Skip to main content
NCBI home page
JAMA Network logoLink to JAMA Network
. 2024 Aug 26:e244123. Online ahead of print. doi: 10.1001/jamainternmed.2024.4123

Physician EHR Time and Visit Volume Following Adoption of Team-Based Documentation Support

Nate C Apathy 1,✉, A Jay Holmgren 2, Dori A Cross 3
PMCID: PMC11348094  PMID: 39186284

Key Points

Question

How do physician visit volume, documentation time in the electronic health record (EHR), overall EHR time, and EHR time outside scheduled hours change for physicians adopting team-based documentation support?

Findings

In this national longitudinal cohort study of 18 265 ambulatory physicians, the adoption of team-based documentation support (ie, coauthored notes) was associated with significant increases in visit volume and decreases in documentation time in the EHR, including after-hours EHR time. Physicians with less than 40% of note text authored by another team member did not realize any time savings following documentation support.

Meaning

For high-intensity adopters, documentation support can increase visit volume while substantially reducing physician EHR burden.

Abstract

Importance

Physicians spend the plurality of active electronic health record (EHR) time on documentation. Excessive documentation limits time spent with patients and is associated with burnout. Organizations need effective strategies to reduce physician documentation burden; however, evidence on team-based documentation (eg, medical scribes) has been limited to small, single-institution studies lacking rigorous estimates of how documentation support changes EHR time and visit volume.

Objectives

To analyze how EHR documentation time and visit volume change following the adoption of team-based documentation approaches.

Design, Setting, and Participants

This national longitudinal cohort study analyzed physician-week EHR metadata from September 2020 through April 2021. A 2-way fixed-effects difference-in-differences regression approach was used to analyze changes in the main outcomes after team-based documentation support adoption. Event study regression models were used to examine variation in changes over time and stratified models to analyze the moderating role of support intensity. The sample included US ambulatory physicians using the EHR. Data were analyzed between October 2022 and September 2023.

Exposure

Team-based documentation support, defined as new onset and consistent use of coauthored documentation with another clinical team member.

Main Outcomes and Measures

The main outcomes included weekly visit volume, EHR documentation time, total EHR time, and EHR time outside clinic hours.

Results

Of 18 265 physicians, 1024 physicians adopted team-based documentation support, with 17 241 comparison physicians who did not adopt such support. The sample included 57.2% primary care physicians, 31.6% medical specialists, and 11.2% surgical specialists; 40.0% practiced in academic settings and 18.4% in outpatient safety-net settings. For adopter physicians, visit volume increased by 6.0% (2.5 visits/wk [95% CI, 1.9-3.0]; P < .001), and documentation time decreased by 9.1% (23.3 min/wk [95% CI, −30.3 to −16.2]; P < .001). Following a 20-week postadoption learning period, visits per week increased by 10.8% and documentation time decreased by 16.2%. Only high-intensity adopters (>40% of note text authored by others) realized reductions in documentation time, both for the full postadoption period (−53.9 min/wk [95% CI, −65.3 to −42.4]; 21.0% decrease; P < .001) and following the learning period (−72.2 min/wk; 28.1% decrease). Low adopters saw no meaningful change in EHR time but realized a similar increase in visit volume.

Conclusions and Relevance

In this national longitudinal cohort study, physicians who adopted team-based documentation experienced increased visit volume and reduced documentation and EHR time, especially after a learning period.

This national longitudinal cohort study of ambulatory physicians analyzes how electronic health record (EHR) documentation time and visit volume change following the adoption of team-based documentation approaches.

Introduction

The digitization of care delivery has burdened physicians with substantial electronic health record (EHR) work. Physicians spend nearly half of their day working in the EHR performing desktop medicine and log excessive time outside of clinic hours. This burden has raised rates of burnout and turnover, resulting in increased costs, exacerbating supply constraints, and lowering quality of care.

Clinical documentation makes up the plurality of physicians’ time spent in the EHR, and evidence suggests this time can be reduced without adversely impacting care quality. One common issue is note bloat, where notes become unnecessarily long with duplicative or clinically irrelevant information, adding to the time burden. As a result, developing strategies and interventions to reduce documentation burden has become a national policy priority. However, federal policy initiatives reducing documentation requirements for billing have demonstrated limited efficacy. EHRs offer efficiency tools like note templates or shortcuts, but these tools can lead to bloated notes and flattened treatment narratives. Physicians could benefit from systemic changes to EHR workflows, such as customizing software interfaces or reducing administrative tasks, but these largely lie outside practices’ control and have not yet been broadly realized. Actionable solutions are needed to alleviate documentation burden without compromising clinical productivity and reducing patient access to care.

Perhaps the most common approach involves shifting documentation tasks from the physician to a team member or medical scribe. Despite their popularity, evidence of the impact of team-based documentation on EHR burden is limited. Prior evaluations have occurred within single institutions, among relatively few physicians, often using self-reported outcomes. Given the cost and long-term staffing commitment required to implement team-based documentation, a more robust examination of documentation support, including considerations of how physicians reallocate documentation time savings to other EHR tasks, is necessary to guide organizations considering these programs. It is also critical to assess the conditions that most reliably foster meaningful changes to physicians’ EHR burden. For research findings to be maximally useful, organizational leaders need to know the time horizon required for new team support structures to take effect. It is also important to assess variation in changes across levels of team-based support intensity (ie, how much of the note is written by the nonphysician team member[s]). It may be the case that there exists a minimum viable level of support below which physicians realize no gains to visit volume or efficiency and do not realize meaningful reductions in EHR time.

