A Clinical Decision Support System for Addressing Health-Related Social Needs in Emergency Department: Defining End User Needs and Preferences

Abstract

Background Health-related social needs (HRSNs) are the unmet social and economic needs (e.g., housing instability) that affect individuals' health and well-being. HRSNs are associated with more emergency department (ED) visits, longer stays, and worse health outcomes. More than a third of ED patients have at least one HRSN, yet patients are rarely screened for HRSNs in the ED. A clinical decision support (CDS) system with predictive modeling offers a promising approach to identifying patients systematically and efficiently with HRSNs in the ED.

Objective This study aimed to identify ED clinician and staff preferences for designing and implementing an HRSN-related CDS system.

Methods A multistep, user-centered design study involving qualitative semistructured interviews, observations of ED workflows, and a multidisciplinary design workshop.

Results We conducted 16 semistructured interviews with ED clinicians and staff. Following the interviews, three research team members observed ED workflows, focusing on patient entry and clinician and staff usage of the electronic health record (EHR) system. Finally, we conducted a 3-hour multidisciplinary design workshop. An HRSN-related CDS system should be visually appealing, color-coordinated, and easily accessible in the EHR. An HRSN-related CDS system should target a select group of ED patients (to be discharged from the ED) and highlight a select set of critical HRSN issues early in the workflow to adjust clinical care adequately. An HRSN-related CDS system should provide a list of actions and the ability to notify the clinical team if the patient's HRSNs were addressed.

Conclusion The user-centered design identified a set of specific preferences for an HRSN-related CDS system to be implemented in the ED. Future work will focus on implementing and refining the CDS system and assessing the rates of changes in clinical care (e.g., rates of referrals) to address patient HRSNs in the ED.

Keywords

health-related social needs - clinical decision support system - emergency department - user-centered design

Background and Significance

Health-related social needs (HRSNs) are patient's unmet social and economic needs (e.g., housing instability, homelessness, and nutrition insecurity) that contribute to poor health.[1] HRSNs are associated with repeated emergency department (ED) visits, longer ED visit times, increased health care utilization, increased costs, and poor patient outcomes.[1] [2] [3] [4] [5] Due to reporting requirements from the Centers for Medicare and Medicaid Services, The Healthcare Effectiveness Data and Information Set (HEDIS) quality measures from the National Committee on Quality Assurance, and recent Joint Commission accreditation requirements, HRSN screening is a critical topic for health care organizations.[6] [7] [8] Estimates vary, but studies show that up to two-thirds of ED patients experience at least one HRSN, which may directly impact their care and outcomes.[9] [10]

The ED poses unique challenges to routine and systematic HRSN screening compared to other settings,[11] [12] with unsystematic, costly, and potentially biased results.[9] [13] [14] [15] [16] ED clinicians report hesitancy in administering HRSN screening.[9] [13] [15] [17] Likewise, ED patients are often unwilling to disclose—in whole or in part—their HRSNs for fear of negative impacts on their care.[17] [18] [19] [20] [21] Additionally, the high volume of ED care creates workflow and time constraints.[9] [13] [15] Subsequently, ED clinicians and staff report that HRSN information is inconsistent and often not incorporated into clinical care.[17] Effective identification of patients with HRSN requires workflow redesign with buy-in from clinicians and patients.

Applying health information technology, such as the Clinical Decision Support (CDS) system, can support effective ED workflows for identifying patients with HRSNs.[22] [23] [24] A CDS system using algorithms can address numerous ED challenges, such as time constraints, lack of integrated data systems, and insufficient staff training.[25] [26] [27] ED physicians often underutilize patients' prior HRSN screening results. However, leveraging robust data from longitudinal electronic health records (EHRs) and health information exchanges can help identify patients needing screening or referrals.[26] [28] A data-driven CDS system would promote universal screening by avoiding missed screening due to selective questionnaire administration or patient nonresponse. This approach is not entirely new, as ED physicians already use algorithms for conditions like sepsis and opioid use disorder.[29] [30] [31]

Nevertheless, little evidence exists on designing and implementing an HRSN-related CDS system while accounting for critical factors affecting the uptake of information technology interventions. This paper describes the ED clinician and staff needs and preferences for an HRSN-related CDS system to support identifying patients with HRSNs in the ED. Our findings contribute to developing and evaluating CDS systems with predictive modeling to effectively address patients' HRSNs in the ED and other settings.

Objectives

We conducted a multistep, user-centered design study to identify ED clinician and staff preferences for designing and implementing an HRSN-related CDS system. A user-centered design approach minimizes design errors and future usability issues by incorporating end-user perspectives at every stage of development.[32] [33] The user-centered design approach has been shown to enhance the CDS system's uptake and adoption by addressing end-user goals.[34] [35] Our goal was to identify end-user needs and preferences, such as informational layout and organization, to be a foundation for developing an HRSN-related CDS system for EDs.

Methods

Overview

Our multistep study consisted of semistructured qualitative interviews with ED clinicians and staff, observations of ED workflows, and a multidisciplinary design workshop (see [Fig. 1]). This study is part of a larger project to improve patient HRSN screenings in the ED using informatics tools (AHRQ-1R01HS028008). The Indiana University Institutional Review Board approved this study.

