Use of the THERMOCOOL SMARTTOUCH catheter for ablation of atrial fibrillation: the relationship between hospital procedure volume, re-admissions, and economic outcomes

Abstract

Objective: The purpose of this study was to examine the relationship between hospital volume of prior THERMOCOOL SMARTTOUCH catheter use and health and economic outcomes among hospitalized patients with atrial fibrillation (AF) undergoing ablation using this device.

Materials and methods: Patients aged ≥18 years with a primary diagnosis of AF undergoing ablation treatment using the THERMOCOOL SMARTTOUCH catheter between January 2014 and June 2016 were identified from the Premier hospital database with the first date of such a procedure being defined as the index date. Hospital volume of prior THERMOCOOL SMARTTOUCH catheter use was determined during the 12-month pre-index period, and was classified into five groups: no volume (0), low volume (1–50), mid volume (51–100), high volume (101–150), and very high volume (≥151). Outcomes, including length of stay (LOS; for inpatient procedure only), hospital costs (total, hospital pharmacy, supply), and all-cause re-admission were evaluated. A generalized estimating equation (GEE) with exchangeable correlation structure was used to examine the impact of hospital volume on LOS, hospital costs, and re-admissions controlling for hospital clustering and other covariates.

Results: The study population included 640 hospitalized AF patients. The adjusted mean LOS was significantly shorter in very high-volume hospitals than hospitals with no volume (mean LOS 2.30 vs 4.33 days; p = .0377). As volume increased, the mean adjusted supply cost tended to decrease, although these changes emerged as non-significant. The 12-month all-cause re-admission was significantly lower among patients undergoing ablation in low (Odds ratio [OR] = 0.27; confidence interval [CI] = 0.08–0.85) and mid (OR = 0.12; CI = 0.02–0.61) volume hospitals compared to hospitals with no volume.

Limitations: Study results may not be generalizable to all US hospitals.

Conclusions: Among AF patients undergoing ablation, increased hospital volume of prior THERMOCOOL SMARTTOUCH catheter use was associated with shorter LOS and a lower likelihood of all-cause re-admission.

Keywords:

• Atrial fibrillation
• catheter ablation
• THERMOCOOL SMARTTOUCH catheter
• procedure volume
• costs
• LOS

Introduction

Atrial fibrillation (AF) is the most common form of arrhythmia¹. In the US, more than 3 million individuals have been diagnosed with AF, and this number is expected to increase to 7.5 million by 2050². Based on the duration of episodes, AF may be classified as paroxysmal (terminates <7 days within onset), early persistent (episodes occur for >7 days and <3 months), persistent (episodes occur for >7 days), long-standing persistent (episodes occur for >12 months), or permanent (episodes continue and attempts to restore sinus rhythm are ceased)³. AF is considered a significant risk factor for stroke⁴, thereby necessitating timely management. Drug therapy is typically the first-line treatment for individuals with AF, with drugs used for rhythm control or for rate control⁵.

Patients may become refractory or intolerant to pharmacotherapy treatments for AF. For such individuals, catheter ablation is considered an effective treatment strategy. According to the recent consensus guidelines for the management of patients with AF, catheter ablation use is recommended for patients who are refractory or intolerant to at least one class I or III anti-arrhythmic drug and have symptomatic paroxysmal AF (Class I, Level of Evidence A), symptomatic persistent AF (Class IIa, Level of Evidence B-NR), or symptomatic long-standing (>12 months) persistent AF (Class IIb, Level of Evidence C-LD). The guidelines suggest that catheter ablation may also be used as first-line therapy for patients with symptomatic paroxysmal AF or persistent AF³. Several studies, both non-randomized observational studies as well as randomized studies, have found pulmonary vein isolation with radiofrequency (RF) ablation to be significantly more effective than anti-arrhythmic drug treatment in reducing re-occurrence of AF and hospitalization in the 1-year follow-up period, and improve quality-of-life when used as first-line therapy in symptomatic AF^3,6.