To address these gaps, our study uses a national sample of physicians and a difference-in-differences design to answer 3 research questions regarding changes following the adoption of team-based documentation support. First, how does documentation support change physician visit volume, efficiency, documentation, and time spent in the EHR? Second, do changes intensify over time as interventions mature? Third, do changes vary according to how much support is being provided, as measured by the proportion of clinical note text written by nonphysician team members? Our results provide the first large-scale, precise estimates, to our knowledge, of changes in documentation burden, visit volume, and overall EHR time for physicians who choose to adopt team-based documentation support, and directly inform ongoing discussions of policy and organizational interventions for reducing physicians’ EHR burden.

Methods

Study Design

In this national longitudinal cohort study, we used a staggered adoption difference-in-differences study design to estimate changes in EHR time and visit volume following the adoption of team-based documentation support among physicians who adopted these practices. Our study period ran from September 2020 through May 2021, with weekly observations for each physician in the sample. Our data include all US-based ambulatory physicians (22 specialties, 369 organizations) that used Epic Systems’ EHR during this period. The data were deidentified, and thus have been deemed exempt from IRB review with a waiver of informed consent.

Data Collection

We used data provided by a leading ambulatory EHR vendor in the US. The data were derived from Epic’s Signal analytics platform, which measures ambulatory physician usage of the EHR primarily for efficiency and training purposes. These data, referred to as EHR usage metadata provide detailed measures of physician EHR use (eg, time spent in the EHR, note length) and of productivity (eg, visit volume, visit mix). These data also capture the volume of nonphysician team member documentation contributions, offering the unique ability to identify and evaluate changes in our outcomes following the adoption of documentation support. All data were deidentified but included physician specialty and organizational characteristics.

Team-Based Documentation Support Adoption

We measured team-based documentation support as the proportion of aggregate weekly note text in characters for all weekly ambulatory encounters written by someone other than the focal physician. We classified physicians as adopting documentation support during the study period if they demonstrated a 1-time shift in behavior from no documentation support for a minimum of 4 weeks (ie, all note text authored 100% by the physician) to consistent documentation support (some note text authored by other individuals) for at least 4 weeks. This identification strategy is consistent with similar evaluative work using Signal to assess medical scribe adoption. We excluded physicians with sporadic support, those who vacillate between the presence and absence of support over time, as well as physicians with support throughout our study period, as these physicians do not constitute a valid comparison group. This measure is designed to capture a variety of team-based documentation approaches (eg, traditional in-person scribes, virtual scribes, or note coauthorship with nonphysician clinician members of the care team). However, it is unlikely to capture attending physician–resident note coauthorship due to our requirement of a novel onset, 1-time shift to note coauthorship. The only circumstance in which attending physician–resident coauthorship would qualify would be cases in which the attending physician had never coauthored notes with any residents prior to the 1-time shift that we observed. Our comparison group was drawn from the subsample physicians with no documentation support throughout the study period. From this group, we included as comparisons all physicians that shared an organization-specialty pairing with adopter physicians; for example, if adopter physician 1 was a primary care physician, we included all primary care physicians from adopter 1’s same organization as comparison physicians (eFigure 1 in Supplement 1).

Outcome Measures

We evaluated changes in 4 key outcomes following the adoption of documentation support, all observed at the physician-week level: total weekly ambulatory visit volume, time in the EHR spent on documentation, total EHR time, and EHR time outside of scheduled hours. EHR time outside of scheduled hours includes any EHR time occurring prior to 30 minutes before the first appointment of the day and after the last appointment of the day plus a 30-minute buffer. We also evaluated secondary outcomes capturing visit mix (new and established visits, evaluation and management coding levels); efficiency (share of visits closed within the same day and within 2 days); other EHR activity domains (medical record review, orders, and inbox); and documentation composition (note length and characters of manual text entered by the physician).

We constructed all EHR time measures in 3 ways: aggregate weekly time; time per scheduled day; and time per visit. We focused our results and discussion on aggregate weekly outcomes as measures of changes that physicians perceive when documentation behavior and visit volumes may be changing simultaneously.

Intensity of Documentation Support

To stratify physicians by the intensity of documentation support, we first visualized the distribution of the share of nonphysician note text among adopter physicians in the weeks following adoption (eFigure 2 in Supplement 1). This figure demonstrated a bimodal distribution, suggesting 2 levels of support. The median postadoption share of nonphysician note text was 47.1%. To account for the increase in nonphysician note text over time, we set our cutoff for high-support physicians at more than 40% of note text authored by nonphysicians over the study period, producing similarly sized groups (511 low-support physicians, 513 high-support physicians; eFigure 3 in Supplement 1) and minimal category crossover (ie, high-support physicians had very few weeks with less than 40% support).

Statistical Analysis

We first calculated descriptive statistics overall and stratified across adopters and nonadopters for organization and physician characteristics. Next, we calculated average values for all outcome measures across 3 groups: adopters preadoption, adopters postadoption, and comparison group physicians. To estimate average changes following the onset of documentation support for adopting physicians (average treatment effect [ATT]), we ran a series of ordinary least squares regression models, using a 2-way fixed-effects difference-in-differences approach. Each model regresses a given outcome on a binary variable set to 1 for adopter physicians in postadoption weeks and 0 in all other cases. All models include physician and week fixed effects to adjust for time-invariant physician characteristics and secular trends, respectively. Hence, the estimates from these models reflect the average changes in our outcomes after adopting documentation support among adopters only, and these estimates do not generalize to nonadopter physicians.

