You are here

Article Text

Article menu

Download PDFPDF

Original research
A novel approach to improving emergency department consultant response times
  1. Christine Soong1,
  2. Sasha High1,
  3. Matthew W Morgan1,
  4. Howard Ovens2
  1. 1Division of General Internal Medicine, University of Toronto, Toronto, Ontario, Canada
  2. 2Department of Family and Community Medicine, University of Toronto, Toronto, Ontario, Canada
  1. Correspondence to Dr Christine Soong, Mount Sinai Hospital, 600 University Avenue, Room 428, Toronto, Ontario M5G 1X5, Canada; CSoong{at}mtsinai.on.ca

Abstract

Background Emergency department (ED) overcrowding is a threat to patient safety and public health. Availability of specialty consultation to the ED may contribute to overcrowding. We implemented a novel intervention using education, goal setting and real-time performance feedback to improve time to admission for patients referred to general internal medicine (GIM).

Methods Using a time-series design, we examined the effects of a quality improvement intervention on ED wait-times in an academic medical centre. The multifaceted approach included a didactic session for GIM housestaff on medicine triage principles and methods; setting a goal to have disposition decisions and, where appropriate, admission order within 4 h of consultation request; and providing personal data feedback on their performance on this metric to GIM housestaff during their rotation on the inpatient teaching service over a 1-year period. We compared time from consultation request to disposition decision and overall ED length of stay (LOS) for all patients referred to GIM during the intervention period (February 2011–February 2012) with data from the control period (January 2010–January 2011).

Results Mean time from GIM consultation request to admission order entry decreased by 92 min (SD, 5, p<0.05) from 321min in the control period to 229 min in the intervention period. Overall ED LOS for GIM patients decreased by 59 min (SD, 14, p<0.05) for admitted patients from 1022 min in the control period to 963 min in the intervention period, and by 40 min (SD, 13, p<0.05) for all patients referred to GIM. GIM staffing and patient characteristics remained stable across the two periods.

Discussion ED throughput for admitted medical patients improved with a quality improvement initiative involving education, goal setting and performance feedback.

  • Emergency Department
  • Healthcare Quality Improvement
  • Medical Education

https://doi.org/10.1136/bmjqs-2012-001503

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Introduction

    Increased emergency department (ED) wait-times are associated with increased mortality, low patient satisfaction and poor access to services.1–4 In the publicly funded, single-payer healthcare system in Ontario, legislation requires that all hospitals report ED wait-times as part of the ‘pay-for-results’ programme that provides financial incentives for meeting key performance indicators, including time to hospital admission.5 This, in part, has been stimulating interest in decreasing ED wait-times.

    Access to timely specialty consultation can be challenging and can contribute to long ED wait-times while patients wait to see a specialist before disposition.5 In Ontario, the focus has been to reduce time to ED decisions (to either admit or discharge the patient), improve access to inpatient beds, and better integrate care between hospital and community.6 Among the many strategies proposed is one of improving time to specialist consultation. This poses a unique challenge in academic medical centres where students and clinicians in training, who need supervision and require time to formulate plans, conduct initial consultations in the ED. The presence of trainees in the ED has been shown to be associated with increased ED length of stay (LOS).7 ,8

    Few studies have examined educational methods to improve efficiency and flow specifically in the ED.9 Given that a large proportion of inpatients in many hospitals are admitted to general internal medicine (GIM), and access to consultants is a well-documented barrier to patient flow in the ED,1–4 ,6 we chose to focus our quality improvement initiative on GIM consultation request to time to admission decision. We adopted an audit and feedback method for this project, according to results from a recent Cochrane Review that found these strategies had a positive effect on professional practice and outcomes for healthcare professionals.10 Given that in our institution, medical housestaff are the first responders to the ED for specialty consultation requests, we created a unique coaching and feedback intervention targeting GIM residents to reduce time to admission decision at a tertiary, academic medical centre.

    Methods

    This study took place at Mount Sinai Hospital, a 472-bed academic medical centre in Toronto, Ontario, Canada. The emergency department has 30 fixed stretcher locations and handles over 50 000 visits a year. The GIM ward has 84 beds and had 3048 discharges in 2011. The GIM residency programme has 41 residents. All GIM patients are admitted through the ED. GIM housestaff are first-line responders to consultation requests from the ED physician, and determine need for admission.

