Explain the limitations presented by the study population and sample size used. Support your explanation with examples from the study.

B. Discuss the advantages and disadvantages of the type of statistical analysis used. Support your discussion with examples from the study.

C. Identify the limitations of the study design used and explain why those limitations exist.

D. Discuss how the results from this study compare with those of a study listed in the references of your main article. Support your discussion with examples from both studies.

E. Explain how this study contributes to scientific literature. Support your explanation with specific examples.

F. Describe further investigations that could take place because of this study.

Time Motion Analysis: Impact of Scribes on Provider Time ManagementAuthor links open overlay panelHeather A.HeatonMDRonaWangMD, MBAKyle J.FarrellBSOctavia S.RuelasBSDeepi G.GoyalMDChristine M.LohseAnnie T.SadostyMDDavid M.NestlerMD, MSShow moreAdd to MendeleyShareCitehttps://doi.org/10.1016/j.jemermed.2018.04.018Get rights and contentAbstractBackground

Scribes are unlicensed professionals trained in medical data entry. Limited data exist on the impact of scribes on provider time management in the emergency department (ED). Time-motion analysis is a tool utilized in business to capture detailed movements and durations to task completion. It offers a means to categorize how providers allocate their time during a clinical shift.

Objective

Evaluate the impact of scribes on how ED providers spend their time.

Methods

A prospective observational study was conducted to assess scribe impact on provider time utilization. Four research assistants (RAs) observed attending providers on 24 8-h control shifts (without a scribe), and 24 scribed shifts. RAs observed and categorized provider activity. Providers self-reported after-hours documentation times. Two-sample t-tests were used for normally distributed data, and Wilcoxon rank-sum tests were used for skewed data. All tests were two-sided, and p-values < 0.05 were considered statistically significant.

Results

Scribes decreased total documentation time both on shift (mean 55.3 vs. 36.4min, p<0.001) and post shift (mean 42.5 vs. 23.3min, p=0.038). They did not significantly decrease the amount of time spent reviewing the medical records or placing orders, nor did they have an impact on provider time spent at patients' bedside or time spent discussing patient care with team members.

Conclusion

The presence of scribes decreased provider documentation time but did not change the amount of time spent at the bedside or communicating with other team members. Scribes may be a potential strategy to decrease the clerical burden.

• Previous article in issue
• Next article in issue

Keywordsscribesclerical burdenprovider burnoutelectronic medical recordsIntroduction

Clerical burden, including electronic health record (EHR) documentation, reviewing past medical records, and ordering tests and medications, has become a significant burden on emergency department (ED) providers (1). Tasks using a computer interface can take up more than half of a physician's time per shift (2). EHRs serve a promising role in health care quality and safety; however, multiple recent publications outline the limitations and difficulties associated with them, including their innate inefficiencies, time-consuming nature, and disruptiveness when used in the patient encounter 3, 4, 5. Further, literature suggests that physicians have shifted their focus from face-to-face patient care to face-to-screen, with an estimated 43% of a physician's time spent on data entry, compared with 28% on direct patient care (6).

Scribes, nonlicensed health care team members that document the patient history and examination contemporaneously with the clinical encounter, offer a potential solution to the clerical burdens and time constraints felt by ED providers. There is no clear definition of the scope of practice of scribes, and duties vary amongst clinical sites. Scribes keep track of laboratory and radiological findings and record pertinent documentation to improve physician productivity and patient care (3). They do not act independently, but rather function under the supervision of a physician to assist with documentation, retrieval of test results, and support workflow (7).

Several editorials propose the use of scribes as an operational improvement tactic for providers in a variety of health care settings 8, 9, 10, 11. However, rigorous peer-reviewed literature is limited 12, 13, 14, 15, 16, 17, 18, 19, 20. Furthermore, peer-review studies looking at the use of scribes in an ED are even more limited 14, 15, 16, 17, 18, 19, 20. A recent meta-analysis highlighted the difficulty in determining how and when scribes are beneficial to EDs (21).

