Targeted EHR-based communication of diagnostic uncertainty (TECU) in the emergency department: Protocol for an effectiveness implementation trial

Abstract

Background:

At least 37 % emergency department (ED) patients are discharged without a definitive diagnosis, resulting in diagnostic uncertainty. Transitions of care are high-risk periods for patient safety, especially for patients discharged with diagnostic uncertainty. Effective communication between clinicians and patients improves these care transitions. Previous research developed tools to support high quality discharge in the setting of diagnostic uncertainty, including: a communication checklist, a physician training program, and a patient-facing information sheet.

Methods:

This study aims to enhance transitions for patients leaving the ED with diagnostic uncertainty through an electronic health record (EHR)-based strategy. The intervention, the Targeted EHR-based Communication about Uncertainty strategy, integrates use of an Uncertainty Communication Checklist and an Uncertainty Discharge Document into routine ED discharge workflows. The checklist aids clinicians in addressing topics of chief concern for ED patients discharged with diagnostic uncertainty. The patient-centered discharge document explains the concept of diagnostic uncertainty to patients. The pre-post trial seeks to test the intervention’s effectiveness in reducing patient uncertainty and return ED visits compared to standard of care with 300 participants. It also aims to evaluate the intervention’s adoption and implementation barriers and facilitators.

Discussion:

This trial builds on previous research to establish a communication strategy that equips patients with essential information for safe transitions amidst diagnostic uncertainty. The EHR-based strategy seeks to standardize communication and minimize bias and variation. Findings will be disseminated, to inform future trials and improve understanding of patient experiences with diagnostic uncertainty.

Trial registration:

This trial was prospectively registered on 10/9/2024 with ClinicalTrials.gov (# NCT06638021).

1. Background

Diagnostic error is a significant cause of preventable patient harm across healthcare settings. A systematic review extrapolated that diagnostic errors in U.S. Emergency Departments (EDs) are prevalent (7.4 million cases annually) [1]. This report raised awareness of the need for more effective solutions to measure and address diagnostic errors in the ED’s unique and high-risk clinical environment.

Over one-third (37 %) of ED patients are discharged without a pathologic diagnosis [2]. While providing a symptom-based diagnosis at ED discharge is not necessarily a diagnostic error, it may delay diagnosis. Emergency physicians often provide symptom-based diagnoses (e. g., chest pain) after ruling out life-threatening etiologies [3,4]. Patients may receive a more definitive diagnosis along the diagnostic pathway, after further time and testing. Preliminary ED diagnoses occur frequently, but are not necessarily incorrect or inappropriate, particularly since a main emergency care role the identification of immediately life-threatening conditions. However, clinicians’ failure to acknowledge a diagnosis as preliminary, explain the limitations of the ED evaluation, and provide next steps may result in delay (or lack) of diagnosis. Despite the frequency of diagnostic uncertainty, ED clinicians receive minimal guidance on how to discuss it with patients [5]. Consequently, many clinicians do not discuss diagnostic uncertainty with patients, leaving gaps in patient understanding of next steps [6].

In response to this gap, our team worked with patients and international experts in communication, diagnostic uncertainty, and education to develop the Uncertainty Communication Checklist to guide ED discharge conversations about diagnostic uncertainty [7]. Conversations were designed to be supported by an accompanying Uncertainty Discharge Document, which reinforces key messages from the Uncertainty Communication Checklist. We additionally created the Uncertainty Communication Education Module to train clinicians on using the communication checklist; in our subsequent trial, the education module was effective in improving clinician uncertainty communication in a simulated environment [8,9].

A gap remains in translating these communication strategies into clinical care and demonstrating the impact of improved communication on patient care. This study builds upon our prior work by implementing a strategy via the electronic health record (EHR) – the Targeted EHR-based Communication about Uncertainty (TECU) - that integrates the Uncertainty Communication Checklist and Uncertainty Discharge Document into ED discharge. The study primarily focuses on effectiveness outcomes and secondarily explores implementation with an overall goal of facilitating higher quality discharge for ED patients with diagnostic uncertainty. Specifically, we aim to test the TECU’s effectiveness in management of patient uncertainty compared to usual care. The TECU has been designed to increase patients’ understanding of the presence of diagnostic uncertainty and helping them to manage the uncertainty about the significance of their symptoms.