    We established a team in collaboration with the ED and GIM to improve the flow of admitted patients through the ED. Internal data review of patients admitted to GIM from the ED revealed that patients were waiting an average of greater than 5 h for a decision on admission after a consultation was requested. During the preintervention period, a member of the improvement team (CS) observed housestaff to understand workflow and response to ED requests. Different residents on call for ED GIM consultations were shadowed during the course of their call day. Shadowing occurred during weekdays, between 8:00  and 17:00. The residents were aware they were being observed but were not given details of the study or the purpose of the observation. We collected information on general workflow including time to response to ED consultation requests, and the nature of the response to the consultation. Specifically, we recorded what happened after a resident received a consultation request, how long it took to present to the ED to evaluate the patient, and when a disposition was determined. There was heterogeneity in the approach to ED consults and several factors contributed to delays in consultant response. We observed the following factors affecting timely response to ED consultation: low prioritisation of responses to ED consultations when housestaff were challenged with managing multiple clinical and teaching responsibilities, and serial processing whereby admission was deferred until a junior housestaff completed a full assessment. Informal discussions with housestaff after the observation period revealed root causes of delayed response to ED consultations included a belief that patients in the ED were already closely monitored in an acute care area and that they were ‘safe’ and actively managed, which led to prolonged delays in the ED prior to consultation evaluation and ultimately management. In addition, there was a cultural practice of waiting for a junior member of the team to complete a full history and physical before entering admission orders. Given these findings, the improvement team concluded that a multifaceted approach was needed with a focus on awareness of and education on harms of overcrowding and delays in management. The three main components of the intervention consisted of (1) educating GIM housestaff on rapid triage through a didactic education session, (2) setting goals and expectations for ED response and admission times and (3) providing feedback of personal admission times for each GIM resident every 2 weeks.

    Triage coaching intervention

    A hospitalist physician (CS or SH) performed triage coaching for the GIM housestaff. Medicine trainees attended a didactic session at the start of each rotation (every 2 months) on the inpatient service. This introduction provided rationale behind reducing ED wait-times and relevant patient safety concerns associated with ED overcrowding. Next, an overview of principles of rapid medicine triage was presented with examples of scenarios requiring admission to hospital. Existing data on time from consultation request to admission order entry was provided, and a 1 h target (up to 4 h if disposition was unclear) was established. The educational session lasted less than 20 min. The final component of the intervention involved providing trainees with personal performance feedback. Each senior admitting resident was given his/her consultation request to admission times every 2 weeks with comparative mean data of all GIM patients.

    Study design

    Using a time series design, we compared the relevant institution-level data from the intervention period (February 2011–February 2012) with data from the control period (January 2010–January 2011). Our institutional Research Ethics Board approved this study.

    Data sources

    Administrative personnel collected relevant data from institutional databases, including patient volume, demographics, GIM inpatient mortality and cardiac arrest rates, ED admission times and LOS, admission diagnoses and overall GIM LOS. Registered nurses in the ED electronically entered the consultation request time in the institution's internal database when the ED physician completed a request for consultation. Disposition decision is based on time of admission order entry or discharge from ED. All admission orders are entered electronically in a computer order entry system, which provided us with the time of admission. For admitted GIM patients, the time from consultation request to admission decision is calculated by subtracting the time of consultation request from the time of admission order entry. The hospital's internal data performance measurement office monitors data validity as part of the institution's quality improvement process.

    Outcome measures

    Time from GIM consultation request to admission order entry was the primary outcome of interest. We also assessed overall ED LOS for all patients referred to GIM including those not admitted, and overall GIM inpatient LOS, mortality and cardiac arrest rates. Qualitative feedback from ED and GIM providers occurred at monthly scheduled debriefing meetings.