To assess variation in changes following documentation support over time, we use an event study regression approach that estimates weekly changes via an interaction term capturing weeks relative to treatment (−30 to 30, with adoption of team-based documentation support beginning at week 0) and a binary indicator for treatment group (adopter or nonadopter). This approach estimates the difference in our outcomes between adopters and comparison physicians in each week leading up to and following adoption. Estimates in the weeks following adoption allow us to assess variation in the ATT over time, while null estimates in the weeks prior to adoption support the assumption of parallel trends on which difference-in-differences inference relies (ie, no trend differences in the preadoption period between adopters and nonadopters). We ran our event study models on all constructions of our outcomes. Finally, to analyze the extent to which changes in our outcomes following documentation support vary by support intensity, we ran stratified regression models for each of the 2 adopter subgroups (high-intensity adopters and low-intensity adopters). From these models, we estimate average changes following the adoption of documentation support for each subgroup as well as event study estimates to examine how the ATT estimate changes over time. All 2-sided regression models used a cutoff of P < .05 for statistical significance, and we used heteroskedasticity robust standard errors clustered at the individual physician level to account for autocorrelation. All analyses were conducted in R statistical software, version 4.3.2 (R Project for Statistical Computing).

Recent developments in econometric methods for evaluations of interventions with staggered adoption have provided numerous options for reestimation of ATTs that correct for the potential bias introduced by varying observation periods of treated and comparison units. As a sensitivity test, we reestimated our combined and high-support models for our 4 primary outcomes using the method proposed by Callaway and Sant’Anna, which ensures that only never-treated units serve as comparisons and corrects for the relatively long time span over which our adoption occurs via doubly robust standard errors (eMethods in Supplement 1).

Results

Study Population

The final sample included 18 265 physicians, with more than 669 721 total physician-week observations. A total of 1024 adopter physicians were identified (eFigure 4 in Supplement 1), with the remaining 17 241 physicians serving as nonadopter comparisons (Table). Overall, the sample consisted of 57.2% primary care physicians, 31.6% medical specialists, and 11.2% surgical specialists. Physicians varied in their practice setting with 40.0% practicing in academic settings and 18.4% in outpatient safety-net settings.

Table. Descriptive Statistics of Physician and Organization Study Sample.

Characteristic Overall Adopters Preadoption Postadoption Nonadopters
No. of physicians 18 265 1024 NA NA 17 241
Primary outcomes, mean (SD)
Total visits/wk 41.5 (28.2) 41.5 (27.3) 45.7 (29.0) 41.4 (28.2)
Active time in notes, min/wk 256.7 (206.3) 246.6 (201.6) 230.8 (201.1) 258.0 (206.5)
Active time in EHR, min/wk 707.6 (475.6) 660.4 (448.6) 637.5 (436.8) 711.1 (477.3)
Active time outside scheduled hours, min/wk 127.5 (146.5) 127.7 (148.4) 128.6 (141.7) 127.5 (146.6)
Physician specialty, %
Primary care 57.2 36.2 58.5
Medical specialty 31.6 49.3 30.5
Surgical specialty 11.2 14.5 11.0
Organizational characteristics, %
Size
<25 Physicians 0.4 2.1 0.4
25-50 Physicians 2 6.2 1.8
51-200 Physicians 21.3 32.1 20.7
>200 Physicians 76.2 59.6 77.2
Type
Hospital and clinic facilities 86.1 86.1 86.1
Ambulatory only 8.8 6.4 8.9
Othera 5.1 7.4 5.0
Status
Academic 40.0 44.7 39.7
Teaching 55.4 55.4 55.4
Pediatric 1.0 2.9 0.9
Safety net 18.4 16.9 18.5
Community hospital 36.9 44.7 36.4
Religiously affiliated 13.1 12.6 13.1
US census region
Midwest 31.2 26.8 31.5
Northeast 26.9 29.1 26.7
South 19.2 22.5 19
West 22.7 21.7 22.8
Open in a new tab

Abbreviation: EHR, electronic health record; NA, not applicable.

a

Examples of other organization types include minute clinics and mobile clinics.

Average Changes Following the Adoption of Documentation Support

Difference-in-differences analyses show that the average physician who adopted documentation support saw a 6.0% increase in total visit volume (2.5 visits/wk [95% CI, 1.9-3.0]; P < .001), a 9.1% decrease in documentation time (23.3 min/wk [95% CI, −30.3 to −16.2]; P < .001), a 4.1% decrease in overall EHR time (28.9 min/wk [95% CI, −38.6 to −19.2]), and a 5.1% decrease in active EHR time outside clinic hours (−6.5 min/wk [95% CI, −10.0 to −3.0]) (Figure 1). After adopting documentation support, progress note length increased (431 characters/note [95% CI, 333-528]), while physician-contributed manual text decreased (−59 characters/note [95% CI, −73 to −45]). All estimates and relative changes are presented in eTable 1 in Supplement 1.

Figure 1. Relative Changes Following Adoption of Team-Based Documentation Support for Primary Outcomes.

Figure 1.

Open in a new tab

Estimates of relative changes for each outcome are derived from 2-way fixed-effects difference-in-differences regression models. Relative changes express the absolute change estimate as a proportion of the overall average for a given outcome. Models included physician and week fixed effects. The unit of observation is the physician-week, and the standard errors are clustered at the physician level. See eTable 1 in Supplement 1 for absolute and relative change estimates as well as per-day and per-visit normalized outcomes. EHR indicates electronic health record.

