Article, Emergency Medicine

Minimal impact of an electronic medical records system

a b s t r a c t

Background: Electronic medical records (EMRs) implementation in hospitals and emergency departments (EDs) is becoming increasingly more common. The purpose of this study was to determine the impact of an EMR sys- tem on patient-related factors that correlate to ED workflow efficiency.

Methods: A retrospective chart review assessed monthly census reports of all patients who registered and were treated to disposition during conversion from paper charts to an EMR system. The primary outcome measure- ment was an analysis of the time of registration to discharge or total ED length of stay as well as rate of those who left without being seen, eloped, or left against medical advice. These data were recorded from 3 periods, for 18 months: before installation of the EMR system (pre-EMR), during acclimation to the EMR, and post accli- mation (post-EMR).

Results: A total of 61626 individual patient records were collected and analyzed. The total ED length of stay across all patient subtypes was not significantly affected by the installation of the hospital-wide EMR system (P = .481); however, a significant decrease was found for patients who were admitted to the hospital from the ED (P b

.00001). The percentage of patients who left without being seen between the pre-EMR and post-EMR periods was 1.8% and 2.7%, respectively, representing a significant increase (P b .0001). The number of patients who left against medical advice did not change across the periods of the present investigation (P N .25).

Conclusions: Installation of a hospital-wide EMR system had minimal impact on workflow efficiency parameters in an ED.

(C) 2015

Introduction

The implementation of electronic medical record (EMR) systems in hospitals and emergency departments (EDs) is becoming increasingly more common because of government reimbursement of such prod- ucts. Eligible hospitals can qualify to receive incentive payments under both Medicare and Medicaid if they use certified EMR technology [1]. Hospitals must demonstrate that implementation has improvED patient care by reporting 13 required core objectives and 15 clinical quality measures [1]. Because this is a relatively recent movement, there is a paucity of information on the impact of implementation of EMR in the ED and the effects on factors such as total ED length of stay as well as the total number of patients who leave before disposition, in- cluding Left without being seen , eloped, and left against medical advice (AMA).

Recent studies have shown contradictory effects of EMR systems and computerized documentation on the clinical practice of medicine. Po- tential benefits of EMR in medical practice include easy and Accurate documentation, reduction in Medical errors, incorporating standards

? Support: This study was a retrospective analysis of patient records and did not require equipment or funding support.

* Corresponding author at: Department of Biological Sciences, Youngstown State Uni- versity, One University Plaza, Youngstown, OH 44555.

E-mail address: [email protected] (J.M. Tall).

of care, improving quality of care, and improved billing or reimburse- ment [2,3]. There have been little actual data on these perceived bene- fits. In a recent study of computerized physician Order entry, the odds of mortality increased after implementation [2]. Unintended conse- quences or disadvantages of EMR include increased documentation time, interruption in clinical workflow, system errors in patient care, and additional interruptions in medical work [3]. Changes in communi- cation practices and patterns and an overdependence on technology may also play a significant role in unintended errors [2].

The Associated costs of an EMR system not only include the actual pur- chase, installation, and maintenance of the system but also the potential for future loss of productivity and efficiency [2]. Activities decreasing overall patient care time include increased time documenting, increased interruption, and increased physical movement of providers between pa- tient care and computer areas [4]. Potential loss of revenue includes de- creased productivity, increased rate of LWBS, increased time to disposition, increased ED crowding, and increased time spent on diver- sion. Although the loss of productivity is most evident with initial imple- mentation of the EMR program, decreased productivity may negatively impact patient satisfaction and prompt patients to select a competing ED for future care [2].

