Review article: Leaving the emergency department without being seen

REVIEW ARTICLE

Review article: Leaving the emergencydepartment without being seenMarcus Kennedy,1 Catherine E MacBean,2 Caroline Brand,3,4 Vijaya Sundararajan3 andDavid McD Taylor5

1Adult Retrieval Victoria, Metropolitan Ambulance Service, 2Emergency Department and 3ClinicalEpidemiology and Health Service Evaluation Unit (CEHSEU), Royal Melbourne Hospital, 4Centre forClinical Excellence in Patient Safety, Monash University, and 5Emergency and General MedicineResearch, Austin Hospitals, Melbourne, Victoria, Australia

Abstract

Patients who leave the ED without being seen (LWBS) are unlikely to be satisfied with thequality of the service provided and might be at risk from conditions that have not beenassessed or treated. We therefore examined the available research literature to inform thefollowing questions: (i) In patients who attend for ED care, what factors are associated withthe decision to LWBS? (ii) In patients who attend for ED care, are there adverse healthoutcomes associated with the decision to LWBS? (iii) Which interventions have been usedto try to reduce the number of patients who attend for ED care and LWBS? From theavailable literature, there was insufficient evidence to draw firm conclusions; however, theliterature does suggest that patients who LWBS have conditions of lower urgency andlower acuity, are more likely to be male and younger, and are likely to identify prolongedwaiting times as a central concern. LWBS patients generally have very low rates ofsubsequent admission, and reports of serious adverse events are rare. Many LWBS patientsgo on to seek alternative medical attention, and they might have higher rates of ongoingsymptoms at follow-up. Further research is recommended to include comprehensive cohortor well-designed case–control studies. These studies should assess a wide range of relatedfactors, including patient, hospital and other relevant factors. They should compare out-comes for groups of LWBS patients with those who wait and should include cross-sectoraldata mapping to truly detect re-attendance and admission rates.

Key words: did not wait, emergency medicine, leave without being seen, patient discharge, risk.

Correspondence: Dr Marcus Kennedy, Adult Retrieval Victoria, Metropolitan Ambulance Service, PO Box 2000, Doncaster, Vic. 3108,Australia. Email: [email protected]

Marcus Kennedy, MB BS, FACEM, FRACGP, DA(UK), DipIMC(RCSEd), Director; Catherine E MacBean, BA(Hons), Research Assistant; CarolineBrand, MB BS, BA, MPH, FRACP, Director; Vijaya Sundararajan, BA, MD, MPubHlth, FACP, Senior Medical Advisor; David McD Taylor, MD,MPH, DRCOG, FACEM, Director of Emergency and General Medicine Research.

doi: 10.1111/j.1742-6723.2008.01103.xEmergency Medicine Australasia (2008) 20, 306–313

© 2008 The AuthorsJournal compilation © 2008 Australasian College for Emergency Medicine and Australasian Society for Emergency Medicine

mailto:[email protected]

Introduction

The population of interest for this review were patientswho present to the ED and leave without being seen(LWBS) by a doctor.

Objectives

To review all available and relevant literature in orderto inform the following:In patients who attend the ED for care,• What factors are associated with the decision to

LWBS?• Are there adverse health outcomes associated with

the decision to LWBS?• Which interventions have been used to try to reduce

the number of patients who attend for ED care andLWBS?

Methods

Literature search

A review of the literature was conducted, focusing onthe ED patient population who LWBS. Searches forEnglish-language medical literature published between1990 and 2006 were conducted using the MEDLINE,CINAHL and PUBMED databases and the CochraneDatabase of Systematic reviews. Searches for ‘grey lit-erature’ were also conducted on various government,research and public websites. The final keyword searchstrategy employed included emergency and any of thefollowing: did not wait, dnw, drop-out, elope, leave, leavewithout being seen, leaving without being seen, leftwithout being seen, lwbs, left without treatment, lwot,lwt, wait, walk-out.

