Missing the boat: odds for the patients who leave ED without being seen

Emergency Departments (ED) not only provide care to patients with critical and life threatening emergencies, but also look after round the clock to those who have acute yet stable medical illnesses [1‐3]. The resultant ED overcrowding which was first described twenty years ago, has now become a well-established barrier in access to health care [4‐6]. The problem is exacerbated in low income countries by utilization of ED as a primary access point to the healthcare especially on weekends and after hours for less urgent conditions [1]. However, the balance is now tilting towards high acuity patients, ED boarding of admitted patients, and hospital occupancy as a cause of ED overcrowding rather than influx of non-urgent patients [4, 5, 7]. ED overcrowding not only reduces patient satisfaction but it also increases the number of patients that leave without being seen by a physician (LWBS) [3, 7]. Large number of these patients may not find appropriate care elsewhere and therefore a critical treatment opportunity is missed by the health system. The percentage of LWBS patients has been recognized as a proxy indicator of ED performance and overcrowding [8, 9].

A number of studies from high income countries with well-established primary health care system have reported a variable number of LWBS which ranges between <1% to 20% of all ED visits [10‐14]. It has been suggested that patients who LWBS are at an increased risk of morbidity and mortality; however, a more recent administrative follow-up demonstrated these patients are at a lower risk of hospitalization and death than triage-matched controls [15‐19]. Several factors have been found as being associated with LWBS, such as low acuity illness, young age, and male sex and prolonged waiting time [20‐24]. Additionally, the triage time, previous ED visits, seasonal variation, access to primary care, diversion status and ED overcrowding also have significant impact on LWBS [10, 25‐29]. A literature review of local published research from Pakistan showed no study documenting the characteristics of patients who leave ED without being seen by a physician in this region.

Emergency Medicine as a specialty is still in its infancy in Pakistan [30‐32]. Our department was the first one to be established back in 2008. Over the years, we have observed an increase in patient volume as well as acuity. The ED had expanded to 46 beds but the hospital beds remained the same which brought in the issues of overcrowding, left without being seen patients and ED through put issues. Therefore, this study is aimed at defining the LWBS population in a tertiary care hospital while determining percentages and factors associated with LWBS as we do not know the characteristics of our patients who are leaving. This baseline information will be critical in developing evidence based interventions aimed at improving the health care management of such patients and consequently reducing the morbidity and mortality resulting from leaving.

A total of 38,762 patients were triaged from April to December 2010 at AKU - ED and were included in our study; 13866(35.33%) were admitted. Total 5,086 patients left the ED without getting medical care during this period, giving an overall rate of over 13% over 9 month period.

There were significant differences between those patients receiving medical care and those who left before treatment (Table 1). Percentage of LWBS visits for females was slightly higher than males but the relationship between sex of patient and LWBS visits was not significant in the multivariable regression model (Tables 2 & 3). For triage level assignments, proportion of patients in P3 categories was highest (15.7%) who left without being seen, which means that they were seven times more likely to be in a LWBS visit compared to P1 & P2 patients. The length of stay and the percentage of patients leaving were also increased with increased number of P1 and P2 patients (Figure 1). LWBS percentages seem to vary with time of the day and were more than 20% in the night shift (11 pm to 7 am) compared to about 4% in the morning shift (7am to 3pm). This finding was found to persist in the regression analysis, which revealed a 2.6 times higher odds of an LWBS visit if the patient presented to the ED in the night shift compared to the morning hours. Another important predictor of LWBS visit is the diversion status of the ED at the time of presentation. Patients visiting during the ED diversion hours are 1.5 times more likely to have a LWBS visit than when diversion status is off (19.8% vs. 9.8% during off-diversion). Sex and day of the week on which the patient presented showed an association with LWBS visits at the univariate level, but this relationship was not found after adjustment with other factors in the multivariable model (Tables 2 & 3). Percentage of LWBS is more in female (13.75) patients as compared to male (12.58).

We also observed a difference in percentage of LWBS over the total study duration, being lowest in April (6.4%) and highest in the September to October period (up to 19%) (Table 2). This pattern was consistent in the regression model which showed higher odds of LWBS visits in October (marginally significant), November and December (Table 3).

Median waiting time for pediatric patients was 154 minutes and 171 minutes for adults who left. Patients with a waiting time of over 180 minutes had 26 times higher odds of leaving compared to those who waited for less than 30 minutes. This relationship is consistent in the adjusted multivariable model (Table 3). Relationship of wait time with age, triage category diversion status and shift of the day had been shown in Figure 2.

The top seven presenting complaints of patients with LWBS visits are shown in Figure 3. Most of the cases were of fever, non-specific complaints, abdominal pain, and vomiting/ diarrhea. Co-morbid was identified in 12.6% of patients with LWBS visits. Total patients who returned to hospital within 48 hours were 181 in LWBS group vs. 251 among the patients who were discharge; which means overall 3.6% of the LWBS patients vs. 1.11% of discharged patients needed to revisit in ED for medical care. Among the LWBS, 77 (1.5%) and 6(0.26%) in the discharge group required admission to the inpatient units (Table 4).