Emergency department crowding results from an imbalance be- tween patient demand and the ED’s supply of resources. Crowding re- sults in long wait times, high LWBS rates, and increased ED boarding times [5]. Inadequate capacity to care for patients has a negative impact

http://dx.doi.org/10.1016/j.ajem.2015.02.022

0735-6757/(C) 2015

664 J.M. Tall et al. / American Journal of Emergency Medicine 33 (2015) 663666

Table 1

The 3 periods associated with the implementation of a new hospital-wide EMR system and total LOS for all patients who registered and were treated to Disposition decision in the ED

Group description

Group period

Group name

Total ED census (no. of registered patients)

Total ED LOS (min)

Before the installation of the EMR system

02/01/11 to 08/20/11

Pre-EMR

21905

192 (114)

During acclimation to the EMR system

08/21/11 to 01/31/12

EMR-ACC

16792

197 (167)

After acclimation to the EMR system

02/01/12 to 08/31/12

Post-EMR

22929

196 (159)

Total ED LOS data are given in minutes and expressed as the mean and SD.

on patient outcomes, such as delays in medication delivery, higher com- plication rates, and increased mortality [5]. The imbalance between sup- ply and demand is potentially exacerbated by increased time of documentation and inefficiencies with EMR use.

The purpose of this study was to determine the impact of an EMR system on patient-related factors associated to ED workflow efficiency. Data were analyzed from 3 distinct phases associated with the imple- mentation of a new EMR system in the hospital, representing a total pe- riod of 18 months: before the installation of the EMR system (pre-EMR), during the hospital-wide acclimation to the EMR (EMR-ACC), and after the acclimation period to the EMR (post-EMR).

Methods

We performed a retrospective data review assessing monthly census reports of all patients who registered and were treated to disposition during conversion from paper charts to an EMR system. This study was conducted in an ED of a community hospital with 219 licensed beds and was approved by the hospital system’s institutional review board. The ED is a level III trauma center with 16 acute care beds and 5 minor care or “fast-track” beds, with an annual volume of 39000 pa- tients per year. There is a 4-year American Osteopathic Association accredited emergency medicine residency program with 8 residents in total. During the period of the present investigation, there were 8 emer- gency physicians on staff, ranging in age from 33 to 58 years as well as ro- tating resident physicians from family practice, internal medicine, and orthopedic surgery programs. Emergency department patients are seen in the ED by an attending/resident physician team approximately 50% of the time and by an attending alone 50% of the time. At the time, the ED staff did not include any physician assistants or nurse practitioners.

Before the installation of the EMR system, all ED documentation was completed on a paper template system (T sheets, T-System, Inc, Dallas, TX) along with telephone dictation. Patient progression through the ED treatment process was as follows: physician orders were handwritten on a T sheet located in the patient medical chart, the chart was given to the department secretary who entered the orders into the computer, the chart was placed into a rack, and a color indicator notified the nurs- ing staff that physician orders were made and pending. As laboratory and radiology results became available, this information automatically populated the computerized patient Tracking system (McKesson Corp, San Francisco, CA).

The EMR “go live” date for the entire hospital, including the ED, was August 21, 2011. The EMR system selected and installed hospital-wide was the EpicCare electronic medical record system, hereafter referred to as Epic. Before implementing Epic, each physician and resident phy- sician was required to attend a total of 6 hours of classroom instruction- al sessions to learn the new documentation system. On the “go live”

date and for several weeks after, EMR “trainers” were available in the ED for additional assistance.

The primary outcome measurement was an analysis of total ED LOS in minutes. Data were analyzed from 3 distinct periods associated with the implementation of a new EMR system in the hospital, each phase representing a 5.5- to 6.5-month period: (1) pre-EMR, (2) EMR-ACC, and (3) post-EMR (Table 1). The pre-EMR and post-EMR periods repre- sent the same amount of time as well as the same time of year, from February through August. The total LOS was the period from registration to patient discharge from the ED. Before Epic, the time of registration and time of discharge was handwritten on each patient medical chart and entered into a computerized spreadsheet (Microsoft Excel 2010, Redmond, WA) by the nursing manager on a daily basis. After the instal- lation of Epic, the time of registration to discharge, total ED LOS, is gen- erated and collated automatically. All patients who registered and were treated to disposition determination were included in this study and further categorized into the following subgroups: treated in the ED and discharged (T & D), admitted to the hospital (admit), or transferred to another facility (transfer). Although data from patients who regis- tered and left the ED (LWBS, eloped, and AMA patients) were recorded and reported, these patients were excluded from the LOS analyses. These data were used to determine if the implementation of an EMR system affectED patient flow through and efficiency in the ED.

