; the greatest change observed across both time periods was that of risk resolution. Finally, when controlling for key demographic variables (e.g., gender, age, neighborhood deprivation) and prior utilization, in general, social risks were not related to utilization; however, we observed significant associations between housing insecurity and quality and increased use of expensive, emergent services six months later.

Risk domain stability

Across individual social risks, most were stable over time with a minority either improving (risk resolving) or worsening (new risk reported). The proportion experiencing changes ranged from 15–20% (food insecurity, housing quality problems) to 6–11% (trouble paying for utilities and housing insecurity). These findings are largely consistent with the broader literature that suggests losing one’s housing is quite rare in the general population, and even rapid rehousing interventions can require up to 120–143 days for resolution [28, 29]. Trouble paying for utilities may be more stable for different reasons. Specifically, there is some literature to suggest that people prioritize paying for rent and food, and utilities are routinely deprioritized [30, 31]. However, trouble paying for utilities was rare in this sample perhaps because these individuals received healthcare subsidies, which may free up financial resources to pay for other material needs, or because respondents could access utility assistance benefits for which income eligibility requirements may be more generous than Medicaid and SNAP [32]. Conversely, food insecurity is notoriously labile [33, 34], given fluctuations in income and/or challenges managing fixed/limited incomes, a problem that has been exacerbated by the COVID-19 pandemic [35,36,37].

The general stability of social risk has implications for screening cadence and intervention evaluation design. Risks were mostly stable over three and six months, which suggests that screening every six months is likely sufficient to sensitively detect emergence or resolution of social risks in this population. This recommendation is consistent with several other studies [38], acknowledging that some scholars advocate for annual screening, which reduces the possibility of over-screening [39], may be more patient-centered, and improve health system feasibility. These data suggest that time to follow-up in intervention evaluations may appropriately be set to three or six months since social risks in this population appear relatively stable in the absence of intervention. However, this study is one of few longitudinal studies of social risk and some domains (e.g., housing quality) may be more naturally labile particularly among marginalized populations.

Social adversity profile stability

Although the social adversity profiles were somewhat stable, many people did transition between groups from the initial survey to three or six months. For example, nearly half of those endorsing a risk but no need initially reported their risks resolved three months later, whereas far fewer of those initially in the “no risk and no need” group endorsed risk three months later and most in this group stayed free of social risk. This pattern was almost identical across six months of observation. Importantly, a portion of those endorsing risk but no need initially went on to desire assistance at three and six months; however, this portion was much smaller across the six-month timeframe. This may suggest that desire for assistance is more labile, and/or that individuals were able to connect with resources by six months, which we were unable to track in this initiative. In support of the former interpretation, as noted above, it is common in the literature to report on a phenomenon in which far fewer people who endorse social risk want help. Reasons for not wanting help from the healthcare system include stigma, privacy concerns, mistrust in the healthcare system, perception that this is not the role of the healthcare system, or concern about disclosing a social need for which the healthcare system cannot provide assistance [38, 40, 41]. To the latter point, it might be that social health interventions may take six months or more to resolve need, and perhaps longer to resolve the underlying social risk. However, much of this is speculation in absence of having a true baseline or intervention group for comparison.

These findings might inform social health screening and intervention initiatives. First, these data suggest that screening could focus initially on identifying social risk, as a measure of underlying social adversity, rather than identifying those with desire for assistance, which may be associated with psychological constructs such as readiness to change. This approach sets the stage for health systems to pair screening with social risk-informed care (or social risk adjusted care [42]) in which patients’ social context is considered carefully by physicians and care teams in collaboratively generating their care plan [43], consistent with precision medicine [44]. Although this is not (yet) standard practice, this approach would warrant screening for risk to ensure those realities inform shared decision making even for members who do not wish for health systems to target their social issues directly, and especially in cases where resources do not exist to resolve risks. Conversely, screening focused on desire for assistance with social risk presents an obvious starting place for health systems eager to connect patients with resources, but given its lability, may warrant more frequent assessment to be responsive to patients’ preferences.

Social risk predicts utilization

Among members receiving health insurance subsidies, we observed two statistically significant prospective associations with utilization. Specifically, we found that housing insecurity was associated with increased use of emergency department visits (with an almost three-fold relative risk compared to those without housing insecurity) and housing condition problems predicted increased use of urgent care services. Housing interventions thus have the potential to decrease use of expensive emergent services [45,46,47]; however, it is important to note that we did not control for possible confounders such as serious persistent mental illness that may drive housing challenges and utilization issues. These findings should be considered preliminary.