2. Methods

2.1. Setting

The research is being conducted at Thomas Jefferson University Hospital (TJUH), an urban academic hospital located in Philadelphia, Pennsylvania, United States of America. TJUH had 70,942 ED visits, with 53,007 ED discharges in 2022. IRB approval was obtained from Thomas Jefferson University, with reliability agreements established via Smart IRB with Northwestern and Washington Universities.

2.2. Conceptual Frameworks/Process Redesign Frameworks

Conceptual frameworks were used to guide both intervention design as well as study design. The intervention is informed by Cognitive Load Theory. The study planning, implementation and evaluation is guided by the Reach, Efficacy, Adoption, Implementation, Maintenance (RE-AIM) and Practical Robust Implementation and Sustainability Model (PRISM) frameworks.

Cognitive Load Theory suggests that working memory has a limited capacity, which can be overwhelmed when managing complex information, but managed by instruction adaptation and breaking problems down into smaller parts [10-12]. The decision to develop the TECU intervention as an EHR alert that automatically triggers based on predefined criteria was made to reduce the cognitive load on clinicians by removing the need for them to independently identify diagnostic uncertainty for each patient at the time of discharge. Additionally, the Uncertainty Communication Checklist is provided as a link within the EHR alert to reduce the clinician cognitive load during these discharge conversations by providing an easily-accessible structured framework to guide the conversations.

The RE-AIM framework will guide the planning and evaluation of implementing the intervention [13]. RE-AIM outcomes reflect major considerations that should be addressed early in the process of developing a solution for a specific problem [14]. Initiating communication at the time of discharge with diagnostic uncertainty increases the opportunity to reach most of the target patient population. A hybrid type 1 effectiveness implementation design, informed by RE-AIM, enables an initial assessment of TECU’s effectiveness while also capturing key data on adoption, provider perception, and process outcomes. This information will be crucial for informing scalability and maintenance (although maintenance will not be evaluated in this study). Investigators will explore barriers and facilitators to adoption and implementation from both clinician and patient perspectives.

The Practical Robust Implementation and Sustainability Model complements RE-AIM by emphasizing contextual domains and guiding assessment of implementation barriers and facilitators [15]. This model will help categorize determinants of implementation in routine care and will inform future scaled implementation.

2.3. Study design

The study employs a hybrid type 1 effectiveness implementation design with a pre-post trial of the TECU at an urban academic ED. Hybrid type 1 effectiveness implementation study designs primarily focus on assessing the effectiveness of a clinical intervention and secondarily aim to understand the context of implementation [16]. The pre-post methodology was selected rather than randomization at the patient level or clinician level because a) the intervention is delivered by clinicians, so patient level randomization would result in contamination and b) ED care delivery is team-based, thus clinician level randomization would also result in contamination.

The effectiveness assessments of the hybrid trial will be accomplished through patient surveys and EHR review whereas the implementation assessments will be accomplished through clinical adoption observations, clinician surveys and interviews, and patient interviews. The primary goal is to determine the TECU effectiveness in improving patient management of uncertainty, with a secondary goal of understanding the context for implementation. This is a non-randomized trial, intervention allocation is based on presentation date to the ED, and the study phase (pre-TECU versus post-TECU). ED clinicians will receive training in diagnostic uncertainty communication and the TECU intervention before the post-TECU phase begins.

2.4. Participants

Patient eligibility criteria include: 1) age 18 or older, 2) English-speaking, 3) assigned one of a pre-defined list of symptom-based diagnoses as their primary ED discharge diagnosis, 4) planned for ED discharge, 5) able to complete follow-up assessments at 2 and 30 days, and 6) TECU Best Practice Advisory (BPA) activation delivered to clinician during discharge process (see intervention below). Exclusion criteria include: 1) clinically unstable, psychologically impaired or intoxicated, or otherwise unable to provide informed consent, 2) admitted to hospital, 3) previously enrolled, or 4) more than five ED visits in the past 30 days. Other than study team members, all emergency medicine clinicians at TJUH, including attending physicians, resident physicians, and Advance Practice Providers, will be eligible for inclusion in the clinician surveys and clinician focus groups.

2.5. Recruitment and consent

We seek to enroll 150 patients pre-TECU implementation and 150 patients post-TECU implementation. During both periods, eligible patients will be identified by a research coordinator through Epic notification each time an eligible patient is being discharged. In the pre-TECU period this notification is only visible to the RA team, but does not display to the clinicians. Eligible, interested patients will complete the informed consent process. During the consent discussion, staff will inform patients the study purpose is to evaluate how clinicians communicate with discharged ED patients with a symptom-based diagnosis. A subgroup of post-TECU patients will also be invited to participate in qualitative interviews and complete verbal consent.

