Abstract
PURPOSE Scribes are increasingly being used in clinical practice despite a lack of high-quality evidence regarding their effects. Our objective was to evaluate the effect of medical scribes on physician satisfaction, patient satisfaction, and charting efficiency.
METHODS We conducted a randomized controlled trial in which physicians in an academic family medicine clinic were randomized to 1 week with a scribe then 1 week without a scribe for the course of 1 year. Scribes drafted all relevant documentation, which was reviewed by the physician before attestation and signing. In encounters without a scribe, the physician performed all charting duties. Our outcomes were physician satisfaction, measured by a 5-item instrument that included physicians’ perceptions of chart quality and chart accuracy; patient satisfaction, measured by a 6-item instrument; and charting efficiency, measured by time to chart close.
RESULTS Scribes improved all aspects of physician satisfaction, including overall satisfaction with clinic (OR = 10.75), having enough face time with patients (OR = 3.71), time spent charting (OR = 86.09), chart quality (OR = 7.25), and chart accuracy (OR = 4.61) (all P values <.001). Scribes had no effect on patient satisfaction. Scribes increased the proportion of charts that were closed within 48 hours (OR =1.18, P =.028).
CONCLUSIONS To our knowledge, we have conducted the first randomized controlled trial of scribes. We found that scribes produced significant improvements in overall physician satisfaction, satisfaction with chart quality and accuracy, and charting efficiency without detracting from patient satisfaction. Scribes appear to be a promising strategy to improve health care efficiency and reduce physician burnout.
- medical scribes
- electronic health records
- work satisfaction
- patient satisfaction
- efficiency
- primary care physicians
- randomized controlled trial
INTRODUCTION
Electronic health records (EHRs) have radically transformed the practice of medicine. Driven by federal meaningful use incentives and penalties,1,2 more than 95% of US hospitals and 56% of office-based physicians have adopted EHRs.3,4 Electronic health records hold promise to improve patient safety, quality of care, physician efficiency and performance, patient-physician communication, patient participation, cost of care, and health outcomes.5–9 There is also growing evidence, however, that in their current state, EHRs are associated with decreased physician productivity and revenue,10 negative patient-physician interactions and relationships,11 and widespread physician dissatisfaction.12–14
More than one-half of all US physicians experience burnout, with primary care physicians having one of the highest rates.15 Among the largest contributors to burnout is a growing clerical workload.16–18 For every hour physicians provide direct face time to patients, 2 more hours are spent on EHR and desk work.19 Many physicians leave most charting to the end of the day, and spend 1 to 2 hours each night working on the EHR.19
One strategy to decrease clerical burden is the use of scribes. Scribes are nonlicensed team members trained to document patient encounters in real time under the direct supervision of a physician.20 Scribes do not act independently but may assist with charting, recording laboratory and radiology results, and supporting physician workflow with EHR data entry.21 Although the use of scribes as physician extenders in emergency departments has been reported as early as the 1970s, it is only recently that the popularity of scribes has skyrocketed.22 Scribes are currently being used in more than 1,000 hospitals and clinics across 44 states.23 It is estimated that by 2020, there will be 100,000 scribes in the United States, or 1 scribe for every 9 physicians.23
Despite the increasing presence of scribes, methodologically rigorous studies regarding their impact are lacking. Two systematic reviews found, using data from observational studies, that scribes may improve revenue, patient and physician satisfaction, productivity, efficiency, and the quality of patient-physician interactions.24,25 There has been no randomized controlled study of scribes, and few studies have been undertaken in the primary care setting. Given that most office visits are to primary care physicians,26 research in this setting is particularly warranted.
METHODS
Design
Physicians were randomly assigned to 1 week practicing with a scribe then 1 week without a scribe for the course of 52 weeks. Randomization at the physician-week level was chosen instead of randomizing at the level of patients, as variations in length of patient appointments posed challenges for proper allocation of scribes across patients and was too disruptive to normal clinic flow. We also chose not to randomize at the physician level, as the small number of physicians included in this study would not properly protect against imbalance in the scribe and no-scribe groups. During the week in which a physician was assigned a scribe, the scribe attended all appointments and drafted all relevant documentation, including the history and physical findings, objective examination findings, laboratory and radiology results, assessment and plan, and patient instructions. The physician reviewed the note, attested to its accuracy, and signed it before the chart was closed. During the week in which the physician was not assigned a scribe, the physician performed all charting duties. The EHR used was the out-patient version of Epic (Epic Systems Corporation).
