BACKGROUND AND OBJECTIVES:

Although clinical settings are increasingly screening for social determinants of health, essential questions about optimal screening remain. We aimed to assess primary care contexts of individuals choosing not to answer questions about health-related social needs and to compare screening question response with subsequent use of resource information.

METHODS:

We compared caregiver responses to an electronic survey administered during a child’s emergency department visit and through telephone follow-up 2 weeks later by responses to questions about health-related social needs (no social needs endorsed, ≥1 endorsed, none endorsed but ≥1 question not answered).

RESULTS:

Of 146 respondents, 42 (29%) endorsed ≥1 health-related social need. Additionally, 19 (13%) endorsed no social needs but did not answer ≥1 question. Compared with those denying all social needs and those endorsing ≥1 social need, respondents who did not answer social needs screening questions reported longer duration since their child’s last primary care visit, lower perceptions of primary care, and less social support. For the 61 respondents participating in the 2-week follow-up survey, reported use of a community resource packet was 37% among those who had reported a social need, 26% among those who had denied all social needs, and 0% among those who had not answered ≥1 social needs questions.

CONCLUSIONS:

Clinicians and systems implementing screening for health-related social risks should plan for individuals who choose not to respond to specific items and may also wish to consider strategies that do not rely on screening and disclosure, particularly in communities known to have high prevalence of social needs.

What’s Known on This Subject:

Screening for health-related social needs allows for tailored advice based on identified needs. However, risk of potential miscategorization due to nonresponse and false-negatives is not fully understood.

What This Study Adds:

One in 8 individuals chose not to answer ≥1 social needs screening questions, and these individuals reported poorer perceptions of primary care. Of parents who subsequently used resource information, less than half had endorsed health-related social needs.

Health-related social needs (including material hardships, such as housing instability, food insecurity, and transportation barriers) are associated with adverse health outcomes during childhood and beyond.18  Identifying social needs among patients and communities may allow clinicians to tailor medical care to individual social circumstances, to directly address unmet social needs, and to better understand and advocate for population health needs.2,9  Patient-level interventions informed by social needs screening have demonstrated improved connection to social services through referrals to community resources, patient navigators, and community health workers, with a subset of studies also demonstrating improved health outcomes,1014  leading clinicians, practices, health systems, and payers to increasingly incorporate screening for social needs in clinical practice.2,1517 

With this broader adoption of screening, it is important to clarify the accuracy and limitations of screening tools and implementation decisions. Ideally, implementation should be guided both by current evidence and by family preferences, with the goal of creating family-centered processes that support shared decision-making and respect for patient autonomy.1820  This requires, however, being explicit about choices to be made. First, with several instruments available to screen for social determinants of health,2127  a choice must be made about which to use. Some of these tools have been validated against longer instruments,26  and others have been tested for their ability to modify visit-level outcomes (eg, foster discussion of social needs, increase referral to community resources, and reduce unmet family-reported needs),22  but in recent systematic reviews, authors note that none have had all dimensions of validity assessed.2729  An additional choice to be made is about the mode of administration, with evidence to date suggesting that electronic screening and written screening are more acceptable to parents than verbal screening.3032  For those considering adopting electronic screening into clinical practice, 2 additional specific questions about programming and interpretation within electronic platforms require evidence to guide implementation.

A first question for those seeking to implement electronic screening of health-related social needs is precisely how to address potential nonresponse during data collection. Specifically, should the electronic survey force a response for each question, allow respondents to skip questions, or allow respondents to specifically declare a decision not to answer a specific question?33  Although these latter options could enhance respect for patient autonomy, the frequency with which families might choose to skip questions is not clear, nor is the optimal interpretation of nonresponse. However, some platforms develop computer-generated summaries of responses for clinician review,33  requiring additional decisions about how nonresponses should be integrated into such summaries.

A second question is how to respond to individuals who deny all health-related social needs. As screening becomes more automated, the opportunity to use screening to open conversation may diminish, increasing the need to be thoughtful about the potential for false-negatives.34  Parameters such as sensitivity and specificity have not been consistently reported for social needs screeners.28  As a result, rates of false-negatives and the potential benefit of intervention among patients screening negative are not clear. Interest in this question is informed by domestic violence research, in which high rates of nondisclosure with screening because of multiple underlying reasons3538  have led to the development of alternative strategies employing universal interventions that do not rely on screening.39  Whether universal provision of counseling or resources should be considered in lieu of screening for health-related social needs is an important but under-asked question.34,40 

Underlying both of these questions is a recognition that we lack adequate understanding of the needs of those not endorsing health-related social needs on screening instruments (either those skipping questions or actively denying all social needs). To address these knowledge gaps, we examined reported social needs among caregivers of children presenting with nonurgent emergency department (ED) visits, focusing on the subset of health-related social needs related to economic stability or material hardship. We focused on this population because of previous literature revealing high levels of social needs.41,42  In our analysis, we examined the frequency and the associated health care contexts of respondents skipping or declining to answer social needs questions. We specifically compared those skipping or declining to answer social needs questions with those endorsing social needs and those denying all social needs to help guide future clinical decisions regarding this population. We also assessed family-reported use of universally provided community resource information at the 2-week follow-up among families endorsing and not endorsing social needs.

We examined caregiver responses to a survey administered electronically during a child’s pediatric ED visit and caregiver responses to a follow-up telephone survey. Ethical approval for this research was provided by the University of Pittsburgh Institutional Review Board.

