Blackwell Science, LtdOxford, UKEMMEmergency Medicine Australasia1035-68512005 Blackwell Publishing Asia Pty Ltd 2005175434442Original ArticleWalkout of patients from NSW EDM Mohsin et al.
Emergency Medicine Australasia (2005) 17, 434–442
ORIGINAL RESEARCH
Factors associated with walkout of patients from New South Wales hospital emergency departments, Australia Mohammed Mohsin,1 Lis Young,2 Sue Ieraci,3 and Adrian E Bauman4 1 Centre for Research, Evidence Management and Surveillance, Division of Population Health, Sydney South West Area Health Service, Liverpool BC, 2Simpson Centre for Health Services Research, Liverpool Health Service, Liverpool, 3Sydney South West Area Health Service, and 4Department of Public Health and Epidemiology, School of Public Health, University of Sydney, Sydney, New South Wales, Australia
Abstract Objective
To explore the association between daily patient numbers, sociodemographic, clinical and system characteristics and walkout rates of patients from New South Wales (NSW) public hospital ED (without seeing a medical officer).
Methods
This was a secondary analysis of the NSW Emergency Department Data Collection, between 1 January 1999 and 31 December 2001, with a total of 4 356 323 ED attendances. Bivariate and multiple logistic regression analyses were performed to explore the relationship of walkout from ED and explanatory variables.
Results
About 5.7% of the patients attending in NSW public hospital ED left without seeing a doctor. The results from bivariate and multivariate analyses revealed that walkout rates significantly varied by sociodemographic and clinical characteristics of the patients. Patient groups that more likely to walk out were those aged 15–44 years, from a nonEnglish-speaking background, Aboriginal, with lower socioeconomic status, with no private health insurance coverage and longer waiting times for triage. Patient volume in the ED showed a significant positive association with walkout rates.
Conclusions
After adjusting for triage category (urgency of presentation) and triage time, patients from lower socioeconomic backgrounds and those without private health insurance coverage were more likely to leave ED without treatment. These results have important implications for Health Services. Future strategies aimed at minimizing walkouts from public hospital ED should prioritize and target factors identified in the present study.
Key words:
emergency department, patient volume, triage category, walkout.
Correspondence:
Dr Mohammed Mohsin, Centre for Research, Evidence Management and Surveillance, Division of Population Health, Sydney South West Area Health Service, Liverpool BC, NSW 1871, Australia. Email:
[email protected] or
[email protected]
Mohammed Mohsin, MSc (Stats), MSc (Demography), PhD, Biostatistician; Lis Young, FAFPHM, RCAP, Clinical Coordinator; Sue Ieraci, MB BS, FACEM, Staff Specialist in Emergency Department Bankstown and Sydney South West Area Advisor in Emergency Medicine; Adrian E Bauman, MPH, PhD, FAFPHM, Professor of Public Health and Epidemiology.
Walkout of patients from NSW ED
Introduction Public hospital ED across Australia provide one of the few options for unscheduled, acute medical care.1,2 During the past decade, public hospital ED have become increasingly crowded.3–6 Studies in Australia and elsewhere have found that the increase in the use of ED has led to longer waiting times and because of this some patients leave the ED without being seen by a doctor.7– 19 Patients who leave an ED before receiving medical care are an important group because it is not understood whether clinical outcomes are compromised as a result.15,16 As long as the potential for adverse outcomes has not been disproven, this group remains important in terms of risk management.17–19 If these patients are
Table 1.