The study was conducted from July 2015 to June 2016. Four physicians and 2 scribes participated in the study, which was undertaken in a family medicine clinic associated with a large academic medical center in Northern California. All physicians were board-certified in family medicine and had an average of 6 years of practice experience. None had prior experience working with scribes. As part-time clinicians, each physician in the study had 4, 4-hour clinic sessions per week when data were collected. Both scribes were college graduates who completed an 80-hour training course administered by a commercial scribe company (Elite Medical Scribes). One scribe was assigned to 2 physicians in the first 6 months of the study; in the second 6 months, that scribe was assigned to the other 2 physicians. This allowed us to test for any learning effects that may have occurred in the physician-scribe dyads.
Physician Satisfaction
Physician satisfaction was measured by a self-administered 5-item questionnaire. Answers were recorded using a 7-point Likert scale, with a value of 1 indicating strong disagreement (strongly dissatisfied) and 7 indicating strong agreement (strongly satisfied). Physicians were offered 1 questionnaire after each 4-hour clinic session. For data analyses, we dichotomized each answer into strongly satisfied vs non–strongly satisfied (7 vs 1 to 6) because of skewness in results. In sensitivity analyses, we tested alternate ways to characterize the outcome by dichotomizing scores from 1 to 5 and 6 to 7. To investigate the effect of scribes on aspects of physician satisfaction, each item was assessed using its own fixed-effects logistic regression equation with the physician questionnaire as the unit of analysis and accommodating multiple observations per physician. We adjusted for whether the interaction was new so we could test any learning effects over time within physician-scribe dyads. Specifically, we investigated whether a physician paired with a scribe had significantly lower satisfaction scores during the first quarter than during the second quarter that the same physician and scribe were paired. We adjusted for multiple hypothesis testing using the conservative Bonferroni correction, resulting in an α of .01.27
Patient Satisfaction
Patient satisfaction was measured using a shortened, validated, 6-item questionnaire designed for the primary care setting.28 Each patient was asked to complete the questionnaire immediately after the appointment. To encourage completion, questionnaires were made anonymous. Answers were recorded using a 7-point Likert scale, with 1 indicating strong disagreement (strongly dissatisfied) and 7 indicating strong agreement (strongly satisfied). Each response was dichotomized into strongly satisfied (7) vs non–strongly satisfied (1 to 6) because of skewness of the distribution. In sensitivity analyses, we tested alternate ways to characterize the outcome, specifically dichotomizing scores from 1 to 5 and 6 to 7. We investigated each item separately using its own fixed-effects logistic regression equation with the patient questionnaire as the unit of analysis, clustering questionnaires within physician. All tests were evaluated against a Bonferroni-corrected α of .007.
Charting Efficiency
Physician efficiency was measured as the time to chart close, which is calculated as the time from appointment start to the physician signing the chart note, which is marked by timestamps in the EHR. Industry standards (Medicare documentation guidelines)29 state that charts should be completed within 48 hours; therefore, we dichotomized time to close chart into 48 hours or less vs more than 48 hours. We ran fixed-effects logistic regression with chart as the unit of analysis, accommodating clustering of charts within physician.
This study was exempted from formal review by the Institutional Review Board of Stanford University School of Medicine.
RESULTS
Physician Satisfaction
A total of 361 physician satisfaction questionnaires were completed, for a 73.1% response rate (Table 1). Physicians were more likely to complete a questionnaire when a scribe was present (53.2%) than when a scribe was not (46.8%). Scribes produced significantly higher physician satisfaction in all aspects of care and charting (Tables 2–4). Physicians who worked with a scribe had 10.75 the adjusted odds of expressing high satisfaction with their clinic that day, 3.71 the adjusted odds of having enough face time with patients, and 86.09 the adjusted odds of expressing high satisfaction with the amount of time they spent charting (all P <.001). Scribes increased physician satisfaction with the quality and accuracy of their charts. Physicians reported 7.25 the adjusted odds of being satisfied with their chart quality when a scribe was present (P <.001). There was no difference in satisfaction with quality when the physician-scribe dyad was new vs established (P= .451). Physicians reported 4.61 the adjusted odds of being satisfied with chart accuracy when a scribe was present (P < .001). Physicians did report being less satisfied with chart accuracy when the physician-scribe dyad was new (adjusted OR = 0.39) vs established, but results were not significant using a Bonferroni-corrected α of .01 (P = .019). There was no difference in significance of the impact of scribes on physician survey results when we dichotomized the responses into 1 to 5 vs 6 to 7.