We developed a 59-item survey to assess health-related social needs, social support, perceptions of primary care, trust in health care providers, and experiences in seeking care (Supplemental Information); this analysis was focused on responses to social needs items. To assess health-related social needs, we synthesized items inquiring about health-related social needs from multiple existing measures (WECARE [well-child care visit, evaluation, community resources, advocacy, referral, education] and PRAPARE [Protocol for Responding to and Assessing Patients' Assets, Risks, and Experiences]21,22 ). For each of these items, we allowed respondents to skip the question and also included the option “I choose not to answer this question.” We did not inquire about desire for assistance with specific needs because we instead provided all participants with community resource information. We assessed social support with questions from the Medical Outcomes Study social support survey.43  We assessed perceptions of primary care and trust in physicians with validated surveys.44,45  We also inquired about the last primary care visit and immunization status. The survey was reviewed by a group of primary care physicians, ED physicians, and social workers; edited to a sixth-grade reading level; and piloted with 2 young adults.

We administered the survey to caregivers of children presenting to an academic pediatric ED between February 2019 and June 2019 using an electronic survey tool (Qualtrics, Provo, UT). Caregivers were eligible if they accompanied a child aged 5 years or younger who received a low-acuity triage score (level 4 or 5) during weekday business hours. We selected young children presenting with nonurgent ED visits because of previous literature revealing high levels of poverty and social needs.41,42  Caregivers unable to participate in English or Spanish were excluded. Research assistants identified potentially eligible caregivers and invited them to participate in a survey to understand the experiences of caregivers in seeking care and resources for their children. Potential participants were advised that responses were confidential and would not be shared with the clinical team and were advised to tell their child’s medical team directly any information they wished the clinical team to know. After informed consent was obtained, participants were given an electronic tablet for survey completion. The research assistant paused the survey for patient care and resumed the survey once patient care was completed. The median survey administration time was 16 minutes.

A resource packet was developed by one investigator (K.M.G.) in collaboration with an outpatient pediatric care coordinator and an inpatient children’s hospital social worker. This 5-page community resource list included resources organized by potential health and social needs and summarized at a 10th-grade reading level, including health insurance, primary care, immunizations, high-quality online health information, transportation assistance, home nurse visiting programs, food and nutrition resources, utility assistance, child supplies (eg, diaper bank), housing, child care, and legal assistance. After completing the survey, all survey participants were provided with this resource list by research assistants who were unaware of survey responses. Research assistants used a standard script to remind participants that the medical team would not know survey information, that the research assistant also did not know their responses, and that many families find the resources in the resource packet useful.

Survey participants were offered the choice of opting in to a follow-up phone call 2 weeks after the survey to assess primary care access and use, social needs, and community resource use at that time. The telephone survey included 6 yes or no items about subsequent health care use, health care needs, social needs, and resource use. To assess resource use, caregivers were asked whether they used any of the information about community resources provided during their child’s ED visit. Five additional open-ended questions addressed experiences in contacting resources and opportunities to make community resource information more useful for families. The investigator completing follow-up calls (K.M.G) was blinded to previous survey responses. Caregivers were also given the option of opting in to review the child’s immunization status through the state immunization registry. We attempted to contact each parent who opted in to the follow-up phone call at least 3 times.

We examined caregiver participant characteristics using descriptive statistics. Next, we examined responses to individual health-related social needs questions, categorized as (1) social need endorsed, (2) social need denied, and (3) question declined or skipped. On the basis of responses to individual items, each participant was also categorized as one of the following: (1) at least one social need endorsed, (2) all social needs denied, and (3) no social needs endorsed but at least one question declined or skipped. We selected these categories because the course of action regarding follow-up conversation and offering services was potentially clearer for those in the first 2 categories relative to the third.

To examine the primary care contexts of respondents by social needs response category, we compared trust in physicians and perceptions of primary care across response categories and used simple linear regression to test for statistically significant differences. We compared categorical variables (time since last primary care visit, immunization status, and social support) using χ2 tests or Fisher’s exact tests.

Among participants consenting to follow-up contact whom we were able to reach, we assessed rates of parent-reported community resources use through χ2 tests. We also estimated the proportion of respondents reporting use of community resource information who had and had not previously endorsed ≥1 social need. Finally, we synthesized themes in responses to open-ended telephone follow-up questions about community resource packet use.

In total, 168 caregivers were approached, with 157 agreeing to participate (93%) and 146 completing the survey (89% of those approached). Among those who declined participation, the 2 most common reasons given were not having time (2 of 11) and being too stressed (2 of 11). Participants were primarily mothers (81%) and had multiple children in their household (68%; Table 1). Half of participants identified as white and 31% as Black or African American. Nearly two-thirds of participants were insured through Medicaid or the Children’s Health Insurance Program (CHIP).