In RF ablation, lesion formation remains the critical factor in the isolation of pulmonary veins. Until a few years ago, this process was based on electrical diminution and impedance drop during the application of the catheter^7,8. With the introduction of contact force (CF) catheters, which provide real-time information on catheter to tissue interaction, there has been a significant improvement in the creation of durable lesions⁸. The use of CF technology has been found to be associated with reduced pulmonary vein reconnection⁹, which is a critical marker of ablation success.

The role of volume in influencing outcomes for RF ablation, especially for next generation RF catheters with CF technology, has not been explored in a real-world setting. This study aims to understand the relationship between hospital volume of prior THERMOCOOL SMARTTOUCH catheter use and outcomes among patients with AF who underwent RF ablation using this device.

Methods

Data source

A retrospective analysis of the Premier hospital database was performed for the purpose of the study. The Premier hospital database contains complete clinical coding, hospital cost, and patient billing data from more than 600 hospitals geographically dispersed throughout the US.

Study sample

Patients who underwent RF ablation as a primary procedure for the treatment of AF using the THERMOCOOL SMARTTOUCH catheter during an inpatient or outpatient hospital admission between January 1, 2014 and June 30, 2016 were identified. The starting period of January 1, 2014 was chosen as it relates to the launch date of the THERMOCOOL SMARTTOUCH catheter, while the end period of June 2016 related to the availability of the most recent Premier data at the time of study analysis. The first observed hospital admission during this time period was defined as the index hospital admission. Eligible patients were required to be at least 18 years of age at the time of the index admission. Patients with a catheter ablation procedure performed during the 12-months prior to the index admission were excluded from the study. Only those patients who underwent an index procedure in hospitals that had consistently provided data (inpatient and outpatient) in the Premier database in the 12-month pre-index period were considered for inclusion. Additionally, inclusion was restricted to patients who underwent the index procedure in hospitals that had consistently provided data (inpatient and outpatient) in the Premier database between January 2014 and the index procedure date. Patients with a diagnosis or procedural code for either implantation of a pacemaker or implantable cardioverter defibrillator during the 12-months prior to or at the time of the index admission were also excluded from the study analysis¹⁰. Patients who underwent one or more of the following procedures: surgical ablation, valvular procedures, or left atrial appendage occlusion during the 12-months prior to or at the time of the index admission were excluded.

Patient and hospital characteristics

Patient characteristics including age, gender, race and marital status were evaluated at the index hospitalization. Patient clinical characteristics including the Charlson Comorbidity Index (CCI) score and select comorbid conditions, including congestive heart failure (CHF), peripheral vascular disease (PVD), myocardial infarction (MI), cerebrovascular disease, chronic pulmonary disease (COPD), diabetes, and hypertension, were also evaluated. Other characteristics at the index admissions, including payer type and year of procedure, were also evaluated. Hospital characteristics including teaching status, hospital region, hospital bed size, hospital costing method, hospital prior total ablation volume (excluding THERMOCOOL SMARTTOUCH^® Catheter), and the specialty of the physician performing the procedure were also determined. The hospital volume of prior THERMOCOOL SMARTTOUCH catheter usage was determined during the 12-month pre-index period, and was the primary independent variable of interest. Based on prior THERMOCOOL SMARTTOUCH catheter usage, hospitals were classified into five volume groups: no volume (0 prior procedures), low volume (1–50 prior procedures), mid volume (51–100 prior procedures), high volume (101–150 prior procedures), and very high volume (≥151 prior procedures).

Healthcare and economic outcomes measurements

Study outcome measurements included hospital length of stay (LOS), hospital costs (total, pharmacy, supply), and all-cause re-admission. LOS and hospital costs (total, pharmacy, supply) outcomes were evaluated based on the index hospital admission for ablation procedure among patients with AF. LOS was only evaluated for inpatient hospital admissions, while the total hospital cost, hospital pharmacy cost, and supply cost were evaluated based on both inpatient and outpatient hospital admissions data. All-cause inpatient re-admissions were measured in the 12-month period after the discharge from the index admission. Re-admissions were evaluated only among those patients who underwent the index ablation procedure between January 2014 and June 2015 to allow for 1 year of follow-up, and with hospital providers that had contributed inpatient data on a 12-month post-index continuous basis to the Premier hospital database.

