Our QI project was conducted at Vancouver General Hospital (VGH), Vancouver, Canada, a tertiary referral centre performing approximately 700–800 cardiac surgical cases annually, between December 2020 and November 2021. At VGH, postoperative care of patients undergoing cardiac surgery takes place initially in the CSICU by anesthesiologists, CSICU nurses, and cardiac surgeons, and later on the cardiac surgery ward by cardiac surgeons, nurse practitioners, and ward nurses. Our initial QI team was composed of individuals with an interest in QI work from anesthesia, nursing, and pharmacy.

Postoperative AF was defined as “clinically important POAF,” defined as AF or flutter that results in angina, heart failure, or symptomatic hypotension, or that requires treatment with a rate-controlling drug, antiarrhythmic drug, or electrical cardioversion. The baseline rate of POAF in patients undergoing cardiac surgery at VGH was established using data collected by Heart Services British Columbia, a central database that collects outcome data for cardiac surgery in the province of British Columbia, and retrospective analysis of discharge records from the year preceding our study—1 January 2019 to 31 December 2019. According to Heart Services BC data, the rate of POAF in patients undergoing cardiac surgery at VGH was 33.0%.

A Project Ethics Community Consensus Initiative Screening Tool was completed, and this QI project was deemed minimal risk to patients.Footnote 1 As per the Tri-Council Policy Statement 29 and endorsed by the Vancouver Coastal Health Research Institute, a formal research ethics board approval was therefore not required for this QI initiative.

Prior to implementation, the goals of our project were discussed among members of our cardiac anesthesia, cardiac surgery, pharmacy, and cardiac nursing teams. All of these stakeholders were supportive. The plan was to develop a stepwise preprinted order set (PPO) that would support the clinician in identifying patients at elevated risk of POAF and then prescribe appropriate POAF prophylaxis. The first step of the PPO would risk stratify the patient using an evidence-based POAF risk scoring system, after which the second step would employ a protocolized prescription of beta blockers and/or amiodarone based on the patient’s calculated risk. Buy-in from clinicians was achieved by making the action of prescribing prophylaxis for POAF as easy as possible. We did this by making the PPO as clear and concise as possible to ensure a high completion rate. We also asked all bedside nurses in the CSICU to ensure that the PPO was completed for every patient to ensure a high compliance rate.

Preimplementation planning phase

We developed a 16-point POAF risk stratification system to be applied to all cardiac surgery patients (Fig. 1). Our POAF scoring system was composed of both patient and surgical risk factors, as recommended by the SCA/EACTA Practice Advisory.1 We also performed a literature review comparing the POAF incidence according only to surgical subtype. The POAF risk scoring system was designed to categorize all patients as either “normal risk” or “elevated risk,” in accordance with the SCA/EACTA Practice Advisory.1 We also developed an amiodarone prescribing protocol for patients deemed to be at elevated risk for POAF. The risk scoring system and the amiodarone dosing protocol were developed following a literature review. This review highlighted 25 relevant studies, five of which referred to risk analysis for POAF and 19 of which addressed the question of amiodarone dosing and timing for POAF prophylaxis.10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32 After developing the PPO, we held focused group meetings with all stakeholders, including representation from anesthesiology, cardiac surgery, nursing, and pharmacy, to finalize the PPO. This collaborative approach was designed to maximize clinical staff buy-in. The dosing protocol, similar to that for POAF treatment, was reviewed and approved by the VGH institutional Pharmacy Safety Committee. Contraindications to the use of amiodarone were included in PPO as well as baseline blood work to monitor for hepatic complications from amiodarone administration.

Fig. 1

Preprinted order set for prevention of postoperative atrial fibrillation in cardiac surgery. Page 1 contains the POAF risk stratification scoring template and prophylaxis prescribing tool. Reverse side of single page PPO (Page 1A) describes suggested dosing, contraindications, and reasons to discontinue prophylaxis.

bpm = beats per minute; COPD = chronic obstructive pulmonary disease; CSICU = cardiac surgery intensive care unit; eGFR = estimated glomerular filtration rate; HR = heart rate; LVEF = left ventricular ejection fraction; POAF = postoperative atrial fibrillation; POD = postoperative day; PPO = preprinted order set; TTE = transthoracic echocardiography

The dosing and timing of amiodarone administration was considered in the preplanning phase. Studies have shown that both pre- and postoperative administration of amiodarone were effective in preventing POAF after surgery.25,26 Therefore, the PPO indicated that amiodarone prophylaxis for elevated risk patients could begin within the first 24 hr of admission to the CSICU postoperatively. This would provide an opportunity for the anesthesiologist, surgeon, and CSICU clinicians to make collaborative serial assessments on the risk vs benefit of starting amiodarone for a particular patient. The amiodarone protocol consisted of an iv loading dose of amiodarone 150 mg iv over one hour followed by an amiodarone iv infusion at 1 mg·min–1 for six hours, then a 0.5 mg·min–1iv infusion for 18 hr. After the 18-hr iv loading regimen, oral amiodarone 200 mg bid was prescribed for seven days (or until hospital discharge), to include the most high-risk period for developing POAF.

