ECEva Chang

Eva Chang

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Eva Chang has a total of 20 co-authors. Here are the top co-authors by number of shared publications:

  • Amanda Luff

    Advocate Aurora Research Institute, Advocate Health, Milwaukee, WI, USA

    3 shared publications
  • Nancy D. Berkman

    RTI International, Research Triangle Park, North Carolina

    3 shared publications
  • Rania Ali

    RTI-University of North Carolina Evidence-Based Practice Center, Research Triangle Park, USA

    3 shared publications
  • Julie Seibert

    RTI–University of North Carolina Evidence-based Practice Center and RTI International, Research Triangle Park, and North Carolina Department of Health and Human Services, Division of Mental Health, Developmental Disability and Substance Abuse Services, Raleigh, North Carolina (J.S.).

    2 shared publications
  • Nancy D. Berkman

    Retired, RTI International, Research Triangle Park, NC, USA

    2 shared publications
  • Rasha Khatib

    Advocate Aurora Research Institute, Advocate Health, Milwaukee, USA

    2 shared publications

Latest Publications

  • Open Access

    Downstream Healthcare Use Associated with Primary Care In-Person and Telemedicine Visits in 2022

    by Eva Chang, Amanda Luff, Thomas Moen, Edmund Fernandez, Rasha Khatib

    Background: More real-world evidence is needed to understand how telemedicine in primary care affects downstream healthcare use. Objective: To assess the impact of primary care visit modes on prescription orders and follow-up healthcare use. Design: Retrospective cohort study of adult primary care visits in a large Midwestern healthcare system from January to December 2022. Visits were categorized as in-person, video, or audio-only. Inverse probability of treatment weights (IPTW) was used to balance baseline patient demographic and visit characteristics. Participants: In total, 993,029 patients with 2,195,735 primary care visits. Main Measures: Prescription orders and healthcare use within 30 days (follow-up primary care visit, emergency department (ED) visit, and hospital visit). Generalized linear mixed models with IPTW were used to assess associations between visit mode and outcomes. Key Results: Telemedicine accounted for 3.5% of adult primary care visits (3.0% video and 0.5% audio-only). After adjustment, compared to in-person visits, telemedicine visits were associated with more antibiotic orders (video adjusted odds ratio (aOR) = 1.31, 95%CI = 1.25–1.37; audio-only aOR = 1.22, 95%CI = 1.09–1.37). Audio-only visits were also associated with less prescription orders overall (aOR = 0.80, 95%CI = 0.73–0.87) while no difference was observed between video and in-person visits. Telemedicine visits were also associated with more follow-up primary care visits (video aOR = 1.14, 95%CI = 1.08–1.19; audio-only aOR = 1.41, 95%CI = 1.27–1.56) and hospital visits (video aOR = 1.32, 95%CI = 1.18–1.49; audio-only aOR = 1.59, 95%CI = 1.29–1.96). Audio-only visits were associated with more ED visits (aOR = 1.33, 95%CI = 1.11–1.59) while no difference was observed in video visits. Conclusions: Telemedicine visits in primary care were associated with more antibiotic orders, follow-up primary care visits, and more hospital visits compared to in-person visits; audio-only visits were also associated with less prescriptions and more ED visits. Differences in outcomes by primary care visits mode may reflect how clinicians and patients differentially use video and audio-only visits.

  • Top Cited Papers

    Implementation of Social Needs Screening and Intervention in Primary Care

    by Eva Chang, Iridian Guzman, Nicole Glowacki, Rasha Khatib

    Health care systems are investing significant resources in social needs screening and intervention programs. To understand characteristics contributing to implementation of social needs screening and intervention programs in primary care. CINAHL, Cochrane, Ovid, PubMed, and Scopus (January 2015 to April 2025). U.S.-based programs using structured tools to screen adult patients for at least 1 Healthcare Effectiveness Data and Information Set measure (food insecurity, transportation, and housing insecurity) and addressing social needs in primary care settings. Program characteristics; screening and intervention implementation processes; and patient screening, intervention, clinical, and health care use outcomes. The review included 23 studies. Seventeen reported on screening outcomes, 11 in populations characterized by a particular condition or health care use and 6 in the general population. Programs with the highest percentage of patients screened focused on fewer than 500 patients with a particular condition or health care use and received support from additional staff or volunteers (4 of 17 studies). Of patients screened, 10.1% to 100% reported a social need. Eleven studies reported on receipt of assistance or resources, with a higher percentage of patients receiving assistance or resources among programs that targeted a smaller population. Few studies reported clinical and health care use outcomes, with mixed findings. Few studies had complete reporting of screening and intervention rates and outcomes. Program characteristics and other screening and intervention processes varied across and within studies. Social needs programs focused on smaller, targeted populations were more likely to screen and assist a higher percentage of patients. Programs with adequate staffing may also screen a higher proportion of patients. However, evidence is mixed, particularly for clinical and health care use outcomes. Considerable differences among screening and intervention programs preclude simple suggestions for universal implementation. Live Well Intramural Pilot Grant Program. (PROSPERO: CRD42023431151).


