ORIGINAL RESEARCH

Forecasting the effect of physician assistants in a pediatric ED Quynh Doan, MDCM, MHSc, PhD; William Hall; Steven Shechter, PhD; Niranjan Kissoon, MD; Sam Sheps, MD, MSc; Joel Singer, PhD; Hubert Wong, PhD; David Johnson, MD

ABSTRACT Background: Most pediatric ED visits are for nonemergent problems. Physician assistants are well trained to manage these patients; however, their effect on patient flow in a pediatric ED is unknown. Objectives: To compare the effect on key pediatric ED efficiency indicators of extending physician coverage versus adding PAs with equivalent incremental costs. Methods: We used discrete event simulation modeling to compare the effect of additional physician coverage versus adding PAs on wait time, length of stay (LOS), and patients leaving without being seen. Results: Simulation of extended physician coverage reduced wait times, LOS, and rates of leaving without being seen across acuity levels. Adding PAs reduced wait times and LOS for high-acuity visits, and slightly increased the LOS for low-acuity visits. Conclusions: With restricted autonomy, PAs mainly benefitted the high-acuity patients. Increasing the level of PA autonomy was critical in broadening the effect of PAs to all acuity levels. Keywords: pediatric ED, healthcare costs, provider services, workforce, physician assistants, discrete event simulation

Overcrowding in North American EDs is a growing con concern and is believed to be the result of converging and competing factors.1,2 Pediatric EDs have also experienced problems with overcrowding, resulting in a decreased quality of care for patients.3-6 The most frequent cause of overcrowding in general EDs is boarding of admitted patients.7 In contrast, a 2010 survey of Canadian pediatric ED directors and nursing staff suggested Quynh Doan is an assistant professor in the Department of Pediatrics at the University of British Columbia in Vancouver. William Hall is a research associate at the Centre for Clinical Epidemiology and Evaluation at the University of British Columbia. Steven Shechter is an associate professor in the Sauder School of Business at the University of British Columbia. Niranjan Kissoon is vice president of medical affairs at British Columbia Children’s Hospital in Vancouver and a professor of acute and critical care and global child health at the University of British Columbia. In the university’s School of Population and Public Health, Sam Sheps and Joel Singer are professors and Hubert Wong is an associate professor. David Johnson is a professor in the Department of Pediatrics and Pharmacology and Physiology

that high volumes coupled with inadequate staffing levels lead to prolonged waiting times.8 Waiting time is a factor that may contribute to some patients and families (0.2% to 7.6% of all pediatric ED visits) leaving without having been seen by a physician.9-11 These patients may have conditions that can lead to serious adverse events if left untreated.12 Therefore, overcrowding not only is inconvenient for families but can also have adverse health consequences. Nevertheless, pediatric EDs treat a relatively higher proportion of nonurgent and low-complexity cases (35% to 70% of cases) and have lower patient admission rates than general EDs.13-16 Efforts in pediatric EDs to improve patient flow should focus on the efficient management of large volumes of patients with low acuity. Innovative staffing models, such as the use of physician assistants (PAs), are well accepted in the United States.17-21 However, PAs are relatively new to the healthcare system in Canada, and although their skills and scope of practice are well suited for the pediatric ED, the range of clinical presentations PAs would be involved with and their level of autonomy in a Canadian pediatric ED may be limited at first.22,23 Nonetheless, experience in the United States and a few Canadian provinces suggests that PAs (who require less training, and are less costly than physicians) may become a cost-effective solution to pediatric ED overcrowding.24 Discrete event simulation (DES) modeling is commonly used in healthcare systems to model and subsequently optimize resource allocation.25,26 Although many simulation studies have examined staffing level changes on health systems, none have evaluated the effect of PAs on patient crowding, compared with the effect of additional physicians in pediatric EDs.27-29 To address this gap, we used a at the University of Calgary in Calgary, Alberta. The authors have disclosed that this study received a local seeding grant from the Innovations in Acute Care and Technology cluster of the Child and Family Research Institute. Dr. Doan was also supported with a doctoral award by the Canadian Child Health Clinician Scientist Program while completing this study as part of her dissertation work. Acknowledgment: The authors would like to thank the research student volunteers who conducted the direct observations and data collection permitting the building of the DES model. DOI: 10.1097/01.JAA.0000451860.95151.e1 Copyright © 2014 American Academy of Physician Assistants