Survey Questionnaire Completion
Physician and Patient Questionnaire Results, Unadjusted
Physician Satisfaction, Unadjusted Results
Physician Satisfaction, Adjusted Results
Patient Satisfaction
A total of 1,475 patient satisfaction questionnaires were completed for an 87.7% response rate (Table 1). Patients were more likely to complete a questionnaire when a scribe was present (54.8%) than when a scribe was not (45.2%). In adjusted analyses, there were no significant differences in any aspect of patient satisfaction between physician visits in which a scribe was or was not present (Table 2 and Table 5). Satisfaction across patient questionnaires, however, was high with or without a scribe, with more than 91% of patients in either group reporting being highly satisfied with their care. There was no difference in significance of the impact of scribes on patient survey results when we dichotomized the responses into 1 to 5 vs 6 to 7.
Patient Satisfaction, Adjusted Results
Charting Efficiency
Scribes improved time to close chart. In unadjusted analyses, 28.5% of charts that were drafted by physicians were closed in 48 hours relative to 32.6% of charts drafted by scribes. In adjusted analyses, scribed charts had 1.18 the adjusted odds of being closed within 48 hours compared with physician-only charts (P = .028) (Table 6).
Charting Efficiency, Adjusted Results for Less Than 48 Hours to Close Chart
DISCUSSION
To our knowledge, we have undertaken the first randomized controlled trial evaluating the effects of medical scribes. We found that scribes significantly improved physician satisfaction across all measured aspects of patient care and documentation. Scribes improved physician-perceived chart quality and chart accuracy, as well as charting efficiency as measured by the likelihood of closing a chart within 48 hours.
When working with a scribe, physicians were much more satisfied with how their clinic went, the length of time they spent face-to-face with patients, and the time they spent charting. These findings suggest that scribes may have a protective effect on physicians’ well-being. Implementation of team documentation is an important component of achieving the Quadruple Aim,30 a patient-centered approach to care that also emphasizes improving the work life of physicians. Spending less time on documentation frees up the physician to pursue direct clinical care and care coordination, thus enhancing joy of practice and preventing burnout. In academic centers, scribes provide faculty physicians more time to teach medical students and residents.31
We found that not only were physicians satisfied with the quality and accuracy of charting done by scribes, they were more satisfied with scribed charts than with their own. This finding is consistent with a study suggesting that scribed notes are of higher quality than physician-only notes.32 Patient encounters in primary care are often highly complex; scribes enable physicians to capture all the important details in the note while communicating effectively with the patient in the room.
During a typical day in the ambulatory setting, 49% of physician time is spent on EHR and desk work, whereas only 27% is spent face-to-face with patients.19 Physicians can use EHR shortcuts, such as copy and paste,33 but these actions are associated with a risk of documentation error that can jeopardize patient safety.34,35 In addition, documentation competes with panel management and EHR inbox completion. It is estimated that the average primary care physician receives 76.9 EHR inbox notifications daily, requiring an investment of approximately 66.8 minutes per day.36 Eliminating the burden of writing notes affords more time for physicians to attend to the tasks of panel management during, not after, their workday.
Our study found no difference in patient satisfaction between visits with or without a scribe, perhaps because of ceiling effects; patients expressed high satisfaction both during visits with and without a scribe. Nevertheless, we found that the presence of a scribe did not decrease patient satisfaction. This finding has been found in other nonrandomized studies, even in settings as sensitive as a urology practice.37
Our study is the first to evaluate charting efficiency in a randomized controlled manner. We found that scribed charts were more likely to be closed within 48 hours compared with charts completed by physicians alone. Charts that are completed in a timely manner allow patient data to be accessed by other physicians in the health care system, which is particularly important to safety and effective care coordination. Charts completed in a timely manner may also be more accurate than those completed multiple days after the patient’s visit.