To the best of our knowledge, this is the first study to describe the characteristics of Pakistani patients who left emergency department without being seen by a physician from a tertiary care hospital. We have described the relationship of LWBS with age, triage category, day of week, and shift of day, diversion status and waiting time. We found that in this study LWBS were 13% which is comparable to other international data (1.0 – 15%) but higher than the benchmark set by USA (1.7%) [16, 17, 33].

Although the sex of those who LWBS does not have significant effect in a multiple regression model, it appears that age of a patient had a profound impact (Table 3). The odds of leaving for male patient who is 20–40 years of age is 17 times more than a patient at extremes of age, regardless of severity of illness. Children were found to be at a lower risk of being left, this may show increased sensitivity and nonspecific sign and symptoms towards extremes of age that gives them priority over other age groups [15, 23, 34]. Proportion of LWBS are higher in females contrary to international data may be because in our community structure females have the responsibility of taking care of all the household things as well their health is not given as much priority because of existing inequities in our communities [35, 36].

It was observed that higher proportions of low acuity (98.2%) patients with less severe illnesses like fever, upper respiratory tract infection (URTI), acute gastroenteritis were leaving. Research has shown that LWBS and acuity has a dose- response relationship; with 15.2% of non-urgent patients leaving as compared to 0.1% of critical patients [10, 13, 37]. The fact that most of those patients who left although had low acuity illnesses yet they required some work up or treatment e.g. abdominal pain or diarrhea with dehydration, highlights the importance of accessibility of urgent care settings or short stay units. This could be assessed by the percentage of subsequent return visit in the ED after leaving. The return visits in our study are found to be 3.6% much higher than the internationally reported numbers (1.2%) in a USA study with 1.5% requiring hospital admission subsequently [38].

A high number of walk-in patients such as those with fever or URTI utilize the ED mostly in after hours, and usually spend a long time in waiting because of their relatively stable condition. This fact also emphasizes the need for creating structures such as fast track Clinics or urgent care centers that cater the high influx of patients with seasonal illnesses who need not be referred to a tertiary care hospital for treatment and a separate patient care area for elderly patients [39‐42].

From the results, it appears that patients who are asked to wait for a longer time period are also more likely to leave than those who are assigned bed within a relatively shorter time span. The odds for leaving in this study are 0.2% with every10 minutes increase in waiting time. Although the accuracy of waiting time duration is uncertain, percentages have been described in other studies [17, 22, 27, 29, 37, 43‐45]. Probable reasons could be either patients got tired of waiting, seek advice in another healthcare facility or they felt better and left [15, 20, 46, 47].

The contributory factors for LWBS are overcrowding due to high patient influx and boarded patients in ED, lack of awareness among general population regarding ED utilization as well as inefficient primary health care facilities [20, 39, 48‐52]. This crowding result into prolong waiting hours and ultimately increased rate of LWBS. In our institution because of lack of availability of inpatient beds in high acuity areas these patients who are either critically ill or intubated have to stay in the ED at times for more than 24–48 hours before their final disposition. The situation further worsen when more and more critical patients continue to land in the emergency department with limited resources like nursing staff and beds available. It is a proven phenomenon that when ED was crowded and on diversion there was 2.26 times risk of leaving the ED. Similar results have been reported by TL Viet and K V Rhodes that ED crowding increases the LWBS rate [21].

Increased percentages of LWBS during weekend or night shift and seasonal variations gives insight into epidemics such as dengue fever, inadequate outpatient services on the weekends and after hour’s utilization of ED services for minor illnesses [16, 21, 51, 53‐56]. Our data had demonstrated a sudden increase in LWBS patients in the third quarter that coincide with the dengue epidemic of 2010 in Pakistan [57, 58]. A strong seasonal variation with highest LWBS (up to 70%) in winters is also found in other studies [45].

There are certain limitations of this study. First data were collected retrospectively. Secondly the study was conducted in a single tertiary care private hospital therefore results may not be generalized. Our department is the first in Pakistan to practice a defined triage system which started recently. Very little is known about the reliability and validity of the triage at our institute. This is the first ever analyzed data from AKUH-ED. Follow up studies are needed to address this issue in detail.

The cross sectional design of study did not enable us to follow the clinical outcome of LWBS patients in detail. Additional studies are required to determine subsequent morbidity and mortality as well as other hospital factors affecting the percentage of LWBS. As all the patients are not the registered patients at AKUH, so the return visit of all the patients who had been triaged cannot be traced for any adverse outcome. This was the first reported data so we haven’t studied the different age group characteristics separately. Subsequent studies on pediatric, adult and geriatric patients are needed to further elaborate their characteristics and factors affecting their decision of leaving.

In this study, we found certain factors were strongly associated with LWBS such as age, low triage acuity, weekend, and night shift presentation and prolonged waiting times. ED Diversion was also associated with higher odd of leaving. This is only a single center data from a private tertiary care hospital and figures could be different in other public or private settings.

Strategies should be designed to shorten the waiting time and additional outpatient facilities such as fast track clinics to reduce the burden of these patients from ED and avoid possible bad outcome in this population who miss the opportunity of health care provision due to weak primary care facilities.