    Statistical analysis

    Descriptive statistics were used to summarise characteristics of the patients and outcome measures. Continuous variables were compared using a t test, and categorical variables using a χ2 test. Statistical process control chart was used to monitor improvement over time. Control limits were set at 3 ς (ie, 3 SD) with eight consecutive points above or below the mean suggestive of special cause variation (equivalent to p<0.01).11 Total stretcher time saved in the ED was calculated by multiplying the mean reduction in time of ED LOS per patient by total number of patients referred to GIM in the intervention period. A p value<0.05 was considered statistically significant. All analyses were conducted using STATA 10.X (Stata Corporation, College Station, Texas, USA) and QI Macros SPC software for Excel 2012.07 (KnowWare International, Denver, Colorado, USA).

    Results

    The number of GIM referrals and admissions from the ED were similar during the two periods (table 1). Patient characteristics and admission diagnoses were also similar between the two periods. The numbers of ED and GIM providers (including residents) were unchanged across the two periods. As a part of this improvement initiative, there was an addition of a hospitalist triage ‘coach’; (role described in ‘Triage coach intervention’ section under Methods) to facilitate housestaff education around flow of patients admitted to GIM. No other interventions affecting ED GIM patient flow were conducted during this period.

    View this table:
    Table 1

    Characteristics and outcomes of general internal medicine referrals*

    There was a statistically significant decrease in the mean time from consultation request to admission order entry after the initiation of the quality improvement intervention (figure 1). This shift has been sustained over 13 months. The mean time from a request for GIM consultation to electronic admission order entry decreased by 92 min (29%; 321 min vs 229min; p<0.001). Overall ED LOS for all patients referred to GIM decreased by 59 min for admitted patients (6%; 1022 min vs 963 min; p<0.05) and by 79 min for patients discharged from the ED (19% reduction; 413 min vs 334 min; p<0.05). The mean ED LOS for all patients referred to GIM (admitted and discharged) decreased by 40 min (4%; 947 min vs 907 min; p<0.05). The reduction in consultation request to admission order entry accounts for the majority of reduction in overall ED LOS for admitted patients, and is shown in figure 2. Total stretcher time saved in the ED during the intervention period was 2246 h, or approximately 6 h per day.

    Figure 1
    Figure 1

    Statistical process control chart of mean time from consultation request to admission order entry.

    Figure 2
    Figure 2

    Mean time spent in emergency department  by interval for admitted general internal medicine  patients.

    Analysis of median time from consultation request to admission orders for individual residents prior to and following personal data feedback revealed that the lowest performing quartile (ie, the slowest group of residents before receiving feedback) had the greatest reduction (median reduction of 55 min) compared with the highest performing quartile which remained relatively stable (median gain of 10 min). The second and third quartiles had a median reduction of 13 and 10 min, respectively.

    No significant concerns or provider dissatisfaction were identified, although some residents indicated there was a perception of increased pressure to quickly determine admission status during the intervention period. Early feedback from ED staff identified an increase in ‘partial’ admission order entry, indicating some housestaff were able to initiate admission orders quicker but unable to complete the full orders necessary for admission. This behaviour was noted at the start of the intervention. The intervention was adapted to emphasise completion of the basic admission order set during the orientation session and audit feedback review resulted in subsequent reduction in incomplete admission order entry.

    Discussion

    Emergency department overcrowding is a threat to patient safety and public health.5 ,12–14 Availability of inpatient beds and access to specialty consultants have been cited as important causes of crowding.6 ,12 ,14 Our study has shown that a simple intervention involving goal-setting, education and performance feedback for GIM residents can result in a significant reduction in time to admission for patients referred to GIM, reduction in ED LOS, and improvement in ED resource availability (as measured by an increase in daily total stretcher time).

    To date, few studies have specifically examined consultant response and admission times as a solution to ED overcrowding, although this is recognised as an important factor in reducing wait-times.5 ,7 ,8 Howell and colleagues examined active bed management by hospitalists as an intervention providing dedicated ED consultant support that improved throughput and decreased ambulance diversion in a single institution.9 ,15 The reduction in total ED LOS was greater than our study at a relative reduction of 21% (98 min). Although the authors were able to show a cost savings associated with decreased ambulance diversion of more than US$1086 per hour with an active bed management system, this did not take into account the additional personnel costs. Staffing a 24 h, 7-days-a-week, active bed management service requires significant human resources and expense, which may be an important deterrent to implementation at other institutions. Similarly, Christmas and colleagues showed the presence of a senior consultant in the ED overnight decreased admission rates and EDLOS.16 While appealing solutions, these studies focused on increasing consultant staffing to the ED, a substantial investment in resources not typically available in many institutions. Other approaches to ED overcrowding include increasing ED personnel,17–21 hospital beds,22–24 patient diversion (eg, redirecting patients in the ED to rapid referral outpatient clinics or to other care facilities),25–28 and process improvement modelling (eg, simulation models examining processes to determine barriers to flow).28 ,29 There are practical implementation issues with such approaches, as considerable effort, facility engineering and resources are required for success.

