Introduction
There has been increasing interest in determining why patients use the emergency department (ED) for non-urgent or low-acuity Canadian Triage and Acuity Scale (CTAS) level 4–5 visits. In Ontario, reducing ED wait times is a key component of the provincial ED and Alternate Level of Care strategies. In the Waterloo Wellington Local Health Integration Network (WWLHIN), low-acuity CTAS 4–5 ED visits are considered to significantly contribute to increased ED wait times. The WWLHIN identified the need to better understand the reasons for CTAS 4–5 patient visits as a key area for future research. A WWLHIN report indicated that “[45%] of total ED visits were non-urgent (CTAS 4–5), [and] these people could have been managed in another setting”. 1 The stated goal of the WWLHIN is to decrease CTAS 4–5 ED utilization by 10% and ensure appropriate use of ED resources. 1 A 2005 Canadian Institute for Health Information (CIHI) study investigating ED wait times found that 57% of ED visits were for less urgent (CTAS 4) or non-urgent (CTAS 5) reasons. 2
The Centre for Family Medicine is a family health team (FHT) in the region of Waterloo comprised of family physicians, interdisciplinary health care providers, support staff, and medical learners (residents and medical students). At the time of the study, the team consisted of 16 family physicians who had a total patient roster of about 20,000 patients. During 2011, exact clinic hours of operation varied by physician. Most physicians worked Monday to Friday, and clinics ran between 8 am and 5 pm. After-hours clinics were available Monday to Thursday, 5 pm to 7 pm; Friday, 3 pm to 5 pm; and Saturday, 10 am to 1 pm. During 2011, there was no after-hours clinic available on Sunday. Outside of clinic hours, patients had access to a telephone health advisory service (THAS); and a resident and staff physician were on call if THAS nurses had questions that required a physician. There was variability among the 16 physicians in the FHT in terms of same day and next day bookings. In 2012–2013, the FHT took steps toward advanced access by reducing the time to next available appointment and ensuring that physicians kept more appointment slots available for same day bookings.
In 2011, the WWLHIN had a patient population of 775,000, with 277,880 ED visits or 366.6 ED visits per 1,000 patients.Reference Glazier, Zagorski and Rayner 3
Currently, very little is known about ED use by patients rostered to FHT in Ontario. A recent report comparing primary care models in Ontario showed that there was typically one visit per FHT patient per year in 2008/09 and 0.9 visits per FHT patient per year in 2009/10. 4 By better understanding why FHT patients present to the ED, opportunities can be identified for primary care quality improvement that mayhelp reduce low-acuity visits to EDs in Ontario.
The primary aim of this study was to determine the characteristics of CTAS 4–5 patients and develop a predictive model of low-acuity ED utilization. The secondary aim was to compare this information with characteristics of CTAS 1–2 patients and develop a predictive model of high-acuity ED utilization. The study also sought to determine what factors were predictive of leaving without being seen (LWBS).
Methods
This retrospective correlational study explored characteristics of low-acuity ED utilization by patients enrolled in an FHT and factors predictive of low-acuity ED utilization. Research Ethics Board approval was obtained through Hamilton Health Sciences and McMaster University, Faculty of Health Sciences.
At the Centre for Family Medicine, ED face sheets are scanned into the electronic medical record (EMR) and labeled as “emergency physician/emergency medicine.” The term “emergency physician/emergency medicine” was searched in which the occurrence date was later than January 1, 2011 and earlier than December 31, 2011 (using the search feature in Practice Solutions Software Inc., 2002–2012 - PS Suite 5.1; 243-A, Jul 20, 2012). The last recorded ED visit in 2011 was used to extract data. The total number of ED visits during 2011 per patient was recorded. The sample included all Centre for Family Medicine FHT patients with ED visits in 2011 (n=1580); there were a total of 2230 visits.
The dependent variable was low-acuity (CTAS 4–5) ED visits.Reference Beveridge, Clarke and Janes 6 – Reference Bullard, Unger and Spence 8 Independent variables included: age, sex, time of day, day of the week, presenting complaint, number of ED visits, diagnosis, and result of visit.
The presenting complaint was recorded as it was written on the ED record. The data were coded using the Canadian Emergency Department Information System (CEDIS) Presenting Complaint List (V2.0 – April 2012)Reference Moe, Bailey and Oland 9 , and the presenting complaints were used for analysis. The final diagnosis was recorded in a similar manner, and the data were coded using the Canadian Emergency Department Diagnosis Shortlist (CED-DxS) (V.2.0 – April 2012) 10 , and the diagnoses were used for analysis.