The impact of scribes must be critically examined to inform health administrators and physicians considering employing scribes. Additional research in task substitution and workflow efficiency may aid hospital administrators and medical practitioners seeking to enhance daily work productivity. A study method known as "time motion analysis" (TMA) can be applied to identify how scribes affect practitioners' workflow. TMA is a method that systematically breaks down a clinical shift into individual functional components through direct and continuous observation. Thus, a practitioner's work shift is categorized into time units that are then assigned to routine clinical tasks usually encountered during the shift. Although this methodology has been described once before in relation to scribes in the clinic setting, we are the first to apply TMA to ED scribes (13). Using TMA, we evaluated and compared how ED practitioners spent their time on a shift, with and without a scribe on their team.

MethodsStudy Design and Setting

This study was conducted at an academic ED that hosts an Emergency Medicine residency training program. Our ED includes several distinct treatment areas, or pods, one of which is dedicated to pediatric patients. We manage 75,000 patient visits annually, 82% of whom are adults (age>17years). On average, 35% of adult patients and 13% of pediatric patients are admitted.

To standardize this study, we limited our evaluation to a single area of the ED. This area manages adult patients with Emergency Severity Index levels of 2 through 5, and is staffed by an attending provider and a nurse practitioner or a physician assistant every Tuesday-Friday. These shifts are predictably busy, comprise similar patient populations, and provide similar provider experiences. Staffing Saturday, Sunday, and Monday varies, and therefore these days were excluded a priori. Triage nurses assign patients to areas and rooms on an acuity-based, first-come-first-served basis, and providers have no input on which patients are assigned to their area. To minimize bias, the study's three physician investigators did not work in this area during the study period. Our study utilized a prospective cohort design.

The study was deemed exempt by our Institutional Review Board.

Selection of Participants

ED attending physicians and ED scribes were observed from January 31, 2017 to April 21, 2017. Scribe staffing followed an allocation pattern developed independently from the providers' schedule, with no preference given to specific providers or scribes. The pattern ensured balance between the scribe (intervention) and nonscribe (control) groups in times and days of the week, with an equal number of scribe days and nonscribe days in this study.

Intervention

Scribes were recruited and trained through an in-house program with a defined curriculum developed by a physician with prior experience implementing scribe programs (22). The scribes were largely prehealth students hired as temporary employees for expected 1- to 2-year periods. Each scribe provided 1-to-1 support to an attending physician for the entirety of the physician's shift. Scribe experience ranged from 6months to 2years. Scribes in our institution accompany physicians into the patient room during the initial encounter to document the history of present illness; review of systems; past medical, social, and surgical history; and physical examination. After the initial evaluation, they prompt the provider for a dictated medical decision-making section. Throughout the patient's ED stay, scribes document re-evaluations of the patient, pertinent laboratory and radiology findings, and disposition discussions with discharge instructions as appropriate.

Methods and Measurements

Four research assistants (RAs) were hired to observe and record activities throughout an entire shift. Prior to initiation of the study, the RAs participated in observation shifts with the Principal Investigator to practice using the data collection tool and to assure reliability in data categorization and collection.

A tablet-based time recorder was used for real-time capture of all activities during the teams' 8-h shift. If the provider was still at work when the RA's 8-h shift was complete, additional after-shift documentation time was self-reported by providers and electronically mailed to the RA at the end of the work day.

Outcomes

Data collected included shift date and time, provider demographics, and total time spent on the following categories:

1.

Time spent interacting with the EHR, including:

a.

Order entry

b.

EHR reading/review, including Past Medical History and tests/imaging results

c.

Documentation entry

2.

Time spent at the patient bedside

3.

Time spent discussing care with other team members, including:

a.

Direct patient care discussions with other providers

b.

On-shift education to other staff

4.

Other time, including:

a.

Personal time (including breaks, eating, and non-work-related conversations)

b.