Fig. 1 Overview of methods. ED, emergency department.

Semistructured Qualitative Interviews

A detailed description of our recruitment and analytical approach is presented elsewhere and in [Supplementary Material S1] (available in the online version).[17] We recruited full-time ED clinicians and staff from a 300+ bed public hospital system with more than 100,000 annual ED visits in the mid-Western United States. We used a pragmatic sampling strategy to recruit diverse participants across ED clinicians and clinical staff. We reached thematic saturation with 16 interviews. Thematic saturation is when no new themes emerge, and the researchers are confident after repeatedly observing similar data instances with well-developed themes. Each participant gave verbal informed consent before the interview. Participants were compensated with a $50 gift card for their time.

Our research team designed a semistructured interview guide using patient HRSN literature, the 5 Rights of CDS framework,[34] and the Contextual Information Model[35] (see [Supplementary Material S1] [available in the online version] for a complete interview guide). The first half of the interview asked participants about the accessibility, collection methods, and utilization of HRSN information in clinical ED care. The second half of the interview guide asked for participants' insights on the CDS system's design using the 5 Rights framework and Contextual Information Model. The 5 Rights framework guides the CDS design, implementation, and evaluation. It includes the following CDS system requirements: (1) provision of key data elements (right information); (2) the appropriate member for the information (right person); (3) a preferred point in the workflow (right time); (4) how the information could be accessed (right channel); and (5) how end users receive the information (right format). The Contextual Information Model describes individual end-users' perceptions of fit, that is, the level of congruence between the intervention, user, organizational culture, and workflow.[36] [37] We pilot-tested our interview guide, assessing its length and content, with a nurse practitioner, a social worker, and two patient members from our advisory panel.

From December 2022 to May 2023, at least two team members (an experienced qualitative researcher, a notetaker, and occasionally a third observer: O.M., A.T.H., C.M., and J.R.V.) conducted interviews using an online meeting platform. The interviewers had no prior personal or professional relationships with study participants. The study team (O.M., A.T.H., C.M., J.R.V., and C.A.H.) reviewed the data regularly as they were acquired and continued collecting data until no new themes emerged, indicating thematic saturation.[38] [39] [40] Participants provided demographic and professional details via a web survey before the interviews.

We used a modified thematic analysis to analyze interview transcripts. Initial screenings of three transcripts ensured the adequacy of interview questions. We developed a codebook, tested its reliability on three transcripts, and refined it collaboratively (O.M., A.T.H., C.M., and J.R.V.). Next, two members (O.M. and C.M.) independently coded each transcript, resolving discrepancies to form a consensus on themes and quotes. Axial coding was subsequently used to distill overarching themes, with final themes agreed upon through team discussion. To ensure the robustness of our qualitative analysis, we practiced reflexivity, provided detailed descriptions, and considered alternative explanations. We used Dedoose software version 8.2 for analysis. Our advisory panel validated the findings.

Observations of Emergency Department Workflow

After completing qualitative interviews, a research team member (J.R.V.) with two research staff (L.S. and P.A.) observed ED workflows. The team had opportunistic conversations with one physician, two registration desk members, one registered nurse (RN), and one patient navigator. The three members took field notes independently by systematically capturing information from the observation of the ED and conversations on patient entry, clinician and staff usage of the EHR system, ED workflow, opportunities for collecting and acting on HRSN-related information, and other pertinent observations. After the observation, the team members consolidated their notes through a joint discussion to produce a synopsis document for further use in the design workshop.

Design Workshop

Based on the qualitative interview data and observations described above, we identified several informational layouts and organization preferences for an HRSN-related CDS system. We conducted a 3-hour design workshop to refine the end-user preferences and produce low-fidelity CDS system prototypes. The following individuals participated in the workshop: ED clinicians (n = 4), an ED registered nurse (n = 1), a chief medical information officer for the health system (n = 1), a patient navigator working in the ED (n = 1), an EHR programmer (n = 1), user experience (UX) designer (n = 1), and our research team (n = 2) comprised an informatics researcher (J.R.V.) and research staff (L.S.). The patient navigator is a care team member assigned to assist the patient in navigating the health care system (e.g., explaining the procedures and ED workflow) and connecting to postdischarge care. A patient navigator was included in the workshop to provide key insights into the ED patient journey. Participants completed an informed consent form and were compensated for their time.

One team member (J.R.V.) opened the workshop with a brief presentation outlining the study's purpose and key qualitative interview findings. Next, participants were divided into two small groups. One team member was assigned to moderate the small group discussion on brainstorming solutions to CDS system design questions centered on the 5 Rights Framework. Participants also sketched low-fidelity CDS system prototypes with guidance from the research team members and a UX designer. Following small group discussions, each group presented their responses and design preferences with a concluding discussion on the preferred CDS system information layout and organization.

After the workshop, we qualitatively analyzed detailed notes and sketches produced by workshop participants. One team member who was not present at the workshop reviewed notes from two notetakers and sketches and produced an initial summary (O.M.). Another team member reviewed and compared the summary to the workshop's notes and takeaways (J.R.V.). These two team members produced a merged summary through a consensus process to arrive at a final list of informational layouts and CDS system organization.