TABLE 1

Respondent Characteristics

n (%)
Total 146 (100) 
Relation to child  
 Mother 117 (81) 
 Father 21 (15) 
 Other (eg, grandmother) 6 (4) 
Respondent age, y  
 <25 34 (23) 
 25–44 105 (72) 
 >45 6 (4) 
Respondent race  
 African American or Black 45 (31) 
 Asian American 3 (2) 
 White 72 (50) 
 Other (eg, biracial) 9 (6) 
 Prefer not to say 16 (11) 
Respondent ethnicity  
 Hispanic 6 (4) 
 Not Hispanic 118 (81) 
 Prefer not to say 21 (14) 
Respondent education attainment  
 Less than high school or GED 7 (5) 
 High school or GED 40 (28) 
 Some college 41 (28) 
 Completed college or additional postgraduate training 57 (39) 
Child age  
 0–11 mo 25 (17) 
 12–23 mo 35 (24) 
 2 y 28 (19) 
 3 y 25 (17) 
 4 y 15 (10) 
 5 y 18 (12) 
Child insurance  
 Medicaid or CHIP 90 (62) 
 Commercial or employer based 40 (28) 
 No insurance at this time 3 (2) 
 Other 12 (8) 
Household children  
 1 46 (32) 
 2–3 82 (57) 
 ≥4 16 (11) 
Usual mode of travel to pediatric clinic  
 Bus 16 (11) 
 Car that belongs to family or friend 16 (11) 
 Car that belongs to me 101 (69) 
 Other (eg, walking) 3 (3) 
 Taxi, jitney, rideshare 10 (7) 
n (%)
Total 146 (100) 
Relation to child  
 Mother 117 (81) 
 Father 21 (15) 
 Other (eg, grandmother) 6 (4) 
Respondent age, y  
 <25 34 (23) 
 25–44 105 (72) 
 >45 6 (4) 
Respondent race  
 African American or Black 45 (31) 
 Asian American 3 (2) 
 White 72 (50) 
 Other (eg, biracial) 9 (6) 
 Prefer not to say 16 (11) 
Respondent ethnicity  
 Hispanic 6 (4) 
 Not Hispanic 118 (81) 
 Prefer not to say 21 (14) 
Respondent education attainment  
 Less than high school or GED 7 (5) 
 High school or GED 40 (28) 
 Some college 41 (28) 
 Completed college or additional postgraduate training 57 (39) 
Child age  
 0–11 mo 25 (17) 
 12–23 mo 35 (24) 
 2 y 28 (19) 
 3 y 25 (17) 
 4 y 15 (10) 
 5 y 18 (12) 
Child insurance  
 Medicaid or CHIP 90 (62) 
 Commercial or employer based 40 (28) 
 No insurance at this time 3 (2) 
 Other 12 (8) 
Household children  
 1 46 (32) 
 2–3 82 (57) 
 ≥4 16 (11) 
Usual mode of travel to pediatric clinic  
 Bus 16 (11) 
 Car that belongs to family or friend 16 (11) 
 Car that belongs to me 101 (69) 
 Other (eg, walking) 3 (3) 
 Taxi, jitney, rideshare 10 (7) 

Because of nonresponse to individual items, totals may not add to 146 for each category. GED, general education diploma.

Among respondents, 42 (29%) endorsed ≥1 social need (Table 2). The most commonly endorsed social needs were food insecurity (12%) and transportation difficulties (17%). Of those endorsing ≥1 social need, 24% (10 of 42) also skipped ≥1 question. A total of 104 individuals did not endorse any social needs. Of these, 85 (58% of total sample) denied all social needs. The remaining 19 individuals (13% of total sample) did not endorse any social needs but also did not answer at least 1 social needs question. Of these 19 individuals, most (15 of 19; 79%) specifically selected the response “I choose not to answer this question” at least once, with the remaining 4 instead leaving at least one question blank. Of those not answering questions but also endorsing at least one social need, 10% (1 of 10) omitted responses to multiple questions. Of those not answering questions and also not endorsing social needs, 42% (8 of 19) omitted responses to multiple questions.

TABLE 2

Health-Related Social Needs Response Category and Individual Question Responses

Across All QuestionsAll Social Needs DeniedAt Least 1 Social Need EndorsedNo Social Needs Endorsed but at Least 1 Question Not Answered
Combined responses, n (%) 85 (58) 42 (29) 19 (13) 
Individual question responses No Yes Not answered 
Food insecurity, n (%) 117 (80) 17 (12) 12 (8) 
Housing instability, n (%) 130 (89) 9 (6) 7 (5) 
Unmet utility needs, n (%) 123 (84) 5 (3) 18 (13) 
Unmet phone needs, n (%) 124 (85) 4 (3) 18 (13) 
Unmet health care needs, n (%) 123 (84) 5 (3) 18 (13) 
Unmet medication needs, n (%) 126 (86) 2 (1) 18 (13) 
Unmet clothing needs, n (%) 122 (84) 6 (4) 18 (13) 
Unmet transportation needs, n (%) 111 (76) 25 (17) 10 (7) 
Across All QuestionsAll Social Needs DeniedAt Least 1 Social Need EndorsedNo Social Needs Endorsed but at Least 1 Question Not Answered
Combined responses, n (%) 85 (58) 42 (29) 19 (13) 
Individual question responses No Yes Not answered 
Food insecurity, n (%) 117 (80) 17 (12) 12 (8) 
Housing instability, n (%) 130 (89) 9 (6) 7 (5) 
Unmet utility needs, n (%) 123 (84) 5 (3) 18 (13) 
Unmet phone needs, n (%) 124 (85) 4 (3) 18 (13) 
Unmet health care needs, n (%) 123 (84) 5 (3) 18 (13) 
Unmet medication needs, n (%) 126 (86) 2 (1) 18 (13) 
Unmet clothing needs, n (%) 122 (84) 6 (4) 18 (13) 
Unmet transportation needs, n (%) 111 (76) 25 (17) 10 (7) 

Compared with those denying all social needs and those endorsing at least one social need, caregivers who did not answer at least one social needs question were less likely to have seen their child’s primary care provider in the last month (24% vs 55% among those with at least one social need and 57% among those with no social needs; P = .004; Table 3). Individuals who did not answer at least one social needs question had a lower perception of their primary care (total score: 68.8) than either those endorsing at least one social need (total score: 70.9) or those denying all social needs (score: 79.8; P = .01). Among specific subscales, individuals who did not answer at least 1 social needs question had significantly lower scores than the other groups for perceptions of their child’s providers’ contextual knowledge of their child (67.0 vs 75.1 endorsing and 83.8 denying social needs; P = .01) and comprehensiveness of care (66.1 vs 77.0 endorsing and 82.5 denying social needs; P = .02). Trust in providers was statistically similar across all groups (P = .10). Finally, caregivers who did not answer social needs questions were less likely to endorse social support compared with caregivers who denied all social needs or endorsed at least one social need (both P = .001).