Statistical analysis

Means and standard deviations were reported for continuous variables, and counts and percentages were reported for categorical variables. The relationship between the hospital volume of prior THERMOCOOL SMARTTOUCH catheter use and study outcomes including LOS and hospital costs (total, pharmacy, supply) were assessed using a Generalized Estimating Equations (GEE) model under an exchangeable correlation structure with log link and negative binomial distribution for LOS and gamma distribution function for cost. The GEE model accounted for potential hospital-level clustering and other covariates. Survey logistic regression analysis was used to assess the impact of hospital volume of prior THERMOCOOL SMARTTOUCH catheter use on all-cause re-admissions. Covariates in the GEE and logistic regression models included age, gender, race, marital status, CCI score, select comorbid conditions (CHF, PVD, MI, COPD, diabetes, hypertension), payer type, year of procedure, hospital teaching status, hospital region, hospital bed size, hospital costing method, hospital prior total ablation volume (excluding THERMOCOOL SMARTTOUCH catheter), and physician specialty. Additionally, sensitivity analyses were conducted wherein hospitalized AF patients with either a primary procedure code for ablation, primary procedure code for electrophysiology (EP) testing (with secondary ablation), and those with primary procedure cardiac mapping (with secondary ablation) were included in the analyses and compared to patients with a primary code for ablation, which had been used in the default analysis.

Results

Patient attrition

Patient attrition is shown in Table 1. The final study population consisted of 640 patients (at 37 hospitals) with a primary diagnosis of AF who underwent a primary ablation procedure using the THERMOCOOL SMARTTOUCH catheter during the index inpatient or outpatient hospital admission.

Table 1. Patient attrition.

Step	Criteria	Hospitals	Patients
		n	n
1	Patients ≥18 years with a “THERMOCOOL SMARTTOUCH” catheter device record on the date of the index ablation procedure (between January 2014–June 2016) for atrial fibrillation (AF)	51	1,391
2	Patients with the index procedure at hospitals that provided data (inpatient and outpatient) for at least 12-months during the pre-index period in the Premier database	45	1,165
3	Patients with the index procedure at hospitals that provided both inpatient and outpatient continuous enrollment record between January 2014 and the index ablation procedure year/month	43	1,134
4	Exclusion of patients with a diagnosis or procedural code for implantation of a pacemaker or implantable cardioverter defibrillator, surgical ablation, valvular procedure, and left atrial appendage occlusion during the 12-month pre-index period or at the time of the index procedure	43	1,060
5	Final study sample after excluding patients who had any other catheter device listed (besides “THERMOCOOL SMARTTOUCH”) during the index procedure	37	640

Patient and hospital characteristics

Patient and hospital characteristics are presented in Table 2. In terms of age and gender distribution, roughly 41% were 60–69 years of age, and 66% were male. Almost 82% of the sample were white. The most common payer type was Medicare/Medicaid (50%). Almost 70% of ablations occurred in an outpatient setting. Hospital volumes of prior THERMOCOOL SMARTTOUCH catheter use for AF catheter ablations are presented in Table 3.

Table 2. Patient and hospital characteristics.

