Does autotext usage decrease documentation time among resident physicians? A retrospective analysis of EHR usage data
Data files
Jun 26, 2025 version files 6.40 KB
-
Autotext_residents_analysis_for_dryad.do
1.85 KB
-
curated_mimic_autotext_data.csv
511 B
-
README.md
4.04 KB
Abstract
Objective
Usage of autotext or “dotphrases” is ubiquitous among provider workflows in electronic health records (EHRs). Yet little is known about the impact of these tools in inpatient settings and among resident physicians. We aimed to evaluate the association between autotext usage and documentation time among resident physicians in an academic medical center using the Cerner® EHR.
Dataset Description
The association between auto text executions and documentation time per patient seen for 705 resident physicians rotating at a large academic medical center from July 2021 to June 2023 was analyzed via linear regression after controlling for specialty, post-graduate year (PGY), provider gender, and patient volume.
NOTE: The dataset in this study cannot be shared publicly due to the risk of identifying subjects who constitute a vulnerable population and may be known personally to members of the research community (physicians in training). Inclusion of details of gender, department, and year in training, which pose a risk of allowing subjects to be identified, is integral to the analysis of the data, so the data cannot be published in a meaningful form without these details. Accordingly, a short sample dataset that retains the data structure but with randomly generated values that mimic the actual data is provided instead.
The dataset used in this study was prepared from raw data downloaded from the Cerner Lightson and Cerner Advance toolkits and aggregated at the level of an individual resident physician over an academic year. As the study covers two academic years, 2021-2022 and 2022-2023, there are two entries for some resident physicians who were at the institution during both academic years. The dataset includes a randomly generated anonymous identifier for each provider, as well as demographics on department, gender, and PGY, and data on autotext usage, patient volume, documentation time per patient seen, and total EHR time per patient seen.
Results
There was no significant overall association between autotext executions per patient seen and documentation time per patient seen in specialties using Dynamic Documentation as their primary workflow (β=-0.1 min per autotext execution per patient seen, 95% CI -0.6 to 0.5 min, p=0.79). However, there was increased documentation time among residents with no autotext usage compared to residents who used autotext, and this effect was mediated by the use of personalized autotexts. Specialty, PGY, gender, and patient volume were significant determinants of documentation time.
Discussion
Efforts to decrease documentation time among resident physicians should encourage autotext adoption but should not be focused on the promotion of autotext usage alone. Further research should address the questions of identifying other determinants of documentation time, autotext design standards, and how autotext usage affects measures of note quality.
Conclusion
Autotext adoption decreases documentation time among resident physicians, but among those who adopt autotext, higher levels of usage show no benefit.
Dataset DOI: 10.5061/dryad.xksn02vt5
Description of the data and file structure
Short synthetic sample intended to mimic data used in retrospective study of autotext usage and documentation time among resident physicians. A synthetic dataset is presented because the actual dataset used in this study could not be sufficiently anonymized to share publicly due to the nature of the data (see Abstract.)
The dataset originally used in this study was prepared from raw data downloaded from the Cerner LightsOn and Cerner Advance toolkits and aggregated at the level of an individual resident physician over the course of an academic year. As the study covers two academic years, 2021-2022 and 2022-2023, there are two entries for some resident physicians who were at the institution during both academic years. The dataset includes a randomly generated anonymous identifier for each provider, as well as demographics on department, gender, and PGY, and data on autotext usage, patient volume, documentation time per patient seen, and total EHR time per patient seen. The data presented in this repository is designed to mimic the structure of this data accordingly.
Files and variables
File: curated_mimic_autotext_data.csv
Description:
Variables
- anon_id: Anonymous identifier for each resident physician. For instance, rows 4 and 5 have the same anon_id because they represent the same individual, an internal medicine resident, across two academic years, 2021-2022 (ac_year =1, when this individual was a PGY-1) and 2022-2023 (ac_year =2, when this individual was a PGY-2.)
- department: Name of department in which each resident physician is training.
- dept: Numerical index of department for use in regressions.
- PGYyear: Post-graduate year of individual resident in each row.
- ngender: numerical proxy for gender (0 = male gender, 1 = female gender.)
- DocTime: Documentation time per patient seen in minutes averaged over the academic year.
- AutoTxt: Autotext executions per patient seen averaged over the academic year.
- ac_year: academic year (1 = 2021-2022, 2 = 2022-2023.)
- sum_patientsseen: Total patients seen over the course of the academic year under consideration.
- n: Random variable representing assignment (1 = yes, 2 = no) for row to be analyzed for the purposes of analyses such as Kruskal-Wallis tests in which resident with data in multiple years had to have one year of data dropped
- drop: variable equal to 1 when in department of Emergency Medicine or Anesthesia, 0 otherwise, as residents in these departments were dropped from analyses as described in the manuscript.
- d_private: Dummy variable equal to 1 when there was any personal autotext use for the resident physician in the academic year being analyzed in this row, 0 otherwise.
- d_public: Dummy variable equal to 1 when there was any system autotext use for the resident physician in the academic year being analyzed in this row, 0 otherwise.
- d_pri_public: Dummy variable equal to 1 when there was any autotext use whatsoever for the resident physician in the academic year being analyzed in this row, 0 otherwise.
- nonadopt: Dummy variable equal to 1 for autotext nonadopters, 0 otherwise (1 - d_pri_public.)
- ActualTime: Total EHR time per patient seen in minutes averaged over the academic year.
Code/software
All analyses conducted using STATA 17.0 (College Station, Texas, United States.) Attached file “autotext residents_analysis for dryad.do” is a STATA .do file giving code that replicates the crucial analyses reported in tables and paragraphs in the main texts when used with the full dataset.
Access information
Other publicly accessible locations of the data:
- none
Data was derived from the following sources:
- Cerner LightsOn and Cerner Advance toolkits, licensed to Kaleida Health, Buffalo, NY.