Inclusion and exclusion criteria

Inclusion criteria were that articles were in English,published between 1990 and 2006 and whose primepurpose pertained directly to patients who LWBS.Articles were excluded if they focused solely on privateor military hospitals or if they used definitions of LWBSthat were inconsistent with our definition.

One researcher (CM) was responsible for initiallyreviewing the medical databases and grey literature.Articles were provisionally selected on the basis of titleswith relevance to the topic. Eligibility for inclusion was

decided by three members of the research team on thebasis of the abstract of each article.

Information was extracted from each article in refer-ence to the following characteristics: purpose of study,country, study period, nature of emergencies present-ing, study design, sampling procedures, response rates,ethics committee authorization, percentage of patientswho LWBS, LWBS variables considered, main findings,outcomes, details of interventions, hospital/workforcecharacteristics, terminology used, definitions provided,analyses used and relevant study weaknesses.

Discussion

Key findings are summarized in Box 1 and Table 1.

How many patients LWBS?

LWBS rates from metropolitan public hospitals inVictoria, Australia, from the years 1999 to 2003, variedfrom 1.1% to 10.1% but averaged 5.4%, 5.8%, 6.0% and5.8% for the years 1999/2000, 2000/2001, 2001/2002 and2002/2003, respectively.1

International LWBS rates were generally comparablewith Australian rates with the exception of very lowrates in the two Asian studies.2,3 Rates observed in theUSA were variable, with one very low rate of 0.844 andtwo relatively high rates of 14.9%5 and 15%.6 ThreeCanadian studies7–9 reported rates between 1.4% and3.57%, and studies from the UK10,11 reported rates of3.26% and 7.2%. Rates from Hong Kong2 (0.36%)and Taiwan3 (0.1%) were markedly lower than thoseobserved elsewhere. These differences might reflect dif-ferences in culture, ED structure or service delivery.

Generally, lower rates were observed in paediatricsettings and ranged from 1.7% to 5.5%.8,12–14

A study conducted in the Los Angeles County byStock et al.15 in 1994 revealed an overall rate of 4.2%.This was the only study found to include a largenumber of sites (30) and both private and public hospi-tals; the public and private hospitals recorded rates of7.3% and 2.4%, respectively. Another indication of thevariability of rates according to different factors can beseen within the same hospital setting, in a 5 year studyreported by Kyriacou et al. in 1999.16 The study evalu-ated the relationship between waiting times to see aphysician, rates of LWBS and a series of continuousquality improvement interventions. The rates observedin this single hospital setting ranged from 2.65% to7.03%.

Patients leaving without being seen

307© 2008 The AuthorsJournal compilation © 2008 Australasian College for Emergency Medicine and Australasian Society for Emergency Medicine

Which patients LWBS?

Demographic characteristicsA small number of studies directly examined the rela-tionship between the patients’ demography and their

likelihood to LWBS. Mohsin et al.17 found that patientswho were male, younger, English-speaking, uninsuredor with a lower socioeconomic status were more likely toLWBS. Baker et al.18 also reported similar sex patterns.In particular, they found that those who LWBS weremore likely to be male than female (8.3% vs 5.8%),younger (with an odds ratio of 0.978 for each year ofage), ambulant (8.7% of ambulant patients as opposedto 3.5% of those who arrived by ambulance) and self-referred (7.4% vs 3.1% of those who were GP referred).The Baker et al.18 study also demonstrated that thesepatients present at times when waiting times are longestand that time of presentation, rather than individualpatient characteristics, seems to be the most powerfulpredictors of time that LWBS patients would wait.Recent studies conducted by Ding et al.19 and Goodacreet al.11 support these findings. Baibergenova et al.20

reported a strong correlation (r = 0.62) between LWBSrate and median ED length of stay, a sensitive measureof overall ED patient flow performance.