Characteristic	Count	Percentage
Age category
18–49 years	66	10.32
50–59 years	143	22.34
60–69 years	262	40.94
≥ 70 years	169	26.41
Gender
Female	217	33.91
Male	423	66.09
Race
Non-white	118	18.44
White	522	81.56
Marital status
Married	385	60.16
Others	255	39.84
Payer type
Medicare/Medicaid	322	50.31
Commercial	263	41.09
Other	55	8.59
Procedure year/half
2014 half 1	57	8.91
2014 half 2	188	29.38
2015 half 1	226	35.31
2015 half 2	125	19.53
2016 half 1	44	6.88
Inpatient/outpatient hospitalization indicator
Inpatient	194	30.31
Outpatient	446	69.69
Charlson Comorbidity Index (CCI) group (score)
1 (Score 0)	312	48.75
2 (Score 1–2)	270	42.19
3 (Score ≥3)	58	9.06
Comorbid conditions
Myocardial infarction	38	5.94
Congestive heart failure	148	23.13
Peripheral vascular disease	34	5.31
Cerebrovascular disease	10	1.56
Chronic pulmonary disease	106	16.56
Diabetes	128	20
Hypertension	441	68.91
Teaching status
No	300	46.88
Yes	340	53.13
Hospital costing method type^a
Procedural	513	80.16
Ratio of cost to charge (RCC)	127	19.84
Hospital bed size
0–299 beds	120	18.75
300–499 beds	242	37.81
500+ beds	278	43.44
Physician specialty
Cardiovascular specialty	488	76.25
Internal medicine	57	8.91
Other	95	14.84
Hospital region
Northeast	150	23.44
South	399	62.34
West-Midwest	91	14.22
Prior ablation volume (excluding TCST)
1 (indicating ≤100 prior volume)	155	24.22
2 (indicating 101–200 prior volume)	209	32.66
3 (indicating 201–300 prior volume)	158	24.69
4 (indicating ≥300 prior volume)	118	18.44

TCST, THERMOCOOL SMARTTOUCH catheter.

^a Hospital costing method type is indicative of the costing methodology used by the hospital. When the hospital has a cost accounting system and uses Relative Value Units to estimate cost, it is classified as “Procedural”. Hospitals that do not have a cost accounting system or if they do have a cost accounting system but do not use Relative Value Units are classified as “Ratio of cost to charge (RCC)”.

Table 3. Hospital volume of prior THERMOCOOL SMARTTOUCH catheter use.

Volume of THERMOCOOL SMARTTOUCH use^a	Count	Percentage
No volume (0)	37	5.78
Low volume (1–50)	234	36.56
Mid volume (51–100)	174	27.19
High volume (101–150)	107	16.72
Very high volume (≥151)	88	13.75

^a Numbers listed in parentheses reflect the total number of ablation procedures performed by hospital using the THERMOCOOL SMARTTOUCH catheter among patients with atrial fibrillation in the 12-month pre-index period.

Study outcome measurements

Length of stay (LOS)

In the bivariate analysis, no significant differences (p = .9531) in LOS emerged by hospital volume of prior THERMOCOOL SMARTTOUCH catheter use (Table 4). However, in the adjusted GEE analysis, a significant difference in LOS emerged by hospital volume of prior THERMOCOOL SMARTTOUCH catheter use after controlling for study covariates. For inpatient admissions, as hospital volume of prior THERMOCOOL SMARTTOUCH catheter use increased, the mean adjusted LOS (days) decreased (Figure 1). The mean adjusted LOS was 4.33 days for hospitals with no prior volume, and 2.3 days for hospitals with very high prior volume (p = .0377) of ablations performed using the THERMOCOOL SMARTTOUCH catheter.

Figure 1. Regression adjusted index hospital admission length of stay (LOS) by hospital volume of prior THERMOCOOL SMARTTOUCH catheter use, n = 194. Length of stay (LOS) outcome assessed only for those patients who underwent ablation using the THERMOCOOL SMARTTOUCH catheter in an inpatient setting. Results based on GEE model under an exchangeable correlation structure with negative binomial distribution and log link function. *p = .0377.

Table 4. Bivariate analysis comparing mean length of stay (LOS) and cost (total, pharmacy, supply) by hospital volume of prior THERMOCOOL SMARTTOUCH catheter use.