Total ED LOS measurements are expressed as the mean and SD. The number of patients in each subgroup (T & D, admit, and transfer) and patients who registered but left the ED (LWBS, AMA, eloped) is expressed as the total number and percentage of ED census. Univariate analysis of variance (ANOVA) testing was used to determine if total ED LOS was significantly affected by the implementation of an EMR across all periods (pre-EMR, EMR-ACC, post-EMR). An additional univariate ANOVA was used to determine if total ED LOS was affected across all pa- tient subgroups (T & D, admit, and transfer). Bonferroni post hoc pairwise comparisons were used to determine within-group differences for total ED LOS, with a significance level correction for multiple com- parisons. Multiple 2-sample t test between proportions were performed to determine whether there was a significant difference between the percentage of patients who LWBS, left AMA, or eloped during the 3 pe- riods (Statistical Calculator; StatPac, Bloomington, MN). Analysis of var- iance and post hoc pairwise comparison data were analyzed using SPSS version 20 (IBM SPSS, Armonk, NY) statistical software. For all inferen- tial analyses, a P <= .05 was considered statistically significant.

Results

A total of 61626 individual patient records were collected and ana- lyzed during 3 periods: pre-EMR, n = 21905; EMR-ACC, n = 16792; post-EMR, n = 22 929 (Table 1). The ED census across Time groups

Table 2

The 3 periods associated with the implementation an EMR system and total LOS for all patients who registered, subcategorized by disposition decision

Group name

No. of T & D patients (% of census)

T & D patients total ED LOS (min)

No. of admit patients (% of census)

Admit patients total ED LOS (min)

No. of transfer patients (% of census)

Transfer patients total ED LOS (min)

Pre-EMR

17,238 (78.8%)

157 (97)

3686 (16.8%)

343 (97)

306 (1.4%)

303 (126)

EMR-ACC

13,190 (78.5%)

163 (122)

2784 (16.6%)

339 (232)a

301 (1.8%)

356 (298)a

Post-EMR

17,718 (77.3%)

161 (116)

3920 (17.1%)

339 (222)a

388 (1.7%)

324 (236)

Total ED LOS data are given in minutes and expressed as the mean and SD.

a P b .001 vs pre-EMR period.

J.M. Tall et al. / American Journal of Emergency Medicine 33 (2015) 663666 665

varied; however, the distribution of patient subgroups was highly con- sistent (ie, the number of patients who were treated and discharged was 78.8%, 78.5%, and 77.3% during pre-EMR, EMR-ACC, and post- EMR, respectively; Table 2).

The mean total ED LOS were 192 (SD, 114), 197 (SD, 167), and 196 (SD, 159) minutes before Epic, during the acclimation period to Epic, and after acclimation, respectively (Table 1). The total ED LOS was not significantly affected by the installation of the hospital-wide EMR sys- tem (P = .481). To determine if the implementation of an EMR system affected particular patient subgroups, an ANOVA was performed and re- vealed that total ED LOS was significantly affected among the 3 periods and 3 patient populations (P b .001). Post hoc pairwise comparisons found that the EMR system did not significantly affect the total ED LOS for patients who were treated then discharged (pre-EMR 157 [SD, 97] vs EMR-ACC 163 [SD, 122], P = .130; pre-EMR 157 [SD, 97] vs post-

EMR 161 [SD, 116], P = 1.000; Table 2). A significant decrease in total ED LOS was found between the pre-EMR (343; SD, 97) and EMR-ACC (339; SD, 232; P b .001) and pre-EMR (343; SD, 97) and post-EMR

(339; SD, 222; P b .001) for patients who were admitted to the hospital from the ED (Table 2). A significant increase in total ED LOS was found between pre-EMR (303; SD, 126) and during the period of acclimation to the new EMR system (356; SD, 298) for patients who were trans- ferred to another facility (P b .001; Table 2).