    Our study targeted consultant response times to address one aspect of ED overcrowding. Without increasing staffing, we were able to improve admission times and modestly decrease the overall ED LOS for GIM inpatients. Each aspect of the intervention can be implemented without additional resources. Triage education may be incorporated into existing resident orientation, and performance feedback can often be generated for individual providers through institutional electronic medical record systems. Each component of the intervention was likely equally important in attaining the results, however, there was great interest among the trainees regarding feedback with personal performance data. In particular, housestaff were intrigued by their personal ‘admission times’ in comparison with their peers. In the current climate of performance measurement and peer-to-peer comparison, this intervention highlighted the utility of such measures in changing physician behaviour. Of note, the greatest impact appeared to be among the residents with the longest admission times given this group experienced the largest improvements after receiving personal feedback. The impact of the feedback was likely minimised in the ‘faster’ resident quartile because they likely already implemented necessary change in practice and were performing at ‘top speed’. Some residents reflected that there was increased pressure to make decisions quickly, but recognised that triage skills were important in preparing them for ‘real world’ practice in our current pay-for-performance environment. Our intervention is simple to implement without increased cost, and provides valuable practical skills to trainees. The strategy to affect behavioural change through education of quick triage assessments and admission determination, and the use of real-time performance data to reinforce and sustain change can be applied to a broad range of provider groups and practice situations. Due, in part, to this intervention, our institution has the lowest mean ED LOS for all admitted patients (5.3 h) in comparison with peer hospitals (range of 5.6–8.1 h).30

    Several limitations of this study should be considered. First, as the residents were not randomised to the intervention, potential confounders, such as patient acuity, or pressure from ED staff to determine disposition may have influenced outcomes. However, important factors, such as patient volumes and characteristics, bed capacity and admission diagnoses were not significantly different between the two study periods. Further, no other throughput interventions or staffing changes occurred on the GIM ED consultation service during this study that may have impacted the results. Using statistical process control chart rules, we were able to demonstrate a significant change in the intervention period unlikely due to normal cause variation. Second, this study took place at a single academic medical centre, and results may not be generalisable to other institutions, particularly ones where ED physicians determine admission status. However, the intervention may be applied to any provider group, regardless of who determines admission to hospital. Third, two different physicians performed triage coaching which may have resulted in subtle variations in education, although teaching material and performance data collection remained the same, thereby minimising any significant impact on learners. Fourth, quality improvement studies may be subject to the ‘Hawthorne effect’(31) where initial observed positive results wane with time; however, our results were maintained for 1-year postintervention. We plan to continue our efforts and measurement to determine if the effects are sustained. Fifth, while we experienced a reduction in wait-times, our post-intervention ED LOS may still be considered long in comparison with international standards, thereby limiting generalisability. Last, we were not able to capture all potential unintended consequence of the intervention; however, patient safety event reporting volumes remained stable throughout the study.

    In summary, a simple, low-cost intervention focused on goal setting, education and performance feedback can alter physician behaviour, and improve consultant response times to the ED, resulting in positive effects on throughput and overcrowding.

    Acknowledgments

    The authors would like to thank Hapiloe Byaruhanga and Joanne Bon in the Office of Quality and Performance Measurement for their assistance in data acquisition.