Volume of THERMOCOOL SMARTTOUCH use^a	LOS^a^,^b (Mean, SD)	Total cost^c^,^d (Mean, SD)	Pharmacy cost^c^,^d (Mean, SD)	Supply cost^c^,^e (Mean, SD)
No volume (0)	3.17 (2.33)	$24,799.69 ($7,080.32)	$473.14 ($901.95)	$10,696.22 ($4,630.45)
Low volume (1–50)	4.01 (4.09)	$25,111.71 ($9,984.00)	$788.50 ($1,149.64)	$9997.05 ($4425.49)
Mid volume (51–100)	3.98 (4.74)	$25,743.44 ($8,873.98)	$644.21 ($849.37)	$10,110.20 ($3,215.66)
High volume (101–150)	3.34 (3.00)	$25,580.33 ($10,277.17)	$1,110.86 ($1,174.86)	$8,759.23 ($2,614.26)
Very high volume (≥151)	4.56 (7.05)	$19,039.31 ($11,697.33)	$940.30 ($1,468.91)	$8,804.77 ($3,371.13)

^a Length of stay (LOS) outcome assessed only for those patients who underwent ablation using THERMOCOOL SMARTTOUCH catheter in an inpatient setting (n = 194).

^b p = .9531.

^c Total cost, pharmacy cost, and supply cost assessed for the entire sample, i.e. patients who underwent ablation using THERMOCOOL SMARTTOUCH catheter in an inpatient or outpatient setting (n = 640).

^d p < .0001.

^e p = .0001.

Hospital costs (total, hospital pharmacy, supply)

The bivariate analysis revealed significant differences in the mean total hospital cost, mean pharmacy cost, and mean supply cost by hospital volume of prior THERMOCOOL SMARTTOUCH catheter use (Table 4). Mean total cost and mean supply cost significantly decreased with an increase in volume, while mean pharmacy cost increased with volume. After adjusting for key covariates, no significant difference was observed in the mean total hospital cost, pharmacy cost, and supply cost by hospital volume of prior THERMOCOOL SMARTTOUCH catheter use. In the adjusted analysis, as hospital volume increased, supply costs decreased, although the decrease was not significant. Although not significantly different, hospitals with high volume of prior THERMOCOOL SMARTTOUCH catheter use had a lower supply cost ($8,051 vs $9,939; p = .0523) as compared to hospitals with no volume (Table 5).

Table 5. Regression adjusted index hospital admissions cost (total, pharmacy, supply) by hospital volume of prior THERMOCOOL SMARTTOUCH catheter use, n = 640.

Hospital volume of prior THERMOCOOL SMARTTOUCH catheter use	Estimate^a	Adjusted mean	Standard error of mean	Lower mean	Upper mean	p-value
Total hospital cost
No volume	REF	$22,881	$1,461.70	$20,188	$25,933	–
Low volume	0.0378	$23,763	$1,177.39	$21,564	$26,186	.5035
Mid volume	0.0381	$23,769	$1,140.53	$21,635	$26,113	.5588
High volume	0.0443	$23,917	$1,628.43	$20,929	$27,332	.6168
Very high volume	–0.0409	$21,963	$1,279.47	$19,593	$24,620	.611
Hospital pharmacy cost
No volume	REF	$614.71	$105.12	$439.65	$859.48	–
Low volume	–0.0628	$577.27	$87.3461	$429.13	$776.56	.4984
Mid volume	–0.1809	$512.97	$67.8078	$395.89	$664.67	.2159
High volume	–0.0837	$565.34	$76.7921	$433.20	$737.79	.602
Very high volume	0.2166	$763.38	$126.59	$551.55	$1,056.56	.2645
Supply cost
No volume	REF	$9,939.55	$1,089.87	$8,017.38	$12,323	–
Low volume	–0.0887	$9,095.76	$697.89	$7,825.80	$10,572	.1899
Mid volume	–0.1078	$8,924.13	$655.95	$7,726.81	$10,307	.1229
High volume	–0.2107	$8,051.54	$645.74	$6,880.38	$9,422.07	.0523
Very high volume	–0.1202	$8,813.50	$793.29	$7,388.10	$10,514	.2261

^a A negative estimate value indicates a decrease in cost, while a positive estimate value indicates an increase in cost relative to the reference.

