Presentation
PS13 - The Role of Electronic Health Record Data and Display in Promoting Diagnostic Equity: Insights from Clinicians and Health Equity Experts
SessionPoster Session 1
DescriptionBackground: Diagnostic error is often a symptom of health inequities in the clinical space, in which socially disadvantaged groups are disproportionately at risk for inaccurate, delayed, or missed diagnoses. Electronic Health Records (EHRs) often fall short in supporting equitable diagnosis due to poor usability, alert fatigue, and an overload of inactionable data. EHRs frequently fail to facilitate access to social, economic, and environmental determinants of health that may be critical to achieving accurate and equitable diagnosis and care. This work explores how healthcare providers and health equity experts conceptualize, experience, and use patient-level EHR data in the diagnostic process with a focus on diagnostic error-related data access and representation. These insights represent a foundational step towards optimizing EHR data collection and display, to prioritize equitable diagnostic practices.
Methods: We recruited healthcare providers (MD, DO, or APP) (n=10) and health equity experts (n=2) to participate in 60-minute semi-structured virtual interviews. Tailored interview guides were developed, focusing on a set of patient-level variables commonly associated with diagnostic equity. Questions focused on understanding how health and patient-level variables are utilized to support equitable diagnostic practices. Variables were organized into seven key categories commonly associated with health disparities: demographics (race/ethnicity, age, sex/gender, cultural and linguistic preferences, health literacy, marital status), socioeconomic status (income, education, insurance, employment, financial hardship, food insecurity), lifestyle factors (tobacco use, substance use, physical activity), clinical indicators (chronic conditions, healthcare utilization, adherence to preventative care guidelines, mental health diagnoses, substance use diagnoses), social support (safe housing, family/friends at home, caregiver support), geographic factors (social vulnerability index (SVI), area deprivation index (ADI)), and neighborhood and built environment (housing stability, transportation). Clinical participants were asked if and how they use these variables in the diagnostic process, how well these variables are captured in their EHR systems, and whether the data is accurate and actionable. Health equity experts were asked to provide best-practice perspectives on the integration of these variables into the diagnostic process. All interviews were recorded, transcribed, and analyzed using a rapid, thematic approach conducted by two researchers (GF and LS) to identify emergent themes and insights.
Results:
12 interviews were conducted between June to July 2024 with data collection ongoing. By March 2025, it is anticipated that the sample will increase to approximately 40 participants. 10 (83%) were providers and 2 (17%) were health equity experts. Most participants identified as women (n=10, 83%). Of the n=10 providers, most worked in primary care (n=6, 60%).
Findings by variable category:
Demographics: While many participants considered race when reviewing patient demographics, they expressed concerns about perpetuating race-based medicine, emphasizing the need for culturally sensitive patient care. Despite race and ethnicity being captured in the EHR, participants noted inaccuracies and inconsistent documentation. Health equity experts highlighted that race and ethnicity often serve as proxies for health disparities and systemic inequalities, advocating for more granular, self-reported data. Age, considered crucial for screening and treatment, was considered highly accurate due to its link to birthdate, though one health equity expert cautioned that biological age may vary based on factors such as stress or trauma. While sex was viewed as critical for screenings and treatment, gender data was frequently inaccurate, particularly for transgender and non-binary individuals. Sexual orientation was deemed important for sexual health conversations, but EHR accuracy was variable. Participants considered patient health literacy essential for effective communication, but its capture lacked standard measurement tools, leaving providers to rely on subjective evaluation. Marital status, while potentially offering insights into social support systems, was often outdated or incomplete, raising concerns about accuracy.
Socioeconomic Status: While recognized as potentially informative, income was not typically documented in the EHR, with providers often inferring it based on insurance status or patient conversations. One health equity researcher cautioned that income alone does not fully capture financial hardship, as debt and other factors are often excluded. Education level was often linked to health literacy, echoed by a health equity expert, and faced similar challenges of inconsistent capture and update in the EHR. In contrast, insurance information was considered readily accessible and accurate, playing a vital role in guiding treatment options and connecting patients with health and social resources. Employment status, financial hardship, and food insecurity, though potentially insightful for tailoring treatment plans, were frequently unreliable or missing in the EHR.
