ABSTRACT
Introduction The emergency department (ED) is under pressure to meet length of stay (LOS) metrics for care in the ED. An aspect that we propose affects LOS is the order for urine sample collection and subsequent urinalysis (UA) as both are time consuming steps. This project's primary goals are to determine if ordering a UA increases LOS and how often UA contributes to clinical decision-making and/or disposition decisions in the ED. Secondary objectives were to identify factors that contribute to the ordering of a UA and to decipher if LOS was more impacted in patients who were discharged vs. admitted to the hospital. Methods Retrospective chart review was conducted of patients who presented to our ED in April 2016 during 12 consecutive days. Data were abstracted onto a data collection sheet with the abstractor blinded to study hypotheses. Variables included whether a UA was ordered, times of UA order and result, who ordered the UA (mid-level provider [MLP] vs. physician), whether the UA was cancelled, whether the UA result influenced clinical decision-making (based on the medical decision-making section of the physician chart) or disposition decision, LOS, age, and gender. Descriptive statistics and multivariable regression analysis were used to analyze relationships between the variables collected and their influence on LOS. Results The overall median LOS was 157 minutes, with an interquartile range (IQR) of 81 to 246 minutes. For discharged patients, it was 142 minutes, with an IQR of 46 to 236 minutes. For admitted patients, it was 177 minutes, with an IQR of 118 to 260 minutes. Amongst admitted patients, multivariable regression analysis demonstrated that the following factor was associated with increased LOS: being seen first by the provider-in-triage (PIT) then physician in main ED (p < 0.0001). Amongst discharged patients, multivariable regression analysis demonstrated that the following factors were associated with increased LOS: being seen first by the PIT then physician in main ED (p = 0.0296), being seen by MLP only (p < 0.0001), having a UA ordered (p = 0.0005), being seen on weekend (p = 0.0166), and being an older patient (p = 0.0475). The UA was cancelled in 9% of our patients, and in 60% of cases, these UAs were ordered by the PIT. Patient disposition decision was made prior to UA resulting in 60 cases (25%). The UA was used in clinical decision-making in 118 cases (66%). The following predictor factors were associated via univariate analysis with using a UA for decision-making: being female (p = 0.0050, 95% CI: 0.0068-0.378), being an older patient (p < 0.0001, 95% CI: -0.010 to -0.004), being first seen by the PIT and then a physician (p = 0.0486, 95% CI: 0.0048-0.1555), and discharged patients (p < 0.0001, 95% CI: -0.6749 to -0.4487). Conclusion Our results suggest that having a UA ordered increased ED LOS, especially in patients who are ultimately discharged. In our ED, routine UAs are ordered more often by MLPs than physicians. A routine UA may not impact clinical decision-making up to 33% of the time, nor alter disposition decision one out of four times. Given that 9% have the test eventually cancelled, one should reconsider the utility in ordering routine UAs in ED patients, as they increase LOS and place an additional burden on the patient and the ED personnel.
ABSTRACT
Objective The objective of this study was to improve sepsis bundle compliance via an educational intervention in our emergency department (ED). Methods This was a before and after study. Historical data on sepsis bundle compliance was obtained from our quality officer. Data were collected for 30 consecutive days to compare sepsis bundle compliance rates before and after the intervention. Descriptive statistics were compiled, and the z-test for proportions was used to calculate statistical significance. The intervention was two-fold: 1) a bright yellow card with sepsis criteria listed was posted on all ED workstation computers and 2) there was a daily email blast for one month with "sepsis facts." These email blasts were short pearls that highlighted the importance of recognizing and treating sepsis. Results The sepsis bundle compliance rates in the month prior to the intervention was 38%. In the month during the targeted intervention, the compliance rate increased to 56%. There was a statistically significant increase in bundle compliance rates during the intervention (p=0.0399). We also administered a survey to the ED attendings and residents following the completion of the study to assess whether they perceived that our intervention was helping them increase compliance with ordering the sepsis bundle. The response rate was 94%. To the question "Did you feel the sepsis cards placed on the workstations make you more likely to consider sepsis earlier in patients under your care in the emergency department?" 70% answered agree or strongly agree. To the question "Were you more likely to order the sepsis bundle after receiving the daily "Sepsis Facts"?" 29% were neutral while 59% answered agree or strongly agree. Finally, to the question "Did you feel the sepsis cards and "sepsis facts" help you improve the care of Septic patients in the emergency department?" 76% answered agree or strongly agree. Conclusion Sepsis criteria reminders and email blasts highlighting the importance of treating and recognizing sepsis can improve compliance with sepsis bundle ordering within the emergency department.