JAAPA Journal of the American Academy of Physician Assistants

Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

www.JAAPA.com

35

ORIGINAL RESEARCH

DES model of our pediatric ED to compare the effect on key efficiency indicators of extending physician coverage versus adding PAs with equivalent incremental costs. DES allows for modeling of discontinuous systems, and can emulate actual events in healthcare systems. It employs statistical calculations and theoretical distributions to deliver outcomes of a particular process.25,30-32 Once built, a model can be used to test the effect of changing one or a series of variables and to determine the effect of these changes on outcomes. In the case of ED modeling, outcomes can include patient wait times, leaving without being seen rates, and LOS. Using DES, we sought to forecast the differential effect of extending physician coverage versus introducing PAs on measures of flow in a Canadian pediatric ED. MATERIAL AND METHODS System description and conceptual model building A DES model was built based on the pediatric ED at British Columbia Children’s Hospital (BCCH). BCCH is the only tertiary care pediatric ED in British Columbia, with about 40,000 patient visits per year. The BCCH pediatric ED functions as a two-track system using the Canadian Triage and Acuity Scale (CTAS). Patients triaged to CTAS levels 1 to 3 are seen in the high-acuity track; those triaged to CTAS levels 4 and 5 are managed in the low-acuity track, commonly called “fast track.” Separate teams, composed of an ED pediatrician and physician trainees, staff each track. The low-acuity track is open 14 hours per day (late morning to late evening); the high-acuity track is open 24 hours a day. In addition, a third ED physician (flow physician) is present in the evening 7 days a week during the busy season (October to May, winter-spring), and Friday to Sunday the rest of the year (June to September, summerfall). The flow physician manages patients in either track depending on the longest queue, and does not have clinical teaching responsibility for trainees during his or her shift. Data collection Information on system layout and operating procedures was collected through consultation with pediatric ED staff and direct observation of patient flow from registration to discharge. The system was reviewed with the pediatric ED chief physician and nurse manager to ensure accurate representation of patients’ courses through the pediatric ED. Additional information such as distribution of patient triage levels and leaving without being seen rates was obtained from the Provincial Health Services Authority (PHSA) support services. Research assistants carried out direct observation using stopwatches to time all the processes in a patient visit, from pretriage to discharge. Observations were conducted during a sample of days, evenings, and nights, across all days of the week and during all four seasons from July 2010 to June 2011. More than 1,000 hours of data were collected, with hundreds of observations for integral data points. For each process observed, timed durations were fitted to theoretical probability distributions using the input analy36