TABLE 3

Association Between Health-Related Social Needs Response Category, Primary Care Contexts, and Social Support

All Social Needs DeniedAt Least 1 Social Need EndorsedNo Social Needs Endorsed but at Least 1 Question Not AnsweredP
Participants, n 85 42 19 — 
Last visit to PCP, n (%)    .004* 
 <1 mo ago 47 (57) 22 (55) 4 (24) — 
 2–5 mo ago 26 (32) 13 (33) 5 (29) — 
 ≥6 mo ago 9 (11) 5 (13) 8 (47) — 
Immunization status, n (%)     
 Up to date per parent report (n = 144) 80 (95) 34 (81) 15 (83) .02* 
 Up to date per review of state immunization system (n = 94a49 (86) 19 (76) 9 (75) .42 
Perceptions of primary care, score (SD)     
 Total 79.8 (16.4) 70.9* (20.9) 68.8* (19.3) .01*,b 
 Continuity subscore 55.3 (21.3) 56.3 (26.9) 56.1 (19.4) .97 
 Access subscore 79.7 (19.0) 63.3* (25.4) 73.3 (22.2) .001* 
 Contextual knowledge subscore 83.8 (22.1) 75.1 (24.3) 67.0* (24.4) .01* 
 Communication subscore 89.5 (16.8) 79.4* (22.8) 81.6 (23.2) .02* 
 Comprehensive subscore 82.5 (20.4) 77.0 (25.1) 66.1* (28.0) .02* 
 Coordination subscore 74.9 (27.6) 64.7 (34.0) 63.2 (28.6) .11 
Trust in provider (n = 137), score (SD) 21.4 (4.0) 20.2 (4.3) 19.3 (4.0) .10 
Social support (n = 143), n (%)     
 Tangible support 73 (86) 25 (61) 9 (53) .001* 
 Emotional and/or informational support 78 (92) 29 (69) 10 (63) .001* 
All Social Needs DeniedAt Least 1 Social Need EndorsedNo Social Needs Endorsed but at Least 1 Question Not AnsweredP
Participants, n 85 42 19 — 
Last visit to PCP, n (%)    .004* 
 <1 mo ago 47 (57) 22 (55) 4 (24) — 
 2–5 mo ago 26 (32) 13 (33) 5 (29) — 
 ≥6 mo ago 9 (11) 5 (13) 8 (47) — 
Immunization status, n (%)     
 Up to date per parent report (n = 144) 80 (95) 34 (81) 15 (83) .02* 
 Up to date per review of state immunization system (n = 94a49 (86) 19 (76) 9 (75) .42 
Perceptions of primary care, score (SD)     
 Total 79.8 (16.4) 70.9* (20.9) 68.8* (19.3) .01*,b 
 Continuity subscore 55.3 (21.3) 56.3 (26.9) 56.1 (19.4) .97 
 Access subscore 79.7 (19.0) 63.3* (25.4) 73.3 (22.2) .001* 
 Contextual knowledge subscore 83.8 (22.1) 75.1 (24.3) 67.0* (24.4) .01* 
 Communication subscore 89.5 (16.8) 79.4* (22.8) 81.6 (23.2) .02* 
 Comprehensive subscore 82.5 (20.4) 77.0 (25.1) 66.1* (28.0) .02* 
 Coordination subscore 74.9 (27.6) 64.7 (34.0) 63.2 (28.6) .11 
Trust in provider (n = 137), score (SD) 21.4 (4.0) 20.2 (4.3) 19.3 (4.0) .10 
Social support (n = 143), n (%)     
 Tangible support 73 (86) 25 (61) 9 (53) .001* 
 Emotional and/or informational support 78 (92) 29 (69) 10 (63) .001* 

PCP, primary care provider; —, not applicable.

a

Immunization status in state immunization system is reported only for children whose parents consented and who had information available in the state immunization system.

b

For continuous variables, the P value is reported for significance of the parameter within the regression model, with astericks in results columns indicating the value(s) varying significantly (p<0.05) from the reference category (all social needs denied).

*

P < .05.

Overall, 66% of participants agreed to a follow-up call. Those who did not answer social needs questions were the least likely to agree to a follow-up call (42%; Table 4). Of all agreeing to a follow-up call, 64% were reached (n = 61; 42% of overall sample). Of those reached, 37% of individuals who had endorsed at least one social need reported they had used or had immediate plans to use community resource information. Of those who had denied all social needs, 26% reported that they had used or had immediate plans to use community resource information. Of those who had not answered social needs questions, none reported that they had used or had immediate plans to use community resource information. Among the 61 participants who completed a follow-up call, only 44% of those who reported use of universally provided community resource information would have been identified as having a health-related social need on the basis of their responses to survey screening items.