Results

We completed 16 interviews with ED clinicians and staff. Interview participants were mostly female and White and, on average, had 7 years of experience practicing medicine (see [Table 1] with participant demographic information). Our design workshop participants had similar demographic characteristics (see [Table 2]). [Table 3] presents the illustrative quotes for each theme from the qualitative interviews.

Table 1 Demographics of qualitative interview participants (n = 16)

Characteristics	ED providers (n = 8)	ED staff (n = 8)	Total (n = 16)
Gender
Female	50.0	87.5	68.7
Male	50.0	12.5	31.3
Transgender	0.0	0.0	0.0
Race/ethnicity
Asian	12.5	0.0	6.3
Black or African American	0.0	37.5	18.7
Hispanic	0.0	12.5	6.3
Multiple/other	25.0	12.5	18.8
White	62.5	37.5	50.0
Age (mean, SD)	37.8 (7.2)	41.4 (10.9)	39.6 (9.1)
Work experience (mean, y)	7.6 (8.2)	6.1 (5.2)	6.85 (6.7)

Abbreviations: ED, emergency department; SD, standard deviation.

Table 2 Demographics of user-centered design participants (n = 8), excluding the research team

Characteristics	Count, percentage
Participant type
ED clinicians	4 (50)
ED registered nurse	1 (12.5)
Chief medical information officer	1 (12.5)
ED patient navigator	1 (12.5)
An EHR programmer	1 (12.5)
Gender
Female	2 (25)
Male	6 (75)
Race/ethnicity
Asian	0
Black or African American	2 (25)
Hispanic	0
White	6 (75)
Work experience (mean years, SD)	9.1 (2.1)

Abbreviations: ED, emergency department; EHR, electronic health record; SD, standard deviation.

Work experience was calculated using only data from ED clinicians, registered nurses, and a patient navigator.

Table 3 Themes from qualitative interviews with emergency department clinicians and staff (N = 16) with illustrative quotes

Theme	Description and representative quotes
1. The HRSN-related CDS system must stand out in the EHR to ensure proper use.
ED clinician 1.A	A one-liner, that can highlight red and bold, like, an abnormal vital sign with because it's not a life or a death thing, but you want it extremely visible, and that when someone knows that they want to look for it, they'll know where to find it in an instant. (#1)
1.B	Something very straightforward and simple. Something that's not going to be, nothing loud, I guess. Nothing that's going to be like, you know, startling and abrupt and, you know, not red. Um, but yeah. Something that's going to be like more muted and is just part of the chart that you're looking at. (#3)
1.C	Probably just a list of what their concerns were. A number is gonna be hard for me to know what could be with that (#5).
1.D	Numbers are meaningless. Because we do a lot of other calculators and comorbidities where we get a number, heart score we get a number and so numbers without a tool to tell us what those numbers mean would be pretty useless. And so, I guess a flag system of holy smokes, this patient is the highest risk we can be versus I'm not too worried about this patient. (#6)
1.E	If we had guidelines as to what to do with each number, or like each set of numbers of ranges of numbers, then like maybe a numerical rating would be able to do that, and maybe setting the expectation that like we can only address one or two of these, and in the ED visit (#8).
1.F	I would not use a numeric value to represent, these issues, because… And I get it the computer needs an algorithm to think and process through things, but it would be more useful as a provider if there were those boxes where you would put those numbers and it would either be like green or red or green, yellow, red, or something and I can take my mouse and highlight over that box and it says, “Patient homeless,” or “Patient uninsured, unemployed,” and so, or “Patient substance abuse,” “Patient food scarcity.” That would be more specific. (#7)
1.G	If it's more than one or two clicks on an electronic health record it will not get used (#6).
ED staff 1.H	Give me a, a scale, so I can have an idea. Because anything absolute, like, I have to go and recheck it, because of the population we serve. But if I got a range of a scale, I can deal with it (#14).
1.I	I would want to know who's gathering the information. What determines a score? (#13)
1.J	I would probably guarantee that over 50 percent of patients we see is going to ping this algorithm.(#15)
2. The HRSN-related CDS system should appear at the beginning or end of the workflow.
ED clinician 2.A	Definitely early, yeah. Earlier than later (#2)
2.B	Right after the triage is completed, the initial triage, because that gives us the most time to be able to get the ball rolling on services and getting people consulted and know, notifying people that these, this patient needs additional assistance (#9)
2.C	You wouldn't wanna put that tool in the discharge area of the workflow because by the time you get there, the patient that, you know—we're trying to move patients to rooms, so, uh, the patient's gonna be leaving the room (#4).
2.D	The hard part with stuff like this is if you do it on the front end you either anchor or you get biased to actually why they're there versus trying to do everything else with the patient. So if I'm too worried about their social stuff up front I could see an environment where I'm neglecting why they're actually there. If they're [patient] is there for their sprained ankle they want to address that first, um, or their sore throat or their chest pain. And so if you overload them with other stuff or if you overload the docs with other stuff too early it decreases the value or, it can affect the clinical outcome I would guess (#6)
2.E	I think it would be the beginning and the end, because it may affect your initial assessment somewhat, and it may—it will definitely affect discharge for anything, whether from the ED or from inpatient (#3)
ED staff 2.F	In that tab for us (social work) that it just popped up, like, if I pulled up a, a patient and it just popped up there, I would know one, two, three, four, and I could address that before I even go in there, to try to readdress it, and I could take a resource, if I needed to, that could save me time, as well (#14)
3. The HRSN-related CDS system should be accessible to all care team members.
ED clinician 3.A	In our Epic secure Chat to be able to link us the patient's chart to a social worker for them to read and, you know, like, “Hey, I'm reaching out because of these concerns. This is what the patient is concerned about.” This is my conversation with them, that way they don't have to chart dive to see what was going on and then they can follow-up from there (#9).
3.B	The [health system] nurses are really great at like truly caring for their patients, and so I think part of that is—and a lot of the times they'll check me. They'll say “_____, you forgot X, Y, and Z. They can't get home” or “They can't afford this.” And like—so I think them having that information is super important. I also sometimes think that having registration have that information is also important, or at least access to care, because then they can try and see whether they can get them in with a sooner financial counselor, or figuring out kind of how to help them jump—jump start getting registered for insurance and things of that sort, so they can get plugged into the healthcare system. So I don't think it needs to be limited by any means because I think that data helps everybody (#8).
4. The HRSN-related CDS system can have different EHR locations.
ED clinician 4.A	That would be something that would be able to catch our eye on the disposition tab. (#6)
4.B	You could put an alert. What you can do is make it on the—there's an area on the left side with, where vital signs are placed, or—actually, not in the demographics area, no one ever clicks that. But next to the vital signs. (#1)
4.C	Sort of a sidebar—something highlighted (#4)
4.D	I think the track board would help (#12)
ED staff 4.E	That tab for us that it just popped up, if I pulled up a patient and it just popped up there, I would know one, two, three, four, and I could address that before I even go in there, to try to readdress it, and I could take a resource, if I needed to, that could save me time, as well (#14)