Acuity and urgency of patients illnessesStudy results are difficult to interpret given the lack ofinternational uniformity in triage criteria and assess-ments. Although findings are somewhat mixed, itappears that patients who LWBS generally have condi-tions of a less acute and less urgent nature. Most of thestudies reviewed did not include categorized acuity ortriage breakdowns for patients who did wait to be seen.

Goldman et al.8 found no LWBS children in category1 or 2 of the Canadian ED Triage and Acuity Score, 15%in category 3 and 85% in categories 4 or 5. Similarly,Browne et al.12 found no LWBS children in category 1 or2 of the Australian National Triage Scale and 2.5%,62.2% and 35.3% of LWBS patients in categories 3, 4and 5, respectively. In the UK, Goodacre et al.11 foundthat, as triage category reduced in urgency, patientswere much more likely to LWBS.

Hospital characteristicsStock et al ’s15 cross-sectional analysis of LWBS ratesfrom 30 Los Angeles County ED revealed a consider-ably higher LWBS rate from public hospitals (7.3%)than private hospitals (2.4%) and an overall rate of4.2%. A number of other hospital characteristics wereassociated with higher LWBS rates, including being ateaching hospital or a trauma centre, having long esti-mated waiting times and a high percentage of patientswithout health insurance.

Box 1. Summary of results

Twenty-six articles met the selection criteria forinclusion. Although insufficient evidence exists todraw absolute conclusions, the literature doessuggest the following:

Patients who LWBS generally have conditionsof lower urgency and lower acuity.Patients who LWBS generally have longer waittimes.Demographic characteristics of LWBS patientssuggest that these patients are more likely to bemale and younger.Putative associations have been proposed tolink various hospital and temporal factors, forexample, levels of overcrowding and the timeof ED presentation.Patients identify various reasons why theydecide to LWBS. However, the issue of pro-longed waiting times clearly emerges as acentral concern.Many LWBS patients go on to seek alternativemedical attention subsequent to their ED atten-dance. However, limited data exist to allowanalysis of the rate, characteristics or outcomesof such patients.LWBS patients generally have low rates of sub-sequent hospital admission.Reports of serious adverse events are rare inLWBS patients.LWBS patients might have higher rates ofongoing symptoms at follow-up, when com-pared with controls.There is inadequate information to drawconclusions regarding patients who LWBS inregional and rural areas.Only two interventions specifically designed tolower LWBS rates were published and one wasunsuccessful.Challenges and limitations have been identifiedfor research into LWBS rates, in particularthose associated with obtaining adequatesample sizes and response rates from surveyedpatients.

M Kennedy et al.

308 © 2008 The AuthorsJournal compilation © 2008 Australasian College for Emergency Medicine and Australasian Society for Emergency Medicine

ED activity and crowdingWeiss et al.5 demonstrated that patients were morelikely to LWBS as the ED became overcrowded. Theyemployed a six-item overcrowding scale, the NationalEmergency Department Overcrowding Scale, and founda correlation of 0.665 between overcrowding and LWBSrates. Hobbs et al.21 developed statistical models to iden-tify and quantify the contributions of various factorsthat have an impact on LWBS rates. Hospital-basedfactors investigated included the number of patientsseen and admitted from each area of their ED. Multi-variate analysis revealed the total number of patientsbeing cared for in the main ED to be the most powerfulpredictor of LWBS rates. Other variables included thetotal number of trauma and resuscitation patients andthe total number of observation unit admissions to thehospital. Moshin et al.22 reported similar findings of astrong relationship (r = 0.753) between the total number

of patients treated by the ED on a particular day andLWBS rate.

Finally, a recent study in California investigatedhospital-based factors in relation to LWBS rates.Polevoi et al.23 examined percentage of ED bedcapacity, acuity of ED patients, length of stay ofdischarged patients in the ED, patients in the EDawaiting an inpatient bed, inpatient floor capacity,intensive care unit capacity and the characteristics ofthe attending physician in charge. They concludedthat a significant relationship exists between ED occu-pancy and LWBS rates once ED occupancy is greaterthan 100%.