TABLE 4

Association Between Social Risk Response Category, Willingness to Receive a Follow-up Call, and Use of Resources

All Social Needs DeniedAt Least 1 Social Need EndorsedNo Social Needs Endorsed but at Least 1 Question Not Answered
Survey sample, n 85 42 19 
Agreed to follow-up call, n (%) 57 (67) 31 (74) 8 (42) 
Contacted for follow-up call (of those who agreed to follow-up call), n (%) 35 (61) 19 (61) 7 (88) 
Used or immediate plan to use resources reported (of those who were contacted for follow-up call), n (%) 9 (26) 7 (37) 0 (0) 
All Social Needs DeniedAt Least 1 Social Need EndorsedNo Social Needs Endorsed but at Least 1 Question Not Answered
Survey sample, n 85 42 19 
Agreed to follow-up call, n (%) 57 (67) 31 (74) 8 (42) 
Contacted for follow-up call (of those who agreed to follow-up call), n (%) 35 (61) 19 (61) 7 (88) 
Used or immediate plan to use resources reported (of those who were contacted for follow-up call), n (%) 9 (26) 7 (37) 0 (0) 

In open-ended comments, most participant comments were positive, with aspects specifically valued by respondents including that the resource packet was comprehensive and organized (Table 5). However, several respondents reported either not remembering receiving a packet (5 of 61) or having lost it (3 of 61). Relatedly, several suggested that an electronic version would be better because “I could keep it for future reference.” Additional opportunities for improvement included a request for a cover page to offer more privacy, reducing the total amount of information provided, and providing the information at a less stressful time (eg, a well-child visit). No one expressed displeasure at having been given the information, with even those who stated they had no need for it providing positive comments: “If it was something I needed, it would have been a good reference.” Finally, 4 respondents reported that they had provided the resource packet to a friend, family member, or client.

TABLE 5

Recommendations Regarding Universal Resource Packet

Example Statements
Positive feature theme  
 Nonspecific positive “Perfect; great information; convenient.” 
 “If it was something I needed, it would have been a good reference.” 
 Comprehensive “Everything was on there that needs to be on there.” 
 “Was complete, had good information.” 
 “It was fantastic to have on one place for families.” 
 Useful “Easy to use, the phone numbers were listed.” 
 Organized “Very organized.” 
 “Everything was organized.” 
 Ability to share “Gave some resources to a friend.” 
 Plan to keep for future use “Kept the paper for future use.” 
Ideas for improvement  
 Prefer electronic version “I’m always on my phone.” 
 “Maybe I could look at a Web site.” 
 “E-mail may have been a better route.” 
 Prefer electronic version for future reference “If it was an electronic form, an E-mail would be better, I could keep for future reference. Some people lose the paper.” 
 “Forgot about packet, may not have forgot if emailed.” 
 Privacy “Cover sheet to make more private.” 
 Too much information “There was a lot of information, maybe having one sheet that is most useful.” 
 Difficult to receive information during acute illness “If it was presented at a well-visit would have been able to be more organized and would have been able to pay more attention to the paper.” 
Example Statements
Positive feature theme  
 Nonspecific positive “Perfect; great information; convenient.” 
 “If it was something I needed, it would have been a good reference.” 
 Comprehensive “Everything was on there that needs to be on there.” 
 “Was complete, had good information.” 
 “It was fantastic to have on one place for families.” 
 Useful “Easy to use, the phone numbers were listed.” 
 Organized “Very organized.” 
 “Everything was organized.” 
 Ability to share “Gave some resources to a friend.” 
 Plan to keep for future use “Kept the paper for future use.” 
Ideas for improvement  
 Prefer electronic version “I’m always on my phone.” 
 “Maybe I could look at a Web site.” 
 “E-mail may have been a better route.” 
 Prefer electronic version for future reference “If it was an electronic form, an E-mail would be better, I could keep for future reference. Some people lose the paper.” 
 “Forgot about packet, may not have forgot if emailed.” 
 Privacy “Cover sheet to make more private.” 
 Too much information “There was a lot of information, maybe having one sheet that is most useful.” 
 Difficult to receive information during acute illness “If it was presented at a well-visit would have been able to be more organized and would have been able to pay more attention to the paper.” 

As more health systems navigate optimal strategies for addressing social determinants of health, we aimed to examine 2 questions related to broader implementation of screening. First, we sought to characterize nonresponse to social determinant questions. We found that ∼1 in 8 individuals chose to decline or skip ≥1 of these questions and that nonresponse was associated with decreased parental perceptions of primary care, decreased use of primary care, and lower levels of social support. Second, we compared identification of social needs through screening questions to subsequent parent-reported use of community resource information. We found that questions only identified less than half of those who subsequently reported use of universally provided resource information. These findings have implications for those planning to implement and interpret social determinant screening in pediatric health systems.

First, for those designing electronic surveys to collect social determinant data, our results suggest that a reasonable proportion of potential respondents may use “I choose not to answer” options if provided. This option is specifically incorporated throughout the PRAPARE survey but not explicitly part of other social determinant screening tools. Also, we noted that the single item with the highest number of individuals choosing not to answer was a multiprong “choose all that apply” question, suggesting that questions about single topics may be more acceptable to respondents who are making choices about disclosure of needs. Within our results, the “I choose not to answer” response was chosen more often than skipping the question entirely, suggesting that including this option, perhaps while forcing a response to each item, may appropriately address respect for autonomy in data collection.

Second, for those interpreting electronic survey data on social needs screening data, deciding how to interpret nonresponse is important. Our results give reason to be concerned about caregivers who skip or choose not to answer screening questions because their social support and primary care connections may be less strong than those either endorsing social needs or denying all social needs. Specifically, these individuals appear to have fewer social supports and appear to be less well connected to primary care, with a difference in the overall perception of primary care score of similar magnitude as the difference observed previously between children with and without health insurance.44  These individuals were also less likely to be open to a follow-up call, and none of those reached from this group disclosed any use of social resources during the follow-up call. This suggests that those declining to answer questions may have greater discomfort in disclosing their medical or social needs. Although this leaves open questions about the level of need and the best way to offer support to this group, at minimum, it may be valuable for these individuals to be flagged by electronic surveys so that clinicians can pursue follow-up conversations (rather than assuming these individuals have no social needs, for example).