Several limitations of this study should be considered. First, this was a small sample of people receiving health insurance subsidies on the healthcare exchange within a single large integrated, healthcare system in Southern California and, thus, the results might not generalize to other regions or subpopulations. Second, not all social risk domains were assessed, such as interpersonal violence. Third, the sample was not restricted to people struggling with homogeneous health conditions, which could dilute relations between social risk and utilization that are important to understand. Fourth, structural vulnerability [48], and systems of oppression, may lead to both social risk and high/low healthcare use, and we were unable to directly explore these influences. Fifth, several statistical models were run, which may elevate the risk of Type I error in reporting. Finally, we were not able to track whether this health system was systematically providing social risk interventions that could explain the cases of improvement observed.


Overall, most social risks were stable over three or six months while desire for assistance was less stable. Given this, the question remains: How frequently should we be asking our patients about their social conditions? Until a business case for social needs targeted care is made, health systems will struggle to staff more frequent screening. More research is needed to determine how social risk and needs evolve naturally to guide healthcare system efforts to intervene upstream.

Availability of data and materials

The study was part of ongoing quality improvement efforts. As such, it was deemed not human subjects research after administrative review by the KP IRB; informed consent was not required. Because of this quality improvement status, the datasets generated for the current study are not publicly available.



Accountable Health Communities


Electronic Health Records


International Review Board

KP :

Kaiser Permanente


Prevalence-adjusted Kappa statistic

RR :

Relative risk

SDoH :

Social determinants of health


The Social Needs Network for Evaluation and Translation


  1. Marmot M. Social determinants of health inequalities. Lancet. 2005;365(9464):1099–104.

    Article  PubMed  Google Scholar 

  2. Braveman P, Egerter S, Williams DR. The social determinants of health: coming of age. Annu Rev Public Health. 2011;32(1):381–98.

    Article  PubMed  Google Scholar 

  3. Artiga S, Hinton E. Beyond health care: the role of social determinants in promoting health and health equity. Health. 2019;20(10):1-13.

  4. Goldstein D, Holmes J. 2011 Physicians' daily life report. Robert Wood Johnson Foundation; 2011.

  5. Kennedy BP, Kawachi I, Glass R, Prothrow-Stith D. Income distribution, socioeconomic status, and self rated health in the United States: multilevel analysis. BMJ. 1998;317(7163):917–21.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  6. Xiao Q, Berrigan D, Matthews CE. A prospective investigation of neighborhood socioeconomic deprivation and self-rated health in a large US cohort. Health Place. 2017;44:70–6.

    Article  PubMed  PubMed Central  Google Scholar 

  7. Ota A, Yatsuya H, Nishi N, Okuda N, Ohkubo T, Hayakawa T, et al. Relationships among socioeconomic factors and self-rated health in Japanese Adults: NIPPON DATA2010. J Epidemiol. 2018;28(Suppl 3):S66–72.

    Article  PubMed  PubMed Central  Google Scholar 

  8. Sulander T, Pohjolainen P, Karvinen E. Self-Rated Health (SRH) and Socioeconomic Position (SEP) among urban home-dwelling older adults. Arch Gerontol Geriatr. 2012;54(1):117–20.

    Article  PubMed  Google Scholar 

  9. Koch CG, Li L, Kaplan GA, Wachterman J, Shishehbor MH, Sabik J, et al. Socioeconomic position, not race, is linked to death after cardiac surgery. Circ Cardiovasc Qual Outcomes. 2010;3(3):267–76.

    Article  PubMed  Google Scholar 

  10. Valtorta NK, Hanratty B. Socioeconomic variation in the financial consequences of ill health for older people with chronic diseases: a systematic review. Maturitas. 2013;74(4):313–33.

    Article  PubMed  Google Scholar 

  11. Sterling S, Chi F, Weisner C, Grant R, Pruzansky A, Bui S, et al. Association of behavioral health factors and social determinants of health with high and persistently high healthcare costs. Prev Med Rep. 2018;11:154–9.

    Article  PubMed  PubMed Central  Google Scholar 

  12. Holzer J, Canavan M, Cherlin E, Bradley E. Health hot spots: mapping hospital costs and social determinants of health. Open J Prev Med. 2014;04(09):717–22.

    Article  Google Scholar 

  13. Hatef E, Searle KM, Predmore Z, Lasser EC, Kharrazi H, Nelson K, et al. The impact of social determinants of health on hospitalization in the veterans health administration. Am J Prev Med. 2019;56(6):811–8.

    Article  PubMed  Google Scholar 

  14. Marmot M, Allen J, Bell R, Bloomer E, Goldblatt P. WHO European review of social determinants of health and the health divide. Lancet. 2012;380(9846):1011–29.

    Article  PubMed  Google Scholar 

  15. Walker RJ, Smalls BL, Campbell JA, Strom Williams JL, Egede LE. Impact of social determinants of health on outcomes for type 2 diabetes: a systematic review. Endocr. 2014;47(1):29–48.