Temporal factorsA number of temporal factors have been suggested toimpact upon LWBS rates.

Table 1. Publications reviewed

Author Country Published Sites Design % LWBS

Lee et al.27 Australia 2006 1 Prospective data collection and questionnaire. 6.7Mohsin et al.22 Australia 2005 55 Secondary data analysis – retrospective observational 5.7Fry et al.24 Australia 2003 1 Cross-sectional descriptive 7.9Browne et al.12 Australia 2001 1 Prospective (case–control) surveillance 5.5Mohsin et al.17 Australia 1998 5 Retrospective observational 4.9Hanson et al.16 Australia 1994 1 Review of complaints 1.7Rowe et al.28 Canada 2006 2 Prospective case–control 4.5Goldman et al.8 Canada 2005 1 Case–control 3Monzon et al.9 Canada 2005 1 Prospective case–control 3.57Fernandes et al.7 Canada 1994 2 Prospective cohort (cross-sectional) 1.4Lee et al.2 Hong Kong 1998 1 Prospective 0.36Liao et al.3 Taiwan 2002 1 Cross-sectional 0.1Goodacre and Webster11 UK 2005 1 Cohort 7.2Khanna et al.10 UK 1999 1 Prospective follow-up 3.26Ding et al.19 USA 2006 1 Pair-matched case–control 6.4Kronfol et al.25 USA 2006 1 Prospective observational 3.65Chan et al.26 USA 2005 1 Before–after intervention trial 7.7 (pre)

4.4 (post)Polevoi et al.23 USA 2005 1 Modified case–crossover design 1.8Weiss et al.5 USA 2005 1 Prospective observational 14.9Arendt et al.4 USA 2003 1 Retrospective observational 0.84Hobbs et al.21 USA 2000 1 Retrospective cohort analysis 7.4Kyriacou et al.16 USA 1999 1 Prospective time study analysis 7.03, 6.18, 6.55

2.65, 3.97, 4.57, 7dos Santos et al.13 USA 1994 1 Prospective follow-up (case–control) 3.9Stock et al.15 USA 1994 30 Cross-sectional 4.2 (all hospitals)

7.3 (public)Baker et al.18 USA 1991 1 Case–control follow-up 8.2Bindman et al.6 USA 1991 1 Observational cohort 15.0



Elapse time to LWBS. The length of time patientselapsed before LWBS shows marked variability.Studies from Hong Kong2 and Taiwan3 reported that56% and 58% of patients, respectively, who LWBS didso within 60 min of registration. Furthermore, 96% ofall LWBS patients left within 2 h in the Hong Kongstudy. Khanna et al.10 reported that 58% of patients leftwithin the first 2 h, with an average time of 2.44 h.Bindman et al.6 reported that only 5% of LWBS patientsleft in the first hour and that the average elapse time toLWBS was 3 h and 30 min. Baker et al.18 reported a 6.4 haverage wait time for LWBS patients.

Time, day and season of presentation. Other temporalfactors have been associated with increased LWBSrates. These include night shifts,7 arrival between mid-night and 04.00 hours8,17 and attendance in winter.14

However, in order to determine the relative influence ofeach of these temporal factors, in isolation from con-founding variables such as hospital staffing levels,multiple regression analysis would be required. Thislevel of statistical analysis is not available in studiesperformed to date.

Why do patients LWBS?

Over half the studies in this review included some sortof analysis of why patients chose to LWBS. These tookthe form of retrospective enquiries by either written ortelephone survey and within 48 h to 4 weeks of the EDpresentation.