www.JAAPA.com

sis function on ARENA 10.0 Simulation Software (Rockwell Software, Sewickley, Pa.). The frequency and duration of other physician activities were recorded in a similar fashion to the direct patient care processes. Other physician activities included bedside teaching, discussions with nursing staff, consultations with other physicians, and breaks. These were labeled as “personal activities” (breaks and nonpediatric ED patient-related discussions or paperwork), or “pediatric ED operations and patient care-related activities.” Patient assessment, review with physician, and reassessment times for physician trainees were collected during direct observation shifts. The number of trainees present on any given shift was variable and a sample of the 2010 trainee pediatric ED schedule was used in the model. DES model building and validation The simulation model was built using ARENA 10.0 simulation software. Theoretical time distributions based on aforementioned data for each observed process were used to populate the model. The model was verified by examination of its output for reasonableness, and by using stress tests to ensure that all modules and software functions were adequate. Stress testing consists of examining how the model design responded to extraordinary conditions. If a model is adequately designed, it will respond as anticipated even under unrealistically extreme conditions, and the simulation would terminate due to software capacity limits, rather than design malfunction. The baseline simulation model (without additional physician coverage or PAs) was validated by comparing the average total annual visit numbers, waiting time, LOS, and leaving without being seen rates from five iterations of one simulated year to those from 2010 administrative data. Pediatric ED visits triaged to CTAS 1 are infrequent, and the priority is to assess, resuscitate, and stabilize the child. Because of the emphasis on management, collected time stamps for administrative purposes often are entered after clinical stabilization and often are inaccurate. We therefore opted a priori to validate the model comparing outcome measure summaries of only CTAS 2 to 5. Scenario testing and outcome measures The first scenario (additional ED pediatrician coverage) was simulated by adding 1,450 physician hours/year to the physician schedule in the baseline model. These additional hours would cost about $307,000 Canadian (about $279,000 US) and represent the additional coverage suggested by the hospital administration for volume surges. These additional hours were added into the model to provide an extra 6-hour shift each day during the busiest periods of the year (winterspring), and busiest hours of the day (afternoon and evening). The second scenario (adding PAs to the clinical team) simulated adding a provider who could assess low-acuity patients, but had to review cases with physicians. This conservative approach reflects the results of two previous studies involving PAs and Canadian physicians practicing Volume 27 • Number 8 • August 2014

Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

Forecasting the effect of physician assistants in a pediatric ED

in Canadian pediatric EDs. These surveys reported participants’ selection of conditions thought to be appropriate for PAs to manage either with direct physician supervision or independently.22,23 More than 80% of physicians and PAs identified three clinical conditions (ear pain, upper respiratory infections, and insect bites) that accounted for 21% of all CTAS 4-5 visits (about 5,900 patients) that did not need direct physician support. Based on these data, our model required PAs to review 79% of randomly assigned cases with a physician. The same budget ($307,000 Canadian) would pay for 5,750 hours of PA time (rate obtained through personal communications with the past president of the Canadian PA Association). These PA hours were added to the baseline model over the whole year, but favored the busiest TABLE 1.

days of the year and busiest hours of the day. In this scenario, the model allowed the physician assigned to the low-acuity track to assess patients in either the low- or high-acuity track depending on where the queue of patients was longer, and return if a patient needed to be reviewed with a trainee or a PA in the low-acuity track. Because PAs are not licensed to work in British Columbia, their service times (assessing patients and reviewing cases with physicians) could not be observed directly. A comprehensive literature review also failed to identify service times for PAs practicing in pediatric EDs. Given this absence of quantified service times, we talked to PAs working in Manitoba and Washington state EDs, who stated that the duration of PA assessments was comparable to that of assessments performed by physicians. This asser-

Model validation results Baseline model output

CTAS

Mean n visits/ year (rounded)

Mean n visits during winterspring (rounded)

Administrative data Mean n visits during summerfall (rounded)