Varying Changes Over Time

Event study regressions illustrated intensifying changes over time postadoption for both total weekly visits and documentation time (Figure 2). The average change in total visits more than 20 weeks after adoption was 4.5 additional visits per week, a 10.8% relative increase, compared to an overall average change of 2.5 visits per week (Figure 2A). A similar pattern was observed for weekly documentation time, with an estimated 41.5 fewer minutes in weeks 20 and beyond (16.2% decrease) compared to the average change of 23.3 fewer minutes (Figure 2B).

Figure 2. Event Study Estimates for Primary Outcomes.

Figure 2.

Open in a new tab

Shown are differences over time in primary outcomes between physicians who adopted documentation support and those who had no documentation support throughout the study period. Estimates are from 2-way fixed-effects event study regression models with indicators for each week preceding and following adoption, indicators for adopters, and the interaction between these terms. Data points indicate coefficient estimates, and bars represent the 95% CIs of the differences in outcomes between adopters and nonadopters leading up to and after support adoption. The vertical dashed lines represent the first week of support adoption, with year 1 serving as the reference year for estimates in each panel. The unit of observation is the physician-week. EHR indicates electronic health record.

Varying Changes by the Intensity of Support Adoption

In stratified models examining changes separately between high-intensity documentation support adopters and low-intensity adopters, both groups of support adopters saw increases in visit volume, but only physicians classified as high-intensity adopters realized any reductions in measures of EHR active time (Figure 3). High-support physicians reduced weekly time in notes for the full postadoption period by 53.9 minutes (95% CI, −65.3 to −42.4; P < .001), a 21.0% decrease compared to a null estimate for low-support physicians (−3.1 min/wk [95% CI, −11.3 to 5.1]) (eTable 2 in Supplement 1). Following the learning period, weekly time in notes was reduced by 72.2 minutes, a 28.1% decrease. High-intensity adopters also were the only physicians to realize significant decreases in time outside scheduled hours (−12.9 min/wk [95% CI, −17.5 to −8.4]; 10.1% decrease). As in the aggregate findings, stratified event study regressions illustrated intensifying changes in documentation time, EHR time, and time outside scheduled hours for high-support physicians but consistent null estimates for low-support physicians (Figure 4). Regression models using the Callaway and Sant’Anna estimation approach yielded qualitatively similar estimates both overall and among high adopters for weekly visit volume and time in notes (eTable 3 in Supplement 1). Dynamic estimates illustrating increasing changes over time were also consistent in these models (eFigure 5 and eFigure 6 in Supplement 1).

Figure 3. Comparison of Changes Stratified by Team-Based Documentation Support Intensity for Primary Outcomes.

Figure 3.

Open in a new tab

Estimates for each outcome are derived from 2-way fixed-effects difference-in-differences regression models, with separate models run for high-support (>40% of note text authored by other team members) and low-support physicians. Models included physician and week fixed effects. The unit of observation is the physician-week, and the standard errors are clustered at the physician level. See eTable 2 in Supplement 1 for a comparison of all absolute change estimates as well as per-day and per-visit normalized outcomes. EHR indicates electronic health record.

Figure 4. Event Study Estimates Stratified by Team-Based Documentation Support Intensity for Primary Outcomes.

Figure 4.

Open in a new tab

Shown are differences over time in primary outcomes for physicians adopting low amounts of team-based documentation support and physicians adopting high amounts of team-based documentation support (>40% of note text authored by other team members). All estimates are in comparison to nonadopter physicians. Each panel contains estimates from 2-way fixed-effects event study regression models with indicators for each week preceding and following adoption, indicators for high-support or low-support physicians, and the interaction between these terms. Data points indicate coefficient estimates, and bars represent the 95% CIs of the differences in outcomes between adopters and nonadopters following the onset of team-based documentation. Year 1 serves as the reference year for all estimates in each panel. The unit of observation is the physician-week. EHR indicates electronic health record.

Discussion

In this national longitudinal cohort study using difference-in-differences analyses, we found that physicians who adopted team-based documentation support experienced a durable decrease in EHR time, concentrated in less time spent documenting. Documentation support gains take time to fully realize: we observed the largest changes after accounting for a learning period of 20 weeks postadoption. We found that reduced EHR time from documentation support only occurs for physicians characterized as high-intensity adopters. Our results, to our knowledge, are the first multi-institutional, national-scale estimates of changes following the adoption of team-based documentation support among physicians who adopt these workflows, and the circumstances under which these approaches reduce physician documentation burden and increase clinical visit volume. Our results help to quantify the potential changes that may be possible for physicians who choose to adopt team-based documentation strategies, especially at high levels of intensity.

For high-support physicians, we found a 21% decrease in time spent writing notes (approximately 1 hour less time spent documenting per week) and a 10% decrease in EHR time outside scheduled hours. These are large relative changes that are likely to be both operationally meaningful and salient. For comparison, national-scale studies of reduced documentation requirements put in place in 2021 via changes to evaluation and management billing found no immediate changes in documentation time or note length and only a 4.1% relative decrease in documentation time after 1 year. As these policies mature, organizational leaders should consider coupling national efforts with programs to maximize reductions in documentation burden. The fact that time savings only occur for high-support physicians suggests a threshold of support required for physicians to realize EHR time efficiencies. More limited support (ie, low-support physicians) may indicate the incomplete transfer of responsibility and/or training, insufficient organizational support, or other imperfections of program implementation. Operational tracking of the share of note text authored by others could be useful for identifying and addressing implementation challenges that have limited the use of collaborative documentation and are inhibiting the full realization of the benefits. The challenge of durably changing documentation practices is underscored by our finding of a roughly 5-month learning curve. Such a long-term change, consistent with existing literature on medical scribes, suggests that trust and familiarity grow over time as adopter physicians and their teams refine collaborative workflows.