The number of patients who registered and left the ED as LWBS, AMA, or eloped was recorded (Table 3); however, these patients were excluded from the time-related total ED LOS statistical analyses. Multiple 2-sample t test between proportions were performed to determine whether there was a significant difference between the percentage of patients who LWBS, left AMA, or eloped. These analyses demonstrated that, during ac- climation to an EMR system, the percentage of patients who LWBS signif- icantly increased (pre-EMR 1.8% vs EMR-ACC 2.1%, P = .034), whereas the percentage of patients who eloped (pre-EMR 0.5% vs EMR-ACC 0.3%, P =

.002) significantly decreased. After the acclimation period to the new EMR system, the percentage of patients who LWBS continued to signifi- cantly increase (EMR-ACC 2.1% vs post-EMR 2.7%, P b .001). This trend produced a significant increase in the percentage of patients who LWBS between the pre-EMR and post-EMR periods, 1.8% and 2.7%, respectively (P b .001). The number of patients who left AMA did not change across the periods of the present investigation (Table 3).

Discussion

Installation of an EMR system had minimal impact on workflow effi- ciency parameters in an ED. The current investigation found that total ED LOS was not significantly different after installation of a hospital- wide EMR system. After the installation of Epic, the mean ED total LOS was 192 minutes, and after full acclimation to the EMR system, the mean was 196 minutes. This lack of change in total ED LOS cannot be at- tributed to seasonal or census variation because these 2 periods repre- sent the same time of year, from February through August, and practically the same ED census. These data indicate that the EMR neither decreased nor increased the overall quantity of time a patient spent in

Table 3 The 3 periods associated with the implementation an EMR system and data from patients who registered and left the ED

the ED, unlike a recent investigation that found, after an adjustment pe- riod, that the ED LOS times improved as compared with pre-EMR [6].

The mean LOS times did not change during the 18 months of this study, but the SD in time increased after Epic installation. The greater variability in the total LOS pre-EMR vs post-EMR was observed in pa- tients who are admitted to the hospital from the ED and patients who await transfer to a different facility. Factors that increase Patient boarding in the ED are often beyond the control of the ED staff, and im- plementation of an EMR would not provide considerable assistance to improve these issues. In addition, the trend of an increased LOS variabil- ity was not as sizeable across periods for patient who were treated and discharged from the ED, a patient group whose LOS is generally more regulated by the ED staff themselves.

To determine if there was a learning curve, increased LOS, associated with the installation of the EMR, data from the 5.5-month acclimation period to Epic were included in the analyses. The present study found that ED LOS slightly increased after initial EMR installation (pre-EMR, 192 minutes, and EMR-ACC, 197 minutes), similar to the results report- ed by Ward et al [6]. The acclimation time represents the phase of EMR implementation when the ED staff as well as the entire hospital was gaining proficiency with the Epic. Because no significant difference was found between the EMR-ACC and post-EMR periods, these data support the following: (1) the 6 hours of classroom instructional ses- sions on using the EMR system before the “go live” date were sufficient, and (2) the additional assistance from the EMR “trainers” in the ED for several weeks after the installation of the EMR was useful. The present findings are consistent with a previous study by Mayer et al [7] that found no significant learning curve or difference in LOS among interns learning an EMR system in an ED.