Mean n visits/year

Mean n visits during winter-spring

Mean n visits during summer-fall

1

252

110

141

252

113

139

2

4,688

2,424

2,264

4,646

2,402

2,244

3

13,523

7,426

6,096

13,525

7,416

6,109

4

19,074

10,218

8,856

19,056

10,188

8,868

5

2,364

1,232

1,131

2,362

1,237

1,125

CTAS

LOS min

1

226.7

235.2

219.6

237.8

282.9

201.1

2

315.4

322.9

307.2

315.0

321.6

307.9

3

259.2

268.1

248.4

249.2

257.8

238.8

4

176.2

182.1

169.4

170.2

176.4

163.0

5

154.8

164.4

144.3

144.9

148.7

140.6

CTAS

Waiting time

Waiting time winter-spring

Waiting time summer-fall

Waiting time

Waiting time winter-spring

Waiting time summer-fall

LOS winter-spring LOS summer-fall

LOS min

LOS winter-spring LOS summer-fall

1

11.8

13.5

10.4

52.4

57.3

47.9

2

66.9

72.4

61.1

70.5

74.1

66.6

3

114.5

122.6

104.7

114.6

123.2

104.3

4

107.6

114.2

99.9

106.0

113.1

98.1

5

103.9

114.1

92.6

103.9

110.5

97.0

CTAS

n leaving without being seen

Proportion leaving without being seen

n leaving without being seen

Proportion leaving without being seen

1

0

0

0

0

2

0

0

0

0

3

168

1.2%

166

1.2%

4

592

3.1%

594

3.1%

5

355

15.0%

369

15.6%

JAAPA Journal of the American Academy of Physician Assistants

Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

www.JAAPA.com

37

ORIGINAL RESEARCH

tion was supported by data from a review of an urgent care clinic.33 We therefore assumed that time distributions for PA assessments in the model were similar to physician assessment times collected during direct observation. The time required for a PA to review a case with a physician was estimated to be similar to the time required to review a junior trainee’s work. Thus, the same time distribution used for case review among trainees in the baseline model was also used for case review by PAs in the second scenario. Sensitivity analyses We ran modified versions of the scenario with PAs to explore the effect of varying five factors: • The physician assigned to the low-acuity track remained there regardless of the length of queue. • The proportion of PA cases requiring review by a physician varied between 10% and 100%. • PA double coverage was simulated in the evenings (the busiest time for the pediatric ED) by overlapping shifts; in the original second scenario, a maximum of one PA was assigned to be on shift at any given time. • The number of patients a PA could assess before having to review with a physician was increased from zero to two. • PA initial assessment times were increased from being the same distribution as ED pediatricians to mirror that of physician trainees. TABLE 2.

CTAS

RESULTS Model verification and validation The pediatric ED baseline model responded to stress testing adequately, and model integrity was maintained. The output summaries mirrored those acquired from the administrative database with the exception of outcome measures from CTAS 1 level patients, which were assumed to be inaccurate in the administrative database as outlined previously. We also stratified the baseline model outputs across the two main seasonal periods (winter-spring and summer-fall), and found that the leaving without being seen rates, waiting time, and LOS remained comparable between the model output and the administrative data (Table 1). Strategy scenario testing Outputs from the first scenario model with extended flow physician coverage are shown in Table 2. Extending the flow physician coverage achieved

Model outcome measures summary and differences from the baseline model for first alternative scenario model (extended flow physician coverage) Proportion leaving without being seen (95% CI)

Absolute difference in % leaving without being seen (95% CI in difference)

Relative (%) reduction

1

0

2

0

3

0.5% (0.3%, 0.7%)

0.8% (0.5%, 1.0%)

60%

4

2.2% (1.9%, 2.4%)

1% (0.7%, 1.3%)

31%

5

11% (9.53%, 12.47%)

5% (2.4%, 7.5%)

27%

CTAS

Overall LOS min (95% CI)

1

219 (213.2, 224.9)

2

Absolute reduction in LOS in min (95% CI)

Relative (%) reduction LOS

7.7 (2.4, 12.9)

3.4%

293.5 (290.5, 296.4)

21.9 (19.8, 24)

6.9%

3

216.4 (211.7, 221.1)

42.8 (39.5, 46.2)

16.5%

4

154.6 (151, 158.2)

21.5 (19, 24.1)

12.2%

5

133.4 (129, 137.8)

21.4 (18.3, 24.5)

13.8%

CTAS

38

Outcome analyses Each modified scenario model was run for a simulated year, and the average output of five iterations was compared against the baseline model. Descriptive statistics were used to report absolute differences in mean waiting time, LOS, and leaving without being seen rates between each of the scenarios and the baseline model stratified by pediatric CTAS levels.