These reductions in EHR time correspond to a simultaneous increase in visit volume, which may be particularly important for health systems evaluating the cost-effectiveness of team-based documentation and policymakers looking to address supply constraints in the US health care system. Conservatively, team-based documentation support yielded a 6.0% increase in weekly visits, an increase near estimates of medical scribe costs across specialties. Especially since we find visit volume increases for both low-support and high-support physicians, it is possible that these findings reflect mandatory increases in visit volume to cover the cost of medical scribes or other team-based approaches. However, crisis levels of physician burnout and stress suggest that organizational leaders proceed cautiously in mandating any reallocation of saved EHR time to increased visit volumes. Importantly, if these mandates led to anticipatory changes in visit volume among adopters, we would observe these changes in adopters’ behavior in the weeks prior to adoption. We see very minimal but suggestive evidence of this in the visit volume event study estimates (Figure 2A).

Limitations

This study should be interpreted with some limitations in mind. First, while we used a staggered adoption difference-in-differences framework for estimation, adopters of documentation support likely differ in unobserved ways from nonadopters. Our event study estimates illustrate parallel preadoption trends for our outcomes; however, it is possible that adopter physicians reacted differently to shocks arising postadoption, in which case our estimates would be subject to some degree of selection bias. Second, we use data from a single EHR vendor, Epic Systems. Epic is the largest EHR vendor in the US, with large ambulatory and hospital market share. Epic clients include a variety of organizations like safety-net hospitals, academic medical centers, and ambulatory physician offices, but may be larger and better resourced than many independent physician practices. Third, because we use deidentified EHR usage metadata, we do not observe which team member contributes to shared documentation or the factors of specific shared documentation approaches that lead to reliably high adoption. This limits our ability to provide proscriptive guidance on what specific aspects of team-based documentation approaches (other than high adoption) are important for success. Future research should aim to explore the specific implementation details including staffing models and organizational policies that make high adoption of team-based documentation possible and in turn influence outcomes. Our measure does have the unique ability to capture clinical collaboration and the extent of shared note contribution. Past studies have randomized physicians to medical scribe support but have not directly measured documentation behavior. Further, we exclude sporadic support physicians from analysis to isolate the onset of new, intentional, and permanent changes in documentation practices. Fourth, our data do not include note text, so we do not observe documentation quality. Although our results found longer progress notes and suggest potential note bloat, other studies have not found deterioration in note quality when comparing team-based to physician-only documentation. More detailed exploration of note text is required to determine the conditions under which less documentation time and/or shorter notes is a normatively desirable outcome or has the potential to compromise care quality. Furthermore, medically scribed notes almost certainly introduce trade-offs vis-à-vis note quality and editorial responsibilities for physicians that do not exist in sole-authorship workflows. The long-term sustainability of team-based documentation has not been robustly explored; understanding when and why physicians might choose to reverse their decision to adopt this support would provide important additional insights.

Conclusions

Federal policy efforts to reduce documentation burden have been at best modestly successful via usability requirements for EHR developer certification and reimbursement policy. Professional medical and informatics societies have promoted other clinician-level technology solutions like efficiency tools such as templated text, which exhibit diminishing returns and a tendency toward note bloat. The findings from this national longitudinal cohort study suggest that team-based documentation support can be an important complement to these efforts. We stress that collaborative documentation approaches should not crowd out continued policy and organizational efforts to return documentation to a lower burden, value-added aspect of clinical work. For example, recent developments in artificial intelligence (AI) tools and their application to clinical practice may provide similar support for documentation at a lower marginal cost. However, AI tools will likely require similar learning periods to maximize efficiency and effectiveness, and the expected magnitude of these gains is unclear. As a result, staff support for documentation may be a practical short-term solution to address burden and capacity constraints. The results of our study may serve as a benchmark for comparison with other documentation burden interventions, including AI tools, and suggest that team-based documentation strategies are the rare intervention that has the potential to simultaneously reduce physician burden and boost visit volume when adopted sufficiently. Future studies should examine how to optimize team-based documentation and apply AI tools to this challenge, as well as other collaboration opportunities like ordering and inbox management.

Supplement 1.

eMethods. Estimation of Average Treatment Effects with Staggered Adoption

eFigure 1. Sampling diagram

eFigure 2. Distribution and median of team-authored note content postadoption, all adopter physicians

eFigure 3. Distribution and median of team-authored note content postadoption, stratified by support intensity

eFigure 4. Adoption plot of physicians adopting documentation support

eTable 1. Regression Estimates of Changes After Adoption of Documentation Support, all outcomes

eTable 2. Comparison of Estimates by Support Intensity

eTable 3. Comparison of TWFE and Callaway & Sant’Anna Estimates, Primary Outcomes

eFigure 5. Dynamic Treatment Effect Estimates, Callaway & Sant’Anna Modeling Estimates, Total Weekly Visits

eFigure 6. Dynamic Treatment Effect Estimates, Callaway & Sant’Anna Modeling Estimates, Total Weekly Time in Notes

jamainternmed-e244123-s001.pdf (1,015KB, pdf)
Supplement 2.