After evaluating for normality, basic descriptive statistics were used to describe demographics, time of day, result of visit, and presenting complaints of the targeted population of low-acuity (CTAS 4–5) ED use compared to high-acuity (CTAS 1–2) ED use.
Variables that were commonly typed rather than handwritten were preferentially selected in the scanned copies of the ED record, since extracting variables from the handwritten ED record was difficult. As a result, some variables were missing (i.e., diagnosis and result of visit [disposition]). Variables typically typed (i.e., computer-entered rather than handwritten) include sex, age, day of the week, time of day, and CTAS level.Reference Lin and Worster 5 After initial descriptive statistics were completed, selected variables were tested for predictive value with univariate and multivariate logistic regression models. Variables were selected based on clinical relevance: age (≥65 versus <65 years of age), sex (male versus female), time of visit (between 7 am and 9 pm versus between 9 pm and 7 am), and the number of visits (≥3 versus <3 visits). Adjusted odds ratios were calculated for CTAS levels 4-5 based on age cutoffs, while controlling for sex, day of the week of the visit, and whether the patient was a frequent ED user with ≥3 ED visits in one year. Adjusted odds ratios were calculated for CTAS 1–2 patients based on age cutoffs, while controlling for sex and whether the patient was a frequent ED user.
Predictors of theoretical relevance were tested by univariate logistic regression models and incorporated into the final multivariate logistic regression model to generate odds ratios predicting low acuity and high acuity ED utilization and LWBS. All analyses were performed using SAS statistical software version 9.3 (SAS Institute, Cary, NC).
Results
In 2011, there were 1580 patients who visited the ED. Females accounted for 49.9% of the study population; males accounted for 49.8%. The patient ages ranged from 1 month to 99 years of age, with a mean age of 43 (Figure 1). CTAS 1–3 accounted for 56% (n=884) of visits, CTAS 4–5 accounted for 24% (n=387) of visits, and 20% (n=309) of visits had no CTAS level recorded (Table 1).
Analysis of time of day of ED use showed a first peak at 11 am and second peak at 7 pm for CTAS 4–5 level patients (Figure 2). There was no significant difference in ED use by day of the week, however, there was a slight increase in ED use on Sundays and Mondays (Figure 3).
The primary reason for patient presentation to the ED was cardiovascular complaint, followed by orthopedic and gastrointestinal complaints (Figure 4). The majority of patients (n=1,200) had one visit to the ED during the study period, 221 patients had two visits, and 74 patients had three visits. The highest ED use by a single patient during the study period was 25 visits.
The leading diagnostic categories for CTAS 4–5 patients were injury, poisoning, and other (n=121). The next most common diagnostic category was diseases of the musculoskeletal system and connective tissue (n=56) (Figure 5).
The result of each ED visit and final disposition was determined (Figure 6). The majority (71.8%) of patients were discharged home from the ED (n=1135). Admissions accounted for 4.4% (n=69) of ED visits, 4.6% (n=72) of patients LWBS, and 1.1% (n=17) of patients were placed in a clinical decision unit (CDU) with the expected disposition of “discharge”, after obtaining diagnostic test results during a period of observation. A small number of patients were transferred to another institution, 0.6% (n=9). Patients leaving against medical advice accounted for 0.3% (n=4) visits. During the study period, one ED patient death was recorded (0.1%). No disposition was recorded for 17.3% (n=273) visits.
Age greater than 65 years was associated with lower ED use by patients triaged to CTAS 4–5 (OR 0.61 [0.44–0.83], p=0.0017) (Table 2). Differences between sexes, day of the week, and total number of visits by a patient were not found to be statistically significant predictors in the CTAS 4–5 patient group.
* Adjusted Odds Ratio: controlling for sex, day of the week of the visit, and whether the patient was a frequent ED user.
Multivariate logistic regression analysis showed that age > 65 also predicted ED use in CTAS 1–2 patients (OR 1.75 [1.28–2.38], p=0.0005). Differences between sexes, day of the week, and total number of visits by a patient were not found to be statistically significant in the CTAS 1–2 patient group (Table 3).
* Adjusted Odds Ratio: controlling for sex and whether or not the patient was a frequent ED user.
Patients who went to the ED during the day were less likely to LWBS compared to patients who went at night (OR 0.70 [0.53–0.91], p=0.0173). Patients who went to the ED at night were 1.4 times more likely to LWBS compared to patients who went to the ED during the day (Table 4).