Abbreviations: CDS, clinical decision support; ED, emergency department; EHR, electronic health record; HRSN, health-related social need.

Qualitative Interviews

The HRSN-Related CDS System Must Stand Out in the EHR to Ensure Proper Use

Participants strongly opposed receiving a single number as a notification. Instead, they wanted to receive a number accompanied by numeric thresholds indicative of the severity of the patient HRSNs with colored ranges consistent across different populations. Alternatively, they preferred seeing patient HRSNs displayed as a chart (e.g., a wheel) with individual HRSNs as sections. Participants stressed that the HRSN-related CDS system should be easily accessible (i.e., a few clicks) and visible through a color-changing, passive, noninterruptive alert in the patient's chart (see 1.A–1.G illustrative quotes in [Table 3]).

The HRSN-Related CDS System Should Appear at the Beginning or End of the Workflow

Participants differed substantially in their views on the HRSN-related CDS system's location in the workflow. Some participants preferred to see the HRSN information early in the workflow, such as during the opening of the EHR or initial triage. Others preferred having the HSRN information at the beginning and end of the encounter. One participant advised against putting the HRSN score at the discharge, as it would be too late. Another participant wanted to see HRSN information in the disposition tab towards the end of the encounter; presumably, this would prevent clinicians from forming a biased opinion about a patient's situation and instead allow them to focus on the presenting medical concerns rather than the patient's HRSNs. Participants did not comment on whether workflow location should vary by clinical team role (see 2.A–2.E illustrative quotes in [Table 3]).

The HRSN-Related CDS System Should Be Accessible to All Care Team Members

Participants wanted the HRSN-related CDS system, potential actions, and resource list to be shareable via Epic (the EHR) Secure with other care team members (residents, nurses, social work, finance, registration, and patient navigators; see 3.A and 3.B illustrative quotes in [Table 3]).

The HRSN-Related CDS System Can Have Different EHR Locations

Participants offered different EHR locations for an HRSN-related CDS system, including the tracker board, disposition tab, sidebar, patient demographics section, vital signs section, triage notes, or a sidebar/storyboard (see 4.A–4.E illustrative quotes in [Table 3]).

Emergency Department Walk-through Observations

Two observations were relevant to the development of an HRSN-related CDS system. First, upon arrival, patients were registered in the EHR system using their name and date of birth and linked to medical record numbers, typically before providing any ED care. Second, the observed ED clinicians and staff customized their screen views within the EHR to limit exposure to existing patient HRSN and demographic information. HRSN information already available to them was often initially obscured by the ED tracker board or ED floor maps.

Design Workshop

Building on the identified CDS system information layout and organization (see 1.A–1.G and 2.A–2.F illustrative quotes in [Table 3]), participants drew images depicting how they prefer to review HRSNs captured in the CDS system. Participants recommended a color-coordinated bar chart system with a detailed list of potential issues resulting in patients being flagged as having high HRSN (e.g., financial instability, transportation issues), as they may have a differential impact on the care decisions and processes. For example, colors could be used to create contrasts: the critical issues could be highlighted (such as colored in red) to draw the end user's attention, with less critical issues depicted in another color, like green.