Data Sharing Statement

jamainternmed-e244123-s002.pdf (6.5KB, pdf)

References

  • 1.Gawande A. Why doctors hate their computers. The New Yorker. Published online November 5, 2018. Accessed April 9, 2019. https://www.newyorker.com/magazine/2018/11/12/why-doctors-hate-their-computers
  • 2.Overhage JM, McCallie D Jr. Physician time spent using the electronic health record during outpatient encounters: a descriptive study. Ann Intern Med. 2020;172(3):169-174. Published online January 14, 2020. doi: 10.7326/M18-3684 [DOI] [PubMed] [Google Scholar]
  • 3.Tai-Seale M, Olson CW, Li J, et al. Electronic health record logs indicate that physicians split time evenly between seeing patients and desktop medicine. Health Aff (Millwood). 2017;36(4):655-662. doi: 10.1377/hlthaff.2016.0811 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Rotenstein LS, Holmgren AJ, Downing NL, Bates DW. Differences in total and after-hours electronic health record time across ambulatory specialties. JAMA Intern Med. 2021;181(6):863-865. Published online March 22, 2021. doi: 10.1001/jamainternmed.2021.0256 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Adler-Milstein J, Zhao W, Willard-Grace R, Knox M, Grumbach K. Electronic health records and burnout: time spent on the electronic health record after hours and message volume associated with exhaustion but not with cynicism among primary care clinicians. J Am Med Inform Assoc. 2020;27(4):531-538. doi: 10.1093/jamia/ocz220 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Gardner RL, Cooper E, Haskell J, et al. Physician stress and burnout: the impact of health information technology. J Am Med Inform Assoc. 2019;26(2):106-114. doi: 10.1093/jamia/ocy145 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Shanafelt TD, West CP, Sinsky C, et al. Changes in burnout and satisfaction with work-life integration in physicians and the general US working population between 2011 and 2020. Mayo Clin Proc. 2022;97(3):491-506. doi: 10.1016/j.mayocp.2021.11.021 [DOI] [PubMed] [Google Scholar]
  • 8.Sinsky CA, Shanafelt TD, Dyrbye LN, Sabety AH, Carlasare LE, West CP. Health care expenditures attributable to primary care physician overall and burnout-related turnover: A cross-sectional analysis. Mayo Clin Proc. 2022;97(4):693-702. Published online February 25, 2022. doi: 10.1016/j.mayocp.2021.09.013 [DOI] [PubMed] [Google Scholar]
  • 9.Holmgren AJ, Downing NL, Bates DW, et al. Assessment of electronic health record use between US and non-US health systems. JAMA Intern Med. 2021;181(2):251-259. Published online December 14, 2020. doi: 10.1001/jamainternmed.2020.7071 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Downing NL, Bates DW, Longhurst CA. Physician burnout in the electronic health record era: are we ignoring the real cause? Ann Intern Med. 2018;169(1):50-51. Published online May 8, 2018. doi: 10.7326/M18-0139 [DOI] [PubMed] [Google Scholar]
  • 11.Holmgren AJ, Rotenstein L, Downing NL, Bates DW, Schulman K. Association between state-level malpractice environment and clinician electronic health record (EHR) time. J Am Med Inform Assoc. 2022;29(6):1069-1077. Published online March 10, 2022. doi: 10.1093/jamia/ocac034 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Apathy NC, Hare AJ, Fendrich S, Cross DA. Early changes in billing and notes after evaluation and management guideline change. Ann Intern Med. 2022;175(4):499-504. doi: 10.7326/M21-4402 [DOI] [PubMed] [Google Scholar]
  • 13.Reducing administrative burden. American Medical Association. Accessed March 6, 2023. https://www.ama-assn.org/practice-management/reducing-administrative-burden
  • 14.American Medical Informatics Association . AMIA 25×5—reducing documentation burden to 25% of current state in five years. Accessed March 6, 2023. https://amia.org/about-amia/amia-25x5
  • 15.Maisel N, Thombley R, Overhage JM, Blake K, Sinsky CA, Adler-Milstein J. Physician electronic health record use after changes in US Centers for Medicare & Medicaid Services documentation requirements. JAMA Health Forum. 2023;4(5):e230984. doi: 10.1001/jamahealthforum.2023.0984 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Rule A, Hribar MR. Frequent but fragmented: use of note templates to document outpatient visits at an academic health center. J Am Med Inform Assoc. 2021;29(1):137-141. doi: 10.1093/jamia/ocab230 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Cimino JJ. Putting the “why” in “EHR”: capturing and coding clinical cognition. J Am Med Inform Assoc. 2019;26(11):1379-1384. doi: 10.1093/jamia/ocz125 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Apathy NC, Rotenstein L, Bates DW, Holmgren AJ. Documentation dynamics: note composition, burden, and physician efficiency. Health Serv Res. 2023;58(3):674-685. Published online November 7, 2022. doi: 10.1111/1475-6773.14097 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Cross DA, Holmgren AJ, Apathy NC. The role of organizations in shaping physician use of electronic health records. Health Serv Res. 2024;59(1):e14203. Published online July 12, 2023. doi: 10.1111/1475-6773.14203 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Sieja A, Whittington MD, Patterson VP, et al. The influence of a Sprint optimization and training intervention on time spent in the electronic health record (EHR). JAMIA Open. 2021;4(3):ooab073. doi: 10.1093/jamiaopen/ooab073 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Sieja A, Markley K, Pell J, et al. Optimization sprints: Improving clinician satisfaction and teamwork by rapidly reducing electronic health record burden. Mayo Clin Proc. 2019;94(5):793-802. doi: 10.1016/j.mayocp.2018.08.036 [DOI] [PubMed] [Google Scholar]