12-month all-cause readmissions

Bivariate analysis comparing the occurrence of all-cause re-admission in the 12-month post-index period by hospital volume of prior THERMOCOOL SMARTTOUCH catheter use did not reveal any significant difference (Chi-square, p = .1455). However, after adjusting for key patient and hospital characteristics, increasing hospital volume of prior THERMOCOOL SMARTTOUCH catheter use was associated with a lower all-cause re-admission risk for patients undergoing both inpatient and outpatient procedures. The 0–12-month all-cause re-admission risk was significantly lower among patients undergoing ablation in low (Odds ratio [OR] = 0.27; confidence interval [CI] = 0.08–0.85; p = .0265) and mid (OR = 0.12; CI = 0.02–0.61; p = .0107) volume hospitals compared to hospitals with no volume (Table 6). Re-admissions among patients undergoing ablation in hospitals with high (OR = 0.280; CI = 0.020–3.944) or very high volume (OR = 0.219; CI = 0.031–1.563) were not significantly different than hospitals with no volume.

Table 6. Likelihood of re-admissions during the 12-month post-index period based on hospital volume of prior THERMOCOOL SMARTTOUCH catheter use, n = 354^a.

Hospital volume of prior THERMOCOOL SMARTTOUCH catheter use	Odds ratio	Confidence interval	p-value
No volume	REF	–	–
Low volume	0.270	0.085–0.858	.0265
Mid volume	0.120	0.203–0.611	.0107
High volume	0.280	0.020–3.944	.3457
Very high volume	0.219	0.031–1.563	.1300

^aAll-cause readmission was studied only for those patients who underwent index ablation using THERMOCOOL^® ST Catheter in a hospital that had continuous 12-month post-index inpatient data available in the Premier database

Sensitivity analyses

For sensitivity analysis, we considered a broader sample of patients including those with primary ablation procedure, those with a primary procedure code of EP testing (with secondary ablation), and those with primary procedure code for cardiac mapping (with secondary ablation). In this extended sample of patients, hospital volume of prior THERMOCOOL SMARTTOUCH catheter use was observed to be significantly associated with pharmacy and supply cost. Higher volume was associated with lower pharmacy and supply costs (for procedures in either inpatient or outpatient setting). However, hospital volume of pre-index THERMOCOOL SMARTTOUCH catheter use was not significantly associated with LOS (among inpatient admissions; n = 213) and 12-month all-cause re-admission in this extended sample of patients. It must be noted though that the adjusted mean LOS decreased with increasing volume, from 3.67 days for hospitals with no volume to 2.46 days for hospitals with very high volume. Similarly, the likelihood (OR) of re-admission was <1 for hospitals with low volume (OR = 0.52; CI = 0.22–1.21), mid volume (OR = 0.45; CI = 0.14–1.38), high volume (OR = 0.61; CI = 0.10–3.54), and very high volume (OR = 0.76; CI = 0.21–2.72) as compared to hospitals with no volume in this extended sample.

Discussion

Several studies have found the use of CF-sensing technology for RF ablation, such as the THERMOCOOL SMARTTOUCH catheter, to be associated with significant reduction in AF recurrence as compared to the use of a conventional catheter group^8,9,11–15. In their prospective parallel cohort study of drug refractory patients with highly symptomatic paroxysmal AF, Andrade et al.¹¹ found lower recurrence of arrhythmia (88% vs 66%; p = .047) among patients undergoing ablation using a CF-guided catheter (Biosense Webster, Inc., Irvine, California [BWI], THERMOCOOL SMARTTOUCH catheter) as compared to patients who underwent ablation using a non-CF catheter (BWI, THERMOCOOL catheter). In a prospective non-randomized single-center interventional study, Marijon et al.¹² found lower AF recurrence (10.5% vs 35.9%, p = .04), lower RF ablation time (45.2 min vs 22 min, p = .01), lower fluoroscopy exposure time (20.1 min vs 26.7 min, p < .01), and lower total radiation dose (41.6 min vs 56.7 min, p = .02) associated with use of CF-guided catheters (BWI THERMOCOOL SMARTTOUCH catheter) vs non-CF catheters (BWI, THERMOCOOL EZ Steer Catheter) among patients with paroxysmal AF. This study by Marijon et al.¹² was the first non-randomized study to show the significant advantage associated with the use of CF catheter vs a non-CF catheter. In their meta-analysis to assess the impact of CF catheters on clinical and procedural outcomes for AF ablation, Shurrab et al.¹³ found lower recurrence rate and shorter procedure and fluoroscopic time to be associated with the use of CF catheters as compared to conventional catheters. These studies reflect the significant patient and procedural benefits associated with CF technology in RF ablation.

