Changes in Secure Messaging After Implementation of Billing E-Visits by Demographic Group (2024)

Research Letter

Health Informatics

A. JayHolmgren,PhD, MHI1; Carrie K.Grouse,MD2; ArisOates,MD3; et al JulieO’Brien,MD3; Maria E.Byron,MD1,4

Author Affiliations Article Information

Beginning in 2020, the Centers for Medicare & Medicaid Services and private payers expanded billing options for asynchronous patient portal messages, known as e-visits. Billing requirements include medical decision-making and a minimum of 5 minutes of clinician time over 7 days. Several large health systems recently implemented billing messages as e-visits,¹ possibly in response to increased message volume after the COVID-19 pandemic began in March 2020.² Early research showed that the potential to receive a bill was associated with a small decrease in overall message volume.³ However, little is known regarding demographic characteristics of patients with reduced messaging. It is critical to understand whether the association of e-visit billing with patient messaging differed by race and ethnicity, language, age, or payer.⁴ We assessed changes in patient-initiated messaging volume after e-visit billing implementation across patient demographics at a large health system in a primarily fee-for-service environment.

This cohort study was deemed exempt from review and consent by the University of California, San Francisco (UCSF) Institutional Review Board for using deidentified, aggregate data and follows STROBE reporting guidelines. We analyzed e-visit billing at UCSF Health, which began clinician-driven e-visit billing on November 14, 2021, with patients informed via the portal before sending a message that it may result in a bill and provided information about cost sharing.

To capture the association of e-visit implementation with patient messaging rates across demographic groups, we measured total patient medical advice request messages weekly from October 1, 2020, to August 20, 2022, with November 14, 2021, as the implementation date. The sample period ended on August 20, 2022, to avoid confounding given that UCSF implemented an unrelated change in the portal interface the following week. We measured total weekly messages by patient race and ethnicity (as indicated in the electronic health record), preferred language, payer category (commercial, California state Medicaid [Medi-Cal], Medicare, or other [primarily self-pay]), and age. Messages include those sent by patients or designated proxies. We used 2-tailed t tests with unequal variances to assess differences in mean weekly messaging before and after e-visit implementation. All analyses were conducted in Stata version 17.0 (StataCorp).

Among 5 558 460 messages, mean weekly messages decreased by 2.0% over 94 weeks. Across race and ethnicity groups, the largest reductions in mean weekly messages after e-visit implementation were among Latinx (−323.4 messages; 95% CI, −584.6 to −62.3 messages [−5.3%]; P = .02) and Asian (−506.5 messages; 95% CI, −940.5 to −72.5 [−5.1%]; P = .02) patients (Table).

Across payers, patients in the other category saw a reduction of −256.5 messages (95% CI, −458.4 to −54.7 messages [14.4% ]; P = .01). Patients ages 18 to 49 years saw the largest reduction across age groups (−2172.4 messages; 95% CI, −3083.0 to −1261.9 messages [−10.3%]; P < .01), while patients aged 66 to 80 and older than 80 years had increased messages. Trends were stable throughout the postimplementation period (Figure).

This cohort study found several differences in the change in messaging across patient demographic characteristics. Patients with self-pay and those aged 18 to 49 years saw the largest decreases in mean weekly message volume, and Latinx and Asian patients saw a slightly larger decrease than other racial and ethnic groups. Some patients, including those older than 65 years, had increased messaging volume. While it is reassuring that we did not find evidence of differential associations with messaging for Medicaid or older patients, groups with lower rates of prepandemic messaging,^5,6 the larger reduction in messages for some patient groups warrants further investigation.

Limitations include data from a single health system and an inability to assess care delivered outside of UCSF, associations with patient outcomes, or causality. Our aggregate data allowed us to assess only high-level changes in messaging across groups, which may reflect changes in patient behavior after they learn of the possibility of billing for messaging. We could not observe whether a patient received an e-visit bill or faced out-of-pocket costs. Future research should investigate the association of billing for technology-enabled care with health across patient demographics.

Accepted for Publication: June 13, 2024.

Published: August 9, 2024. doi:10.1001/jamanetworkopen.2024.27053

Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2024 Holmgren AJ et al. JAMA Network Open.

Corresponding Author: A. Jay Holmgren, PhD, MHI, Division of Clinical Informatics and Digital Transformation, University of California, San Francisco, 10 Koret Way, Office 327A, San Francisco, CA 94131 (a.holmgren@ucsf.edu).

Author Contributions: Dr Holmgren had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: All authors.

Acquisition, analysis, or interpretation of data: Holmgren, Oates, Byron.

Drafting of the manuscript: Holmgren.

Critical review of the manuscript for important intellectual content: All authors.

Statistical analysis: Holmgren, Grouse.

Administrative, technical, or material support: Oates, Byron.

Conflict of Interest Disclosures: Dr Holmgren reported receiving grants from the American Medical Association, Healthcare Leadership Council, and Office of the National Coordinator for Health Information Technology outside the submitted work. No other disclosures were reported.

Data Sharing Statement: See the Supplement.