Lifestyle Factors: Participants emphasized the importance of considering lifestyle factors in patient diagnosis, particularly substance use. However, the dynamic nature of these behaviors and the time limitations of patient-provider interactions necessitate frequent updates to the EHR, which was not consistently practiced. A health equity researcher recommended using broader time scales, such as monthly rather than weekly use intervals, to improve accuracy in substance use reporting. Physical activity, while generally acknowledged as important for health and guiding patient treatment and lifestyle recommendations, was also inconsistently and inaccurately documented in the EHR.
Clinical Indicators: Chronic conditions were central to guiding treatment decisions, with the EHR's problem list serving as the primary repository for these data. However, participants had concerns about data accuracy. These conditions were typically documented under medical history or medication lists, potentially limiting their visibility in the context of broader health equity considerations. Healthcare utilization was often viewed as a proxy for care access and care-seeking behaviors, yet it lacked a structured field in the EHR, leading providers to rely on informal documentation such as notes and appointment histories.
Social Support: Access to social support was critical to consider within the diagnostic process, particularly for older adults, as it affects a patient’s ability to manage their care, attend appointments, and recover from surgery or illness. Providers shared that knowing whether a patient has a caregiver or family/friends at home can influence discharge planning or the recommendation of additional support services. However, capturing information about social support and safe housing relied heavily on patient conversations and lacked standardized documentation in the EHR. This information was often recorded in unstructured fields, such as notes, making it difficult to find and update in the EHR.
Geographic Factors: While the Area Deprivation Index (ADI) and Social Vulnerability Index (SVI) were acknowledged as valuable tools for understanding community-level health disparities, they were rarely used in individual diagnostic decisions. Health equity experts suggested focusing on access to local resources, such as sidewalks, grocery stores, and local community, rather than relying solely on geographic data.
Neighborhood and Built Environment: Capturing neighborhood and build environment data such as housing stability and transportation remained a significant challenge. Although these factors were recognized as important for understanding patient context, they were not always formally captured in the EHR. Instead, providers typically relied on patient conversations and addressed these factors only when directly relevant to the immediate medical concern.
Discussion: To better support diagnostic equity at the point of care, there is a need for improved EHR display, connectedness, and functionality. In addition to capturing accurate patient-level information, EHR systems may best serve providers and patients by providing actionable insights and facilitating access to essential community resources to address patients’ social needs. By improving the conditions under which providers search, gather, and rely on information in the EHR, we may come closer to achieving equity in diagnosis.
Methods: We recruited healthcare providers (MD, DO, or APP) (n=10) and health equity experts (n=2) to participate in 60-minute semi-structured virtual interviews. Tailored interview guides were developed, focusing on a set of patient-level variables commonly associated with diagnostic equity. Questions focused on understanding how health and patient-level variables are utilized to support equitable diagnostic practices. Variables were organized into seven key categories commonly associated with health disparities: demographics (race/ethnicity, age, sex/gender, cultural and linguistic preferences, health literacy, marital status), socioeconomic status (income, education, insurance, employment, financial hardship, food insecurity), lifestyle factors (tobacco use, substance use, physical activity), clinical indicators (chronic conditions, healthcare utilization, adherence to preventative care guidelines, mental health diagnoses, substance use diagnoses), social support (safe housing, family/friends at home, caregiver support), geographic factors (social vulnerability index (SVI), area deprivation index (ADI)), and neighborhood and built environment (housing stability, transportation). Clinical participants were asked if and how they use these variables in the diagnostic process, how well these variables are captured in their EHR systems, and whether the data is accurate and actionable. Health equity experts were asked to provide best-practice perspectives on the integration of these variables into the diagnostic process. All interviews were recorded, transcribed, and analyzed using a rapid, thematic approach conducted by two researchers (GF and LS) to identify emergent themes and insights.
Results:
12 interviews were conducted between June to July 2024 with data collection ongoing. By March 2025, it is anticipated that the sample will increase to approximately 40 participants. 10 (83%) were providers and 2 (17%) were health equity experts. Most participants identified as women (n=10, 83%). Of the n=10 providers, most worked in primary care (n=6, 60%).