All clinicians exposed to the TECU BPA will be invited to complete an implementation survey soliciting feedback after the trial is complete; consent will be electronic. On survey completion, clinicians can indicate interest in additionally participating in a focus group about their experience with the TECU, which will have a separate consent process.

2.6. Intervention

The TECU BPA was designed following the traditional build, validate, and implement information technology project phases. BPAs are customizable features in the Epic EHR that scan the patient record and display notifications or suggestions based on clinical data. Criteria rules serve as the starting point for the technology automation. In the case of the TECU BPA, rules identify patients with an ED discharge order and compare entered diagnoses against an established dictionary of uncertain symptom-based diagnosis codes. This dictionary was generated using Qlik Sense analytics to analyze EHR data from 72,478 patient discharges at TJUH ED between 6/1/2023 and 6/1/2024. After filtering for symptom-based ICD-10 codes in the R00-R99 range, a base set of 697 unique diagnoses across 7890 visits was identified. Only codes occurring 3 or more times were included in the final list (“grouper list” – see Appendix A) of 258 codes, covering 93 % of the raw set.

In the post-TECU period, the TECU intervention aims to improve counseling and help patients manage uncertainty through two routes.

• Enhanced counseling by the clinician (Uncertainty Communication Checklist). When the BPA is triggered, an EHR alert reminds clinicians to address diagnostic uncertainty with the patient. The alert includes a link to the Uncertainty Communication Checklist, which can be used as a discussion guide at the clinician’s discretion [7].

• Uncertainty Discharge Document. The one-page, plain language Uncertainty Discharge Document reinforces key Uncertainty Communication Checklist messages and is included in the discharge paperwork when the BPA is accepted. The document follows best practices for health literacy, ensuring readability below the 8th grade level. See Fig. 1.

Fig. 1.

Uncertainty discharge document.

Intervention validation included pilot testing and user feedback, confirming the TECU’s feasibility and optimizing its display. The results of the pilot testing and modifications will be published separately.

2.7. Study visits

The primary outcome assessment will occur at patient enrollment (T0) immediately after ED discharge through in-person survey completion. Secondary and exploratory outcomes will be collected about 2 days (T1) and 30 days (T2) after T0. Patient-reported data for T1 and T2 will be collected via survey links delivered by secure e-mail or text, or by phone, depending on patient preference. Patients will be contacted multiple times to minimize loss to follow-up. Participants will receive $105 for complete participation: $25 for enrollment assessments and $40 for each follow-up assessment.

2.8. Measurement

Patient data will be derived from four sources as detailed below.

• Patient Demographic Survey (T0): At enrollment, all patients will complete a socio-demographic and health questionnaire, including age, gender identity, race, ethnicity, self-reported health status (participants rate their general health from excellent to poor) [17], and health literacy (using the Single Item Literacy screener [18] and the Digital Health Care Literacy Scale) [19].

• Patient-reported Outcome Surveys (T0, T1 and T2): Surveys will assess perceived care transitions, anxiety, uncertainty, and self-reported 30-day ED healthcare utilization. Baseline assessment will be conducted during ED discharge. The first follow-up assessment will evaluate the patient experience days after their ED visit whereas the second follow-up assessment determines if time changes patients’ ratings of their care transition or uncertainty level and serves to catalogue repeat ED visits.

• Electronic Health Record (T2): 30-day healthcare utilization will be extracted from the TJUH Epic EHR and the HealthShare Exchange of Southeast Pennsylvania. A chart review at T2 will capture additional care episodes, including primary care or specialist visits, urgent care visits, ED visits, hospitalizations, and Jefferson telephone or EHR-message interactions.

• Patient Individual Interviews (T2): During post-TECU, audio-recorded interviews will explore the patients’ experience with TECU-guided discharge.

Clinician data will be derived from three sources:

• Adoption observations: During post-TECU, at least two random observations will be completed per month (totaling ~24 observations) to determine adoption of the TECU intervention using a structured checklist (e.g. time to discharge completion, patient receipt of paperwork). Observers will collect data from the EHR and from the clinician post-discharge, so that the observation does not impact the clinician’s behavior. Adoption will be further assessed using Epic work bench data to determine the diagnosis entered that triggered the BPA firing, BPA acceptance/denial, Uncertainty Communication Checklist link usage, and the percentage of clinicians using the BPA.

