This study was part of a quality improvement project conducted at Catharina Heart Centre, Eindhoven, the Netherlands—a high-volume, non-academic hospital performing ~600 PVIs annually, including both primary and redo procedures. Consecutive patients treated between December 2023 and May 2024 were included. Data were collected prospectively during routine care, while the analysis was performed retrospectively after completion of both cohorts.

Eligible patients (≥ 18 years, fluent in Dutch) undergoing catheter ablation for AF were included.

All patients underwent ultrasound-guided femoral vein catheter ablation with uninterrupted anticoagulation and activated clotting time > 300 s after transseptal puncture. For first-time catheter ablation, pulsed field ablation was conducted via a 16.8F sheath and a 6F sheath for a diagnostic catheter. Pulsed field ablation was conducted with the Farawave catheter (31 mm, Boston Scientific, Marlborough, MA, USA). Redo catheter ablation utilised 3D mapping systems (ENSITE or CARTO) via 10.7F and 12.1F sheaths. Pericardial effusion was routinely checked post-procedure using ultrasound.

From December 2023 to February 2024, MC was the standard for all consecutive patients, followed by six hours of pressure bandaging and bed rest, and two hours of observation before discharge.

From February to May 2024, CDs became the preferred groin management method. Patients were treated with two suture-mediated Perclose ProGlide devices (Abbott, Plymouth, MN, USA), followed by two hours of bed rest and one hour of observation. Discharge was permitted if no complications were observed. The goal was same-day discharge (SDD).

Eligible nurses were adults (≥ 18 years, cardiology ward staff) and were informed about the study by their department manager.

Primary outcomes were process indicators (hospitalisation duration, procedure time, (prolonged) post-procedure bed rest, unplanned overnight stays, and pain medication use) and patient experience (pain ratings, pain-related complaints, and other complaints). Procedure time was defined as the duration from patient entry into the catheterisation laboratory to exit upon completion of the procedure. Unplanned overnight stays were defined as unscheduled hospital admissions following the procedure and were calculated as a proportion of patients who were not planned for overnight observation. Planned overnight stays, which occurred more frequently in the MC group, were not considered part of this outcome. Process indicators were extracted from electronic medical records.

Patient experience was assessed through a 14-question survey on pain, information delivery, nursing care, discharge process, and an overall hospital stay rating (Supplementary materials 1). The questionnaire was developed by the study team and reviewed by experts. Pain was measured using the Numeric Rating Scale (NRS) [13].

Secondary outcomes focused on clinical outcomes (general and bleeding complications), staff experience, and costs. General complications were defined as hypotension, respiratory desaturation, and neurological complaints (e.g., coordination disorders). Bleeding complications were classified using the Bleeding Academic Research Consortium system, [14], though types 3a, 3b, and 4 could not be subtyped due to missing haemoglobin measurements. Clinical outcomes were extracted from electronic medical records.

Staff experience was assessed through a 21-question nurse-reported survey on perceived workload, task changes, and confidence in SDD (Supplementary materials 2). The survey was developed by the study team and reviewed by experts.

Hospitalisation costs were obtained from the Dutch Healthcare Authority (Nederlandse Zorgautoriteit) Open Data portal, which provides national average prices of all healthcare services, using 2024 price levels [15]. These encompass the full care episode from admission to discharge, including procedural costs. The cost of CDs was calculated using the average Dutch price (€ 150 per device). Personnel costs for catheterisation laboratory and nursing staff was derived from the Dutch national collective labour agreement, respectively € 4842/month (group 55, grade 12) and € 4324/month (group 50 grade 12) [16]. Cardiologists’ income was set at € 150/hour. The analysis aimed to evaluate whether the use of CDs, despite their additional material cost, resulted in a change in total hospitalisation costs per patient compared to the MC group.

Baseline characteristics were defined using the definitions of the Netherlands Heart Registration, [17], which align with the ESC/ACC and American Heart Association guidelines [18]. Extracted variables from electronic medical records included: age, gender, body mass index (BMI), paroxysmal, persistent, or unknown AF, left atrial volume index, previous ablation, CHA2DS2-VASc score, medical history (i.e., cerebrovascular accident, diabetes, hypertension, myocardial infarction, obstructive sleep apnoea, peripheral artery disease, and renal insufficiency), and type of oral coagulant.

The following procedural characteristics were extracted from electronic medical records: operator, CD procedural success, entry and exit times in the catheterisation laboratory, reaccess of the femoral vein, and any related access difficulties.

Data analysis was performed using SPSS version 29 (IBM, Chicago, IL, USA). Normally distributed data were presented as mean ± SD, non-normally distributed data as median (IQR).

The MC group was compared with the CD group by using Fisher’s exact/Chi-square tests for categorical variables, unpaired T‑test for normally distributed numeric data, and Mann-Whitney U test for non-normally distributed data. Data distribution was assessed with histograms and Q‑Q plots.

To account for potential bias related to hospitalisation duration, the interval between hospital admission and procedure start was compared between groups. Additionally, a subgroup analysis was performed to compare procedure time between groups separately for first-time and redo ablations.

Regression analyses were performed to assess associations between baseline characteristics and primary and secondary outcomes. Additionally, the association between oral coagulant type and operator was examined. Linearity was assessed using scatter plots. The Hosmer-Lemeshow test was used to assess the logistic regression model fit. A p-value p < 0.05 was considered statistically significant.

Eight items of the nursing questionnaire were included in the Cronbach’s alpha; other measured distinct constructs or were unsuitable due to a categorical or open-ended format. Therefore, the mean inter-item correlation was also evaluated.

Mean costs in the MC- and CD group were compared, and average costs savings per patient were noted.

The study followed the Declaration of Helsinki, was approved by the local ethics committee, and participation was voluntary.