  • Open Access

    The role of patient, clinician, and neighborhood characteristics in predicting telemedicine engagement and modality in primary care: a cross-sectional analysis

    by Eva Chang, Amanda Luff

    The ability to use telemedicine may differ by social and clinician characteristics, however, studies examining telemedicine use have predominantly focused

  • Open Access

    Impact of Language Preference and Comorbidities on Telemedicine in Primary Care

    by Amanda Luff, Eva Chang

    Investigating how comorbidity and language preference interact to influence telemedicine use can guide the development of targeted interventions to enhance access for non-English language preference patients. This study aims to address how the relationship between comorbidity status and telemedicine use differs by language preference in a large Midwestern healthcare system.

  • Open Access

    Clinician characteristics and primary care telemedicine use in a large Midwestern healthcare system: A cross-sectional analysis (Preprint)

    by Eva Chang, Amanda Luff

    BACKGROUND While many studies have examined the patient characteristics associated with telemedicine use, few have examined the influence of clinicians on patients’ telemedicine use. OBJECTIVE This study assessed clinician contribution to and characteristics associated with primary care visit mode. METHODS Cross-sectional study of adult (18+ years) primary care patients in a large Midwestern healthcare system in 2022. Our main exposure was clinician characteristics and primary outcomes were primary care visit mode (any telemedicine visits vs. in-person visits only) and telemedicine mode (video vs. audio-only). Logistic regression was used to assess the association between outcomes and clinician characteristics. Pseudo r2 and Bayesian Information Criterion were used to assess changes in variance and model fit. RESULTS In 2022, 3.5% of primary care visits were telemedicine visits and 6% of primary care patients had a telemedicine visit (85% had any video visits and 15% had audio-only visits). Controlling for patient characteristics, patients with clinicians with moderate-sized patient panels (aOR: 1.05, 95%CI: 1.03-1.08), more telemedicine visits (aOR: 4.39, 95%CI: 4.30-4.49), and more patients in racial/ethnic minority groups (OR: 1.04, 95%CI: 1.01-1.07) were more likely to have telemedicine visits. Those with clinicians with large patient panels (aOR: 0.88, 95%CI: 0.86-0.90) and more patients with non-English language preference (NELP; aOR: 0.93, 95%CI: 0.91-0.95) were less likely to have telemedicine visits. Among telemedicine users, patients with only audio-only visits were more likely attributed to clinicians with more telemedicine visits (aOR: 1.33, 95%CI: 1.24-1.42), more patients with NELP (aOR: 1.27, 95%CI: 1.19-1.35), and more elderly patients (aOR: 1.34, 95%CI: 1.26-1.42) and less likely attributed to clinicians who are physicians (aOR: 0.82, 95%CI: 0.76-0.89), with larger patient panels (aOR: 0.47, 95%CI: 0.44-0.50 for ≥1400 patients; aOR: 0.82, 95%CI: 0.77-0.88 for 1000-1399 patients) and with more patients in racial/ethnic minority groups (aOR: 0.75, 95%CI: 0.70-0.81). Adding clinician characteristics improved model fit and explained more variability than patient characteristics alone in the primary care visit mode model. CONCLUSIONS Telemedicine use was associated with clinician type and patient panel characteristics. Clinician characteristics explained greater variability in telemedicine use than patient characteristics alone. Healthcare systems may increase telemedicine uptake by targeting interventions toward clinicians.

  • Open Access

    The relationship between community social risk factors and regional hospital-reported cash, negotiated, and chargemaster prices for 14 common services

    by Eva Chang, Wayne Psek

    Social risk factors are key drivers of the geographic variation in spending in the United States but little is known how community-level social risk factors are associated with hospital prices. Our objective was to describe the relationship between regional hospital-reported prices and social risk factors by price type (chargemaster, cash, commercial, Medicare, and Medicaid). This cross-sectional analysis used newly available hospital-reported prices from acute general hospitals in 2022. The prices were for 14 common services. Prices were winsorized at 98%, wage index-adjusted, standardized by service, and aggregated to hospital service areas (HSAs). For social risk, we used 23 measures across 5 domains of social risk (socioeconomic position; race, ethnicity, and culture; gender; social relationships; and residential and community context). Spearman’s correlation was used to estimate associations between median prices and social risk by price type. Prices were reported from 2,386 acute general hospitals in 45% (1,502 of 3,436) HSAs. Correlations between regional prices and other social risk factors varied by price type (range: -0.19 to 0.31). Chargemaster and cash prices were significantly correlated with the most community characteristics (10 of 23, 43%) followed by commercial prices (8, 35%). Medicare and Medicaid prices were only significantly correlated with 1 measure (all p < 0.01). All price types were significantly correlated with the percentage of uninsured (all p < 0.01). Chargemaster, cash, and commercial prices were positively correlated with percentage of Hispanic residents, residents with limited English proficiency, and non-citizens (all p < 0.05). While regional correlations between prices and social risk factors were weak across all prices, chargemaster, cash, and commercial prices were more like closely aligned with community-level social risk factors than the two public payers (Medicare and Medicaid). Chargemaster, cash, and commercial hospital prices appeared to be higher in socially disadvantaged communities. Further research is needed to clarify the relationship between prices and community social risk factors.