• Clinician surveys: Upon trial completion, we will invite clinicians who have interacted with the TECU to complete a survey (detailed below) exploring implementation barriers.

• Clinician focus groups: To understand clinician engagement with the TECU and their experiences in using it, we will conduct audio-recorded focus groups upon trial completion.

2.9. Effectiveness measures

Since diagnostic uncertainty is non-disease specific, it is difficult to select outcome measures that clearly demonstrate TECU’s impact on post-ED diagnosis or diagnostic accuracy [20]. We initially considered limiting the patient population to specific chief complaints, but ultimately opted for a more patient-centered outcomes measure (e.g., Uncertainty scale) as our primary outcome [21]. The details and collection timepoints of covariates and key outcome measures are outlined in Table 1 and described below.

Table 1.

Data collection schedule.

Outcome	Measurement	T0: Enrollment	T1: 2- Day	T2: 30- Day
Patient outcomes
Uncertainty	The Uncertainty Scale (U-Scale)	X	X	X
Care Transitions	Care Transition Measure (CTM-3)		X	X
	Transitions During Uncertainty (TDU) scale	X	X
Anxiety	The Spielberger State-Trait Anxiety Inventory (STAI) short form	X	X	X
Healthcare utilization
ED Return Visits	ED return visits (for any complaint) within 9 & 30-days post enrollment visit (via self-report and EHR query)			X
Utilization composite	30-day composite count of all healthcare interactions including: ED & clinic visits, admissions, testing appointments, Jefferson telephone or EHR-message interactions (via EHR query)		X

Uncertainty Scale (Primary Outcome):

The Uncertainty Scale is a 30-item measure of patient-level uncertainty related to acute symptoms and independent of underlying disease. Psychometric testing among patients completing the Uncertainty Scale at ED discharge demonstrated high internal consistency (Cronbach’s alpha 0.927, mean item-total correlations 0.53) and concurrent validity based on associations between neuroticism, anxiety, and patient uncertainty related to health events [21]. We secondarily will assess for differences between groups in the “uncertainty related to treatment quality” subdomain because this subdomain has been associated with return ED visits and contains questions closely linked to the care and communication received during the ED visit. See Table 2 for the Uncertainty Scale subdomains and example questions.

Table 2.

Uncertainty scale subdomains and example questions.

Domain	Example Question
Treatment Quality	Q17. I often feel like my doctors don’t give me enough information about my test results.
Diagnosis	Q6. I am not sure if my symptoms are a sign of a serious health problem.
Self-Management	Q3. I have the knowledge and ability to treat my symptoms.
Worries and Concerns	Q10. Feeling nervous about my symptoms is causing emotional distress.
Self-Efficacy	Q23. I don’t know where to go for health when experiencing different kinds of symptoms.
Decision to Seek Care	Q9. I don’t know which symptoms I should go see a medical professional about.
Psychosocial	Q14. I don’t know how to get affordable medical care for my symptoms

Quality of Care Transition:

The Care Transition Measure (CTM)-3, an abbreviated form of the Care Transition Measure (CTM)-15 [20,22], will assess care transition quality. The CTM-3 is useful measure to capture the quality of ED-to-home care transitions to aid quality improvement [23]. The CTM-3 has been associated with outcomes of care after an ED visit, (i.e. ED return visits and medication adherence) [23]. The CTM-3 collection between 2 and 30 days is supported by prior work [22,24].

Transitions During Uncertainty Scale:

The Transitions During Uncertainty Scale was designed and validated by our team to assess the quality-of-care transitions home from the ED, specifically for patients discharged with diagnostic uncertainty. (publication pending).

Spielberger State-Trait Anxiety Inventory (STAI-AD):

The 6-item short-form of this gold-standard measure provides a reliable and sensitive measure of anxiety in relation to an event [25].

Return ED visits and healthcare utilization:

We are tracking ED recidivism as “uncertainty about their condition,” has been noted in prior work as a key driver of ED return visits [26]. ED return rates will be evaluated at 9 days and 30 days. The 9-day ED return window is the only empirically derived timepoint for assessing acute returns potentially related to the initial visit [27]. Recognizing that increased patient awareness of diagnostic uncertainty may lead to more outpatient visits, ED return rates and outpatient follow-up care will be assessed through patient self-report at T2 (30 days) and EHR review.