    References

    1. ↵
      1. Sun BC,
      2. Mohanty SA,
      3. Weiss R,
      4. et al
      . Effects of hospital closures and hospital characteristics on emergency department ambulance diversion, Los Angeles County, 1998 to 2004. Ann Emerg Med 2006;47:309–16.
      OpenUrlCrossRefPubMedWeb of Science
    2. ↵
      1. Miró O,
      2. Antonio MT,
      3. Jiménez S,
      4. et al
      . Decreased health care quality associated with emergency department overcrowding. Eur J Emerg Med 1999;6:105–7.
      OpenUrlCrossRefPubMed
    3. ↵
      1. Richardson DB
      . Increase in patient mortality at 10 days associated with emergency department overcrowding. Med J Aust 2006;184:213–16.
      OpenUrlPubMedWeb of Science
    4. ↵
      1. Redelmeier DA,
      2. Blair PJ,
      3. Collins WE
      . No place to unload: a preliminary analysis of the prevalence, risk factors, and consequences of ambulance diversion. Ann Emerg Med 1994;23:43–7.
      OpenUrlCrossRefPubMedWeb of Science
    5. ↵
      Office of the Auditor General of Ontario—2010 Annual Report (Internet). auditor.on.ca. (cited accessed 12 July 2012). http://www.auditor.on.ca/en/reports_2010_en.htm
    6. ↵
      Improving_access.pdf (Internet). health.gov.on.ca. (cited accessed 12 July 2012). http://www.health.gov.on.ca/en/common/ministry/publications/reports/improving_access/improving_access.pdf
    7. ↵
      1. Jamtvedt G,
      2. Harper M,
      3. Johnston P,
      4. et al
      . Effect of trainees on length of stay in the pediatric emergency department. Acad Emerg Med 2009;16:859–65.
      OpenUrlCrossRefPubMedWeb of Science
    8. ↵
      1. Lammers RL,
      2. Roiger M,
      3. Rice L,
      4. et al
      . The effect of a new emergency medicine residency program on patient length of stay in a community hospital emergency department. Acad Emerg Med 2003;10:725–30.
      OpenUrlCrossRefPubMedWeb of Science
    9. ↵
      1. Venugopal R,
      2. Lang E,
      3. Doyle K,
      4. et al
      . A workshop to improve workflow efficiency in emergency medicine. CJEM 2008;10:525–31.
      OpenUrlPubMed
    10. ↵
      1. Ivers N,
      2. Jamtvedt G,
      3. Flottorp S,
      4. et al
      . Audit and feedback: effects on professional practice and healthcare outcomes. Cochrane Database Syst Rev 2012;(7):1–229.
    11. ↵
      1. Benneyan JC,
      2. Lloyd RC,
      3. Plsek PE
      . Statistical process control as a tool for research and healthcare improvement. Qual Saf Healthcare 2003;12:458–64.
      OpenUrlCrossRef
    12. ↵
      IOM. Future of emergency care—Hospital-based emergency care: at the breaking point. Washington DC: National Academies Press, IOM, 2006.  1–8.
    13. ↵
      1. Plunkett PK,
      2. Byrne DG,
      3. Breslin T,
      4. et al
      . Increasing wait times predict increasing mortality for emergency medical admissions. Eur J Emerg Med 2011;18:192–6.
      OpenUrlCrossRefPubMed
    14. ↵
      1. Bond K,
      2. Ospina MB,
      3. Blitz S,
      4. et al
      . Frequency, determinants and impact of overcrowding in emergency departments in Canada: a national survey. Healthc Q 2007;10:32–40.
      OpenUrlPubMed
    15. ↵
      1. Howell E,
      2. Bessman E,
      3. Kravet S,
      4. et al
      . Active bed management by hospitalists and emergency department throughput. Ann Intern Med 2008;149:804–11.
      OpenUrlCrossRefPubMedWeb of Science
    16. ↵
      1. Christmas E,
      2. Johnson I,
      3. Locker T
      . The impact of 24 h consultant shop floor presence on emergency department performance: a natural experiment. Emerg Med J Published Online First June 12, 2012 doi:10.1136/emermed-2012-201369
    17. ↵
      1. Bucheli B,
      2. Martina B
      . Reduced length of stay in medical emergency department patients: a prospective controlled study on emergency physician staffing. Eur J Emerg Med 2004;11:29–34.
      OpenUrlCrossRefPubMed
    18. ↵
      1. Donald KJ,
      2. Smith AN,
      3. Doherty S,
      4. et al