  • Top Cited Papers
    Open Access

    Patient Characteristics and Telemedicine Use in the US, 2022

    by Eva Chang, Robert B. Penfold, Nancy D. Berkman

    Importance: Telemedicine use was common during the COVID-19 pandemic, expanding many patients’ approaches to accessing health care. Of concern is whether telemedicine access was poorer among higher-needs and disadvantaged populations.Objective: To assess patient characteristics associated with telemedicine use and telemedicine mode and describe telemedicine visit experiences by telemedicine mode.Design, Setting, and Participants: This cross-sectional study included data from the 2022 Health Information National Trends Survey and included US adults with a health care visit. Data were analyzed from May to September 2023.Exposure: Patient characteristics.Main Outcomes and Measures: Any telemedicine visits vs in-person visits only; telemedicine mode (video vs audio-only). Multivariable logistic models assessed patient characteristics associated with telemedicine visits and mode. Bivariate analyses compared telemedicine experiences by mode.Results: The study included 5437 adult patients (mean [SE] age, 49.4 [0.23] years; 3136 females [53.4%]; 1928 males [46.6%]). In 2022, 2384 patients (43%) had a telemedicine visit; 1565 (70%) had a video visit while 819 (30%) had an audio-only visit. In multivariable models, older age (≥75 years: adjusted odds ratio [aOR], 0.63; 95% CI, 0.42-0.94), no internet use (aOR, 0.62; 95% CI, 0.48-0.81), and living in the Midwest (aOR, 0.50; 95% CI, 0.35-0.70) were negatively associated with having telemedicine visits. Female sex (aOR, 1.43; 95% CI, 1.12-1.83), having chronic conditions (aOR, 2.13; 95% CI, 1.66-2.73), and multiple health care visits (2-4 visits: aOR, 1.77; 95% CI, 1.23-2.54; ≥5 visits: aOR, 3.29; 95% CI, 2.20-4.92) were positively associated. Among individuals who used telemedicine, older age (65-74 years: aOR, 2.13; 95% CI, 1.09-4.14; ≥75 years: aOR, 3.58; 95% CI, 1.60-8.00), no health insurance (aOR, 2.84; 95% CI, 1.42-5.67), and no internet use (aOR, 2.11; 95% CI, 1.18-3.78) were positively associated with having audio-only visits. We observed no significant differences in telemedicine use or mode by education, race and ethnicity, or income. Patients’ experiences using telemedicine were generally similar for video and audio-only except more individuals who used audio-only had privacy concerns (20% vs 12%, P = .02).Conclusions and Relevance: In this cross-sectional study of adults with health care visits, many patients, including those with the greatest care needs, chose telemedicine even after in-person visits were available. These findings support continuing this care delivery approach as an option valued by patients. Differences were not observed by most common measures of socioeconomic status. Continued monitoring of telemedicine use is needed to ensure equitable access to health care innovations.


  • Open Access

    Differences in Telemedicine, Emergency Department, and Hospital Utilization Among Nonelderly Adults with Limited English Proficiency Post-COVID-19 Pandemic: a Cross-Sectional Analysis

    by Eva Chang, Teaniese L. Davis, Nancy D. Berkman

    Background: The unprecedented use of telemedicine during the COVID-19 pandemic provided an opportunity to examine its uptake among individuals with limited English proficiency (LEP). Objective: To assess telemedicine use among nonelderly adults with LEP and the association between use of telehealth and emergency department (ED) and hospital visits. Design: Cross-sectional study using the National Health Interview Survey (July 2020–December 2021) Participants: Adults (18–64 years), with LEP (N=1488) or English proficiency (EP) (N=25,873) Main Measures: Telemedicine, ED visits, and hospital visits in the past 12 months. We used multivariate logistic regression to assess (1) the association of English proficiency on having telemedicine visits; and (2) the association of English proficiency and telemedicine visits on having ED and hospital visits. Key Results: Between July 2020 and December 2021, 22% of adults with LEP had a telemedicine visit compared to 35% of adults with EP. After controlling for predisposing, enabling, and need factors, adults with LEP had 20% lower odds of having a telemedicine visit than adults with EP (p=0.02). While English proficiency was not associated with ED or hospital visits during this time, adults with telemedicine visits had significantly greater odds of having any ED (aOR: 1.80, pp<0.001) in the past 12 months. Conclusions: While telemedicine use increased overall during the COVID-19 pandemic, its use remained much less likely among adults with LEP. Interventions targeting structural barriers are needed to address disparities in access to telemedicine. More research is needed to understand the relationship between English proficiency, telemedicine visits, and downstream ED and hospital visits.


  • Open Access

    Characteristics of High-Need, High-Cost Patients

    by Nancy D. Berkman, Eva Chang, Julie Seibert, Rania Ali

    A small percentage of patients may disproportionally account for a large share of health care use and cost. Identifying patients with modifiable health care use is a challenge. Authors review all s... Background: Accurately identifying high-need, high-cost (HNHC) patients to reduce their preventable or modifiable health care use for their chronic conditions is a priority and a challenge for U.S....