Patient Socio-Demographic and Medical Characteristics:

Enrollment assessments will include a socio-demographic and health questionnaire as detailed above.

2.10. Implementation measures

Structured Adoption Checklist:

To assess fidelity, we have developed a process for assessing adoption as detailed above.

Implementation Survey Assessments:

We will focus on summative implementation assessment during the trial to evaluate early implementation outcomes. We will use the Normalisation MeAsure Development questionnaire (NoMAD) [28] to assess clinician perceptions of factors relevant to embedding the TECU intervention that change their work practices and the Clinical Sustainability Assessment Tool (CSAT) to assess organizational capacity for sustainability [29].

Clinician Focus Groups:

Clinician focus groups will be conducted upon trial completion to discuss intervention adoption and implementation within the busy ED practice environment. Additionally, we will explore how TECU exposure potentially triggered clinicians to change patients’ diagnoses from a symptom-based diagnosis (e.g., abdominal pain) to a specific diagnosis (e.g., gastritis).

Patient Individual Interviews:

Upon completion of their 30-day assessment, we will invite a subset of patients during post-TECU to complete a qualitative interview by phone or virtual conferencing about their experience. We will purposively approach 1–2 participants per week, to ensure balanced recruitment across age, gender, race/ethnicity, and discharge diagnosis symptom category. We will continue conduct interviews until thematic saturation [30-32] and data sufficiency [33] are reached.

Analysis Plan.

2.11. Power calculation

The trial sample size is calculated to have sufficient power to detect a difference between the pre and post groups of approximately 6 points on the Uncertainty Scale total score. This difference matches the average within-person change observed in patients surveyed at ED triage and ED discharge [34]. Assuming a standard deviation of 18 points [21], 150 patients per group gives greater than 82 % power to detect this difference (standardized effect size of 0.33) at the two-sided alpha = 0.05 level. Although these calculations do not account for clustering at the clinician level, we anticipate minimal clustering and any loss of power should be offset by the gains from jointly modeling the longitudinal outcomes.

The feasibility evaluation objective is to gather experiential feedback from site clinicians to inform planning for a subsequent TECU modification, sustainability and scaling; we will therefore conduct focus groups with all willing site clinicians independent of thematic saturation. We will additionally conduct patient interviews. We estimate conducting approximately 3–4 focus group sessions (~6–8 individuals/group) and conducting up to 30 patient interviews.

2.12. Effectiveness outcomes analysis

Descriptive statistics will summarize group characteristics at enrollment, and groups will be compared using standard two-group tests, including two-sample t-tests and chi-square tests. The primary outcome is the Uncertainty Scale total score, measured at discharge. We will model all measurement times using mixed effects linear regression, including fixed effects for group (pre-intervention vs. post-intervention), measurement time (discharge, day 2, day 30) and group by time interaction.

As this is a non-randomized trial, we may adjust for potential confounders including time of year, time of day, gender identity, age, race, ethnicity, underlying health conditions, self-reported health status and health literacy, prior experiences with medical uncertainty and ED care if there are differences between groups at the p < 0.2 level in univariate comparisons. We will model correlation among the repeated measurements using an appropriate residual covariance matrix structure, with the structure choice guided by goodness of fit as measured by the Akaike Information Criterion [35].

Clustering by clinician will be accounted for with a random intercept term. Group-specific means will be estimated and comparisons will be made using linear contrasts. Model assumptions of constant residual variance and normality of residuals will be examined graphically. If necessary, the Uncertainty Scale will be transformed to meet model assumptions. The same approach will be used for longitudinal secondary outcomes. ED visits and healthcare interactions will be modeled using Poisson or negative binomial regression with log (follow-up time) as the offset, adjusting for the same covariates as identified for the primary outcome. Incidence rates will be estimated by group, and the incidence rate ratio will be estimated for comparing groups. Incidence of return ED visit by day 9 and day 30 will be modeled using Poisson regression with robust standard errors to allow for estimation of the group-specific risks and relative risk [36]. Clustering by physician will be accounted for using generalized estimating equation methods. We will not perform formal hypothesis testing for secondary outcomes. All group-specific estimates and group comparison estimates will be accompanied by 95 % confidence intervals.