    Article  CAS  Google Scholar 

  16. Viner RM, Ozer EM, Denny S, Marmot M, Resnick M, Fatusi A, et al. Adolescence and the social determinants of health. Lancet. 2012;379(9826):1641–52.

    Article  PubMed  Google Scholar 

  17. Gottlieb LM, Wing H, Adler NE. A systematic review of interventions on patients’ social and economic needs. Am J Prev Med. 2017;53(5):719–29.

    Article  PubMed  Google Scholar 

  18. Finkelstein A, Zhou A, Taubman S, Doyle J. Health care hotspotting - a randomized controlled trial. N Engl J Med. 2020;382(2):152–62.

    Article  PubMed  PubMed Central  Google Scholar 

  19. Johnson TL, Rinehart DJ, Durfee J, Brewer D, Batal H, Blum J, et al. For many patients who use large amounts of health care services, the need is intense yet temporary. Health Affair. 2015;34(8):1312–9.

    Article  Google Scholar 

  20. De Marchis EH, Hessler D, Fichtenberg C, Adler N, Byhoff E, Cohen AJ, et al. Part I: a quantitative study of social risk screening acceptability in patients and caregivers. Am J Prev Med. 2019;57(6S1):S25–37.

    Article  PubMed  PubMed Central  Google Scholar 

  21. Lewis CC, Wellman R, Jones S, Walsh-Bailey C, Thompson E, Derus A, et al. Comparing the performance of two social risk screening tools in a vulnerable subpopulation. J Family Med Prim Care. 2020;9(9):5026–34.

    Article  PubMed  PubMed Central  Google Scholar 

  22. Billioux A, Verlander K, Anthony S, Alley D. 2017. Standardized Screening for Health-Related Social Needs in Clinical Settings: The Accountable Health Communities Screening Tool. NAM Perspectives. Discussion Paper, National Academy of Medicine, Washington, DC.

  23. Henrikson NB, Blasi PR, Dorsey CN, Mettert KD, Nguyen MB, Walsh-Bailey C, et al. Psychometric and pragmatic properties of social risk screening tools: a systematic review. Am J Prev Med. 2019;57(6 Suppl 1):S13–24.

    Article  PubMed  Google Scholar 

  24. Byrt T, Bishop J, Carlin JB. Bias, prevalence and kappa. J Clin Epidemiol. 1993;46(5):423–9.

    Article  CAS  PubMed  Google Scholar 

  25. Sim J, Wright CC. The Kappa statistic in reliability studies: use, interpretation, and sample size requirements. Phys Ther. 2005;85(3):257–68.

    Article  PubMed  Google Scholar 

  26. McHugh ML. Interrater reliability: the Kappa statistic. Biochem Med (Zagreb). 2012;22(3):276–82.

    Article  PubMed  Google Scholar 

  27. Rotoni MA. kappaSize: Sample Size Estimation Functions for Studies of Interobserver Agreement. 1.2 ed: Michael A Rotoni; 2018.

  28. Dizon-Ross E. The Impacts of Homelessness Prevention and Rapid Rehousing on Homeless and Highly Mobile Students. CEPA Working Paper No. 20–08. Stanford center for education policy analysis. 2020.

  29. Schweid R. Invisible Nation. In: Schweid R, editor. Trenton. New Jersey: Rapid Rehousing: University of California Press; 2016. p. 167–79.

    Google Scholar 

  30. Heflin C, London AS, Scott EK. Mitigating material hardship: the strategies low-income families employ to reduce the consequences of poverty. Sociol Inq. 2011;81(2):223–46.

    Article  Google Scholar 

  31. Tach L, Greene SS. “Robbing Peter to Pay Paul”: Economic and cultural explanations for how lower-income families manage debt. Soc Probl. 2014;61(1):1–21.

    Article  Google Scholar 

  32. Low Income Home Energy Assistance Program Sacramento, CA: California Department of Community Services & Development; 2022. [access date 2022 August 19]. Available from:

  33. Loopstra R, Tarasuk V. Severity of household food insecurity is sensitive to change in household income and employment status among low-income families. J Nutr. 2013;143(8):1316–23.

    Article  CAS  PubMed  Google Scholar 

  34. Wolf S, Morrissey T. Economic instability, food insecurity, and child health in the wake of the great recession. Soc Serv Rev. 2017;91(3):534–70.

    Article  Google Scholar 

  35. Gundersen C, Hake M, Dewey A, Engelhard E. Food insecurity during COVID -19. Appl Econ Perspect Policy. 2021;43(1):153–61.