The results of this real-world study showed that the volume of the THERMOCOOL SMARTTOUCH catheter was associated with outcomes, with results indicating a shorter hospital LOS and a lower risk for all-cause 12-month re-admission among hospitals that had used the catheter as compared to hospitals with no volume. It must be noted, though, that the relationship between volume and 12-month re-admission was not linear, and that patients who underwent index ablation at hospitals with high or very high volume did not appear to have significantly lower re-admissions than patients who underwent index ablation at hospitals with no prior volume. Prior studies have also found the success rate of catheter ablation including proportion of patients free of anti-arrhythmic drugs to be higher in higher-volume compared with lower-volume centers^16,17. In their worldwide survey of centers performing catheter ablation procedures for AF, Cappato et al.¹⁶ (2005) found ablation outcomes, i.e. freedom from anti-arrhythmic drugs, to be better in centers with higher-volume of ablation procedure as compared to centers with lower-volumes.

In this study, LOS, which is considered to be a key patient outcome, decreased as the provider procedural volume of THERMOCOOL SMARTTOUCH catheter increased. A reduction in LOS is not only an indicator of improved patient outcomes, but may also indicate economic savings for providers and payers. With medical reimbursement increasingly moving toward bundled payments, shorter LOS may contribute to greater economic savings for providers.

Limitations

There are a few limitations to this study. While the Premier hospital database contains information from a large number of hospitals (∼600) across the US, it is possible that it may not include data records representative of the entire US population of hospitalized AF patients. Some patients may be re-admitted into non-Premier hospitals and, thus, such re-admission data may not be captured in the database. Results may be subject to biases from other confounding variables that were not available in the database. For example, surgeon volume and experience are core variables that could potentially influence outcomes. Due to the limitation of the dataset, we could not control for surgeon experience in our study. It is possible that high or very high-volume centers have physicians which have more experience performing ablation procedures as compared to no volume centers. although we controlled for comorbid conditions, the differential case-mix of patients between centers could also have potentially influenced our results. Additionally, billing and coding errors and missing data could potentially have occurred on database records which may affect the accuracy of study estimates. Lastly, as this is an observational analysis, no causal relationship between the primary independent variable of interest (i.e. hospital volume) and outcomes can be concluded.

Conclusions

The increased prior use of THERMOCOOL SMARTTOUCH catheter use was associated with improved outcomes, including a shorter LOS. Although re-admissions appeared to decrease with increasing volume, this relationship did not appear linear across all volume categories. Considering this relationship between volume and LOS, future studies (clinical trials, retrospective database, etc.) involving the THERMOCOOL SMARTTOUCH catheter should consider the inclusion of procedural volume as a covariate of interest.

Transparency

Declaration of funding

This study was sponsored by Johnson and Johnson.

Declaration of financial/other relationships

AB, IK, LG, RL, and RK are Johnson and Johnson employees. JME peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Previous presentations

Results from the study were presented at the International Society for Pharmacoeconomics and Outcomes Research 6th Latin America Conference, September 15–17, 2017, in Sao Paulo, Brazil.

Acknowledgments

Editorial support was provided by Brandy Menges, PhD, from Novosys Health, which received funding from Johnson and Johnson.