The overwhelming cited reason for patients choosingto LWBS was prolonged waiting times. Patientsreported being ‘tired of waiting’ or that the ‘waiting timewas too long’. A recent Australian study conducted byFry et al.24 included 222 LWBS patients (64% followedup) and revealed that 49% of responding patients leftbecause of the time delay and that 52% of all patientswould have stayed if the waiting time was shorter.Arendt et al.4 asked patients why they left, and 67%of the responding patients said they believed that thewait would be too long or had already been too long.Other studies listed a considerably higher proportion ofpatients concerned about waiting times, including 75%of patients in a Hong Kong-based study2 and 86% ofpatients in an American study.6

A number of other interesting and importantreasons for leaving the ED are apparent. Some LWBSpatients reported that they left the ED because theirsymptoms resolved (37% in the Goldman et al.8 study),they started to feel better (15% in the Monzon et al.9

study) or they felt their problem could wait or hadspontaneously begun to improve (43% in the Khannaet al.10 study).

In contrast to this are a group of potentially greaterconcern – those patients who report leaving because of‘feeling too ill to wait’. In Baker et al.’s paper,18 it wasreported that 53% of LWBS patients felt too sick to waitlonger (the study allowed patients to report multipleresponses). Although other studies reported lower per-centages of people leaving because of feeling too ill towait,9,24 these patients might represent a high-risk groupfor serious adverse events.

Overall, patients often appear to have complexreasons for LWBS.

Outcomes associated with LWBS

Seeking alternate and subsequent medical careMany studies report that approximately half of patientswho LWBS will seek alternate care. This might be withtheir local GP, returning to the same ED, attending adifferent ED or attending another provider. However,the accuracy of the data reported is questionable. Thedata describe only the outcomes of patients who werefollowed up. Large numbers of patients in each of thesestudies were lost to follow-up. The absence of matchedcontrol groups from the population that did wait to beseen further limits the value of these studies.

Baker et al.18 conducted a case–control follow-upstudy and found that 51% of LWBS patients saw aphysician within 1 week. Of these, approximately one-third were seen in a private clinic, one-third returned tothe same ED and one-third went to another ED. Similarto Baker’s findings,18 Bindman et al.6 reported that 55%of patients who LWBS saw a physician within 1 week oftheir ED presentation. Other studies considered a muchtighter time frame. Fernandes et al.7 found that 48% ofpatients saw a physician within 24 h and a further 10%saw a physician between 24 and 72 h. Fry et al.24

reported that 68% of LWBS patients sought medicalattention within 24 h of leaving the ED. Goldman et al.8

reported that 63% of LWBS paediatric patients soughtfurther medical care after the ED visit compared withonly 28% of the control group who were seen. Also in apaediatric setting, dos Santos et al.13 showed that 63%of LWBS patients sought further medical care and that77% of these patients did so within 24 h. Kronfolet al.’s25 findings in a paediatric setting were verysimilar, with a median time to subsequent consultationof 8 h in LWBS patients.

M Kennedy et al.


Health impactIn the case–control study of Bindman et al.,6 LWBSpatients were twice as likely as controls to report theirpain or condition to be worse at follow-up. Similarly,Monzon et al.9 found LWBS patients to be significantlymore likely (19.8%) than non-LWBS patients (8.7%) tofeel worse 2 days after having left the ED.

Serious adverse eventsReports of adverse outcomes appear very rare in theLWBS literature. Monzon et al.9 reported one LWBSpatient who, after presenting to the ED with a psy-chosocial complaint, left and subsequently committedsuicide. Bindman et al.6 also described a complicatedcase that resulted in an appendectomy. With the excep-tion of these instances, no other article reports deaths orserious adverse events. As many of the studies in thisreview have poor response and follow-up rates, there arelarge proportions of LWBS populations for whom out-comes are not known. Consequently, this remains animportant issue requiring further evaluation.

Admissions ratesGenerally, the studies examined in this review reportedrelatively low subsequent hospitalization rates forLWBS patients.8,10 However, the degree of inconsistencyin data trends was very large. No studies included cross-sectoral data mapping to truly detect re-attendance andadmission rates.