Third, we found that the social needs questions we used, which came from multiple tools, only identified a subset of respondents with subsequent parent-reported community resource information use at the 2-week follow-up. For these respondents with no social needs endorsement but subsequent use of community resource information, multiple explanations may contribute: caregivers may have chosen not to disclose social needs at the time of the initial survey, or social needs may have evolved over the subsequent 2 weeks. Authors of previous studies have found a disconnect between reported social needs and reported desire for resources: the authors of one previous study of adult patients found that 5% of patients who screened negative for social needs still requested resources at the time of screening,46  and the authors of another study found that only 15% to 20% of those endorsing social needs desired assistance.47  In another study focused on food insecurity, authors found that nearly half of caregivers who reported food insecurity did not choose to receive resources, whereas 20% of caregivers not reporting food insecurity did choose to receive resources.48  Some current tools specifically include questions about the desire for assistance after identifying social needs14,22 ; a different strategy worth exploring may be asking about the desire for specific types of assistance without first requesting disclosure of needs.

Alternatively, strategies that sidestep endorsement of needs or desire for assistance altogether are worth considering, especially within communities with high prevalence of specific needs, where universal resource provision is another option. Such universal strategies provide scripted messages normalizing needs and then providing resources for all patients. Over the last decade, universal education has grown as an approach for intimate partner violence.39  By providing normalizing messages and universal resources, this approach skips screening questions, thereby lessening the risk of missing false-negatives. Another benefit of this approach is that it has the potential to build community capacity by giving individuals information that can be disseminated to others. This was reported by a handful of respondents during our 2-week follow-up. However, providing education without previous screening questions removes opportunities to tailor resource education and messaging to specific needs, which may result in less effective connection to community resources, missed opportunity to tailor medical management to social needs, and limited ability to track or follow-up on social needs. Another limitation is that bypassing systematic collection of social determinants of health would limit opportunities to understand the prevalence of specific social needs and the structural challenges within a patient panel or community. Finally, interventions that require high resource use may not be feasible to offer universally. Hybrid strategies (eg, universal normalizing messages and resource information, followed by asking if the patient would like assistance connecting with any specific resource) may ultimately allow a blend of low-resource universal intervention with higher-resource tailored intervention. In communities with high prevalence of social needs, universally offering more intensive interventions, such as nurse home visitors or community health workers, may still warrant consideration.

This study had several limitations. Our data were collected from a single academic pediatric ED. This ED serves patients from multiple counties in western Pennsylvania, including rural and urban regions. Individuals identifying as white and African American were well-represented in this cohort, but there was limited representation from other racial and ethnic groups. By recruiting caregivers of children presenting for nonurgent ED care, we focused on a population anticipated to have high social needs. Our sample reported social needs similar to other recent samples of children in primary care, emergency care, or hospital settings31,49  but lower than others.30,32  Surveys were completed in a research setting, and parents were advised that their responses were confidential and would not impact their clinical care. This could have lessened concern about stigma from disclosure of social needs, which may have increased the likelihood of endorsing social needs. On the other hand, this could have lessened the perceived potential benefit of disclosure of social needs, potentially reducing the likelihood of disclosure. We note also that our follow-up sample was smaller than our overall sample, with 60% consenting to follow-up and 60% of those individuals reached, which may bias our findings from the follow-up call if the likelihood of agreeing to a follow-up call was differentially associated with the likelihood of using community resources across the 3 groups.

We found that 1 in 8 individuals chose not to answer ≥1 social needs screening questions, and health systems preparing to screen for health-related social needs should acknowledge and prepare for those who choose not to answer these items, both in survey design and interpretation of responses. Additionally, screening questions only identified 44% of parents who subsequently reported use of resource information 2 weeks later, suggesting a need to improve our ability to identify and support patients and caregivers in need of community resources. Universal education on community resources was acceptable to families and may overcome some limitations of screen-and-refer strategies. Clinicians and systems implementing screening should consider these results when planning responses to negative and incomplete screeners and may also wish to consider strategies that do not rely on screening and disclosure, particularly in communities already known to have a high prevalence of social needs.

Dr Ray designed the study, performed the analysis, interpreted the data, and drafted and critically revised the manuscript; Dr Gitz designed the study, obtained funding, acquired the data, interpreted the data, and critically revised the manuscript; Mr Hu and Mr Davis acquired the data, drafted portions of the manuscript, and critically revised the manuscript; Dr Miller designed the study, interpreted the data, and critically revised the manuscript; and all authors approve the final manuscript as submitted and agree to be accountable for the work.

Dr Gitz’s current affiliation is Children’s Hospital Los Angeles Medical Group, Los Angeles, CA.

FUNDING: Supported in part by a grant from the American Academy of Pediatrics Community Access to Child Health Program (to Dr Gitz). The funders had no role in the design and conduct of the study; the collection, management, analysis, and interpretation of the data; the preparation, review, or approval of the manuscript; or the decision to submit the manuscript for publication.

