Journal club at the Mercy St. Vincent EM Residency takes place once a month, and we discuss up to four potentially practice-changing papers. The discussion includes all residents and is led by senior residents or our research director, Dr. Michael Plewa. These journal club posts are designed to provide you with a deeper dive into the paper, addressing what the investigators did, the strengths and weaknesses, and the outcomes.
Demeester S, Hess RA, Hubbard B, et al. Implementation of a Novel Algorithm to Decrease Unnecessary Hospitalizations in Patients Presenting to a Community Emergency Department With Atrial Fibrillation. Academic Emergency Medicine. 2018;25(6):641-649.
Reviewed by: Breanne Jaqua, DO, MPH, PGY-3
Introduction
Atrial fibrillation is common, and usually treated with hospital admission, but admissions are expensive
No clear consensus on ED management for this condition
Study aims to see if hospital admissions can be reduced (saving money) without compromising patient outcomes
A simple algorithm was created and used to determine whether or not patients with a rapid heart rate (>100 beats per minute) in atrial flutter or atrial fibrillation could be safely discharged for follow up in specialized Afib clinic hosted by cardiologists
What did they do (PICO)?
Type of Study: retrospective, observational, before-after study at a 537-bed academic community ED
Population
Patients presenting with new or recurrent atrial fibrillation or atrial flutter to the ED
All acuities were included in the study
Patients already on anti-coagulation were included in the study
Excluded from the study: alternate primary diagnoses (ie: sepsis), secondary diagnosis of atrial fibrillation, less than 18 years old, pregnant, incarcerated
Intervention
Algorithm: >Patients were admitted if they had any of the following high-risk features: - Hemodynamic instability - Acute heart failure - Acute coronary syndrome - Syncope > Patient’s without high risk features could be sent home with follow up in a novel > Atrial Fibrillation Clinic that was established in conjunction with the cardiology group within 3 business days after discharge from the ED > Synchronized cardioversion was performed if they met CHA2DS2VASc and INR criteria > ED providers were trained to use the algorithm prior to the study start
Medications: > Discharge with either: - Metoprolol 50mg BID or - Diltiazem 120mg or 180mg QD > Rate control IV bolus medications were optional per provider discretion in the ED > IV infusions were discouraged > No anticoagulants started in ED
Goal heart rate prior to discharge: <110 beats per minute
Comparison - Patient cohort gathered in the “pre-intervention phase”
Outcomes - Return ED visits - Adverse events - Admissions
Results
Patient population > 586 patients in control group (algorithm NOT used), 522 patients in intervention group (algorithm used) > No significant differences in patients’ age, sex, race, or acuity > Presence of CHF was the only comorbidity that was significantly different between the 2 groups. (24.6% control group, 34.3% intervention group)
Primary Outcomes > Whole study population: · Admission rates were reduced from 80.4% to 67.4% = absolute reduction of 13% > Sub-group of only lower-acuity patients (Pmort30 scores of 3, 4, or 5): · Admission rates were reduced from 63.6% to 43.7% = absolute reduction of 19.9%
Secondary > ED return visits for any complaint were similar among groups: · Within 3 days: 1.19% control, 1.0% intervention · Within 30 days: 3.8% control, 3.6% intervention > Deaths: none (control & intervention groups) > Of the patients who were discharged home in the intervention group, no one returned for an additional ED visit within 3 days, and only one patient had a repeat ED visit within 30 days > ED length of stay for patients discharged was similar among groups: · Control: 4.73 hours · Intervention: 4.60 hours > 88.5% of patients followed up at the atrial fibrillation clinic within 3 business days
Strengths
Provides an alternative to ED cardioversion which is resource intensive and has adverse events
The algorithm is simple for ED providers to use
Fairly large cohort sizes
Limitations
Admission criteria were chosen by cardiologists based on their clinical experience (ie. not validated)
Retrospective analysis
Before-After study
Study performed at only one hospital (questionable external validity)
Adverse events were inferred, not formally measured, by 3-day and 30-day ED return and readmission rates. Returns visit data was limited to the health care facility that performed the study.
Rhythm control was not part of the study
Discussions & Conclusions
Author’s Conclusions
“Our hospital’s atrial fibrillation initiative highlights the use of a practical, uncomplicated algorithm that demonstrated success in decreasing hospital admissions and preventing return visits to the ED. We believe components of our approach can be adopted and modified by other academic or community hospital systems, leading to improved resource utilization and significant cost savings.”
They recommended additional study of the acuity stratification used to classify patients as high vs low risk
Our Conclusions
Awesome idea: simple algorithm, cost efficient, safe for patients
Would require significant buy-in from the local cardiology groups to establish an atrial fibrillation clinic with such timely follow up.
Editor’s Commentary by Dr. Alex Dzurik:
Rate control rather than rhythm control has often been the mainstay of management of atrial fibrillation in the emergency department as it has been well-established that patients are unreliable in determining afib onset. There is also significant variance in cardiologist opinion of ED cardioversion (some are very much supportive, some are horrified). This algorithm does offer a fairly simple way to implement criteria for safe discharge of people presenting in afib. There are a few significant caveats to note that limits the external validity of this study, however. As this was a pre/post implementation observation trial there is risk of differences in the patient populations studied having demographic differences as a cause of the results rather than the intervention. In fact, this study does show differences in patient population incidence of CHF. CHF is a very important issue to consider as this will significantly alter your consideration for patient risk and severity, and it limits management options. It is well researched that diltiazem should NOT be used in patients with EF <35% as this has demonstrated increased mortality due to the negative inotropy. This difference did not seem to show significant negative change in outcomes in the study, but differences in the cohorts do risk the data being skewed. Additionally, as pointed out by Dr. Jaqua, this requires the infrastructure be in place before implementation of the protocol. The other problem with the external validity is the patient population studied. This is a community hospital setting, and these results may not be generalizable in urban centers where even if the reliable follow up exists, patient adherence to management, follow up, and recommendations can still be low. The study is promising, but definitely requires further study before its ready for prime time as a standard protocol.
Post peer-reviewed by: Alex Dzurik, MD; Academic Faculty, Mercy Health St. Vincent’s Emergency Medicine Residency