• 
• 

Three Part Question

In [a busy adult or mixed (adult and paediatric) Emergency Department], does [having a physician at triage] [improve length of stay (LOS) and flow through department]?

Clinical Scenario

You are the emergency physician in charge of a busy Emergency Department (ED) in a large hospital. The department is experiencing overcrowding and you are investigating possible solutions. You wonder whether placing a senior physician at triage will help to improve flow through the department and reduce patient length of stay (LOS) and ED crowding.

Search Strategy

Medline using the following search strategy:
[exp EMERGENCY SERVICE, HOSPITAL/ (MeSH)] AND [exp TRIAGE/ (MeSH) or triag* (free text) OR “rapid assessment”] AND [doctor* OR consultant* OR intern* OR physician* OR registrar (free text)] AND [time* OR timing OR delay* OR wait* (free text)] AND [Limit to: English Language and Publication Year 1966-2015]

Cochrane Library using the following search strategy:
[[MeSH descriptor: [Emergency Service, Hospital] this term only] OR [MeSH descriptor: [Emergency Medical Services] explode all trees]] AND [[MeSH descriptor: [Triage] explode all trees] OR triage or "rapid assessment"] AND [doctor or consultant or intern or physician or registrar] AND [time or timing or delay or wait] AND [Limit to English Language]

Search Outcome

514 papers were found using the Medline search, of which 498 were irrelevant and 6 were of insufficient quality for inclusion.
55 papers were found using the Cochrane Library search, of which all were either irrelevant or previously found in the Medline search.
10 relevant papers were found using the Medline search and 1 additional study was identified from a systematic review. These are shown in the table below.

Relevant Paper(s)