To gain better understanding of the impact of an EMR on the ED total LOS, patients were subcategorized, based on physician disposition, into 3 groups: T & D, admit, or transfer. The rationale for the re-evaluation and alternate analysis is the recognition that patients who are admitted or transferred from the ED to other facilities can spend significantly more time in the ED as compared with patients who are treated and discharged. Statistically, this is reflected in the considerable SDs ob- served for the mean ED total LOS from each of the 3 periods with sample sizes as sizeable as the present data. The total ED LOS was significantly different across patient subgroups during the 18-month period of the present investigation. A significant decrease of less than 5 minutes in total ED LOS was found for admit patients, after the installation of Epic hospital wide. A significant increase of less than 60 minutes in total ED LOS was found between pre-EMR and EMR-ACC for transfer patients. Interestingly, the analyses revealed that Epic did not significantly affect the total ED LOS for T & D patients. These data support that patients who are treated and discharged spend less time in the ED as compared with patients who are admitted or transferred, despite the installation of an EMR system. In addition, an EMR may have contributed to a very mod- est decrease in the amount of time spent in the ED for patients who are awaiting an inpatient bed.

The lackluster LOS-related effects of the EMR system may be caused by the similar functionality of the new software (Epic) to the previous sys- tem (McKesson). Both systems post patient’s laboratory and imaging re- sults along with color changes indicative to patient/order status. Likewise, both systems display the LOS for a patient, which allows the ED physician to be acutely aware of patients who may be waiting exces-

sively as well as prompt intervention if ancillary studies or nursing care

are delayed. The Epic system provides 2 main differences as compared with the previous computerized system used in the ED: (1) Epic allows the physician to enter orders without the need of another staff member, and (2) Epic provides a more straightforward process to obtain past med- ical history in the patient’s medical record. Both improvements ought to

Group name

No. of LWBS (% of census)

No. of left AMA (% of census)

No. of eloped (% of census)

Pre-EMR

405 (1.8%)a,b

161 (0.7%)

109 (0.5%)d

EMR-ACC

348 (2.1%)c

119 (0.7%)

50 (0.3%)

Post-EMR

624 (2.7%)

178 (0.8%)

101 (0.4%)

These patients were excluded from the LOS analyses.

a P b .001 pre-EMR vs post-EMR.

b P = .034 pre-EMR vs EMR-ACC periods.

c P b .001 EMR-ACC vs post-EMR periods.

d P = .002 pre-EMR vs EMR-ACC.

improve workflow efficiency; however, these factors are seemingly in- consequential as the total ED LOS before and after the installation of Epic did not change. Perhaps using a measure such as total LOS to gauge efficiency in the ED may not be appropriate. If holdups or “bottlenecks”

666 J.M. Tall et al. / American Journal of Emergency Medicine 33 (2015) 663666

are occurring in departments outside the ED, then an EMR system would provide negligible assistance for these matters. Hence, this leads to the type of data results found in the current investigation.

During the 18-month period of the current investigation, a signifi- cant difference was demonstrated in the percentage of patients who LWBS and eloped. Patients who elope were seen by a physician and, without notice, left the ED. The present study found a significant de- crease in the percentage of eloped patients, 0.5% to 0.3%, during the im- plementation of the Epic system, and during this same period, the percentage of patients who registered and left the ED without being seen by a physician increased from 1.8% to 2.1%. During the first 4 weeks of the 5.5-month EMR-ACC period, the ED nursing and physician staff was increased as a preemptive measure during the orientation phase of the new EMR system. This increase in staffing may have been reflected with the reduced percentage of patients who eloped but was reflected neither in a significant improvement in the ED total LOS nor a reduction in patients who LWBS. These patient-related measures re- flect what occurred and cannot predict what could have been; therefore, during the EMR-ACC period, workflow efficiency could have been great- ly hampered if staffing was not increased, the additional assistance from the EMR “trainers” were not in the ED, and the staff did not receive ad- equate software training.

