Parent Perspectives on Documentation and Sharing of Health-Related Social Needs Data

This qualitative study was nested within a quality improvement project in which our team implemented tablet-based HRSN screening and resource referral in 3 inpatient units in an academic quaternary care children’s hospital.^12,16  An opt-out tablet-based HRSN screener is introduced to parents at the time of their child’s admission. The screener (see Supplemental Information) includes screening questions adapted from the validated WE CARE questionnaire across 6 domains: (1) inpatient and outpatient food insecurity, (2) housing concerns, (3) difficulty paying utilities, (4) transportation challenges, (5) parent depressed mood, and (6) intimate partner violence (IPV).¹⁷ 

After caregivers complete the screener, social workers respond to urgent needs such as homelessness, IPV, and inpatient food insecurity. For less urgent needs, caregivers can choose to receive an in-person social work consult, a list of resources tailored to their endorsed needs sent via text or e-mail, or both. All caregivers receive general information about local community-based resources through a local electronic resource map. The current screener was implemented in April 2022. Between April 11, 2022 and November 6, 2023, 2992 parents of hospitalized children were offered the screener and 913 completed the screener. See Supplemental Table 6 for additional demographic characteristics of caregivers who completed the screener.

We recruited caregivers who participated in the HRSN screening and resource referral program. Parents were eligible for the study if they: (1) completed the HRSN screener, (2) were ≥18 years of age, (3) spoke English fluently, and (4) were the legal guardian of the child. We used purposive sampling to include parents with a range of types and number of HRSNs endorsed.

The study was designed to comply with the Consolidated Criteria for Reporting Qualitative Research reporting guideline for qualitative research (Supplemental Table 5).¹⁸  This study was deemed exempt by the institution’s institutional review board.

Potentially eligible parents were identified through a review of a REDCap dataset of parents who participated in HRSN screening. Parents were recruited in person or over the phone by a trained research team member (ZB, MY, SM) within 2 weeks of completing the screener. At the start of the interviews, the research team communicated that there were no right or wrong answers, answers were confidential, and participants could skip questions or stop the interview at any time.

The semi-structured interviews were guided by the health equity implementation framework,¹⁹  which facilitates the consideration of sociopolitical context, health care and organizational context, and patient, provider, and intervention factors in the implementation of programs and projects addressing health equity (Supplemental Fig 1). We adapted the framework to explore the patient and caregiver factors related to the implementation of an HRSN screening and referral program, particularly related to beliefs, preferences, trust, and expectations around the documentation and sharing of HRSN information. We conducted a pilot interview with a parent and iteratively modified the interview guide (Supplemental Information) on the basis of interviewer and parent feedback on emerging themes and the clarity of questions.

The interviews explored parents’ experiences completing the screener, their perspectives on whether and how HRSN should be documented and shared, their opinions on the use of several possible Z codes (International Classification of Diseases, 10th Revision, Clinical Modification [ICD-10-CM] codes ranging from Z55-Z65 that are the encounter reason codes used to document social determinants of health),²⁰  and their beliefs regarding the potential benefits and drawbacks of collecting and sharing HRSN data. The Z codes were presented to them as part of mock after-visit summaries and included ICD-10-CM codes Z59.4 Food insecurity, Z91.89 Lack of access to transportation, Z59.9 Problems related to housing and economic circumstances, and Z62.820 Child affected by parental relationship distress (Supplemental Information).

Interviews were conducted in person or over the phone by a researcher trained in qualitative interviewing (ZB, MY, SM). After the interviews, parents completed a demographic survey sent via text or e-mail. The hospitalized child’s age and insurance status were abstracted from the electronic health record (EHR). Interviews were conducted from June 2022 to February 2023 and lasted an average of 35 minutes. Interviews were continued until data saturation was reached, or the point at which new data repeated what was expressed in previous data.²¹  This was determined through periodic interview team meetings. On completion of the interviews, $25 electronic gift cards were sent to the participants. See the Supplemental Information for more information on the research team.

Interviews were digitally recorded, professionally transcribed, deidentified, and entered into QSR NVivo12 software for data analysis.

Open coding of transcripts was performed by 3 team members (ZB, MY, SM) to identify initial codes. A priori topics from the interview guide informed codebook development deductively, including general codes regarding positive or negative opinions on documentation in the EHR or data sharing with various health care personnel. New codes were added inductively as themes emerged, such as the need for caregiver involvement in decision-making regarding data sharing (Supplemental Table 4). All interviews were coded independently by 2 team members. The degree of interrater agreement was measured by using the $κ$ coefficient. After independent coding, the coders (ZB, MY, SM) identified portions of coded text that had a $κ$ <0.6 (suggesting low agreement in that area of coded text) and resolved discrepancies through discussion and consensus. The final $κ$ coefficient average for each code was 0.98, with a range of 0.91–1.00 across all codes, suggesting excellent agreement.