Relatedly, revising the workflow preferences, participants wanted to be promptly informed about the selective set of critical HRSN-related issues (homelessness, domestic violence, etc.), as they may require drastic changes to clinical care. These critical issues should appear early in the workflow, whereas issues like food insecurity, transportation difficulties, or challenges with follow-up appointments can be displayed at the discharge. This distinction helped clarify contradictory opinions from the qualitative interviews. Participants preferred a targeted application of the CDS system to patients being discharged instead of admitted to the hospital, as ED is the only opportunity to address their clinical and HRSN needs. Accessing the HRSN-related CDS system should trigger a list of potential actions to address patient HRSNs. These actions can include a message to a social worker or patient navigator and a smart phrase template to help document HRSNs.

Refining the preferred EHR access and location findings, participants stated that the HRSN-related CDS system's access and location should be customizable for different care team members. Specifically, RNs and patient navigators preferred the storyboard (a leftmost screen column with a quick overview of patient history). In contrast, ED clinicians favored the tracker board (display of all patients with an assigned clinician, current condition, and work-up progression) and discharge flowsheet (a worksheet used to place discharge orders for prescriptions and referrals). Finally, participants stressed the importance of closing the loop and being informed about whether and how patients receive resources and support. ED clinicians preferred to get notification from social workers or other team members on whether patients' HRSNs were addressed, as it would build trust in the CDS system. This is how one ED clinician stated it: “Being an ED doctor is like being a pilot and you have people jumping out of the plane, and you can't see what happens to them and never hear from them again. I really hope their parachute opened!”

Discussion

Our multistep study identified and validated several information layouts and organization preferences for an HRSN-related CDS system in the ED. Designers of EHR and CDS systems can use our findings when building their systems to address HRSN in ED and other settings.[41] [42]

Our findings indicate that ED clinicians and staff prefer visually appealing (e.g., color-coordinated bar chart), easily accessible HRSN-related CDS systems with a detailed view upon request. These preferences may be driven by current practices in customizing the EHR layout, as we discovered during walk-through observations, where prioritization of the tracker board or other patient information obscures or minimizes the EHR's display of existing HRSN-related patient information. Furthermore, clinician and staff preferences for a visually appealing, easy-to-review CDS system are likely driven by a desire to minimize the cognitive overload experienced in a fast-paced ED environment. Finally, growing awareness of the role of EHR complexities and inefficiencies in clinician burnout further explains why ED clinicians and staff want to improve EHR usability. Clinicians and staff's preference for providing a visual overview with interactive functionality to drill down for details when needed is one of the best practices for EHR design.[43] Additional study is needed to assess whether the visually appealing HRSN-related CDS system with interactive functionality will lead to actual changes in ED care and patient outcomes, such as rates of referrals to address patient HRSNs.

An overarching theme in our findings is ED clinician and staff preference for targeted application of the HRSN-related CDS system. Specifically, clinicians and staff want the CDS system only to target patients to be discharged from the ED and promptly highlight a selective set of critical HRSN-related issues (e.g., homelessness, domestic violence), as they may require drastic changes in clinical ED care. These preferences align closely with the 5 Rights framework by highlighting the importance of CDS systems designed for the right patient at the right time in the workflow with the right information. Likewise, these preferences exemplify using HRSN information to adjust clinical care practices and provide necessary assistance.[44] The targeted application of the CDS system needs to be hardwired into the workflow to be adequately used by ED clinicians and staff, as they may need to remember about this CDS system if it is only employed occasionally and for a subset of patients. Future studies should examine whether noninterruptive or best practice advisory alerts effectively prompt ED clinicians and staff to use HRSN-related CDS systems for a selected set of ED patients.

Notably, additional research is needed to assess whether clinician and staff choice of the target patient population and selective set of HRSNs to be displayed early in the ED care align with which patients and which HRSNs should be addressed in the ED. In other words, we need to assess whether focusing on patients to be discharged from the ED would capture patients with the highest HRSN or patient population who would benefit the most from being identified as having HRSNs in the ED. Similarly, we need to know whether a selective set of critical HRSN-related issues (e.g., homelessness, domestic violence) is sufficient to address a patient's HRSN effectively. Of note, the HRSNs prioritized by the ED physicians and staff in the current study are consistent with those included in the Centers for Medicare and Medicaid Services's quality measures for inpatients.[6] [7]

Two a priori concerns with an analytics-based EHR CDS system within the ED are correct patient identity and timely system performance.[24] [45] [46] Fortunately, our observation of the ED indicated current workflows mitigated these concerns. The robust patient registration process can support the correct identification of patients for linkage to existing information within the EHR. Furthermore, the immediate patient registration and identification confirmation provide ample time for the CDS to present results to end users. Thus, the time between patient registration and actual care would be sufficient to initiate data querying, data processing, and analytical processes necessary for using the CDS system. Future studies should test the HRSN-related CDS systems system's performance and timeliness in delivering HRSN information to end users.