Uncategorized time

Analysis

Continuous times were summarized with means if approximately normally distributed and medians otherwise. p-Values were obtained using two-sample t-tests for normally distributed times, and Wilcoxon rank-sum tests for times that were skewed. All tests were two-sided. Statistical analyses were performed using version 9.4 of the SAS software package (SAS Institute Inc., Cary, NC). Any p-value < 0.05 was considered statistically significant.

ResultsCharacteristics of the Study Subjects

During the study period, 24 shifts with a scribe were observed, and 24 shifts without a scribe were observed. Some physicians were observed more than once due to scheduling done prior to the initiation of the study. Eleven of the scribe shifts and six of the nonscribe shifts included a medical student on the treatment team.

Main Results

Scribes significantly decreased the amount of time spent with the EHR. With no scribe on the treatment team, a median of 179.4min was spent interacting with the EHR. A median of 147.5min was spent with the EHR when a scribe was on the treatment team (p=0.028). Specifically, scribes significantly decreased the amount of time providers spent with both shift documentation (mean 36.4min vs. 55.3min, p<0.001) and postshift documentation (median 17.5min vs. 37.5min, p=0.038). Scribes did not significantly decrease the amount of time spent reviewing the medical records or placing orders (mean 101.3min vs. 87.1min, p=0.15) (see Table1).

Table1. Electronic Health Record (Minutes)

Empty Cell	No Scribe	Scribe	p-Value
Mean	SD	Median	Mean	SD	Median
Reviewing records, order entry	87.1	23.2	83.0	101.3	41.5	90.7	0.15
Shift documentation	55.3	18.8	56.3	36.4	17.0	31.8	<0.001
Postshift documentation	42.5	32.5	37.5	23.3	16.9	17.5	0.038
Total	184.8	46.9	179.4	161.0	43.9	147.5	0.028

Scribes did not significantly affect time spent at the patient bedside (Table2). When subdividing these times, there still was no difference in time spent on the initial interview and examination, or on further interactions such as procedures, re-evaluation, disposition discussions with the patient and family, or patient education. Scribes also failed to affect conversations with care team members, including time spent in conversation with nursing, other ED providers, others (e.g., consulting services), or with medical students (Table3). And finally, scribes did not affect time categorized as miscellaneous/personal (Table4).

Table2. Time at Patient Bedside (Minutes)

Empty Cell	No Scribe	Scribe	p-Value
Mean	SD	Median	Mean	SD	Median
Initial interview, examination	88.5	27.6	86.0	95.8	33.5	94.5	0.41
Procedures, re-evaluation, disposition, education	51.8	35.8	41.0	42.1	27.5	33.7	0.33
Total	140.3	48.7	131.8	138.0	49.0	134.5	0.88

Table3. Conversations with Care Team Members (Minutes)

Empty Cell	No Scribe	Scribe	p-Value
Mean	SD	Median	Mean	SD	Median
With nursing	19.2	9.3	16.5	16.3	8.7	13.5	0.18
With other ED providers	32.1	18.7	31.2	28.6	17.2	28.2	0.51
With others	37.6	16.6	34.3	35.8	17.7	30.6	0.57
Total	88.9	32.9	83.5	80.7	30.0	75.1	0.37
Medical student	48.8	27.5	52.7	43.2	22.8	42.8	0.66

ED=emergency department.

Table4. Miscellaneous Time (Minutes)

Empty Cell	No Scribe	Scribe	p-Value
Mean	SD	Median	Mean	SD	Median
Personal	94.7	60.9	84.9	101.4	55.2	89.9	0.57

Discussion

Scribes significantly impacted the amount of time providers spend interacting with the EHR. Without scribes, ED providers spend approximately one-third of their 8-h shift interacting with the EHR, similar to what has been previously reported in literature. Scribes decreased this time spent by approximately 30%. Also, postshift documentation decreased by nearly 50% when ED providers were paired with scribes. Scribes made no difference in the amount of time spent reviewing the medical record; independent of the presence of a scribe, the provider will need to review old records to assist with the care of the patient. In our facility, scribes do not enter orders, and therefore, no change was seen in electronic order entry. Although not statistically significant, providers paired with scribes did spend slightly more time at the patient's bedside. Furthermore, scribes decreased postshift documentation time, allowing providers to leave in a timelier manner after their shift.