Baseline demographic characteristics of the interviewed sample are reported in Table 1. The 20 interviewed parents were uniformly female, with 55% identifying as Black or African American and 20% as Hispanic or Latino. Half reported a household income of <$40 000 per year, and 65% had a child who was publicly insured. Among the parents interviewed, 25% reported no HRSNs, 60% reported 1 to 2 HRSNs, and 15% reported 3 or more. Parents endorsed needs across each of the 6 domains included in the screener.

Themes centered around 2 domains: (1) documenting HRSN data and (2) sharing HRSN data. Emergent themes and representative quotations are discussed below and detailed in Tables 2 and 3.

Four themes emerged regarding the documentation of HRSN in the EHR.

First, most parents felt comfortable with documenting HRSN data in their child’s EHR, highlighting the connection between social needs and health and the potential to receive needed resources. Parents noted that it may be difficult to provide resources to families if HRSN data cannot be documented and shared in some way with social work or the medical team. A small number of parents felt that HRSN data should not be documented in the child’s EHR, but instead should be stored in a different server to protect privacy.

Second, some parents expressed hesitancy around the documentation of parental mental health and IPV in their child’s EHR. They saw these HRSN as distinct from other HRSN, such as food or housing insecurity, because they felt they were primarily related to the parent. These HRSN were seen as belonging in the parent’s health record rather than their child’s. One parent stated, “Me answering these questions about my mental health should not be on – in my opinion, should not be on his medical history at all. We’re two separate people.”

Third, most parents felt that HRSN data should not be collected and documented without delivering meaningful support to families, noting that it would require families to divulge sensitive data without the receipt of assistance in return. As one parent shared, “I’m kind of torn, honestly. Because I do like to see [HRSN information in the health record] because then you’re aware. Only if you’re able to truly offer help with these things, because, for one, I feel like they’re very – it is invasive…So, you’re being invasive, but can you really help?”

Fourth, when shown examples of ICD-10 Z codes in a mock after-visit summary (Supplemental Information), parents generally felt comfortable with how they were presented and approved of the codes being documented in their child’s EHR. Parents felt that using the Z codes presented as part of the project enabled discreet communication of HRSNs. As one parent stated “Using codes to abbreviate…things, I think it is a really good idea. Just to be a little discreet, but still have that in the system.”

Four themes emerged regarding the sharing of HRSN.

First, parents felt comfortable sharing HRSN data with both social workers and the medical team. Parents felt that social workers should be primary recipients of HRSN data because they are best equipped to provide resources. They also felt that the medical team should receive HRSN information to inform and facilitate a safe discharge, including planning for needed outpatient medical and social support. Many parents noted the connection between unmet HRSN and health outcomes, such as food insecurity affecting the ability to recover from illness, and most felt that it was important for the medical team to be aware of these needs.

Second, despite parents’ general comfort with Z codes, they felt less comfortable with insurance companies receiving HRSN data, largely due to the perception that payors would be unlikely to provide meaningful support. As one parent shared, “I mean, unless if they’re gonna provide the resource I need. But if they’re not, I don’t think they should be knowing of this.”

Third, although parents felt comfortable sharing their needs with social workers and the medical team, some parents identified potential negative impacts of HRSN data sharing on parent and family safety. When asked about sharing one parent’s screener responses with other family members, many participants were concerned that this could cause relationship strain or compromise the responding parent’s safety, especially related to IPV. Some parents worried that documenting and sharing HRSN data could potentially result in increased referrals to CPS when families endorsed resource limitations. A parent wondered, “If I say [I have] any of these issues, will [CPS] be called?…Are my kids in jeopardy of being taken away? What if I answered one of these questions wrong and they make a complaint to [CPS], but it’s not really needed?”.

Finally, because of these concerns and a desire for autonomy, participants expressed the importance of transparency around HRSN data sharing. Many participants noted the importance of respecting parents’ preferences regarding HRSN data sharing and providing opportunities for parents to share those preferences.