shown to have subsequent morbidity as well as being unable to obtain care at sites other than the ED, then Health Services are failing this group of patients. Many patients who sought care from public hospital ED could not identify an alternative setting to see a doctor.9 Despite substantial ED research, little is known about the characteristics of patients who leave an ED without being seen by a doctor. A review of the more recent of these studies is shown in Table 1. Walkout rates differ across countries and within the same geographical boundary. Walkout rates in the USA are quite variable among ED, ranging from 0.84 to 15.0% (Table 1). The most commonly cited reason for leaving ED without treatment was dissatisfaction with the length of time to see the doctor.10,13,16,27,30
Studies regarding rates of ‘walkout’ from ED
Author Australian studies Browne et al.20 Grouse and Bishop21 Green22 Hanson et al.12 Mohsin et al.14 Trinker et al.23 Whitby et al.24 Overseas studies Arendt et al.7 Baker et al.9 Bindman et al.17 Bullard et al.10 Dershewitz and Paichel15 dos Santos et al.11 Das and Wijetunge25 Fernandes et al.16 Gbaanador et al.26 Gibson et al.27 Hobbs et al.13 Khanna et al.28 Kyriacou et al.29 Lee et al.18 Liao et al.30 McNamara31 Sainsbury32 Shaw et al.33 Stock et al.34 Stock et al.35 Weissberg et al.36 McMullan and Veser37
Place of study and no. sites Australia, 1 site, 29 weeks Australia, 1 site, 12 weeks Australia, 1 site, 6 months Sydney, Australia, 1 site Sydney, Australia, 5 sites Australia, Melbourne, 1 site Australia, 10 sites (NSW = 6, SA = 1, QLD = 1, WA = 1, TAS = 1) USA, 1 site USA, 1 site USA, 1 site Canada, 17 sites USA, 1 site USA, 1 site England, 1 site Canada, 2 sites USA, 1 site USA, 1 site USA, 1 site, 21 months UK, 1 site USA, 1 site Hong Kong, 1 site Taiwan, 1 site USA, 1 site USA, 1 site USA, 21 sites USA, 16 sites USA, 25 sites USA, 1 site USA, 1 site
No. patients and age
Walkout rate (%)
18 835, children 6 909 338, Aboriginal 19 862, non-Aboriginal 27 082, paediatric 20 7751, all ages 554, adults 11 169, all ages
5.5 6.7 8.3 6.5 1.7 4.9 7.0 4.5
20 494, all ages 2 259, adults 882, adults 1.49 million 27 230, paediatric 63 500, paediatric 18 750 23 933, all ages 5 769, all ages 26 000, adult 133 666, all ages 3 097 826 22 022 74 485 850 28 774, all ages 48 000, paediatric 52 778, all ages 92 570, all ages 4 868, all ages 18 664, all ages
0.84 8.2 15.0 1.3 1.1 3.9 1.4 1.4 9.3 2.0 7.4 3.3 5.8 0.4 0.1 3.8 1.2 1.6 3.6 4.2 1.4 3.4
NSW, New South Wales; QLD, Queensland; SA, South Australia; TAS, Tasmania; WA, Western Australia.
435
M Mohsin et al.
Little is known about the characteristics of Australian patients who leave an ED without being seen by a doctor.14,21–24 Previous studies were limited to individual communities or hospitals. They have covered short time spans, and specific groups, rather than populations. The purpose of the present study is to profile the ‘walkout’ problem in public hospital ED in one state: New South Wales (NSW), Australia. The present study specifically explore the association between daily patient numbers, sociodemographic, clinical and system characteristics and walkout rates of patients from NSW public hospital ED (without seeing a medical officer). The present study uses large sample population data, to explore factors associated with walkout and generate hypotheses as to their aetiology.
Methods The present study utilized the New South Wales Emergency Department Data Collection (NSW EDDC), during the period 1 January 1999–31 December 2001 (NSW HOIST). The data collection compiles de-identified Emergency Department Information System data from 55 ED that report performance to NSW Health. The participating hospitals comprise 12 Principal Referral (A1), 2 Paediatric (A2), 4 Ungrouped Acute (A3), 14 Major Metropolitan (B1), 9 Major non-metropolitan (B2) and 14 District Hospitals (C1 and C2) (NSW Health 199838). The data are collected by all of these ED in NSW, and reported to NSW Health Department Epidemiology Branch. Patients socioeconomic status (SES) was constructed based on ‘the index of socioeconomic disadvantage’ calculated for NSW 2001 census population by postcode of residence.39 First, all the postcodes for NSW EDDC were ranked by the index of socioeconomic disadvantage, and then grouped into five quintiles of SES from the lowest to the highest. Based on the estimated urgency of the presenting problem, patients attending ED in Australia, are classified (triaged) into one of five categories of the Australasian Triage Scale (ATS), adopted by the Australian Council on Health Care Standards.40,41 After triage, each of the patients might expect to wait a time proportional to their triage category, before being seen by a doctor. Those patients who left the ED without being seen by a doctor are defined as ‘walkouts’. ‘Triage category’ (ATS) is one of the most important data points in ED that aims to ensure that patients are treated in the order of their clinical urgency and that 436
their treatment is appropriately timely. Patients triaged to lower acuity categories might be safe to wait longer for assessment and treatment but might still require hospital admission.41 In the absence of information regarding severity of illness of the patients at the time of their presentation in ED, the present study used ‘triage category’ (or urgency) as proxy for ‘severity of illness’ of the patients. The time between arrival and triage is defined as ‘triage time’ and is calculated as follows: triage time = time of triage − arrival time. Bivariate (cross-tabulations) and multiple logistic regression analyses were performed to explore the relationship of walkout from ED and explanatory variables. The results of the bivariate analyses were expressed as percentages and means, statistical tests of significance such as χ2-test and t-test were used to examine the group differences. Logistic regression analysis explored the independent contribution of each potential explanatory variable on walkout of patients adjusting for other variables. The odds ratios (OR) from logistic regression analysis with their 95% confidence intervals (95% CI), are shown to express the likelihood of walkout for each explanatory variable adjusted for the effects of other variables.