  • 22.Chong AZ, Lee B, Hollenbach K, Kuelbs CL. Disappearing help text: implementing a note-based tool for in-line clinical decision support and note bloat reduction. Appl Clin Inform. 2022;13(5):1033-1039. doi: 10.1055/a-1934-8323 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Kroth PJ, Morioka-Douglas N, Veres S, et al. The electronic elephant in the room: physicians and the electronic health record. JAMIA Open. 2018;1(1):49-56. doi: 10.1093/jamiaopen/ooy016 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Gidwani R, Nguyen C, Kofoed A, et al. Impact of scribes on physician satisfaction, patient satisfaction, and charting efficiency: a randomized controlled trial. Ann Fam Med. 2017;15(5):427-433. doi: 10.1370/afm.2122 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Corby S, Whittaker K, Ash JS, et al. The future of medical scribes documenting in the electronic health record: results of an expert consensus conference. BMC Med Inform Decis Mak. 2021;21(1):204. doi: 10.1186/s12911-021-01560-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Gottlieb M, Palter J, Westrick J, Peksa GD. Effect of medical scribes on throughput, revenue, and patient and provider satisfaction: a systematic review and meta-analysis. Ann Emerg Med. 2021;77(2):180-189. doi: 10.1016/j.annemergmed.2020.07.031 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Pearson E, Frakt A. Medical scribes, productivity, and satisfaction. JAMA. 2019;321(7):635-636. doi: 10.1001/jama.2019.0268 [DOI] [PubMed] [Google Scholar]
  • 28.Mishra P, Kiang JC, Grant RW. Association of medical scribes in primary care with physician workflow and patient experience. JAMA Intern Med. 2018;178(11):1467-1472. doi: 10.1001/jamainternmed.2018.3956 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Corby S, Ash JS, Whittaker K, et al. Translating ethnographic data into knowledge, skills, and attitude statements for medical scribes: a modified Delphi approach. J Am Med Inform Assoc. 2022;29(10):1679-1687. doi: 10.1093/jamia/ocac091 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Ash JS, Corby S, Mohan V, et al. Safe use of the EHR by medical scribes: a qualitative study. J Am Med Inform Assoc. 2021;28(2):294-302. doi: 10.1093/jamia/ocaa199 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Rotenstein L, Melnick ER, Iannaccone C, et al. Virtual scribes and physician time spent on electronic health records. JAMA Netw Open. 2024;7(5):e2413140. doi: 10.1001/jamanetworkopen.2024.13140 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Khullar D. Burnout, professionalism, and the quality of US health care. JAMA Health Forum. 2023;4(3):e230024. doi: 10.1001/jamahealthforum.2023.0024 [DOI] [PubMed] [Google Scholar]
  • 33.Definitive Healthcare . Top 10 ambulatory EHR vendors by 2019 market share. Published May 1, 2019. Accessed August 4, 2021. https://www.definitivehc.com/blog/top-ambulatory-ehr-systems
  • 34.Baxter SL, Apathy NC, Cross DA, Sinsky C, Hribar MR. Measures of electronic health record use in outpatient settings across vendors. J Am Med Inform Assoc. 2021;28(5):955-959. doi: 10.1093/jamia/ocaa266 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Sinsky CA, Rule A, Cohen G, et al. Metrics for assessing physician activity using electronic health record log data. J Am Med Inform Assoc. 2020;27(4):639-643. Published online February 6, 2020. doi: 10.1093/jamia/ocz223 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Rule A, Melnick ER, Apathy NC. Using event logs to observe interactions with electronic health records: an updated scoping review shows increasing use of vendor-derived measures. J Am Med Inform Assoc. 2022;30(1):144-154. doi: 10.1093/jamia/ocac177 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Rule A, Chiang MF, Hribar MR. Using electronic health record audit logs to study clinical activity: a systematic review of aims, measures, and methods. J Am Med Inform Assoc. 2020;27(3):480-490. doi: 10.1093/jamia/ocz196 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Cohen GR, Boi J, Johnson C, Brown L, Patel V. Measuring time clinicians spend using EHRs in the inpatient setting: a national, mixed-methods study. J Am Med Inform Assoc. 2021;28(8):1676-1682. doi: 10.1093/jamia/ocab042 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Ong SY, Jeffery MM, Williams B, et al. How a virtual scribe program improves physicians’ EHR experience, documentation time, and note quality. NEJM Catalyst. 2021. https://catalyst.nejm.org/doi/10.1056/CAT.21.0294 doi: 10.1056/CAT.21.0294 [DOI]
  • 40.Wing C, Simon K, Bello-Gomez RA. Designing difference in difference studies: best practices for public health policy research. Annu Rev Public Health. 2018;39:453-469. doi: 10.1146/annurev-publhealth-040617-013507 [DOI] [PubMed] [Google Scholar]
  • 41.Baker A, Larcker DF, Wang CCY. How much should we trust staggered difference-in-differences estimates. Published online March 29, 2021. doi: 10.2139/ssrn.3794018 [DOI]