To date, only one other published study has evaluated the impact of scribes using TMA. Bank etal. reviewed scribes in a cardiology clinic and found that scribes led to a decrease in direct patient contact time, but an increase in time interacting with patients without a computer (13). Our study found no change in time at the patient bedside.

Limitations

Several limitations to the study exist. First, RAs were not allowed to directly inquire about the tasks providers were performing, as it would have compromised the integrity of the observational study design. Thus, they may have needed to infer from time to time which appropriate category of tasks the provider was undertaking. For example, a provider may have been observed utilizing the EHR, but it may have been unclear whether this was for chart review or order entry. Furthermore, conversation and workflow were often fluid in nature, where a conversation may have started around patient care and subsequently transitioned in and out of educational topics. As such, minor variances in RA interpretation may exist, which may lead to potential errors in categorization and time recordings of work activities. Second, observations were limited only to on-shift hours. Documentation time or patient care that took place after the 8-h shift mark relied on self-reported data. Thus, a small portion of the data in this study depended on providers' best recall rather than captured reports. Third, RAs could only observe interactions and workflow that physically occurred in the patient care area. Should the provider leave the patient care area of the ED, further activities may not be captured by the RAs. For example, the attending workroom, which is equipped with computer work stations, is located outside of the patient care area. Though infrequent, providers may utilize this space during on-shift hours to perform work-related tasks that may be missed by the RAs. Finally, although RAs did not directly interact or communicate with providers, nor were providers made aware of the purpose of the study, the simple presences of RAs on shift may lead to inadvertent provider workflow variations due to the Hawthorne effect.

Conclusion

The presence of scribes decreased overall provider shift documentation time, but did not change the amount of time physicians spent at the patient bedside or communicating with other care team members. For sites where physicians are paid hourly, scribes might be a cost-effective solution to decreasing the hours needed for documentation. These data should be combined with billing, timely chart completion, safety, and outcomes work to understand the total impact of scribes on ED operations. Scribes may be a strategy to decrease clerical documentation burden.

Time Motion Analysis: Impact of Scribes on Provider Time ManagementAuthor links open overlay panelHeather A.HeatonMDRonaWangMD, MBAKyle J.FarrellBSOctavia S.RuelasBSDeepi G.GoyalMDChristine M.LohseAnnie T.SadostyMDDavid M.NestlerMD, MSShow moreAdd to MendeleyShareCitehttps://doi.org/10.1016/j.jemermed.2018.04.018Get rights and contentAbstractBackground

Scribes are unlicensed professionals trained in medical data entry. Limited data exist on the impact of scribes on provider time management in the emergency department (ED). Time-motion analysis is a tool utilized in business to capture detailed movements and durations to task completion. It offers a means to categorize how providers allocate their time during a clinical shift.

Objective

Evaluate the impact of scribes on how ED providers spend their time.

Methods

A prospective observational study was conducted to assess scribe impact on provider time utilization. Four research assistants (RAs) observed attending providers on 24 8-h control shifts (without a scribe), and 24 scribed shifts. RAs observed and categorized provider activity. Providers self-reported after-hours documentation times. Two-sample t-tests were used for normally distributed data, and Wilcoxon rank-sum tests were used for skewed data. All tests were two-sided, and p-values < 0.05 were considered statistically significant.

Results

Scribes decreased total documentation time both on shift (mean 55.3 vs. 36.4min, p<0.001) and post shift (mean 42.5 vs. 23.3min, p=0.038). They did not significantly decrease the amount of time spent reviewing the medical records or placing orders, nor did they have an impact on provider time spent at patients' bedside or time spent discussing patient care with team members.

Conclusion

The presence of scribes decreased provider documentation time but did not change the amount of time spent at the bedside or communicating with other team members. Scribes may be a potential strategy to decrease the clerical burden.