Interventions aimed at reducing the rate ofLWBS patients

This review found only two papers that primarily exam-ined the impact of a study intervention on LWBS rates.In 1992–1993, dos Santos13 et al. in a paediatric EDaimed to determine whether the presence of a patientadvocate affected LWBS rates. The study found thatLWBS rates actually rose from 1.8% to 3.1% followingthe introduction of the ombudsman. As the study wasconducted in one hospital and used only one ombuds-man, it was not possible to draw firm conclusions.

Chan et al.26 studied the impact of a Rapid Entry andAccelerated Care at Triage project, which included arange of triage-based interventions to speed up access tocare. This single site, before and after study, showed animprovement in LWBS rate from 7.7% to 4.4%, rein-forcing the finding of a relationship between waitingtime and LBWS rate.

Limitations

Study designs used to examine LWBS included cross-sectional, cohort studies and simple observationaldesigns. Some studies included only a descriptive analy-sis of LWBS patient populations. Small sample sizesand poor response rates for many studies limited theusefulness of their data. Limited conclusions can bedrawn from findings of such studies. Smaller studiesmight not have the power to adequately assess occur-rences such as hospitalizations and adverse event rates.Patient responses from studies with poor response ratesmight not accurately reflect the LWBS population asa whole.

Some studies chose to include or focus on a verylimited range of variables or factors associated withLWBS, thereby introducing bias.

With the exception of the Stock et al.15 and Mohsinet al.17 studies, all the remainder included one hospitalsite only, and only one study considered the ruralsetting.14

The duration of study periods examined variedgreatly. Some studies were very brief with data collec-tion periods that spanned only days.

The follow-up periods for studies that involved eithertelephone or mail-out follow-up of patients varied. Theyranged from very short periods (48 h) to much longerperiods (4 weeks). There are advantages and disadvan-tages to both shorter and longer follow-up times. First,patient recall might be more reliable with shorterfollow-up times and less reliable after several weeks.However, longer follow-up periods allow for greatercapture of any outcome data from LWBS cases.

Conclusions and recommendations

Is any LWBS rate acceptable?

One issue that remains largely unaddressed in the lit-erature is what constitutes an acceptable or desirableLWBS rate. Generally, the emphasis centres on reducingrates of LWBS without quantifying the issue. If allpatients presenting to the ED genuinely needed emer-gency medical treatment, then we should aim for aLWBS rate of zero. However, this is unrealistic, giventhat some percentage of LWBS patients will arise frompatients for whom other and often better health accessoptions could have been considered in the first instance.Also, the natural history of illness is such that symp-toms that might initially appear urgent or distressing



might resolve or diminish to the extent that the patientno longer considers waiting at an ED appropriate.

As a matter of risk management, it is imperative weunderstand the outcomes for LWBS patients. In particu-lar, the incidence of serious adverse outcomes must beknown before an acceptable rate of LWBS can be deter-mined. This knowledge can also sensibly inform theintensity of strategies to reduce LWBS rates.

There is much scope to further investigate patientswho LWBS, especially the factors associated with theirleaving and their associated health outcomes. Austra-lian and international research suggests associationsbetween LWBS rates and a number of factors. However,the available evidence is limited and somewhatinconsistent. Existing research does highlight someimportant methodological considerations for futureinvestigations in this area. LWBS is a complex, multi-faceted issue, and research needs to be carefullydesigned to incorporate a comprehensive range ofpatient, hospital, temporal and other factors (Box 2).Studies should include not only the LWBS patient popu-lations but also control patients for comparison.