Results General characteristics of the ED patients During the study period 1999–2001, a total of 4 356 323 ED attendances were registered in NSW Hospital ED and the number increased over the years (1 381 893 in 1999 to 1 497 382 in 2001, i.e. 8.4% increase since 1999). Of the total visits 53.8% were made by men, 24.4% by children (0–14 years), 41.2% by young adults (15– 44 years) and 10.7% by the elderly (75 years and older). About 79.2% were uninsured, 17.7% from a nonEnglish-speaking background (NESB) and 2.4% from an Aboriginal background. Most patients (45.7%) visited during the day (08.00–16.00 hours), followed by 39.8% between 16.00 and 23.59 hours, and the remaining 14.5% visits were after midnight. Of the patients, 23.6% were brought to ED by ambulance; this rate was quite high for the elderly (75 years and older) of whom about 64% arrived by ambulance. The greatest number of patients attended Principal Referral Hospital ED (32%), followed by Major Metropolitan ED (23.5%). Out of the total visits, 89% were triaged within 10 min of their arrival. Out of these 0.8% were triaged as immediately life-threatening, 5.8% as imminently life-
Walkout of patients from NSW ED
Factors associated with walkout of patients In 1999–2001, 5.7% of patients attending ED in NSW left without seeing a doctor (244 798 out of 4 314 666; excluding those who were dead on arrival). The rate increased from 5.3% in 1999 to 6.0% in 2001 (Table 2). The results from bivariate and multivariate analyses revealed that walkout rate significantly (P < 0.001) varied by sociodemographic and clinical characteristics of the patients. Among the variables, patients age, Aboriginality, SES, urgency (triage category), waiting time for triage, time of arrival, area of residence and insurance status had relatively strong associations with the walkout of patients from ED. The adjusted OR showed that patients aged 15–44 years were seven times more likely to walk out compared with patients aged 75 years or older (OR = 7.09; 95% CI 6.91–7.29). The results also showed that the walkout rate for the patients aged 0– 4 years was higher than that of those aged 5–14 years (6.1% vs 5.4%) (Table 2). Patients from an NESB were more likely leave ED without treatment than patients from the Englishspeaking background (ESB). The likelihood of walkout among Aboriginal patients was 37% higher than that of non-Aboriginal patients (OR = 1.37; 95% CI 1.33– 1.41). Patients from the lowest SES group had higher rates of walkout than the most socially advantaged patients (OR = 1.39; 95% CI 1.37–1.41). This was consistent with higher rate of walkout for uninsured patients compared with privately insured patients (OR = 1.24; 95% CI 1.23–1.26). The rate of walkout also varied by type of hospital, with higher rate of walkout reported in principal referral and major metropolitan hospitals (6.7%) compared with rural hospitals. This was supported by the higher rate of walkout among urban patients (OR = 1.98; 95% CI 1.93–2.04). As expected, the walkout rate was significantly higher for those who arrived by private/public transport compared with those who were brought to ED by any type of ambulance (OR = 1.52, 95% CI 1.49–1.54). Nonetheless, it is still important to note that 2.4% brought in by ambulance modes were walkouts (Table 2). The pattern of walkout varied by arrival time of day, with higher rate of walkout reported during the night shift and a pronounced drop during office hours. Triage category (urgency) showed a linear relationship to walkout rates, with rate increased as urgency decreased (0.2% for immediately or imminently life-threatening
compared with 9.7% for least urgent). The patients who were triaged as ‘least urgent’ were 58.3-fold more likely to walk out as compared with emergency patients (OR = 58.30, 95% CI 52.48–64.67). Walkout rates were positively associated with increased waiting times for triage; this rate increased from 5.2 (for those who waited less than 10 min for triage) to 13.8% (for those whose waiting time was 30 min or longer) (OR = 2.36; 95% CI 2.31–2.42). The mean waiting time from arrival to triage was significantly longer for walkout patients than nonwalkouts (3.6 vs 5.8 min, P < 0.001). This association between waiting times and walkout rates was consistent across all sociodemographic groups and triage categories. To test the relationship between the number of patients registered on a given day and the rate of walkout on that day, a correlation coefficient was calculated. The correlation coefficient (Pearson’s correlation coefficient) between the number of patients who arrived daily and the rate of daily walkout indicated that patient volume entering the ED showed a significant positive association with walkout of patients (r = 0.753, 95% CI 0.70–0.79, i.e. R2 = 0.567; P < 0.001). An analysis of variation through regression showed that daily patient numbers alone was a significant determinant of walkout and 56.7% of the variation in walkout was due to patient volume (Fig. 1). For patients who stayed in ED
9
8
Percentage of walkout
threatening, 28.4% as potentially life-threatening, 45.3% as potentially serious and 19.7% as less urgent.