Overall waiting time (95% CI)

Absolute reduction in waiting time in min (95% CI)

Relative (%) reduction waiting time

1

8.7 (7.9, 9.6)

3 (2.4, 3.7)

25.8%

2

49.9 (47.7, 52.2)

17 (15.5, 18.5)

25.4%

3

72.6 (68.4, 76.8)

41.9 (39.2, 44.6)

36.4%

4

86.8 (83.5, 90.1)

20.8 (18.7, 22.9)

19.3%

5

82.7 (78.4, 87.1)

21.2 (18.4, 24)

20.4%

www.JAAPA.com

Volume 27 • Number 8 • August 2014

Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

Forecasting the effect of physician assistants in a pediatric ED

significant reduction in waiting time during the winterspring season across all five CTAS levels, with proportionate reduction in LOS during the winter-spring season. No significant changes to the outcome measures were found during the summer-fall season. However, the effect of adding physician coverage during the winter-spring was large enough to affect the annual mean waiting time and LOS in most CTAS levels, and resulted in a reduction of leaving without being seen rates (60% reduction for CTAS 3 level visits, and a 30% reduction for CTAS 4-5 levels visits). Outputs from the second scenario model are shown in Table 3. This model simulated the introduction of a PA to the low-acuity track with physician review of 79% of the PA caseload, in accordance with the estimate given in the physician survey, and sharing of the low-acuity physician with the high-acuity track. This model resulted in meaningful waiting time reductions for patients in CTAS 1 to 3. The waiting time reduction observed for CTAS 4 and 5 was, however, clinically negligible and LOS for CTAS 4 and 5 patients increased slightly. The addition of PAs to the pediatric ED also reduced leaving without being seen rates by 84.4% for CTAS 3 patients, by 18.2% for CTAS 4 patients, and by 6.2% for CTAS 5 patients. Sensitivity analyses By preventing the low-acuity physiTABLE 3.

CTAS

cian from floating between the two tracks, moderate improvement was realized in waiting time outcomes for patients in the low-acuity track and none in the highacuity track. Increasing PA autonomy resulted in consistent and appreciable reductions in mean waiting time, LOS, and leaving without being seen rates for patients in CTAS 4 and 5. When the proportion of cases to be reviewed reached 50%, the PA model would result in a reduction in waiting time for CTAS 4 and 5 comparable to the outcomes of the physician model—a 20% to 35% reduction, while maintaining superior flow improvements for patients in CTAS 1 to 3 as compared with those obtained in the extended flow physician model (Figure 1). Other sensitivity analyses varied the PA schedules to overlap their shifts, changed the PAs’ assessment times to mirror that of trainees, and allowed PAs to assess and accumulate up to two cases before having to review and reassess with a physician. These manipulations did not affect the outcome measures significantly when compared with the original PA model. DISCUSSION The rationale for evaluating the use of PAs in the pediatric ED is to explore models to decrease wait times where physi-

Model outcome measures summary and differences from the baseline model for second alternative scenario model (PA in low-acuity pediatric ED track) Proportion leaving without being seen (95% CI)

Absolute difference in % leaving without being seen (95% CI in difference)

Relative (%) reduction

1

0

2

0

3

0.2% (0.02%, 0.38%)

1.1% (0.9%, 1.3%)

84.4%

4

2.5% (2.28%, 2.8%)

0.6% (0.2%, 0.9%)

18.2%

5

14.1% (12.7%, 13.6%)

0.9% (-1.2%, 2.9%)

6.6%

CTAS

Overall LOS min (95% CI)

Absolute reduction in LOS in min (95% CI)

Relative (%) reduction LOS

1

235.3 (220.5, 250)

-8.6 (-18.1, 0.9)

-3.8%

2

282.7 (280.8, 284.6)

32.6 (31.1, 34.2)

10.4%

3

200.2 (196.8, 203.7)

59.0 (-56.2, 61.9)

22.8%

4

183.8 (182.1, 185.5)

-7.7 (-9.4, -5.9)

-4.4%

5

169.1 (167.9, 170.2)

-14.2 (-11.3, -17.2)

-9.2%

CTAS

Overall waiting time (95% CI)

Absolute reduction in waiting time in min (95% CI)