The PPO prompted the CSICU clinician to commence beta blocker therapy as soon as possible postoperatively for all cardiac surgery patients (Fig. 1). Clinicians were already familiar with prescribing postoperative beta blockers, as it was an informally established practice at our centre, although details about how consistently and how soon after surgery were unknown. Our PPO suggested continuing the patient’s own beta blocker if already on established therapy, otherwise metoprolol 12.5 mg po bid was recommended. An alternative more cardioselective beta blocker, bisoprolol, could be initiated if there were clinical concern for respiratory disease. All patients undergoing cardiac surgery at our CSICU are eventually prescribed beta blockers postoperatively where tolerated; however, it was unknown how soon postoperatively this was being initiated. The PPO therefore highlighted the importance of starting beta blockade as early as deemed tolerable, to provide maximum POAF prophylaxis. Relative contraindications to beta blockade were also listed for consideration by the prescribing clinician (Fig. 1).

Data collection/analysis

Data collection took place in the form of paper audit sheets attached to the patient’s chart and were completed in the CSICU by bedside nurses and on the postoperative cardiac surgery ward by one of two nurse practitioners prior to patient discharge. Paper audit sheets did not contain any identifying information for the patient and were stored in a secured area in the local research office. Patient data were then manually inputted to a secure online database, REDCap®Footnote 2 (Vanderbilt University, Nashville, TN, USA), for secure online storage.

Quality improvement projects that aim to make a process or system produce better results benefit from using “graphical methods of analysis [that] reveal the trajectory of system behaviour over time” such as control charts (or Shewart charts).33 Control charts allow identification of variation due to special causes (as opposed to common causes, which result in random variation in a process or system) through the use of rules.33 By reviewing the incidence of POAF over time in consecutive groups of ten patients, we could observe whether our PDSA cycles were making a difference by following the data chronologically. The control chart for POAF incidence over time was built using an Excel software add on (QI charts, Process Improvement Products, pipproducts.com, San Antonio, TX, USA) according to previously published QI methodology.34 The data were grouped in subgroups of ten consecutive observations to smoothen the variation in the data (e.g., variation in volume of daily cardiac cases over time), while also allowing a reasonable number of observations to analyze. According to previously published QI methodology, five rules were used to analyze the control chart and identify special causes and process changes.34

Plan-Do-Study-Act cyclesPlan-Do-Study-Act cycle 1: 1 December 2020 to 28 February 2021—implementation phase

The first Plan-Do-Study-Act (PDSA) cycle involved a multidisciplinary collaboration to educate clinicians about the importance of POAF prophylaxis, and to review the PPO and gain multidisciplinary approval. Cardiac surgeons and anesthesiologists were informed of our QI initiative and feedback was solicited. An educational drive was conducted with CSICU nursing staff to increase familiarity with the PPO and optimize completion and accuracy of data collection. This was essential in capturing patients in whom prophylactic amiodarone would be suitable. Our initial POAF score threshold for patients at elevated risk, who should receive amiodarone prophylaxis, was 6 or greater during this phase.

Plan-Do-Study-Act cycle 2: 28 February 2021 to 1 June 2021—elevated risk threshold decreased

Upon completion of PDSA cycle 1, a preplanned review of our preliminary data at three months revealed that very few patients met the POAF risk criteria of six points or greater on the risk scoring system, and therefore very few patients qualified for amiodarone prophylaxis (8/150 patients). It also highlighted that, when patients did meet the threshold, the amiodarone protocol was rarely adhered to. A preliminary analysis of the POAF risk scoring system during PDSA cycle 1 revealed a consistent increase in the incidence of POAF to greater than 40% in patients with a POAF score of 2 or greater (Fig. 2). We therefore decreased the POAF risk score to redefine elevated risk as a score of 3 or greater. Improved staff engagement was achieved by discussing the QI project at the Cardiac Anesthesia Divisional Meetings and further nurse education at the bedside. We also posted POAF infographics in clinical areas describing the incidence of POAF, clinical consequences of POAF, and recommendations for initiating amiodarone prophylactically (Electronic Supplementary Material [ESM] eAppendix 1).

Fig. 2

Incidence of postoperative atrial fibrillation (POAF) after cardiac surgery according to the POAF risk score

Plan-Do-Study-Act cycle 3: June 2021 to November 2021—improving early postoperative beta blocker prescription

Summative data analysis at the preplanned six-month time point (i.e., at the end of PDSA cycle 2) showed increased use of prophylactic amiodarone but also highlighted a lack of improvement with early beta blocker administration postoperatively (see Results section below). To understand why, the data collection tool was modified so clinicians would document a reason for withholding beta blockers beyond postoperative day (POD) 1. Interventions intended to expedite beta blocker prescription included a nurse-led bedside checklists for readiness for initiation of beta blockade, and a new CSICU ward round mandate to review each patient's readiness for beta blocker initiation every 12 hr (ESM eAppendix 2). Also, the clinical pharmacist was made responsible for asking the question whether beta blocker restart was appropriate during every patient review on CSICU ward rounds.