This randomized controlled trial was conducted at a single family medicine clinic in an academic medical center. Although our unit of analysis was a physicians’ day or patient encounter, our study’s biggest limitation is the relatively few physicians and scribes. Our findings are positive with respect to physician satisfaction and efficiency, but future randomized studies should be conducted with large sample sizes and across multiple institutions to improve the generalizability of these findings. The physician satisfaction instrument we used measured markers related to joy of practice and was deemed feasible for repeated use, but it was not a validated survey of joy of practice or burnout. Our findings of improved efficiency, as measured by time to chart close, would be strengthened by future work using other objective approaches, such as time and motion analyses. Our data show that physicians reported higher satisfaction with the quality and accuracy of charting when scribes were present; future work should evaluate chart quality in an objective way with blinded observers using a validated instrument. We also found that patient satisfaction was not affected by the presence of a scribe, but we believe that qualitative work would better elucidate patients’ perceptions of scribes. Other worthy avenues of research include evaluating team-based care models using medical assistants or nurses as scribes,38 the effect of scribes on physician productivity and revenue, as well as cost-benefit analyses, which have been described by others39–41 but warrant further research in the primary care setting.
The challenge of modifying physicians’ practices to accommodate EHRs without sacrificing quality of care or quality of physician-patient interactions is not trivial. Some have suggested that scribes are not an appropriate solution, arguing that they are no substitute for better functioning EHRs or may remove some of the pressure on EHR designers to improve their systems.23 We agree that scribes are not a replacement for EHR redesign, but we do consider them an immediate solution that can be implemented while the more onerous and time-consuming problem of EHR redesign is also tackled. We also believe scribes can serve as a complement to a high-functioning EHR, as the latter will still require the mundane capture of information that does not require a physician. By reducing the time that physicians spend on clerical tasks, scribes serve an important function in a multidisciplinary health care team.
Acknowledgments
The authors wish to thank Sangick Chang, MD, MPH, associate dean for primary care; Tim Engberg, RN, MA, vice president for primary care; Juno Vega, RN, clinic manager; and Therese Truong, RN, assistant clinic manager, for their support of the scribe program.
Footnotes
Conflicts of interest: authors report none.
Funding support: This study was supported by a grant to the senior author (S.L.) from the Pisacano Leadership Foundation, the philanthropic foundation of the American Board of Family Medicine.
Disclaimer: The Foundation had no role in the design of the study; the collection, analysis, and interpretation of the data; and the decision to approve publication of the finished manuscript.
Previous presentations: A portion of this manuscript was presented at the Starfield Summit, April 23–26, 2016, Washington, DC; and the Society of Teachers of Family Medicine (STFM) Annual Spring Conference, April 30–May 4, 2016, Minneapolis, Minnesota.
- Received for publication December 6, 2016.
- Revision received April 15, 2017.
- Accepted for publication May 3, 2017.
- © 2017 Annals of Family Medicine, Inc.
References
In this issue
Jump to section
Related Articles
Cited By...
- Enhancing Clinical Documentation Workflow with Ambient Artificial Intelligence: Clinician Perspectives on Work Burden, Burnout, and Job Satisfaction
- Variable Impact of Medical Scribes on Physician Electronic Health Record Documentation Practices: A Quantitative Analysis Across a Large, Integrated Health-System
- Variation in Support for Documentation Among Primary Care Physicians by Gender
- Burnout, Depression, Anxiety, and Stress Among Family Physicians in Kansas: 18 Months into the COVID-19 Pandemic
- Variation in Support for Documentation Among Primary Care Physicians by Gender
- Burnout, Depression, Anxiety, and Stress Among Family Physicians in Kansas: 18 Months into the COVID-19 Pandemic
- Effect of interventions for the well-being, satisfaction and flourishing of general practitioners--a systematic review
- Do Medical Scribes Help Primary Care Providers Respond More Quickly to Out-of-Visit Tasks?
- Quality of Electronic Health Record Data on Which We Stand
- Impact of scribes on emergency medicine doctors productivity and patient throughput: multicentre randomised trial
- Impact of Medical Scribes in Primary Care on Productivity, Face-to-Face Time, and Patient Comfort
- Finance and Time Use Implications of Team Documentation for Primary Care: A Microsimulation
- Physician Burnout: Resilience Training is Only Part of the Solution
- One Year of Family Physicians' Observations on Working with Medical Scribes
- In This Issue: Tools to Help Focus on What is Valuable
- Tethered to the EHR: A Physician Workload Assessment