Finally, we found that clinicians and staff expressed a desire for a simplified path to action to address patient HRSNs and be informed about the outcome of their actions (i.e., a closed-loop system). While not the norm in the case of HRSNs,[47] closed-loop referrals can facilitate care transitions for patients[48] and allow health care organizations to track the effectiveness of these strategies.[49] Our study adds that closed-loop referrals could also supply information about the effectiveness necessary to support end-user trust and buy-in with the CDS. This suggests that any CDS without an adequate system of referrals, services, and tracking may face acceptance challenges. Additional research is needed to identify the best channels for informing ED clinicians on the follow-up care received by ED patients.

Our study has several strengths and limitations. First, we focused on systems to identify and address patient HRSNs in the ED setting more effectively. Addressing HRSNs is an increasing priority for funders, federal agencies, and patient advocacy groups. HRSNs are highly prevalent among the ED population, making it a prime candidate for designing and implementing information technology (IT) interventions to address HRSNs. Effective HRSN-related CDS systems can potentially help millions of patients presenting to the ED with HRSNs. Our findings are also relevant to the broader literature on designing IT interventions to address HRSNs in other settings, such as primary care, nursing homes, and palliative care. Second, we employed a methodologically rigorous multistep study design to comprehensively capture end-user needs and preferences. Still, our data collection was limited to one ED located in the mid-Western United States; thus, the generalizability of our findings to other settings warrants further exploration. Relatedly, our participants may not be sufficiently diverse (mostly White and female), potentially further affecting the generalizability of our findings. Additionally, the number of design workshop participants was relatively small with nursing perspectives potentially underrepresented, warranting additional exploration in future studies.

Furthermore, it remains to be known whether our participants' preferences for information layout and organization improve processes or quality of care—this should be formally tested in future research. Indeed, in our ongoing work, we are translating these findings into decision support implementations in the EHR in the ED settings. We plan to comprehensively examine the effects of these CDS implementations on clinical decisions and patient outcomes.

Conclusion

We identified several information layouts and organization preferences for the HRSN-related CDS system from ED end users. Health care organizations may anticipate increased ED clinician and staff buy-in when incorporating these end-user preferences in designing and implementing CDS systems. A fully realized CDS system can improve ED workflow efficiency by facilitating timely and accurate referrals to address patients' HRSNs, streamlining ED care, and positively impacting care quality and patient outcomes. Additional research is needed to identify how to implement an HRSH (health-related social needs)-related CDS system to ensure sustainability and incorporate the patient perspective into the design of HRSH-related CDS systems.

Clinical Relevance Statement

An HRSN-related CDS system with predictive modeling offers a promising approach to systematically identifying patients with HRSNs in the ED. Following a user-centered design process, we identified several information layouts and organization preferences among ED clinicians and staff. These preferences should inform the design and implementation of CDS systems in the ED.

Multiple-Choice Questions

1. Which of the following was an ED clinician and staff preference for an HRSN-related CDS system to be used in the ED?

   • The CDS system should have a complex interface and target all ED patients.

   • The CDS system should be text-heavy, focus on administrative data, and be used after clinical care.

   • The CDS system should be visually appealing, color-coordinated, easily accessible in the EHR, highlight critical HRSN issues early, and notify the team if HRSNs were addressed.

   • The CDS system should exclusively record patient demographics and be accessible only to the chief medical officer.

   Correct Answer: The correct answer is option c. The CDS system should be visually appealing, color-coordinated, and easily accessible in the EHR. It should also highlight critical HRSN issues early and notify the team if HRSNs were addressed. The HRSN-related CDS system is described as one that is user-friendly in design, visually appealing, and color-coordinated to draw attention to important features. It should be integrated seamlessly into the EHR, target selected ED patients poised for discharge, flag critical HRSN issues at the start of care, and have a mechanism to inform the clinical team about the status of addressed HRSNs. Therefore, option. encompasses all these recommended features.

2. Which of the following was a method used in the study?

   • Focus groups

   • Blind interviews

   • Semistructured qualitative interviews, observations of ED workflows, and a user-centered design workshop.

   • Surveys

   Correct Answer: The correct answer is option c. Our multistep study consisted of semistructured qualitative interviews with ED clinicians and staff, observations of ED workflows, and a multidisciplinary design workshop.

Conflict of Interest

J.R.V. is a founder and equity holder in Uppstroms, LLC, a technology company. No other authors have anything to declare.

Acknowledgments

We also want to thank Lindsey M. Sanner and Philip Adeoye for their assistance with the ED walk-through and user-centered design workshop data collection.

Protection of Human Subjects

The study was performed in compliance with the World Medical Association Declaration of Helsinki on Ethical Principles for Medical Research Involving Human Subjects and was reviewed by Indiana University Institutional Review Board.