In order to adequately address the LWBS problem,targeted research, devoid of previously identified meth-odological flaws, is required. In a proposed study, theprincipal aims would be to accurately determine• The factors that are associated with a patient’s

decision to LWBS• Whether adverse health outcomes are experienced as

a result of LWBS• What initiatives and changes to ED processes are

likely to result in decreasing LWBS ratesSuch a study would be a prospective examination of

patients presenting to the ED and would includepatients who LWBS and a matched control group whodo not LWBS. This design would allow:• A cross-sectional survey of patient perceptions of the

ED reception, triage and waiting processes• A case–control study to accurately determine

reasons why patients LWBSAdequate time frames need to be set to allow for

capture of a large number of LWBS patients, and inorder to maximize the quality of the studies, it is crucialto obtain a high rate of patient follow-up. Furthermore,to ensure that a large-scale, prospective study has goodexternal validity, enrolment of patients from a range ofpublic hospitals would be required, and follow-up wouldneed to be cross-sectoral, capturing interventions andoutcomes in all facets of the health industry (includingcoronial and death registry data) through linkage andinterrogation of appropriate datasets.

Author contributions

MK, design, review of literature, draft review and manu-script preparation. CEM, literature search and prepara-tion of draft manuscript. CB, contribution to design,draft review and manuscript review. VS, contribution todesign, draft review and manuscript review. DMT,design, review of literature, draft review and manu-script review.

Box 2. Potential table of variables for futureresearch

During ED visit under considerationTriage categoryArrival transport modeEthnicity from country of birth, broken downinto the following categories

English-speakingNon-visible minoritiesVisible minorities

Time parameters within the EDArrival date/timeTriage date/timeNurse date/timeDoctor date/time

Date/time parameters of ED visit underconsiderationAgeSexCondition diagnostic groupsSocioeconomic statusProximity to goods and servicesReferral sourceHospital campus codeCompensable statusHospital activity

Three time levelsShift of ED visitWithin the 12 h of ED visitWithin the 24 h of ED visitWithin the 48 h of ED visitWithin the 7 days preceding ED visitWithin the 14 days preceding ED visit

Emergency admissionsElective admissionsShort stay/observation admissions

Before index ED visit under considerationNumber of ED visits in previous 12 monthsNumber of admissions in previous 12 months

M Kennedy et al.


Competing interests

None declared.

Accepted 21 May 2008

References

1. Managing Emergency Demand in Public Hospitals. VictorianAuditor General’s Report. No. 71. Session 2003–2004.

2. Lee KM, Wong TW, Chan R, Lau CC. A study of patients wholeave without notice in an A and E department. Accid. Emerg.Nurs. 1998; 6: 118–21.

3. Liao HC, Liaw SJ, Hu PM, Lee KT, Chen, CM, Wang FL. Emer-gency department patients who leave without being seen by adoctor: the experience of a medical center in northern Taiwan.Chang Gung Med. J. 2002; 25: 367–73.

4. Arendt KW, Sadosty AT, Weaver AL, Brent CR, Boie ET. Theleft-without-being-seen patients: what would keep them fromleaving? Ann. Emerg. Med. 2003; 42: 317–23.

5. Weiss SJ, Ernst AA, Derlet R, King R, Bair A, Nick TG. Rela-tionship between the National ED Overcrowding Scale and thenumber of patients who leave without being seen in an academicED. Am. J. Emerg. Med. 2005; 23: 288–94.

6. Bindman AB, Grumbach K, Keane D, Rauch L, Luce JM. Conse-quences of queuing for care at a public hospital emergencydepartment. JAMA 1991; 266: 1091–6.

7. Fernandes CM, Daya MR, Barry S, Palmer N. Emergency depart-ment patients who leave without seeing a physician: the TorontoHospital experience. Ann. Emerg Med. 1994; 24: 1092–6.

8. Goldman RD, Macpherson A, Schuh A, Mulligan C, Pirie J.Patients who leave the pediatric emergency department withoutbeing seen: a case–control study. CMAJ. 2005; 172 (1): 39–43.

9. Monzon J, Friedman SM, Clarke C, Arenovich T. Patients wholeave the emergency department without being seen by a physi-cian: a control-matched study. CJEM 2005; 7: 108–13.