Relative (%) reduction waiting time

1

7.2 (6.2, 8.2)

4.6 (3.9, 5.3)

38.8%

2

42.2 (40.8, 43.6)

24.7 (23.6, 25.8)

36.9%

3

55.8 (52.1, 59.5)

58.7 (56.3, 61.1)

51.3%

4

99.4 (97.6, 101.2)

8.2 (6.8, 9.6)

7.6%

5

100.9 (98.4, 103.4)

3 (1.1, 4.9)

2.9%

JAAPA Journal of the American Academy of Physician Assistants

Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

www.JAAPA.com

39

Relative reduction in waiting time

ORIGINAL RESEARCH

70% 60% 50%

CTAS 1

40%

CTAS 2

30%

CTAS 3

20%

CTAS 4

10%

CTAS 5

0% -10%

90

70

50

30

10

% PA caseload requiring review and reassessment by the physician FIGURE 1.

Effect of PA autonomy on relative reduction in waiting time

cians are ultimately responsible for care delivered, but are unable to cope with large volumes even when used at full capacity. This study found that additional physician coverage could reduce average waiting time in our tertiary pediatric ED by 20% to 35% across all CTAS level visits. The equivalent investment in PAs could reduce waiting time in a pediatric ED by 35% to 50% for CTAS 1 to 3 but leaves CTAS 4 and 5 unchanged. These figures may underestimate the benefits because they are based on limited autonomy that requires PAs to review 79% of their cases with a physician. If PA autonomy were to be increased, reducing the number of cases requiring review with the attending physician, patients triaged to CTAS 1 to 3 would retain the 35% to 50% reduction in waiting times, and CTAS 4 and 5 patients also would benefit from shorter waiting times. If PAs were to manage and discharge patients without directly involving a physician for half of their caseload, CTAS 4 and 5 patients would benefit from waiting time reductions comparable to those in the extended physician coverage model (that is, 20% to 35% reductions). Although PAs have worked in American EDs with higher autonomy levels, pediatric EDspecific data have yet to be published. We chose to run the simulation with very conservative conditions compared with how PAs may function in well-established roles because PAs are still relatively new to the Canadian civilian healthcare system. We were guided by the results of our two previous surveys of pediatric ED physicians about the degree of direct physician involvement with cases managed by PAs. Other studies have been undertaken to determine PAs’ effect on patient flow within EDs. In 2009, investigators in Ontario used a pre-post design to determine that the odds of achieving wait time benchmarks was 1.9 (95% CI 1.6-2.4) for patients visiting the ED after PA implementation. A similar study in the United States found that the 40

www.JAAPA.com

implementation of PAs in an urgent care unit reduced LOS for patients triaged to a low-acuity level from 127 to 53 minutes and improved patient satisfaction.34 Although these studies found improvements in patient flow after PA implementation, their usefulness is limited by short prepost study periods, failures to address confounders including patient characteristics, and a lack of controls. LIMITATIONS Despite using comprehensive observational data to build our computer simulation model, like any DES model, it cannot account for all of the factors that may affect the efficiency of care delivered in a pediatric ED. For example, whether patients who are seen by a PA have higher return rates than patients who are seen by a physician is an important consideration. However, this could not be observed directly and was not reflected in our model. Another limitation of our model was the lack of assessment and review times for PAs. PAs were not practicing in our pediatric ED; thus, no real service time data were collected. A literature search revealed no publications of PA service times in pediatric EDs, and we had to rely on qualitative data from discussions with practicing PAs to estimate their service times. This method of estimation and assumption is often necessary in computer simulation model construction when introducing entities that have no observable data associated with them.35 The inability to comprehensively compare costs also limited study findings. Although we explicitly kept human resource costs equal, this model did not address other costs associated with care provided in the pediatric ED. These costs can vary by patient CTAS level, LOS, frequency of visits, and ancillary testing (laboratory and radiography) ordered by care providers. Indeed, a systematic review Volume 27 • Number 8 • August 2014

Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

Forecasting the effect of physician assistants in a pediatric ED

found only two studies comparing the overall cost of care delivered by PAs and physicians in general EDs; the studies yielded conflicting results.36 Comprehensive cost comparisons between PAs and emergency pediatricians would also require broader consideration of the differential costs involved in the education of a PA versus a pediatric ED subspecialist. Finally, the absolute generalizability of these findings is limited to pediatric EDs with similar system structures and service times as our pediatric ED. CONCLUSION Our simulations of a pediatric ED found important reduction in waiting time, LOS, and leaving without being seen rates for both scenarios. Further studies and data collection postimplementation are needed to provide definitive recommendations, but our study strongly suggests that using PAs reduces waiting time, LOS, and patients leaving without being seen, and could do so at a reduced cost. JAAPA REFERENCES 1. Rowe B, Bond K, Ospina M, et al. Frequency, determinants and impact of overcrowding in emergency departments in Canada: a national survey of emergency department directors [Technology report No. 67.3]. Ottawa, ON: Canadian Agency for Drugs and Technologies in Health; 2006. http://www.cadth.ca/en/products/health-technology-assessment/publication/621. Accessed April 23, 2014. 2. Derlet RW, Richards JR. Overcrowding in the nation’s emergency departments: complex causes and disturbing effects. Ann Emerg Med. 2000;35(1):63-68. 3. Timm NL, Ho ML, Luria JW. Pediatric emergency department overcrowding and impact on patient flow outcomes. Acad Emerg Med. 2008;15(9):832-837. 4. Kennebeck SS, Timm NL, Kurowski EM, et al. The association of emergency department crowding and time to antibiotics in febrile neonates. Acad Emerg Med. 2011;18(12):1380-1385. 5. Shenoi R, Ma L, Syblik D, Yusuf S. Emergency department crowding and analgesic delay in pediatric sickle cell pain crises. Pediatr Emerg Care. 2011;27(10):911-917. 6. Sills MR, Fairclough D, Ranade D, Kahn MG. Emergency department crowding is associated with decreased quality of care for children. Pediatr Emerg Care. 2011;27(9):837-845. 7. Ospina MB, Bond K, Schull M, et al. Key indicators of overcrowding in Canadian emergency departments: a Delphi study. CJEM. 2007;9(5):339-346. 8. Stang AS, McGillivray D, Bhatt M, et al. Markers of overcrowding in a pediatric emergency department. Acad Emerg Med. 2010;17(2):151-156. 9. Goldman RD, Macpherson A, Schuh S, et al. Patients who leave the pediatric emergency department without being seen: a case-control study. CMAJ. 2005;172(1):39-43. 10. Ng Y, Lewena S. Leaving the paediatric emergency department without being seen: understanding the patient and the risks. J Paediatr Child Health. 2012;48(1):10-15. 11. Gaucher N, Bailey B, Gravel J. Who are the children leaving the emergency department without being seen by a physician? Acad Emerg Med. 2011;18(2):152-157. 12. Gravel J, Gouin S, Carrière B, et al. Unfavourable outcome for children leaving the emergency department without being seen by a physician. CJEM. 2013;15(5):289-299. 13. Ehrlich NJ, Tasmin F, Safi H, et al. Pilot study of ER utilization at Tulsa hospitals. J Okla State Med Assoc. 2004;97(2):64-68. 14. Pileggi C, Raffaele G, Angelillo IF. Paediatric utilization of an emergency department in Italy. Eur J Public Health. 2006;16(5): 565-569.