7
6
5
4
3 3000
3500
4000
4500
5000
5500
6000
Average number of daily patients
Figure 1. Percentage of walkout by average number of daily patients with 95% confidence interval. It is to be noted that this study data set covered a period of 3 years. The average number of patients in each day was calculated and arranged in ascending order (lower to higher) and treated as x-axis, and corresponding percentage of walkout for respective day was plotted on y-axis. Same aggregate data set was used for correlation coefficient.
437
M Mohsin et al.
Table 2. Percentage of walkouts and odds ratios with 95% CI (from multiple logistic regression analysis) for selected characteristics (n = 4 314 666) Selected characteristics
All patients Year of arrival 1999 (reference category)‡ 2001 2002 Sex* Male Female (reference category)‡ Age (years)* 0–4 5–14 15–44 45–64 65–75 75+ (reference category)‡ Country of birth* ESB (reference category)‡ NESB Aboriginal status* Non-Aboriginal (reference category)‡ Aboriginal Insurance status* Insured (reference category)‡ Uninsured Socioeconomic status (SES)* 1st quintile (the lowest SES) 2nd quintile 3rd quintile 4th quintile 5th quintile (highest SES) (reference category)‡ Hospital category* Principal referral and major metropolitan Paediatric specialist (reference category) Ungrouped acute Major non-metropolitan and District one District two Area health services of residence (AHS)* Rural AHS Urban AHS Arrival time* 8.01–16.00 hours (reference category) 16.01–24.00 hours 24.01–8.00 hours Triage category (severity of illness)§ 1 and 2. Immediately/imminently life-threatening (reference category) 3. Potentially life-threatening 4. Potentially serious 5. Less urgent
438
Results from bivariate analyses % Walkout
Results from multiple logistic regression analysis† Adjusted odds ratio (95% CI)
5.7 5.3 5.7 6.0
1.00 1.21 (1.20–1.23)** 1.40 (1.38–1.42)**
5.9 5.5
1.02 (1.01–1.03)** 1.00
6.1 5.4 7.8 4.6 2.2 1.2
5.99 5.21 7.09 4.06 1.88
5.4 6.5
1.00 1.16 (1.15–1.18)**
5.5 7.3
1.00 1.37 (1.33–1.41)**
4.4 6.0
1.00 1.24 (1.23–1.26)**
6.3 5.8 5.5 5.3 5.2
1.39 (1.37–1.41)** 1.29 (1.27–1.31)** 1.20 (1.18–1.22)** 1.10 (1.08–1.12)** 1.00
6.7 4.0 2.9 4.7 3.2
2.01 (1.96–2.06)** 1.00 0.83 (0.80–0.86)** 1.30 (1.27–1.33)** 0.87 (0.84–0.91)**
4.2 6.2
1.00 1.98 (1.93–2.04)**
4.2 7.1 6.4
1.00 1.73 (1.71–1.75)** 1.61 (1.58–1.63)**
0.2
1.00
1.7 7.2 9.7
7.22 (6.50–8.02)** 32.60 (29.36–36.16)** 58.30 (52.48–64.67)**
(5.83–6.17)** (5.06–5.37)** (6.91–7.29)** (3.94–4.17)** (1.82–1.95)** 1.00
Walkout of patients from NSW ED
Table 2.
Cont.
Selected characteristics
Results from bivariate analyses % Walkout
Results from multiple logistic regression analysis† Adjusted odds ratio (95% CI)
2.4 6.6 4.2
1.00 1.52 (1.49–1.54)** 0.94 (0.86–1.01)NS
5.2 7.9 10.6 13.8
1.00 1.26 (1.24–1.28)** 1.66 (1.62–1.70)** 2.63 (2.31–2.42)**
Mode of arrival* Ambulance any type (reference category) Public transport or private car Others Arrival to triage time (min)*