10. Khanna R, Chaudhry MA, Prescott M. Emergency departmentpatients who leave the department without being seen by adoctor. Eur. J. Emerg. Med. 1999; 6: 233–5.

11. Goodacre S, Webster A. Who waits longest in the emergencydepartment and who leaves without being seen? EMJ. 2005; 22:93–6.

12. Browne GJ, McCaskill ME, Giles H, Lam LT, Fasher BJ, Exley B.Paediatric walk-out patients: characteristics and outcomes.J. Paediatr. Child Health 2001; 37: 235–9.

13. dos Santos LM, Stewart G, Rosenberg NM. Pediatric emergencydepartment walk-outs. Pediatr. Emerg. Care 1994; 10: 76–8.

14. Hanson R, Clifton-Smith B, Fasher B. Patient dissatisfaction in apaediatric accident and emergency department. J. Qual. Clin.Pract. 1994; 14: 137–43.

15. Stock LM, Bradley GE, Lewis RJ, Baker DW, Sipsey J, StevensCD. Patients who leave emergency departments without beingseen by a physician: magnitude of the problem in Los AngelesCounty. Ann. Emerg. Med. 1994; 23: 294–8.

16. Kyriacou DN, Ricketts V, Dyne PL, McCollough MD, Talan DA.A 5 year time study analysis of emergency department patientcare efficiency. Ann. Emerg. Med. 1999; 34: 326–35.

17. Mohsin M, Bauman A, Ieraci S. Is there equity in emergencymedical care? Waiting times and walk-outs in South WesternSydney hospital emergency departments. Aust. Health Rev. 1998;21: 133–49.

18. Baker DW, Stevens CD, Brook RH. Patients who leave a publichospital emergency department without being seen by a physi-cian. Causes and consequences. JAMA 1991; 266: 1085–90.

19. Ding R., McCarthy ML, Li G, Kirsch TD, Jung JJ, Kelen GD.Patients who leave without being seen: their characteristics andhistory of emergency department use. Ann. Emerg. Med. 2006;48: 686–93.

20. Baibergenova A, Leeb K, Jokovic A, Gushue S. Missed opportu-nity: patients who leave emergency departments without beingseen. Healthc Policy 2006; 1 (4): 35–41.

21. Hobbs D, Kunzman SC, Tandberg D, Sklar D. Hospital factorsassociated with emergency center patients leaving without beingseen. Am. J. Emerg. Med. 2000; 18: 767–72.

22. Mohsin M, Young L, Ieraci S, Bauman AE. Factors associatedwith walkout of patients from New South Wales hospital emer-gency departments, Australia. Emerg. Med. Australas. 2005; 17:434–42.

23. Polevoi SK, Quinn JV, Kramer NR. Factors associated withpatients who leave without being seen. Acad. Emerg. Med. 2005;12: 232–6.

24. Fry M, Thompson J, Chan A. Patients regularly leave emergencydepartments before medical assessment: a study of did not waitpatients, medical profile and outcome characteristics. AENJ.2003; 6 (2): 21–6.

25. Kronfol RN, Childers K, Caviness AC. Patients who leave ouremergency department without being seen. Pediatr. Emerg. Care2006; 22: 550–4.

26. Chan TC, Killeen JP, Kelly D, Guss DA. Impact of rapid entry andaccelerated care at triage on reducing emergency departmentpatient wait times, lengths of stay, and rate of left without beingseen. Ann. Emerg. Med. 2005; 46: 491–7.

27. Lee G, Endacott R, Flett K, Bushnell R. Characteristics of patientswho did not wait for treatment in the emergency department: Afollow up survey. Accid. Emerg. Nurs. 2006; 14: 56–62.

28. Rowe BH, Channan P, Bullard M et al. Characteristics of patientswho leave emergency departments without being seen. Acad.Emerg. Med. 2006; 13: 848–52.



Documents

Review article: Leaving the emergency department without being seen