• References

• 1 Axelson DJ, Stull MJ, Coates WC. Social determinants of health: a missing link in emergency medicine training. AEM Educ Train 2017; 2 (01) 66-68
• 2 Fink-Samnick E. Managing the social determinants of health: Part I: fundamental knowledge for professional case management. Prof Case Manag 2018; 23 (03) 107-129
• 3 Pruitt Z, Emechebe N, Quast T, Taylor P, Bryant K. Expenditure reductions associated with a Social Service Referral Program. Popul Health Manag 2018; 21 (06) 469-476
• 4 Davis CI, Montgomery AE, Dichter ME, Taylor LD, Blosnich JR. Social determinants and emergency department utilization: findings from the Veterans Health Administration. Am J Emerg Med 2020; 38 (09) 1904-1909
• 5 Anderson ES, Hsieh D, Alter HJ. Social Emergency Medicine: embracing the dual role of the Emergency Department in acute care and population health. Ann Emerg Med 2016; 68 (01) 21-25
• 6 Centers for Medicare and Medicaid Services. Screening for Social Drivers of Health. Centers for Medicare and Medicaid Services Measures Inventory Tool. 2023 . Accessed March 25, 2024 at: https://cmit.cms.gov/cmit/#/MeasureView?variantId=10405&sectionNumber=1
• 7 The Physicians Foundation. Quality ID #487: Screening for Social Drivers of Health. Quality Payment Program. 2022 . Accessed March 25, 2024 at: https://qpp.cms.gov/docs/QPP_quality_measure_specifications/CQM-Measures/2023_Measure_487_MIPSCQM.pdf
• 8 Office of Minority Health. CMS Framework for Health Equity 2022–2032. Centers for Medicare and Medicaid Services. 2022 . Accessed September 12, 2024 at: https://www.cms.gov/files/document/cms-framework-health-equity-2022.pdf
• 9 Wallace AS, Luther B, Guo JW, Wang CY, Sisler S, Wong B. Implementing a Social Determinants Screening and Referral Infrastructure During Routine Emergency Department Visits, Utah, 2017-2018. Prev Chronic Dis 2020; 17 (101205018): E45
• 10 Molina MF, Li CN, Manchanda EC. et al. Prevalence of emergency department social risk and social needs. West J Emerg Med 2020; 21 (06) 152-161
• 11 Yore M, Fockele CE, Duber HC. et al. 2021 SAEM Consensus Conference Proceedings: Research Priorities for Implementing Emergency Department Screening for Social Risks and Needs. West J Emerg Med 2023; 24 (02) 302-311
• 12 Furbacher J, Fockele C, Buono BD. et al. 2021 SAEM Consensus Conference Proceedings: Research Priorities for Developing Emergency Department Screening Tools for Social Risks and Needs. West J Emerg Med 2022; 23 (06) 817-822
• 13 Samuels-Kalow ME, Boggs KM, Cash RE. et al. Screening for health-related social needs of emergency department patients. Ann Emerg Med 2021; 77 (01) 62-68
• 14 Miner JR, Westgard B, Olives TD, Patel R, Biros M. Hunger and food insecurity among patients in an urban emergency department. West J Emerg Med 2013; 14 (03) 253-262
• 15 Shankar KN. Social outreach in the emergency department: are we doing enough?. Ann Emerg Med 2015; 66 (03) 341-342
• 16 Bharmal N, Rennick A, Shideler A. et al. Health-related social needs: which patients respond to screening and who receives resources?. J Gen Intern Med 2023; 38 (12) 2695-2702
• 17 Mazurenko O, Hirsh AT, Harle CA, McNamee C, Vest JR. Health-related social needs information in the emergency department: clinician and patient perspectives on availability and use. BMC Emerg Med 2024; 24 (01) 45
• 18 Pinto AD, Glattstein-Young G, Mohamed A, Bloch G, Leung FH, Glazier RH. Building a foundation to reduce health inequities: routine collection of sociodemographic data in primary care. J Am Board Fam Med 2016; 29 (03) 348-355
• 19 Fichtenberg CM, Alley DE, Mistry KB. Improving social needs intervention research: key questions for advancing the field. Am J Prev Med 2019; 57 (6 suppl 1): S47-S54
• 20 Kusnoor SV, Koonce TY, Hurley ST. et al. Collection of social determinants of health in the community clinic setting: a cross-sectional study. BMC Public Health 2018; 18 (01) 550
• 21 De Marchis EH, Hessler D, Fichtenberg C. et al. Part I: A quantitative study of social risk screening acceptability in patients and caregivers. Am J Prev Med 2019; 57 (6 suppl 1): S25-S37
• 22 Aggarwal CC. Data Classification: Algorithms and Applications. 2015 . Accessed February 10, 2021 at: http://www.crcnetbase.com/isbn/9781466586758
• 23 Vest J, Halverson P, Harris L, Haut D, Menachemi N. Impact of risk stratification and referral to social services on subsequent emergency and hospital utilization. 2020
• 24 Kasthurirathne SN, Vest JR, Menachemi N, Halverson PK, Grannis SJ. Assessing the capacity of social determinants of health data to augment predictive models identifying patients in need of wraparound social services. J Am Med Inform Assoc 2018; 25 (01) 47-53
• 25 Garg A, Boynton-Jarrett R, Dworkin PH. Avoiding the unintended consequences of screening for social determinants of health. JAMA 2016; 316 (08) 813-814