15. Simon HK, Hirsh DA, Rogers AJ, et al. Pediatric emergency department overcrowding: electronic medical record for identification of frequent, lower acuity visitors. Can we effectively identify patients for enhanced resource utilization? J Emerg Med. 2009;36(3):311-316. 16. Sinclair D. Emergency department overcrowding—implications for paediatric emergency medicine. Paediatr Child Health. 2007; 12(6):491-494. 17. Holley JE, Kellermann AL, Andrulis DP. Physician staffing in the emergency departments of public teaching hospitals: a national survey. Ann Emerg Med. 1992;21(1):53-57. 18. Isaacman DJ, Davis HW. Pediatric emergency medicine: state of the art. Pediatrics. 1993;91(3):587-590. 19. Casey MM, Wholey D, Moscovice IS. Rural emergency department staffing and participation in emergency certification and training programs. J Rural Health. 2008;24(3):253-262. 20. Counselman FL, Graffeo CA, Hill JT. Patient satisfaction with physician assistants (PAs) in an ED fast track. Am J Emerg Med. 2000;18(6):661-665. 21. Nyberg SM, Waswick W, Wynn T, Keuter K. Midlevel providers in a Level I trauma service: experience at Wesley Medical Center. J Trauma. 2007;63(1):128-134. 22. Doan Q, Sabhaney V, Kissoon N, et al. The role of physician assistants in a pediatric emergency department: a center review and survey. Pediatr Emerg Care. 2012;28(8):783-788. 23. Doan Q, Piteau S, Sheps S, et al. The role of physician assistants in pediatric emergency medicine: the physician’s view. CJEM. 2013;15(6):321-329. 24. Ducharme J, Alder RJ, Pelletier C, et al. The impact on patient flow after the integration of nurse practitioners and physician assistants in 6 Ontario emergency departments. CJEM. 2009; 11(5):455-461. 25. Gunal M, Pidd M. Discrete event simulation for performance modelling in health care: a review of the literature. J Simul. 2010;4:42-51. 26. Hung GR, Whitehouse SR, O’Neill C, et al. Computer modeling of patient flow in a pediatric emergency department using discrete event simulation. Pediatr Emerg Care. 2007;23(1):5-10. 27. Coats TJ, Michalis S. Mathematical modelling of patients flow through an accident and emergency department. Emerg Med J. 2001;18(3):190-192. 28. Duguay C, Chetouane F. Modeling and improving emergency department systems using discrete event simulation. Simulation. 2007;83(4):311-320. http://sim.sagepub.com/cgi/content/ abstract/83/4/311. Accessed April 23, 2014. 29. White CR, Best JB, Sage CK. Simulation of emergency medical service scheduling. Hosp Top. 1992;70(2):34-37. 30. Babulak E, Wang M. Discrete event simulations: state of the art. In: Goti A, ed. Discrete Event Simulations; 2010:1-10. http:// www.intechopen.com/books/discrete-event-simulations. Accessed April 23, 2014. 31. Jun J, Jacobson S, Swisher J. Application of discrete-event simulation in health care clinics: a survey. J Oper Res Soc. 1999; 50(2):109-123. 32. Jacobson SH, Hall SN, Swisher JR. Discrete event simulation of health care systems. In: Hall R, ed. Patient Flow: Reducing Delay in Healthcare Delivery. 2nd ed. New York, NY: Springer; 2013:273-310. 33. Arnopolin SL, Smithline HA. Patient care by physician assistants and by physicians in an emergency department. JAAPA. 2000;13 (12):39-40. 34. Rodi SW, Grau MV, Orsini CM. Evaluation of a fast track unit: alignment of resources and demand results in improved satisfaction and decreased length of stay for emergency department patients. Qual Manag Health Care. 2006;15(3):163-170. 35. Fishman GS. Discrete-Event Simulation: Modeling, Programming, and Analysis. New York, NY: Springer-Verlag; 2001. 36. Doan Q, Sabhaney V, Kissoon N, et al. A systematic review: the role and impact of the physician assistant in the emergency department. Emerg Med Australas. 2011;23(1):7-15.

JAAPA Journal of the American Academy of Physician Assistants

Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

www.JAAPA.com

41

Forecasting the effect of physician assistants in a pediatric ED.

Most pediatric ED visits are for nonemergent problems. Physician assistants are well trained to manage these patients; however, their effect on patien...
142KB Sizes 2 Downloads 3 Views