Findings by variable category:
Demographics: While many participants considered race when reviewing patient demographics, they expressed concerns about perpetuating race-based medicine, emphasizing the need for culturally sensitive patient care. Despite race and ethnicity being captured in the EHR, participants noted inaccuracies and inconsistent documentation. Health equity experts highlighted that race and ethnicity often serve as proxies for health disparities and systemic inequalities, advocating for more granular, self-reported data. Age, considered crucial for screening and treatment, was considered highly accurate due to its link to birthdate, though one health equity expert cautioned that biological age may vary based on factors such as stress or trauma. While sex was viewed as critical for screenings and treatment, gender data was frequently inaccurate, particularly for transgender and non-binary individuals. Sexual orientation was deemed important for sexual health conversations, but EHR accuracy was variable. Participants considered patient health literacy essential for effective communication, but its capture lacked standard measurement tools, leaving providers to rely on subjective evaluation. Marital status, while potentially offering insights into social support systems, was often outdated or incomplete, raising concerns about accuracy.
Socioeconomic Status: While recognized as potentially informative, income was not typically documented in the EHR, with providers often inferring it based on insurance status or patient conversations. One health equity researcher cautioned that income alone does not fully capture financial hardship, as debt and other factors are often excluded. Education level was often linked to health literacy, echoed by a health equity expert, and faced similar challenges of inconsistent capture and update in the EHR. In contrast, insurance information was considered readily accessible and accurate, playing a vital role in guiding treatment options and connecting patients with health and social resources. Employment status, financial hardship, and food insecurity, though potentially insightful for tailoring treatment plans, were frequently unreliable or missing in the EHR.
Lifestyle Factors: Participants emphasized the importance of considering lifestyle factors in patient diagnosis, particularly substance use. However, the dynamic nature of these behaviors and the time limitations of patient-provider interactions necessitate frequent updates to the EHR, which was not consistently practiced. A health equity researcher recommended using broader time scales, such as monthly rather than weekly use intervals, to improve accuracy in substance use reporting. Physical activity, while generally acknowledged as important for health and guiding patient treatment and lifestyle recommendations, was also inconsistently and inaccurately documented in the EHR.
Clinical Indicators: Chronic conditions were central to guiding treatment decisions, with the EHR's problem list serving as the primary repository for these data. However, participants had concerns about data accuracy. These conditions were typically documented under medical history or medication lists, potentially limiting their visibility in the context of broader health equity considerations. Healthcare utilization was often viewed as a proxy for care access and care-seeking behaviors, yet it lacked a structured field in the EHR, leading providers to rely on informal documentation such as notes and appointment histories.
Social Support: Access to social support was critical to consider within the diagnostic process, particularly for older adults, as it affects a patient’s ability to manage their care, attend appointments, and recover from surgery or illness. Providers shared that knowing whether a patient has a caregiver or family/friends at home can influence discharge planning or the recommendation of additional support services. However, capturing information about social support and safe housing relied heavily on patient conversations and lacked standardized documentation in the EHR. This information was often recorded in unstructured fields, such as notes, making it difficult to find and update in the EHR.
Geographic Factors: While the Area Deprivation Index (ADI) and Social Vulnerability Index (SVI) were acknowledged as valuable tools for understanding community-level health disparities, they were rarely used in individual diagnostic decisions. Health equity experts suggested focusing on access to local resources, such as sidewalks, grocery stores, and local community, rather than relying solely on geographic data.
Neighborhood and Built Environment: Capturing neighborhood and build environment data such as housing stability and transportation remained a significant challenge. Although these factors were recognized as important for understanding patient context, they were not always formally captured in the EHR. Instead, providers typically relied on patient conversations and addressed these factors only when directly relevant to the immediate medical concern.
Discussion: To better support diagnostic equity at the point of care, there is a need for improved EHR display, connectedness, and functionality. In addition to capturing accurate patient-level information, EHR systems may best serve providers and patients by providing actionable insights and facilitating access to essential community resources to address patients’ social needs. By improving the conditions under which providers search, gather, and rely on information in the EHR, we may come closer to achieving equity in diagnosis.
Authors
Event Type
Poster Presentation
TimeMonday, March 314:45pm - 6:15pm EDT
LocationFrontenac Foyer
Digital Health (DH)
Simulation and Education (SE)
Hospital Environments (HE)
Medical and Drug Delivery Devices (MDD)
Patient Safety and Research Initiatives (PS)