• Previous article in issue
• Next article in issue

Keywordsscribesclerical burdenprovider burnoutelectronic medical recordsIntroduction

Clerical burden, including electronic health record (EHR) documentation, reviewing past medical records, and ordering tests and medications, has become a significant burden on emergency department (ED) providers (1). Tasks using a computer interface can take up more than half of a physician's time per shift (2). EHRs serve a promising role in health care quality and safety; however, multiple recent publications outline the limitations and difficulties associated with them, including their innate inefficiencies, time-consuming nature, and disruptiveness when used in the patient encounter 3, 4, 5. Further, literature suggests that physicians have shifted their focus from face-to-face patient care to face-to-screen, with an estimated 43% of a physician's time spent on data entry, compared with 28% on direct patient care (6).

Scribes, nonlicensed health care team members that document the patient history and examination contemporaneously with the clinical encounter, offer a potential solution to the clerical burdens and time constraints felt by ED providers. There is no clear definition of the scope of practice of scribes, and duties vary amongst clinical sites. Scribes keep track of laboratory and radiological findings and record pertinent documentation to improve physician productivity and patient care (3). They do not act independently, but rather function under the supervision of a physician to assist with documentation, retrieval of test results, and support workflow (7).

Several editorials propose the use of scribes as an operational improvement tactic for providers in a variety of health care settings 8, 9, 10, 11. However, rigorous peer-reviewed literature is limited 12, 13, 14, 15, 16, 17, 18, 19, 20. Furthermore, peer-review studies looking at the use of scribes in an ED are even more limited 14, 15, 16, 17, 18, 19, 20. A recent meta-analysis highlighted the difficulty in determining how and when scribes are beneficial to EDs (21).

The impact of scribes must be critically examined to inform health administrators and physicians considering employing scribes. Additional research in task substitution and workflow efficiency may aid hospital administrators and medical practitioners seeking to enhance daily work productivity. A study method known as "time motion analysis" (TMA) can be applied to identify how scribes affect practitioners' workflow. TMA is a method that systematically breaks down a clinical shift into individual functional components through direct and continuous observation. Thus, a practitioner's work shift is categorized into time units that are then assigned to routine clinical tasks usually encountered during the shift. Although this methodology has been described once before in relation to scribes in the clinic setting, we are the first to apply TMA to ED scribes (13). Using TMA, we evaluated and compared how ED practitioners spent their time on a shift, with and without a scribe on their team.

MethodsStudy Design and Setting

This study was conducted at an academic ED that hosts an Emergency Medicine residency training program. Our ED includes several distinct treatment areas, or pods, one of which is dedicated to pediatric patients. We manage 75,000 patient visits annually, 82% of whom are adults (age>17years). On average, 35% of adult patients and 13% of pediatric patients are admitted.

To standardize this study, we limited our evaluation to a single area of the ED. This area manages adult patients with Emergency Severity Index levels of 2 through 5, and is staffed by an attending provider and a nurse practitioner or a physician assistant every Tuesday-Friday. These shifts are predictably busy, comprise similar patient populations, and provide similar provider experiences. Staffing Saturday, Sunday, and Monday varies, and therefore these days were excluded a priori. Triage nurses assign patients to areas and rooms on an acuity-based, first-come-first-served basis, and providers have no input on which patients are assigned to their area. To minimize bias, the study's three physician investigators did not work in this area during the study period. Our study utilized a prospective cohort design.

The study was deemed exempt by our Institutional Review Board.

Selection of Participants

ED attending physicians and ED scribes were observed from January 31, 2017 to April 21, 2017. Scribe staffing followed an allocation pattern developed independently from the providers' schedule, with no preference given to specific providers or scribes. The pattern ensured balance between the scribe (intervention) and nonscribe (control) groups in times and days of the week, with an equal number of scribe days and nonscribe days in this study.