We do not anticipate any missing data at enrollment, and, thus, no missing data for the primary outcome at the primary time of interest. Missing data at 2 and 30 days will not be imputed. Using a mixed effects approach to jointly model all three measurement times provides unbiased group estimates under the assumption that any missing data are missing at random.

2.13. Implementation outcomes analysis

Consistent with RE-AIM, we will combine adoption data with survey data to identify barriers and facilitators and guide TECU design refinement and implementation support tailoring. We will summarize responses to the NoMAD and CSAT using descriptive statistics and comparing results by clinician role and level of TECU engagement. Data will be assessed via Kruskal-Wallis and Mann-Whitney tests (a = 0.05).

Qualitative data from patient interviews and clinician focus groups will be analyzed in NVivo using grounded theory-based inductive coding [37-39]. Transcripts will be analyzed by two independent coders, with a third acting as a tiebreaker before conducting thematic analysis. RE-AIM with PRISM frameworks will map the codes onto the theoretical barriers, facilitators, and implementation process. Observational data from the Structured adoption checklist will be summarized. A timeline of clinical care from ED arrival to discharge will be developed to identify common timepoints in which barriers appear.

3. Discussion

While clinicians are trained to always strive towards diagnosis [40], uncertainty for non-emergent conditions often persists beyond ED care. This uncertainty cannot be eliminated in many cases, however, patients’ responses to ongoing uncertainty can be influenced and potentially modified [26,41-43]. This trial builds on our previous work to establish a communication strategy that ensures patients receive the requisite information to manage uncertainty and support a safe transition home, even when diagnostic uncertainty persists. Historically, no standardized approach existed to communicate with patients about diagnostic uncertainty, despite research showing adequate communication is crucial for safe and high-quality ED discharges [44-49]. Our team developed a patient-centered approach to help emergency clinicians communicate diagnostic uncertainty [7,50-53], demonstrating in a randomized controlled trial success in teaching clinicians to effectively address this issue [8].

This trial will be the first to deliver an uncertainty communication intervention via the EHR. We selected the EHR for this strategy because it can alert clinicians in real-time, ensuring standardized communication, reducing clinician cognitive load, and minimizing variability and bias. Findings from this study will both inform future trials and help scale the intervention. The TECU has been built within EPIC with a goal of enabling dissemination through the shared Epic Community Library.

We made several decisions for the design of this study that involved weighing tradeoffs. Our study utilizes a single group pre-test/post-test design evaluating effectiveness and implementation. While this design lacks the rigor of a randomized trial (e.g. randomized stepped wedge) and might underestimate the impact of other unmeasured temporal changes in emergency care delivery of the course of the study enrollment, we chose it for a few important reasons. First, this is early-stage work and there is need to demonstrate efficacy outcomes before conducting a larger scale study. Second, as noted above, the pre-post design was chosen due to concerns about contamination between groups. We also chose to conduct the study only in English due to resource limitations of both staff and translated study materials. This choice may result in imprecisely estimating the impact of the intervention, as diagnostic uncertainty may be a more difficult concept to explain in the setting of language barriers.

In addition, we opted to include a wide range of symptom-based, uncertain diagnoses (rather than limiting to a few [e.g. “chest pain”, “abdominal pain”]) and to include patients who have had prior ED presentations. This breadth of complaints and participants is both a strength and potential vulnerability of the study. The heterogeneity of the patient population may lead to an underestimation of the impact if there are differences by symptom-based diagnoses or differences based on patient prior visit characteristics; however, if the intervention is successful despite this heterogeneity, it will support future study and implementation. We opted for this more pragmatic approach because heterogeneity of patient conditions and the existence of frequent ED utilizers are realities of emergency care, and we believe the work will be more impactful if the study criterion more closely mirror the ED environment rather than a highly curated subset of patients and conditions. It is simultaneously important to acknowledge for the study that uncertain diagnoses are defined by symptom-based diagnosis codes; however, other “diagnoses” may also be uncertain but would not be captured in this study. Additional limitations include the urban academic ED setting which may not be representative of all EDs and that informing patients of the purpose of the study may impact their responses on the surveys.

4. Conclusion

This trial will assess the effectiveness and implementation of a workflow-integrated solution during ED discharge for patients with diagnostic uncertainty. It aims to improve patient care by enabling improved discharge communication that is focused on helping patients to manage their uncertainty, thus improving care transitions home in the setting of diagnostic uncertainty.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.cct.2025.107910.

Data availability

No data was used for the research described in the article.