The percentage of registered patients who LWBS continued to in-

crease during the 6 months after full acclimation to Epic or post-EMR period. Multiple studies have shown that increased census and higher acuity lead to increased crowding, time to disposition, total LOS, and LWBS rates [8-10]. Although, Emergency Severity Index triage scores were not collected, patients were categorized based upon physician dis- position into 3 groups and, to some extent, related to Acuity level. Those T & D were likely low acuity, admit were likely high acuity, and transfer were likely highest acuity. In the present study, the percentage of ED census for each Patient category was not statistically different among the 3 periods; therefore, acuity level remained consistent. Patients’ will- ingness to wait has been associated with higher acuity, patient loyalty to a certain hospital system, and the need for specialty care [8]. Because the present study did not find a difference in acuity pre-EMR and post-EMR, this cannot explain the increase in LWBS. It is unlikely that a change in loyalty occurred among patients because no major changes in services or availability of specialty care were implemented during the time of the present study. The increasing proportion of LWBS patients is concerning and warrants further investigation. The current study did not collect specific patient demographic information such as sex, age, or medical insurance status. Likewise, time-related data on the average hold or Boarding times in the ED and time on diversion were not collect- ed. These sorts of variables may have been beneficial in determining trends in the patient populations and correlations with ED total LOS as well as patients who leave the ED before disposition.

Conclusions

As more hospitals adopt EMR systems, it is important to understand and make advanced plans for the impact of these changes. The ED, a de- partment that is particularly overcrowded and has limited resources, must be acutely prepared during the installation and implementation of a new EMR system. The purpose of this study was to determine the impact on patient-related factors associated to ED workflow efficiency before the installation of an EMR system, during the hospital-wide orientation to the system, and after the acclimation period. The total ED LOS was not signifi- cantly affected by the installation of the hospital-wide EMR system. A modest decrease in total ED LOS was found for patients who were admit- ted to the hospital, and an increased time in the ED was discovered for transfer patients after the installation of an EMR. A significant difference was demonstrated in the percentage of patients who LWBS and eloped; however, the percentage of patients who left AMA was unchanged. Addi- tional analyses must be undertaken to fully elucidate the short- and long- term impact of an EMR on workflow efficiency.

Acknowledgments

The authors thank Stephanie Kupek, MS, RN, for her assistance with data collection. In addition, we appreciate the time and input from Dr. G. Jay Kerns, Department of Mathematics and Statistics at Youngstown State University, regarding the statistical analyses in the study.

References

  1. Centers for Medicare & Medicaid Services. The Official Web Site for the Medicare and Medicaid electronic health records (EMR) Incentive Programs. Available at http:// www.cms.gov. [Last accessed 12/15/2014].
  2. Handel DA, Hackman JL. Implementing electronic health records in the emergency department. J Emerg Med 2010;38(2):257-63.
  3. Park SY, Lee SY, Chen Y. The effects of EMR deployment on doctors’ work practices: A qualitative study in the emergency department of a teaching hospital. Int J Med In- form 2010;81:204-17.
  4. Abraham J, Kannampallil TG, Reddy MC. Peripheral activities during EMR use in emergency care: A case study. AMIA 2009 Symposium Proceedings, 1-5.
  5. Singer AJ, Thode HC, Viccellio P, Pines JM. The association between length of emer- gency department boarding and mortality. Acad Emerg Med 2011;18:1324-9.
  6. Ward MJ, Froehle CM, Hart KW, Collins SP, Lindsell CJ. Transient and sustained changes in operational performance, Patient evaluation, and medication administra- tion during electronic health record implementation in the emergency department. Ann Emerg Med 2014;63:320-8.
  7. Mayer PH, Yaron M, Lowenstein SR. Impact on length of stay after introduction of emergency department information system. West J Emerg Med 2010;11(4):329-32.
  8. Shaikh SB, Jerrard DA, Witting MD, Winters ME, Brodeur MN. How long are patients willing to wait in the emergency department before leaving without being seen? West J Emerg Med 2012;13(6):463-7.
  9. Murrell KL, Offerman SR, Kauffman MB. Applying lean: Implementation of a rapid triage and treatment system. West J Emerg Med 2011;12(2):184-91.
  10. Kulstad EB, Hart KM, Waghchoure S. occupancy rates and emergency department work index scores correlate with leaving without being seen. West J Emerg Med 2010;11(4):324-8.

Leave a Reply

Your email address will not be published. Required fields are marked *