Intervention

Scribes were recruited and trained through an in-house program with a defined curriculum developed by a physician with prior experience implementing scribe programs (22). The scribes were largely prehealth students hired as temporary employees for expected 1- to 2-year periods. Each scribe provided 1-to-1 support to an attending physician for the entirety of the physician's shift. Scribe experience ranged from 6months to 2years. Scribes in our institution accompany physicians into the patient room during the initial encounter to document the history of present illness; review of systems; past medical, social, and surgical history; and physical examination. After the initial evaluation, they prompt the provider for a dictated medical decision-making section. Throughout the patient's ED stay, scribes document re-evaluations of the patient, pertinent laboratory and radiology findings, and disposition discussions with discharge instructions as appropriate.

Methods and Measurements

Four research assistants (RAs) were hired to observe and record activities throughout an entire shift. Prior to initiation of the study, the RAs participated in observation shifts with the Principal Investigator to practice using the data collection tool and to assure reliability in data categorization and collection.

A tablet-based time recorder was used for real-time capture of all activities during the teams' 8-h shift. If the provider was still at work when the RA's 8-h shift was complete, additional after-shift documentation time was self-reported by providers and electronically mailed to the RA at the end of the work day.

Outcomes

Data collected included shift date and time, provider demographics, and total time spent on the following categories:

1.

Time spent interacting with the EHR, including:

a.

Order entry

b.

EHR reading/review, including Past Medical History and tests/imaging results

c.

Documentation entry

2.

Time spent at the patient bedside

3.

Time spent discussing care with other team members, including:

a.

Direct patient care discussions with other providers

b.

On-shift education to other staff

4.

Other time, including:

a.

Personal time (including breaks, eating, and non-work-related conversations)

b.

Uncategorized time

Analysis

Continuous times were summarized with means if approximately normally distributed and medians otherwise. p-Values were obtained using two-sample t-tests for normally distributed times, and Wilcoxon rank-sum tests for times that were skewed. All tests were two-sided. Statistical analyses were performed using version 9.4 of the SAS software package (SAS Institute Inc., Cary, NC). Any p-value < 0.05 was considered statistically significant.

ResultsCharacteristics of the Study Subjects

During the study period, 24 shifts with a scribe were observed, and 24 shifts without a scribe were observed. Some physicians were observed more than once due to scheduling done prior to the initiation of the study. Eleven of the scribe shifts and six of the nonscribe shifts included a medical student on the treatment team.

Main Results

Scribes significantly decreased the amount of time spent with the EHR. With no scribe on the treatment team, a median of 179.4min was spent interacting with the EHR. A median of 147.5min was spent with the EHR when a scribe was on the treatment team (p=0.028). Specifically, scribes significantly decreased the amount of time providers spent with both shift documentation (mean 36.4min vs. 55.3min, p<0.001) and postshift documentation (median 17.5min vs. 37.5min, p=0.038). Scribes did not significantly decrease the amount of time spent reviewing the medical records or placing orders (mean 101.3min vs. 87.1min, p=0.15) (see Table1).

Table1. Electronic Health Record (Minutes)

Empty Cell	No Scribe	Scribe	p-Value
Mean	SD	Median	Mean	SD	Median
Reviewing records, order entry	87.1	23.2	83.0	101.3	41.5	90.7	0.15
Shift documentation	55.3	18.8	56.3	36.4	17.0	31.8	<0.001
Postshift documentation	42.5	32.5	37.5	23.3	16.9	17.5	0.038
Total	184.8	46.9	179.4	161.0	43.9	147.5	0.028

Scribes did not significantly affect time spent at the patient bedside (Table2). When subdividing these times, there still was no difference in time spent on the initial interview and examination, or on further interactions such as procedures, re-evaluation, disposition discussions with the patient and family, or patient education. Scribes also failed to affect conversations with care team members, including time spent in conversation with nursing, other ED providers, others (e.g., consulting services), or with medical students (Table3). And finally, scribes did not affect time categorized as miscellaneous/personal (Table4).