CHIP

Children’s Health Insurance Program

ED

emergency department

1
Council on Community Pediatrics
.
Poverty and child health in the United States
.
Pediatrics
.
2016
;
137
(
4
):
e20160339
2
Daniel
H
,
Bornstein
SS
,
Kane
GC
;
Health and Public Policy Committee of the American College of Physicians
.
Addressing social determinants to improve patient care and promote health equity: an American College of Physicians position paper
.
Ann Intern Med
.
2018
;
168
(
8
):
577
578
3
Council on Community Pediatrics
.
Community pediatrics: navigating the intersection of medicine, public health, and social determinants of children’s health
.
Pediatrics
.
2013
;
131
(
3
):
623
628. Reaffirmed October 2016
4
Spencer
AE
,
Baul
TD
,
Sikov
J
, et al
.
The relationship between social risks and the mental health of school-age children in primary care
.
Acad Pediatr
.
2020
;
20
(
2
):
208
215
5
Peltz
A
,
Garg
A
.
Food insecurity and health care use
.
Pediatrics
.
2019
;
144
(
4
):
e20190347
6
Clark
RE
,
Weinreb
L
,
Flahive
JM
,
Seifert
RW
.
Infants exposed to homelessness: health, health care use, and health spending from birth to age six
.
Health Aff (Millwood)
.
2019
;
38
(
5
):
721
728
7
Chung
EK
,
Siegel
BS
,
Garg
A
, et al
.
Screening for social determinants of health among children and families living in poverty: a guide for clinicians
.
Curr Probl Pediatr Adolesc Health Care
.
2016
;
46
(
5
):
135
153
8
Baer
TE
,
Scherer
EA
,
Fleegler
EW
,
Hassan
A
.
Food insecurity and the burden of health-related social problems in an urban youth population
.
J Adolesc Health
.
2015
;
57
(
6
):
601
607
9
Gottlieb
L
,
Tobey
R
,
Cantor
J
,
Hessler
D
,
Adler
NE
.
Integrating social and medical data to improve population health: opportunities and barriers
.
Health Aff (Millwood)
.
2016
;
35
(
11
):
2116
2123
10
Beck
AF
,
Cohen
AJ
,
Colvin
JD
, et al
.
Perspectives from the Society for Pediatric Research: interventions targeting social needs in pediatric clinical care
.
Pediatr Res
.
2018
;
84
(
1
):
10
21
11
Kangovi
S
,
Mitra
N
,
Norton
L
, et al
.
Effect of community health worker support on clinical outcomes of low-income patients across primary care facilities: a randomized clinical trial
.
JAMA Intern Med
.
2018
;
178
(
12
):
1635
1643
12
Gottlieb
LM
,
Hessler
D
,
Long
D
, et al
.
Effects of social needs screening and in-person service navigation on child health: a randomized clinical trial
.
JAMA Pediatr
.
2016
;
170
(
11
):
e162521
13
Garg
A
,
Toy
S
,
Tripodis
Y
,
Silverstein
M
,
Freeman
E
.
Addressing social determinants of health at well child care visits: a cluster RCT
.
Pediatrics
.
2015
;
135
(
2
):
e296
e304
14
Hassan
A
,
Scherer
EA
,
Pikcilingis
A
, et al
.
Improving social determinants of health: effectiveness of a web-based intervention
.
Am J Prev Med
.
2015
;
49
(
6
):
822
831
15
Cottrell
EK
,
Dambrun
K
,
Cowburn
S
, et al
.
Variation in electronic health record documentation of social determinants of health across a national network of community health centers
.
Am J Prev Med
.
2019
;
57
(
6 supp1
):
S65
S73
16
Morgenlander
MA
,
Tyrrell
H
,
Garfunkel
LC
,
Serwint
JR
,
Steiner
MJ
,
Schilling
S
.
Screening for social determinants of health in pediatric resident continuity clinic
.
Acad Pediatr
.
2019
;
19
(
8
):
868
874
17
Chisolm
DJ
,
Brook
DL
,
Applegate
MS
,
Kelleher
KJ
.
Social determinants of health priorities of state Medicaid programs
.
BMC Health Serv Res
.
2019
;
19
(
1
):
167
18
Dworkin
PH
,
Garg
A
.
Considering approaches to screening for social determinants of health
.
Pediatrics
.
2019
;
144
(
4
):
e20192395
19
Garg
A
,
Homer
CJ
,
Dworkin
PH
.
Addressing social determinants of health: challenges and opportunities in a value-based model
.
Pediatrics
.
2019
;
143
(
4
):
e20182355
20
Garg
A
,
Boynton-Jarrett
R
,
Dworkin
PH
.
Avoiding the unintended consequences of screening for social determinants of health
.
JAMA
.
2016
;
316
(
8
):
813
814
21
National Association of Community Health Centers
.
PRAPARE: protocol for responding to and assessing patients’ assets, risks, and experiences
.
Available at: www.nachc.org/research-and-data/prapare/. Accessed January 2, 2020
22
Garg
A
,
Butz
AM
,
Dworkin
PH
,
Lewis
RA
,
Thompson
RE
,
Serwint
JR
.
Improving the management of family psychosocial problems at low-income children’s well-child care visits: the WE CARE Project
.
Pediatrics
.
2007
;
120
(
3
):
547
558
23
Billioux
A
,
Verlander
K
,
Anthony
S
,
Alley
D
.
Standardized Screening for Health-Related Social Needs in Clinical Settings: The Accountable Health Communities Screening Tool
.
Washington, DC
:
National Academy of Medicine
;
2017
24
Narayan
A
,
Raphael
JL
,
Rattler
T
,
Bocchini
C
.
Social determinants of health: screening in the clinical setting
.
25
Browne-Yung
K
,
Freeman
T
,
Battersby
M
,
McEvoy
DR
,
Baum
F
.