      . Effect of an on-site emergency physician in a rural emergency department at night. Rural Remote Health 2005;5:380.
      OpenUrlPubMed
    19. ↵
      1. Shaw KN,
      2. Lavelle JM
      . VESAS: a solution to seasonal fluctuations in emergency department census. Ann Emerg Med 1998;32:698–702.
      OpenUrlCrossRefPubMedWeb of Science
    20. ↵
      1. Miró O,
      2. Sánchez M,
      3. Espinosa G,
      4. et al
      . Analysis of patient flow in the emergency department and the effect of an extensive reorganisation. Emerg Med J 2003;20:143–8–discussion148.
      OpenUrlAbstract/FREE Full Text
    21. ↵
      1. Lee-Lewandrowski E,
      2. Corboy D,
      3. Lewandrowski K,
      4. et al
      . Implementation of a point-of-care satellite laboratory in the emergency department of an academic medical center. Impact on test turnaround time and patient emergency department length of stay. Arch Pathol Lab Med 2003;127:456–60.
      OpenUrlPubMedWeb of Science
    22. ↵
      1. Kelen GD,
      2. Scheulen JJ,
      3. Hill PM
      . Effect of an emergency department (ED) managed acute care unit on ED overcrowding and emergency medical services diversion. Acad Emerg Med 2001;8:1095–100.
      OpenUrlPubMedWeb of Science
    23. ↵
      1. Ross MA,
      2. Naylor S,
      3. Compton S,
      4. et al
      . Maximizing use of the emergency department observation unit: A novel hybrid design. Ann Emerg Med 2001;37:267–74.
      OpenUrlCrossRefPubMed
    24. ↵
      1. McConnell KJ,
      2. Richards CF,
      3. Daya M,
      4. et al
      . Effect of increased ICU capacity on emergency department length of stay and ambulance diversion. Ann Emerg Med 2005;45:471–8.
      OpenUrlCrossRefPubMedWeb of Science
    25. ↵
      1. Shah MN,
      2. Fairbanks RJ,
      3. Maddow CL,
      4. et al
      . Description and evaluation of a pilot physician-directed emergency medical services diversion control program. Acad Emerg Med 2006;13:54–60.
      OpenUrlCrossRefPubMedWeb of Science
    26. ↵
      1. Washington DL,
      2. Stevens CD,
      3. Shekelle PG,
      4. et al
      . Next-day care for emergency department users with nonacute conditions. A randomized, controlled trial. Ann Intern Med 2002;137:707–14.
      OpenUrlPubMedWeb of Science
    27. ↵
      1. Scheulen JJ,
      2. Li G,
      3. Kelen GD
      . Impact of ambulance diversion policies in urban, suburban, and rural areas of Central Maryland. Acad Emerg Med 2001;8:36–40.
      OpenUrlPubMedWeb of Science
    28. ↵
      1. Connelly LG,
      2. Bair AE
      . Discrete event simulation of emergency department activity: a platform for system-level operations research. Acad Emerg Med 2004;11:1177–85.
      OpenUrlCrossRefPubMedWeb of Science
    29. ↵
      1. Chin L,
      2. Fleisher G
      . Planning model of resource utilization in an academic pediatric emergency department. Pediatr Emerg Care 1998;14:4–9.
      OpenUrlPubMedWeb of Science
    30. ↵
      Emergency room wait-times. Ontario Ministry of Health and Long-Term Care (Internet). (cited accessed 27 Sep 2012). http://edrs.waittimes.net/en/Data.aspx?LHIN=0&city=&pc=m5g%201×5&dist=25&hosptID=0&str=&view=0&period=0&expand=0
    31. ↵
      1. French JR
      . In: Miller JG. ed. Experiments in social process: a symposium on Social Psychology. McGraw-Hill. Chapter 6, p 82.

    Footnotes

    • Contributors Our manuscript has been read and approved by all authors. Each author provided substantial contributions to (1) conception and design, or acquisition of data, or analysis and interpretation of data; (2) drafting of article or revising it critically for important intellectual content and (3) final approval of the version to be published.

    • Competing interests Dr Ovens receives consultant fees from the government of Ontario for advising on policy related to Emergency Services.

    • Ethics approval Mount Sinai Research Ethics Board.

    • Provenance and peer review Not commissioned; externally peer reviewed.