Author, date and country	Patient group	Study type (level of evidence)	Outcomes	Key results	Study Weaknesses
Cheng I et al 2013 Canada	Non-critically ill ED patients, 8:00-14:30 during 131 weekday shifts across 26 weeks (65 days randomised to intervention and 66 days to control cluster).(n=6,300 pts, 750 seen by doctor). Academic tertiary level hospital.	Cluster randomised controlled trial	- Median ED LOS for patients discharged home from ED	- ED LOS decreased by 28min for high acuity patients (4:01hrs v 4:29hrs) (p=0.01) and 58min for low acuity patients (1:10hrs v 2:08hrs) (p<0.001) receiving intervention.	- There was an extra nurse at triage as well as extra EP, which may account for some benefits seen. - Evenings and weekends not studied. - Single-centre trial.
			- Median ED LOS for patients admitted from ED	- ED LOS for patients referred to other specialties/ admitted did not decrease significantly.
			- LWBS (left without being seen) rate	- LWBS rate decreased by 0.7% (1.5% v 2.2%).
			- Time to investigations (imaging, laboratory)	- Time to investigations all reduced.
Holroyd BR et al 2007 Canada	All adult ED patients, 11:00-20:00 during 42 days (21 control and 21 intervention shift days). (n=5,718 pts). Academic urban tertiary ED.	Randomised controlled trial	- ED LOS	- ED LOS decreased by 36min (4:21hrs v 4:57hrs).	- No power calculation. - Additional member of staff employed for intervention, therefore extra staffing may account for some benefits (rather than specific physician in triage role.) - Single-centre trial.
			- Proportion of patients who LWBS	- LWBS decreased from 6.6% to 5.4% (statistically non-significant).
Russ S et al 2009 USA	All adult ED patients. Physician at triage 11:00-23:00, 5-6 days per week. (n= 10,901 matched pairs of pts). Urban academic tertiary care Medical centre.	Retrospective data collection, comparing patients receiving physician orders at triage with matched control patients not receiving physician orders	- ED LOS	- Median ED LOS increased by 11min for those receiving triage orders.	- Comparison is between patients who do and don’t receive physician orders at triage, rather than all patients who are assessed by a physician at triage. - No randomisation of patients in intervention/control groups - potential bias. - Single centre study.
			- Waiting room time	- Median waiting room time increased by 41min for patients receiving intervention.
			- Time spent in ED bed	- Median time in ED bed decreased by 37min for patients receiving intervention.
Soremekun OA et al 2012 USA	All medium acuity adult ED patients who were subsequently admitted, 11:00-23:00 over 2 year period (12 month pre- and 12 month post-intervention). (n=20,312 pts). Large urban teaching hospital.	Retrospective study - 12 months pre- and post-introduction of physician at triage	- ED LOS	- ED LOS decreased by 13 min (p=0.001).	- Retrospective design. - No randomisation. - Increase in admissions by 14% between the two study periods - Used physician order time on computer which was probably later than time seen, possibly underestimating the effect. - Single centre study.
			- Time to disposition decision	- Median time to disposition decision decreased by 6min (260min v 254min) (p=0.025).
			- Time to physician evaluation and time to physician orders for analgesia, antiemetic, antibiotics, and radiology	- Median times to physician orders all decreased by 16 - 70 min
Jarvis et al 2014 UK	All ED patients (adult and paediatric) between 09:00-17:00, excluding minor injuries, during control phase (54 days) and intervention phase (19 days). (n=4,622 pts). District general hospital ED (major trauma unit).	Prospective, non-randomized observational study	- Time to ‘emergency department ready’ (ED management complete)	- Median time to ‘emergency department ready’ reduced significantly by 53mins (129mins v 76mins) (P<0.0001)	- Point of care testing was implemented alongside rapid physician assessment at triage, therefore cannot analyse impact of doctor at triage alone. - Triage consultant was additional member of staff. - Control and intervention phases were during different seasons – other factors may have affected outcome measures. - Single-centre study.
			- Time to first clinical assessment (Dr or nurse)	- Median time to first clinical assessment reduced by 8mins (12mins v 4 mins) (P<0.0001)
Imperato J et al 2012 USA	All ED patients (adult and paediatric), 13:00-21:00 during 6 months (3 month control and 3 month intervention period). (n=17,631 pts). Small community teaching hospital.	Retrospective study – 3 months pre- and post-introduction of physician at triage	- ED LOS	- Overall median ED LOS reduced by 13 min (3:48hrs to 3:35hrs) (p <0.001). This reduction was greater for admitted patients than for discharged ones (24min vs 7min).	- Retrospective study. - PIT (Physician in Triage) only for 8hr per day (13:00-21:00), whilst data from 24hrs included in results. - Average daily volume was higher during the intervention period and also less experienced staff were used during the intervention, both of which may reduce the effect. - Additional doctor used during intervention – results possibly due to increased staffing. - Single centre study.
			LWBS rate	LWBS rate decreased from 132 to 121 pts (non-significant).
Han JH et al 2010 USA	All ED patients, during 18 week period (9 weeks pre- and 9 weeks post-intervention). Physician triage 13:00-21:00. (n=17,265 pts). Urban academic tertiary care and Level I trauma centre.	Retrospective study - 9 weeks pre- and post- introduction of physician at triage	- ED LOS for admitted and discharged patients	- Median ED LOS overall decreased by 11min for whole cohort, but no significant difference in ED LOS for admitted patients, due to exit block.	- Retrospective data collection. - Triage physician was additional member of staff therefore extra staffing may account for some benefits seen. - Data collected for patients in 24hr periods whilst intervention for only 8hrs per day –Short study period. - Single-centre study.
			- LWBS rate	- LWBS decreased from 4.5% to 2.5%.
Rogg JG et al 2013 USA	All medium-acuity ED patients (adult and paediatric), during 1 year pre- and 3 years post- introduction of START (Supplemented Triage and Rapid Treatment). (n= 180,871 pts (39,142 pre- and 141,729 post-intervention)). Large, urban, academic tertiary care ED and Level 1 trauma centre.	Large retrospective, observational, before-and-after study	- ED LOS - for patients eligible for START intervention	- Median ED LOS decreased by 56min for pts eligible for START.	- Retrospective study. - Changes in ED data across study period – eg. annual ED volume increased by 12% from pre- to post-intervention and increase in nurses. - Single-centre study. - No blinding or randomisation.
			- ED LOS for all ED patients	- ED LOS for all pts (including low/high acuity) decreased by 30mins.
			- LWBS rate	- LWBS rate decreased from 4.8% to 2.9%.
Gray et al 2009 Canada	All ED patients, during 4hr afternoon shifts across 6 weeks (3 control and 3 intervention weeks). (n=5,020 pts). Large urban ED.	Clinical controlled trial	- ED LOS	- ED LOS did not reduce significantly (396min v 409min) (p=0.32).	- Only had physician in triage for 4hr shifts. - Single centre study.
			- LWBS rate	- LWBS rate decreased from 6.3% to 5.7%.
			- Time to see physician	- Mean time to see physician did not reduce (111min v 112min) (p=0.61).
Asha et al 2013 Australia	Ambulant ED pts age >16yrs who were selected by triage nurse as appropriate for consultant assessment, arriving 12:00-18:00 Fri-Mon across 3 months (intervention group). Control group – all ED pts 12:00-18:00 Tues-Thurs in same 3 months. (n=18,962 pts). ED of tertiary referral centre.	Prospective, non-randomised cohort study	- Proportion of pts meeting 4hr target in different subcategories (per whole day, between 12:00-18:00, in admitted pts, in discharged pts)	- Odds of pts meeting 4hr target, when controlled for confounding effects, was 15% higher on intervention days (P < 0.001).	- Pts were initially triaged by a nurse, and then those selected were sent to a SAS (Senior Assessment and Streaming) team including consultant, nurse and intern. Therefore results not strictly/only based on physician at triage. - No randomisation, and intervention was carried out on peak days only - Physician at triage was additional staff member - Single-centre study
			- LWBS rate	- LWBS rate improved by 0.34% (2.5% v 2.84%) (p=0.17); when controlled for confounding LWBS was 28% lower (P=0.003)
Partovi SN et al 2001 USA	All ED patients (adult and paediatric), 09:00-21:00 on 16 consecutive Mondays (8 intervention and 8 control shifts). (n=1,734 pts). Academic, urban hospital.	Prospective controlled trial – non- randomised	- ED LOS	- Mean ED LOS decreased by 82min (363min v 445min).	- Single-centre trial. - No randomisation or blinding used. - US health system very different to NHS.
			- LWBS rate	- LWBS rate decreased by 46% (7.9% v 4.7% patients).