Table2. Time at Patient Bedside (Minutes)

Empty Cell	No Scribe	Scribe	p-Value
Mean	SD	Median	Mean	SD	Median
Initial interview, examination	88.5	27.6	86.0	95.8	33.5	94.5	0.41
Procedures, re-evaluation, disposition, education	51.8	35.8	41.0	42.1	27.5	33.7	0.33
Total	140.3	48.7	131.8	138.0	49.0	134.5	0.88

Table3. Conversations with Care Team Members (Minutes)

Empty Cell	No Scribe	Scribe	p-Value
Mean	SD	Median	Mean	SD	Median
With nursing	19.2	9.3	16.5	16.3	8.7	13.5	0.18
With other ED providers	32.1	18.7	31.2	28.6	17.2	28.2	0.51
With others	37.6	16.6	34.3	35.8	17.7	30.6	0.57
Total	88.9	32.9	83.5	80.7	30.0	75.1	0.37
Medical student	48.8	27.5	52.7	43.2	22.8	42.8	0.66

ED=emergency department.

Table4. Miscellaneous Time (Minutes)

Empty Cell	No Scribe	Scribe	p-Value
Mean	SD	Median	Mean	SD	Median
Personal	94.7	60.9	84.9	101.4	55.2	89.9	0.57

Discussion

Scribes significantly impacted the amount of time providers spend interacting with the EHR. Without scribes, ED providers spend approximately one-third of their 8-h shift interacting with the EHR, similar to what has been previously reported in literature. Scribes decreased this time spent by approximately 30%. Also, postshift documentation decreased by nearly 50% when ED providers were paired with scribes. Scribes made no difference in the amount of time spent reviewing the medical record; independent of the presence of a scribe, the provider will need to review old records to assist with the care of the patient. In our facility, scribes do not enter orders, and therefore, no change was seen in electronic order entry. Although not statistically significant, providers paired with scribes did spend slightly more time at the patient's bedside. Furthermore, scribes decreased postshift documentation time, allowing providers to leave in a timelier manner after their shift.

To date, only one other published study has evaluated the impact of scribes using TMA. Bank etal. reviewed scribes in a cardiology clinic and found that scribes led to a decrease in direct patient contact time, but an increase in time interacting with patients without a computer (13). Our study found no change in time at the patient bedside.

Limitations

Several limitations to the study exist. First, RAs were not allowed to directly inquire about the tasks providers were performing, as it would have compromised the integrity of the observational study design. Thus, they may have needed to infer from time to time which appropriate category of tasks the provider was undertaking. For example, a provider may have been observed utilizing the EHR, but it may have been unclear whether this was for chart review or order entry. Furthermore, conversation and workflow were often fluid in nature, where a conversation may have started around patient care and subsequently transitioned in and out of educational topics. As such, minor variances in RA interpretation may exist, which may lead to potential errors in categorization and time recordings of work activities. Second, observations were limited only to on-shift hours. Documentation time or patient care that took place after the 8-h shift mark relied on self-reported data. Thus, a small portion of the data in this study depended on providers' best recall rather than captured reports. Third, RAs could only observe interactions and workflow that physically occurred in the patient care area. Should the provider leave the patient care area of the ED, further activities may not be captured by the RAs. For example, the attending workroom, which is equipped with computer work stations, is located outside of the patient care area. Though infrequent, providers may utilize this space during on-shift hours to perform work-related tasks that may be missed by the RAs. Finally, although RAs did not directly interact or communicate with providers, nor were providers made aware of the purpose of the study, the simple presences of RAs on shift may lead to inadvertent provider workflow variations due to the Hawthorne effect.

Conclusion

The presence of scribes decreased overall provider shift documentation time, but did not change the amount of time physicians spent at the patient bedside or communicating with other care team members. For sites where physicians are paid hourly, scribes might be a cost-effective solution to decreasing the hours needed for documentation. These data should be combined with billing, timely chart completion, safety, and outcomes work to understand the total impact of scribes on ED operations. Scribes may be a strategy to decrease clerical documentation burden.