In this qualitative study on parents of hospitalized children participating in an HRSN screening and resource referral program, we examined parents’ perspectives on the documentation and sharing of HRSN data with hospital staff and payors. We found that most parents expressed comfort with the documentation of HRSN data in the EHR and approved of the use of ICD-10 Z codes if this information was used to provide them with meaningful support. Most parents viewed social workers and medical teams as the most appropriate recipients of HRSN data; however, few parents felt comfortable with HRSN data being shared with payors. Finally, some parents expressed concerns that documentation and sharing of HRSN data may lead to unwanted or unsafe disclosures and may put parents at risk for referrals to CPS. They desired more transparency and control around who would receive their HRSN data. Taken together, our results contribute to a small but growing body of literature regarding parent perspectives on the documentation and sharing of HRSN.^15,22–28 

Our finding that most parents expressed comfort with the documentation of HRSN data in the EHR aligns with previous research.^15,22,23,25  In the few studies examining parent perspectives on HRSN EHR documentation, results suggest that parents have mixed feelings and would prefer to have transparency and participate in shared decision-making to determine what is documented in the EHR.^15,22,23,25  Our study is among the first to examine parent perspectives on ICD-10 Z codes. Among the parents of hospitalized pediatric patients included in this study, there was general comfort with their use; however, we only asked about a small subset of available ICD-10 Z codes. The use of ICD-10 Z codes will likely accelerate as documentation of HRSN is increasingly tied to reimbursements. Thus, more work needs to be conducted to examine acceptability across a wider range of Z codes. Additionally, we did not explore parent comfort with additional information about their HRSN being included in the chart in social work or medical provider notes. Given our findings about parents’ comfort with more discreet forms of documentation, the authors of future studies should explore parent opinions regarding how HRSN should ideally be documented within EHR notes.

In addition, parents expressed concerns about HRSN EHR documentation and information sharing contributing to inadvertent or unsafe disclosures to the other parent. This concern has been raised in previous studies around the acceptability of HRSN screening and highlights the need for documentation standards that prioritize confidentiality and safety.^29,30  Importantly, data protection may be especially challenging in the pediatric environment because EHRs may not be equipped to protect HRSN data from being shared with adolescent children (who may have portal access) or other legal guardians (who also have access to the child’s chart). To reliably protect parents’ HRSN data, health systems using EHRs need to collaborate with EHR providers and clinical informaticists to identify and address potential vulnerabilities in data documentation and sharing.³¹ 

Consistent with previous work, we found that most parents viewed social workers and medical teams as the most appropriate recipients of HRSN data.¹⁵  However, the authors of previous studies did not examine parents’ perspectives on HRSN data being shared with payors. Few of the participants included in this study were comfortable with their HRSN data being shared with payors, citing the perceived inability of payors to provide meaningful support. This has implications for payors who may increasingly receive HRSN data on their beneficiaries from health systems to facilitate reimbursements. Additionally, some payors are directly screening their beneficiaries. Thus, the acceptability and effectiveness of insurers engaging in social care deserve more investigation.

Finally, our finding that parents expressed concerns that documentation and sharing of HRSN may lead to child welfare referrals aligns with previous work.¹⁵  Parents in previous studies similarly expressed concerns around CPS reports as a result of challenges meeting their basic needs.³²  These findings highlight the need for institutions implementing these programs to proactively track balancing metrics, such as increases in or disparities in referrals to CPS, to detect and address any emerging patterns of bias. It is particularly crucial to track racial inequities in screening rates, provider responses to positive HRSN screens, and successful connection to resources, given well-known issues with provider implicit biases and systemic racism. In addition, as health systems expand HRSN screening, dedicated training of providers around bias related to HRSN screening may help to mitigate unwarranted CPS referrals solely for poverty-related HRSN identified through screening.

The results of this study must be interpreted in the context of its limitations. First, this is a small qualitative study conducted within an academic quaternary care children’s hospital; therefore, generalizability may be limited. Second, selection bias may exist among parents who agreed to participate in interviews. We mitigated this by using purposive sampling to interview parents who did and did not endorse HRSN on the initial screener. However, it is possible that parents with more concerns about HRSN screening, documentation, and data sharing are more likely to opt out of screening; thus, there is the potential that parents with more concerns are underrepresented in this study. Third, we were unable to include parents who speak languages other than English because our screening tool had not yet been translated into other languages. We are now screening parents in 6 different languages, and we acknowledge that eliciting the perspectives of families who speak languages other than English is an important area of future study. Finally, we examined perceptions across a subset of Z codes as focused examples. Thus, more work is needed to examine parents’ perceptions across a broader range of Z codes.

In this qualitative study, parents generally expressed comfort with the documentation and sharing of HRSN within the inpatient health system, but not with payors. Parents highlighted several potential negative impacts of HRSN documentation and data sharing. These results have implications for both health systems and insurers regarding HRSN data management. Future work on HRSN documentation and data sharing should be conducted to further examine perceptions across a wider range of Z codes, perceptions of the roles of payors, and the impacts of broad-based HRSN screening on CPS involvement.

We would like to acknowledge the contributions of every member of the team implementing the HRSN quality improvement program in which this study is nested. We want to thank the patients and families for their contribution to this project.