Comment(s)

All studies found were single centre trials only. Only two adequately randomised controlled trials were found; both of these showed a reduction in ED LOS but one showed no difference in LOS for patients being admitted. The other trials often added an extra doctor to the department during their trial period, making it difficult to know if the effect on LOS was due to early senior triage or increased staffing. Whilst most studies report some reduction in flow measures such as LOS and LWBS (left without being seen) rate, the results were variable. There were also many potential problems with bias and confounding due to the retrospective before and after nature of many of the studies. In systems with significant exit block due to inpatient bed shortages, physician triage is expected to improve safety and quality by treating and investigating patients earlier but it may not consistently reduce LOS.

Clinical Bottom Line

Having a physician at triage may improve flow, LOS and other safety and quality measures, but further good evidence in the form of randomised controlled trials are needed to see if this is a consistent outcome.

References

1. Cheng I et al Implementing wait-time reductions under Ontario government benchmarks (Pay-for-Results): a Cluster Randomized Trial of the Effect of a Physician-Nurse Supplementary Triage Assistance team (MDRNSTAT) o BMC Emergency Medicine 2013, Nov 11;13:17
2. Holroyd BR et al Impact of a Triage Liaison Physician on Emergency Department Overcrowding and Throughput: A Randomized Controlled Trial Acad Emerg Med 2007, Aug;14(8):702-8
3. Russ S et al. Placing physician orders at triage: the effect on length of stay. Ann Emerg Med. 2010, Jul;56(1):27-33.
4. Soremekun OA et al. Impact of physician screening in the emergency department on patient flow. J Emerg Med. 2012 Sep;43(3):509-15
5. Jarvis PRE et al. Does rapid assessment shorten the amount of time patients spend in the emergency department? Br J Hosp Med (Lond.) 2014 Nov;75(11):648-51.
6. Imperato J et al. Physician in triage improves emergency department patient throughput. Intern Emerg Med 2012, Oct;7(5):457-62
7. Han JH et al. The effect of physician triage on emergency department length of stay. J Emerg Med 2010, Aug;39(2):227-33
8. Rogg JG et al. A Long-term Analysis of Physician Triage Screening in the Emergency Department. Acad Emerg Med. 2013 Apr;20(4):374-80.
9. Gray SH, Kingsley SJ, Spence JM. Does having an emergency physician at triage for a 4-hour shift reduce ED length of stay? [abstract] CJEM. 2009;11:272.
10. Asha SE and Ajami A. Improvement in emergency department length of stay using an early senior medical assessment and streaming model of care: A cohort study. Emerg Med Australas. 2013 Oct;25(5):445-51
11. Partovi SN et al. Faculty Triage Shortens Emergency Department Length of Stay. Acad Emerg Med. 2001, Oct;8(10):990-5