* Adjusted Odds Ratio: controlling for sex, time of day and whether or not the patient was a frequent ED user.
Discussion
In 2011, there were 2230 ED visits by 1580 patients, representing approximately 7.9% of the FHT practice roster. Most patients had one visit and were discharged home from the ED. A small percentage of patients had multiple ED visits, which tended to be for high and moderate acuity presenting complaints such as chest pain, shortness of breath, and abdominal pain.
Much discussion in the literature has focused on trying to identify patients with low-acuity presenting complaints to see if they might be safely diverted away from the ED. 11 – Reference Adams 14 A recent study demonstrated limited concordance between presenting complaints and ED discharge diagnoses, suggesting that presenting complaints are unable to accurately identify non-emergent ED use. 11 The study found that only 6.3% of patients presenting to the ED had primary care-treatable diagnoses based on discharge diagnosis and an algorithm. 11
Another study examined reasons for patient presentation to the ED in Edmonton, Alberta.Reference Raven, Lowe and Maselli 12 The study found that many patients made concerted efforts to avoid ED visits; 61% of patients sought alternative care before visiting the ED.Reference Raven, Lowe and Maselli 12 The role of health professionals and systems in driving ED overuse must be considered. Strikingly, 47% of patients who sought alternatives to the ED made contact with a physician’s office.Reference Raven, Lowe and Maselli 12 Patients who called a physician’s office were directed to visit a family physician (3%), a health care professional (9%), or the ED (63%). The study also found that 14% of patients contacted a regional health line.Reference Raven, Lowe and Maselli 12 One study found regional health lines recommended going to the ED (58%), seeing a primary care physician (6%), or seeing another health care professional (3%).Reference Glazier, Zagorski and Rayner 3 Patients who called a physician’s office or regional health line were almost ten times more likely to be directed to visit the ED than a family physician.Reference Glazier, Zagorski and Rayner 3
Limitations of the study include the relatively small sample size and retrospective chart review design. Because the study relied on searching the FHT EMR for scanned ED face sheets, all ED encounters may not have been captured. Another limitation, which may affect generalizability, is that all patients were from a single FHT located in Kitchener, Ontario, Canada. All visits were during 2011, so there might be variation year to year in terms of patterns of ED use. The study did not account for prior health status of patients, which could be a confounding factor. Because not all ED visits were subject to data extraction due to time and resource limitations, it was decided to gather more variables as potential predictors from fewer patients, rather than fewer variables for more patients. This method did not consider that a lack of relationship between reasons for a patient’s ED visits. Finally, the study excluded patients aged 0–4 years of age (Figure 1), because the graph displays the ratio of ED visits compared to numbers of patients of that particular age group in the FHT. The confounding issue in the 0–4 age group is that a number of the physicians provided newborn care to children who were not rostered as part of the FHT. These patients were captured in the EMR census data, which artificially inflated the number of young patients. Data from children aged 0–4 years were included in the predictive models and included when determining the descriptive statistics of the study population.
Further research could make use of a prospective methodology to see if similar patterns of ED utilization are seen across multiple years or at other FHTs throughout Ontario. This methodology could also be used to compare ED utilization by patients in other models of primary care. For patients with multiple visits, a method to randomly sample an ED encounter for analysis might minimize any risk of bias. Future research could also look at the use of the FHT before and after the ED encounter. Additional research could determine if there was a relationship between high and low use of FHT resources and visits to the ED. Another avenue for future research is to compare FHT patients who had ED visits to FHT patients who did not have ED visits. For patients with multiple visits, future work could determine whether there is a correlation within each patient for the urgency of his or her visits.
The study findings should help inform decision-making regarding the provision of care for FHT patients. The findings of this study suggest a number of improvements, such as providing simple laceration repair during clinic and after-hours clinics. The FHT should consider repatriating patients on anticoagulation therapy for venous thromboembolism for point-of-care INR testing, rather than requiring return visits to the ED. Other changes to consider include shifting on call clinics to start at 7 pm to meet peak demand and educating patients about expected wait times in ED for low-acuity complaints. Implementing these suggested changes could provide patients with better alternatives to the ED; these targeted interventions could better serve FHT patients and help reduce low-acuity ED use.
Conclusions
Most low-acuity ED use was by patients under the age of 65, whereas high-acuity ED use was more common among patients with an age greater than 65. Patients are more likely to LWBS during the night (9 pm to 7 am). These findings suggest that patients younger than age 65 are most likely to go to the ED with low-acuity or non-urgent complaints.
Competing Interests: None declared.