Developing a screening tool to recognise social determinants of health in Australian clinical settings
.
Public Health Res Pract
.
2019
;
29
(
4
):
28341813
26
Hager
ER
,
Quigg
AM
,
Black
MM
, et al
.
Development and validity of a 2-item screen to identify families at risk for food insecurity
.
Pediatrics
.
2010
;
126
(
1
).
27
Social Interventions Research & Evaluation Network
.
Social needs screening tools comparison table (Pediatric Settings)
.
28
Sokol
R
,
Austin
A
,
Chandler
C
, et al
.
Screening children for social determinants of health: a systematic review
.
Pediatrics
.
2019
;
144
(
4
):
e20191622
29
Morone
J
.
An integrative review of social determinants of health assessment and screening tools used in pediatrics
.
J Pediatr Nurs
.
2017
;
37
:
22
28
30
Gottlieb
L
,
Hessler
D
,
Long
D
,
Amaya
A
,
Adler
N
.
A randomized trial on screening for social determinants of health: the iScreen study
.
Pediatrics
.
2014
;
134
(
6
).
31
Palakshappa
D
,
Goodpasture
M
,
Albertini
L
,
Brown
CL
,
Montez
K
,
Skelton
JA
.
Written versus verbal food insecurity screening in one primary care clinic
.
Acad Pediatr
.
2020
;
20
(
2
):
203
207
32
Cullen
D
,
Woodford
A
,
Fein
J
.
Food for thought: a randomized trial of food insecurity screening in the emergency department
.
Acad Pediatr
.
2019
;
19
(
6
):
646
651
33
Gold
R
,
Cottrell
E
,
Bunce
A
, et al
.
Developing electronic health record (EHR) strategies related to health center patients’ social determinants of health
.
J Am Board Fam Med
.
2017
;
30
(
4
):
428
447
34
Garg
A
,
Sheldrick
RC
,
Dworkin
PH
.
The inherent fallibility of validated screening tools for social determinants of health
.
Acad Pediatr
.
2018
;
18
(
2
):
123
124
35
Hussain
N
,
Sprague
S
,
Madden
K
,
Hussain
FN
,
Pindiprolu
B
,
Bhandari
M
.
A comparison of the types of screening tool administration methods used for the detection of intimate partner violence: a systematic review and meta-analysis
.
Trauma Violence Abuse
.
2015
;
16
(
1
):
60
69
36
Othman
S
,
Goddard
C
,
Piterman
L
.
Victims’ barriers to discussing domestic violence in clinical consultations: a qualitative enquiry
.
J Interpers Violence
.
2014
;
29
(
8
):
1497
1513
37
Feder
GS
,
Hutson
M
,
Ramsay
J
,
Taket
AR
.
Women exposed to intimate partner violence: expectations and experiences when they encounter health care professionals: a meta-analysis of qualitative studies
.
Arch Intern Med
.
2006
;
166
(
1
):
22
37
38
Reeves
EA
,
Humphreys
JC
.
Describing the healthcare experiences and strategies of women survivors of violence
.
J Clin Nurs
.
2018
;
27
(
5–6
):
1170
1182
39
Miller
E
,
McCaw
B
.
Intimate partner violence
.
N Engl J Med
.
2019
;
380
(
9
):
850
857
40
Krist
AH
,
Davidson
KW
,
Ngo-Metzger
Q
,
Mills
J
.
Social determinants as a preventive service: U.S. Preventive Services Task Force methods considerations for research
.
Am J Prev Med
.
2019
;
57
(
6
supp1
):
S6
S12
41
Malecha
PW
,
Williams
JH
,
Kunzler
NM
,
Goldfrank
LR
,
Alter
HJ
,
Doran
KM
.
Material needs of emergency department patients: a systematic review
.
Acad Emerg Med
.
2018
;
25
(
3
):
330
359
42
Arthur
KC
,
Lucenko
BA
,
Sharkova
IV
,
Xing
J
,
Mangione-Smith
R
.
Using state administrative data to identify social complexity risk factors for children
.
Ann Fam Med
.
2018
;
16
(
1
):
62
69
43
Gjesfjeld
CD
,
Greeno
CG
,
Kim
KH
.
A confirmatory factor analysis of an abbreviated social support instrument: the MOS-SSS
.
Res Soc Work Pract
.
2008
;
18
(
3
):
231
237
44
Seid
M
,
Varni
JW
,
Bermudez
LO
, et al
.
Parents’ perceptions of primary care: measuring parents’ experiences of pediatric primary care quality
.
Pediatrics
.
2001
;
108
(
2
):
264
270
45
Dugan
E
,
Trachtenberg
F
,
Hall
MA
.
Development of abbreviated measures to assess patient trust in a physician, a health insurer, and the medical profession
.
BMC Health Serv Res
.
2005
;
5
:
64
46
Buitron de la Vega
P
,
Losi
S
,
Sprague Martinez
L
, et al
.
Implementing an EHR-based screening and referral system to address social determinants of health in primary care
.
Med Care
.
2019
;
57
(
6
suppl 2
):
S133
S139
47
Gold
R
,
Bunce
A
,
Cowburn
S
, et al
.
Adoption of social determinants of health EHR tools by community health centers
.
Ann Fam Med
.
2018
;
16
(
5
):
399
407
48
Bottino
CJ
,
Rhodes
ET
,
Kreatsoulas
C
,
Cox
JE
,
Fleegler
EW
.
Food insecurity screening in pediatric primary care: can offering referrals help identify families in need?
Acad Pediatr
.
2017
;
17
(
5
):
497
503
49
Vaz
LE
,
Wagner
DV
,
Ramsey
KL
, et al
.
Identification of caregiver-reported social risk factors in hospitalized children
.
Hosp Pediatr
.
2020
;
10
(
1
):
20
28

Competing Interests

POTENTIAL CONFLICTS OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.

FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships to this article to disclose.

Supplementary data