• 26 Bazemore AW, Cottrell EK, Gold R. et al. “Community vital signs”: incorporating geocoded social determinants into electronic records to promote patient and population health. J Am Med Inform Assoc 2016; 23 (02) 407-412
• 27 Cantor MN, Thorpe L. Integrating data on social determinants of health into electronic health records. Health Aff (Millwood) 2018; 37 (04) 585-590
• 28 Vest JR, Adler-Milstein J, Gottlieb LM. et al. Assessment of structured data elements for social risk factors. Am J Manag Care 2022; 28 (01) e14-e23
• 29 Horng S, Sontag DA, Halpern Y, Jernite Y, Shapiro NI, Nathanson LA. Creating an automated trigger for sepsis clinical decision support at emergency department triage using machine learning. PLoS ONE 2017; 12 (04) e0174708
• 30 Chartash D, Paek H, Dziura JD. et al. Identifying opioid use disorder in the emergency department: multi-system electronic health record-based computable phenotype derivation and validation study. JMIR Med Inform 2019; 7 (04) e15794
• 31 Kijpaisalratana N, Sanglertsinlapachai D, Techaratsami S, Musikatavorn K, Saoraya J. Machine learning algorithms for early sepsis detection in the emergency department: a retrospective study. Int J Med Inform 2022; 160: 104689
• 32 Horsky J, Schiff GD, Johnston D, Mercincavage L, Bell D, Middleton B. Interface design principles for usable decision support: a targeted review of best practices for clinical prescribing interventions. J Biomed Inform 2012; 45 (06) 1202-1216
• 33 Luna DR, Rizzato Lede DA, Rubin L. et al. User-centered design improves the usability of drug-drug interaction alerts: a validation study in the real scenario. Stud Health Technol Inform 2017; 245: 1085-1089
• 34 Campbell R. The five “rights” of clinical decision support. J AHIMA 2013; 84 (10) 42-47 , quiz 48
• 35 Callen JL, Braithwaite J, Westbrook JI. Contextual implementation model: a framework for assisting clinical information system implementations. J Am Med Inform Assoc 2008; 15 (02) 255-262
• 36 Goodhue DL, Thompson RL. Task-technology fit and individual performance. Manage Inf Syst Q 1995; 19 (02) 213-236
• 37 Tjan AK. Finally, a way to put your Internet portfolio in order. Harv Bus Rev 2001; 79 (02) 76-85 , 156
• 38 Azungah T. Qualitative research: deductive and inductive approaches to data analysis. Qual Res J 2018; 18 (04) 383-400
• 39 Charmaz K. Constructing Grounded Theory: A Practical Guide through Qualitative Analysis. 1st ed. SAGE Publications Ltd; 2006 . Accessed November 29, 2022 at: http://www.sxf.uevora.pt/wp-content/uploads/2013/03/Charmaz_2006.pdf
• 40 Davies D, Dodd J. Qualitative research and the question of rigor. Qual Health Res 2002; 12 (02) 279-289
• 41 Metaxas A, Hantgan S, Wang KW, Desai J, Zwerling S, Jariwala SP. A framework for social needs-based medical biodesign innovation. Appl Clin Inform 2024; 15 (03) 456-459
• 42 Langevin R, Berry ABL, Zhang J. et al. Implementation fidelity of chatbot screening for social needs: acceptability, feasibility, appropriateness. Appl Clin Inform 2023; 14 (02) 374-391
• 43 Shneiderman B. The eyes have it: a task by data type taxonomy for information visualizations. In: Bederson BB, Shneiderman B. eds. The Craft of Information Visualization. Interactive Technologies. Morgan Kaufmann; 2003: 364-371
• 44 National Academies of Sciences Engineering. . Medicine; Integrating Social Care into the Delivery of Health Care: Moving Upstream to Improve the Nation's Health. The National Academies Press;; 2019.
• 45 Gray GM, Zirikly A, Ahumada LM. et al. Application of natural language processing to identify social needs from patient medical notes: development and assessment of a scalable, performant, and rule-based model in an integrated healthcare delivery system. JAMIA Open 2023; 6 (04) ooad085
• 46 Ben-Assuli O, Vest JR. ED Revisits Forecasting: Utilizing Latent Models BT. In: Bi Y, Bhatia R, Kapoor S. eds. Intelligent Systems and Applications. Springer International Publishing;; 2020: 696-702
• 47 Drewry MB, Yanguela J, Khanna A. et al. A systematic review of electronic community resource referral systems. Am J Prev Med 2023; 65 (06) 1142-1152
• 48 Haynes LA, Casareno C, Fatema S. et al. “What matters to you?”: a participant-centered approach to needs identification and referral to community resources. Maternal Child Health J 2023; 28: 905-914 . Accessed April 26, 2024 at: https://link.springer.com/article/10.1007/s10995-023-03865-4
• 49 Beidler LB, Razon N, Lang H, Fraze TK. “More than just giving them a piece of paper”: Interviews with primary care on social needs referrals to community-based organizations. J Gen Intern Med 2022; 37 (16) 4160-4167

Publication History

Received: 20 May 2024

Accepted: 04 September 2024

Article published online:
18 December 2024