  • 42.Goodman-Bacon A. Difference-in-differences with variation in treatment timing. J Econom. 2021;225(2):254-277. doi: 10.1016/j.jeconom.2021.03.014 [DOI] [Google Scholar]
  • 43.Callaway B, Sant’Anna PHC. Difference-in-differences with multiple time periods. J Econom. Published online December 17, 2020. [Google Scholar]
  • 44.Barry MJ, Tseng CW. Moving to more evidence-based primary care encounters: a farewell to the review of systems. JAMA. 2022;328(15):1495-1496. doi: 10.1001/jama.2022.18346 [DOI] [PubMed] [Google Scholar]
  • 45.Florig ST, Corby S, Devara T, Weiskopf NG, Mohan V, Gold JA. Medical record closure practices of physicians before and after the use of medical scribes. JAMA. 2022;328(13):1350-1352. Published online September 1, 2022. doi: 10.1001/jama.2022.13558 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Rose C, Thombley R, Noshad M, et al. Team is brain: leveraging EHR audit log data for new insights into acute care processes. J Am Med Inform Assoc. 2022;30(1):8-15. Published online October 27, 2022. doi: 10.1093/jamia/ocac201 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Lemieux-Charles L, McGuire WL. What do we know about health care team effectiveness—a review of the literature. Med Care Res Rev. 2006;63(3):263-300. doi: 10.1177/1077558706287003 [DOI] [PubMed] [Google Scholar]
  • 48.Miksanek TJ, Skandari MR, Ham SA, et al. The productivity requirements of implementing a medical scribe program. Ann Intern Med. 2021;174(1):1-7. doi: 10.7326/M20-0428 [DOI] [PubMed] [Google Scholar]
  • 49.Rotenstein LS, Brown R, Sinsky C, Linzer M. The association of work overload with burnout and intent to leave the job across the healthcare workforce during COVID-19. J Gen Intern Med. 2023;38(8):1920-1927. Published online March 23, 2023. doi: 10.1007/s11606-023-08153-z [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Rotenstein LS, Torre M, Ramos MA, et al. Prevalence of burnout among physicians: a systematic review. JAMA. 2018;320(11):1131-1150. doi: 10.1001/jama.2018.12777 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 51.Harry EM, Carlasare LE, Sinsky CA, et al. Childcare stress, burnout, and intent to reduce hours or leave the job during the COVID-19 pandemic among US health care workers. JAMA Netw Open. 2022;5(7):e2221776. doi: 10.1001/jamanetworkopen.2022.21776 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 52.Linzer M, Jin JO, Shah P, et al. Trends in clinician burnout with associated mitigating and aggravating factors during the COVID-19 pandemic. JAMA Health Forum. 2022;3(11):e224163. doi: 10.1001/jamahealthforum.2022.4163 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 53.OCHIN. Accessed March 17, 2022. https://ochin.org/network/members/
  • 54.Rule A, Florig ST, Bedrick S, Mohan V, Gold JA, Hribar MR. Comparing scribed and non-scribed outpatient progress notes. AMIA Annu Symp Proc. 2022;2021:1059-1068. [PMC free article] [PubMed] [Google Scholar]
  • 55.Misra-Hebert AD, Amah L, Rabovsky A, et al. Medical scribes: how do their notes stack up? J Fam Pract. 2016;65(3):155-159. [PubMed] [Google Scholar]
  • 56.American Medical Association . 5 Simple changes to help cut doctors’ EHR burdens. American Medical Association. Published April 30, 2021. Accessed March 6, 2023. https://www.ama-assn.org/practice-management/digital/5-simple-changes-help-cut-doctors-ehr-burdens
  • 57.Owens LM, Wilda JJ, Hahn PY, Koehler T, Fletcher JJ. The association between use of ambient voice technology documentation during primary care patient encounters, documentation burden, and provider burnout. Fam Pract. 2024;41(2):86-91. doi: 10.1093/fampra/cmad092 [DOI] [PubMed] [Google Scholar]
  • 58.Baker HP, Dwyer E, Kalidoss S, Hynes K, Wolf J, Strelzow JA. ChatGPT’s ability to assist with clinical documentation: a randomized controlled trial. J Am Acad Orthop Surg. 2024;32(3):123-129. doi: 10.5435/JAAOS-D-23-00474 [DOI] [PubMed] [Google Scholar]
  • 59.Adler-Milstein J, Redelmeier DA, Wachter RM. The limits of clinician vigilance as an AI safety bulwark. JAMA. 2024;331(14):1173-1174. doi: 10.1001/jama.2024.3620 [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplement 1.

eMethods. Estimation of Average Treatment Effects with Staggered Adoption

eFigure 1. Sampling diagram

eFigure 2. Distribution and median of team-authored note content postadoption, all adopter physicians

eFigure 3. Distribution and median of team-authored note content postadoption, stratified by support intensity

eFigure 4. Adoption plot of physicians adopting documentation support

eTable 1. Regression Estimates of Changes After Adoption of Documentation Support, all outcomes

eTable 2. Comparison of Estimates by Support Intensity

eTable 3. Comparison of TWFE and Callaway & Sant’Anna Estimates, Primary Outcomes

eFigure 5. Dynamic Treatment Effect Estimates, Callaway & Sant’Anna Modeling Estimates, Total Weekly Visits

eFigure 6. Dynamic Treatment Effect Estimates, Callaway & Sant’Anna Modeling Estimates, Total Weekly Time in Notes

jamainternmed-e244123-s001.pdf (1,015KB, pdf)
Supplement 2.

Data Sharing Statement

jamainternmed-e244123-s002.pdf (6.5KB, pdf)

Articles from JAMA Internal Medicine are provided here courtesy of American Medical Association

RESOURCES