    Article  PubMed  Google Scholar 

  36. Niles MT, Bertmann F, Belarmino EH, Wentworth T, Biehl E, Neff R. The early food insecurity impacts of COVID-19. Nutrients. 2020;12(7):2096.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  37. Fitzpatrick KM, Harris C, Drawve G, Willis DE. Assessing Food Insecurity among US Adults during the COVID-19 Pandemic. J Hunger Environ Nutr. 2021;16:1, 1-18.

  38. Byhoff E, De Marchis EH, Hessler D, Fichtenberg C, Adler N, Cohen AJ, et al. Part II: a qualitative study of social risk screening acceptability in patients and caregivers. Am J Prev Med. 2019;57(6):S38–46.

    Article  PubMed  PubMed Central  Google Scholar 

  39. Simon J, Panzer J, Adetoro E, Wright KM, Masinter L, Mohanty N, et al. Frequency of administration of standardized screening questions in federally qualified health centers. JAMA Intern Med. 2021;181(9):1253–5.

    Article  PubMed  PubMed Central  Google Scholar 

  40. De Marchis EH, Alderwick H, Gottlieb LM. Do patients want help addressing social risks? J Am Board Fam Med. 2020;33(2):170–5.

    Article  PubMed  Google Scholar 

  41. Brundage S. Health Affairs Blog [Internet]. Health Affairs; Washington, DC: 2019. Accessed 18 Aug 2022.

  42. National Academy of Sciences Engineering and Medicine. Integrating social care into the delivery of health care: Moving upstream to improve the nation's health. Washington, D.C.; 2019.

  43. Shields-Zeeman L, Lewis C, Gottlieb L. Social and mental health care integration: the leading edge. JAMA Psychiat. 2019;76(9):881–2.

    Article  Google Scholar 

  44. Persmark A, Wemrell M, Zettermark S, Leckie G, Subramanian SV, Merlo J. Precision public health: mapping socioeconomic disparities in opioid dispensations at Swedish pharmacies by Multilevel Analysis of Individual Heterogeneity and Discriminatory Accuracy (MAIHDA). PLoS ONE. 2019;14(8):e0220322.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  45. Parker D. Housing as an intervention on hospital use: Access among chronically homeless persons with disabilities. J Urban Health. 2010;87(6):912–9.

    Article  PubMed  PubMed Central  Google Scholar 

  46. Jaworsky D, Gadermann A, Duhoux A, Naismith TE, Norena M, To MJ, et al. Residential stability reduces unmet health care needs and emergency department utilization among a cohort of homeless and vulnerably housed persons in Canada. J Urban Health. 2016;93(4):666–81.

    Article  PubMed  PubMed Central  Google Scholar 

  47. Moore DT, Rosenheck RA. Comprehensive services delivery and emergency department use among chronically homeless adults. Psychol Serv. 2017;14(2):184.

    Article  PubMed  PubMed Central  Google Scholar 

  48. Bourgois P, Holmes SM, Sue K, Quesada J. Structural vulnerability: operationalizing the concept to address health disparities in clinical care. Acad Med. 2017;92(3):299–307.

    Article  PubMed  PubMed Central  Google Scholar 

Download references


Not applicable.


There are no financial disclosures to report. The Social Needs Network for Evaluation and Translation (SONNET) is funded by Kaiser Permanente National Community Health. Other than the authors, the Kaiser Foundation Health Plan and Colorado Permanente Medical Group had no role in the design, methods, data collection, analysis, or preparation of this paper.

Author information

Authors and Affiliations



CL: Conceptualization, Methodology, Writing – Original draft preparation. SJ: Conceptualization, Methodology, Writing – Original draft preparation; Data visualization. RW: Data curation, Formal Analysis, Writing – Original Draft. AS, LG, MB, ED, and JS: Writing – Reviewing and Editing. The author(s) read and approved the final manuscript.

Corresponding author

Correspondence to Cara C. Lewis.

Ethics declarations

Ethics approval and consent to participate

This study was submitted to the KP International Review Board (IRB) and received a determination that this was “not human subjects research” after administrative review (Human Subjects Assurance Number/Federal Wide Assurance Number: 00002344.) Administrative approval by KP leadership was obtained through the KP Data Use Center formal processes.

Consent for publication

Not applicable.

Competing interests

Adam Sharp began employment at Galvan ( in April 2022 after the research for this project was completed. He has no financial conflicts nor competing interests with this research or manuscript. No other authors have conflicts of interest to report.

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit The Creative Commons Public Domain Dedication waiver ( applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Lewis, C.C., Jones, S.M.W., Wellman, R. et al. Social risks and social needs in a health insurance exchange sample: a longitudinal evaluation of utilization. BMC Health Serv Res 22, 1430 (2022).

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: