Abstract
Background
The commercial use of percutaneous left atrial appendage occlusion with the Watchman device is increasing in the United States. The purpose of this study was to evaluate center-related variation in total hospital costs for Watchman device implantation and identify factors associated with high hospital costs at a national level.
Methods
All adults undergoing elective left atrial appendage occlusion with Watchman were identified in the 2016-2018 National Inpatient Database. Mixed models were used to evaluate the impact of center on total hospital costs, adjusting for patient and center characteristics and length of stay.
Results
A total of 30,175 patients underwent Watchman device implantation at a median cost of $24,500 and demonstrated significant variability across admissions (interdecile range, $13,900-37,000). Nearly 13% of the variability in patient-level costs was related to the center performing the procedure rather than patient factors. Higher-volume centers had lower total costs and demonstrated lesser total cost variation. Centers with low procedural volume, occurrence of procedural complications, congestive heart failure, and length of stay were independent predictors of a high-cost hospitalization. Though complications were associated with increased expenditure, they did not explain the observed cost variation related to the center.
Conclusions
A significant proportion of variation in total hospital cost was attributable to the center performing the procedure. Addressing variability of Watchman-related costs is necessary to achieve high-quality value-based care.
Key Highlights
- 1.
What is already known:
- •
Watchman device implantation is a cost-effective alternative for stroke prevention in atrial fibrillation patients unsuitable for long-term oral anticoagulation therapies.
- •
- 2.
What this study adds:
- •
There are significant differences in costs between hospitals for the Watchman device procedure, highlighting the need for standardizing hospital practices and optimizing costs.
- •
Variation in hospital costs for the Watchman device procedure is influenced by differences in hospital care pathways and resource utilization, with factors such as low procedural volume, procedural complications, congestive heart failure, and longer hospital stays contributing to increased costs.
- •
- 3.
How this study might affect research, practice, or policy:
- •
Our findings demonstrate considerable variability in patient costs for elective left atrial appendage occlusion attributable to the hospital and emphasize the need for further studies analyzing cost variation related to implantation of more contemporary left atrial appendage occlusion devices to develop targeted cost containment strategies.
- •
Introduction
Stroke among atrial fibrillation (AF) patients is a major cause of disability and substantial economic burden. Long-term oral anticoagulation (OAC) has been the mainstay of treatment for prevention of cardioembolic stroke in AF. Percutaneous left atrial appendage occlusion (LAAO) has become an attractive alternative to reduce the risk of stroke in patients with nonvalvular AF when OAC is not suitable or hazardous.
LAAO with the Watchman device (Boston Scientific, Marlborough, Massachusetts) was approved in March 2015 in the United States and is being increasingly utilized for mitigation of thromboembolic risk. While the initial cost of LAAO with the Watchman device is high, it has proven to be a cost-effective treatment strategy compared to vitamin K antagonists and non-vitamin K antagonist OAC therapies. , With the sustained rise in US health care expenditures and increasing emphasis on value-based health care delivery, examination of costs associated with Watchman device implantation is particularly relevant.
To date, variation in Watchman procedure-related expenditures has not been studied. Thus, the purpose of this study was to evaluate center-related variation in total hospital costs for Watchman device implantation and identify factors associated with high hospital costs at a national level.
Methods
Data Source
We performed a 3-year population-based retrospective cross-sectional analysis using national (United States) data from January 2016 to December 2018. The National Inpatient Sample (NIS) database is the largest publicly available all-payer inpatient care database from the United States. It is developed as a part of the Healthcare Cost and Utilization Project (HCUP) and is sponsored by the Agency for Healthcare Research and Quality, available at https://www.hcup-us.ahrq.gov/overview.jsp . The NIS includes data from all nonfederal, short-term, general, and other specialty hospitals in the United States (excluding rehabilitation and long-term acute care hospitals) in the form of deidentified patient information containing demographics, discharge diagnoses, comorbidities, procedures, outcomes, and hospitalization costs. All states that participate in HCUP provide data to the NIS, covering >95% of the US population. The database was designed to include data from a 20% sample of discharges from all participating hospitals. This design of the NIS reduces the margin of error for estimates and delivers more stable and precise estimations. The study was exempt from an institutional review board approval because HCUP-NIS is a publicly available database containing only deidentified patient information.
Study Population
All adults (age ≥18 years) who underwent elective LAAO with Watchman device were identified using the International Classification of Diseases-10th Revision procedure code 02L73DK. Patients with missing data on age, sex, hospitalization costs, and in-hospital mortality were excluded. Furthermore, to reduce the possibility of data duplication, patients with an indicator for transfer to another acute-care facility were excluded.
Variable Definitions
Baseline patient characteristics including age, sex, race, income level, and payer status were defined in accordance with the NIS data dictionary. The previously validated Charlson comorbidity index (CCI) was used to quantify the burden of chronic conditions. In-hospital major adverse events (MAEs) were defined as the composite of mortality, stroke (ischemic or hemorrhagic) or transient ischemic attack, bleeding or transfusion, vascular complications, myocardial infarction, systemic embolization, and pericardial effusion or tamponade requiring pericardiocentesis or surgery. The International Classification of Diseases-10th Revision codes used to define these variables are listed in Supplemental Table 1 . Annual hospital volume was calculated as the total number of elective Watchman device implantation performed at each center. Hospitals were subsequently classified into low volume: ≤15 procedures/year (LVH), medium volume: 16-35 procedures/year (MVH), and high volume: ≥36 procedures/year (HVH) based on their annual case load. Hospitalization costs were generated by application of a hospital-specific cost-to-charge ratio and inflation adjusted to 2018. Total hospitalization costs represent the expenses incurred in the production of hospital services, such as wages, supplies, and utility. However, physician professional fees are not captured by the NIS database. Admission was designated as a high-cost hospitalization if total unadjusted hospitalization cost was in the highest decile.
Outcome
The primary outcome was total hospitalization cost at patient level and its variation related to center-level differences. We also analyzed the variation in MAE attributable to center-level differences due to high correlation between the incidence of complications and hospitalization costs at the patient level. Secondarily, we assessed patient characteristics and predictors of high-cost hospitalization for LAAO with Watchman device.
Statistical Analysis
National estimates were calculated by applying discharge weights. Categorical variables are reported as proportions and compared using Pearson’s chi-squared test. Continuous variables are reported as means with SD or median with interquartile range (IQR), when appropriate. Means and medians were compared using independent samples t-test and Mann-Whitney U test, respectively. Median costs from 2016 to 2018 and between LVH, MVH, and HVH were compared using nonparametric, independent samples Kruskal-Wallis test. A multivariate regression model with high-cost hospitalization status as dependent variable was developed to examine predictive factors. To evaluate the effect of individual center on total hospital costs, a 2-level log-gamma generalized mixed effects model was used with centers as a random effect because of the skewed distribution of cost data ( Supplementary figures 1-5 ). The proportion of total cost variation explained by the random center effect was calculated. SPSS Statistics 25.0 (IBM Corp., Armonk, New York) and R statistical software (R Core Team 2020) were used to perform the statistical analysis. All p values were 2-sided with a significance threshold of <0.05.
Results
A total of 30,175 patients met the study criteria and underwent elective admission for Watchman device implantation at an average of 290 hospitals per year across the United States. The mean age was 76 years, and women constituted 41.7% of the cohort. Less than 20% of the patients were in the highest income quartile, and Medicare was the primary insurer for most patients (89%). Congestive heart failure was the most common comorbidity, and median CCI score was 1 [1-3]. The vast majority of patients (62.5%) underwent Watchman device implantation at a high-volume hospital. A MAE occurred in 4.6% of the study cohort with bleeding/transfusion (2.9%) and vascular complication (2.5%) being the most common events. In-hospital mortality was 0.1% ( Table 1 , Figure 1 ). The rates of MAEs were lower in HVH compared to LVH (4.3 vs. 5.1%, p = 0.016).
| Variable | Count | Summary statistic |
|---|---|---|
| Age (y, mean ± SD) | – | 76.02 ± 7.97 |
| Female (%) | 12,585 | 41.7 |
| White (%) | 25,375 | 84.1 |
| Charlson comorbidity index (median, IQR) | – | 1 [1-3] |
| Comorbidities (%) | ||
| Congestive heart failure | 11,595 | 38.4 |
| Coronary artery disease | 3755 | 12.4 |
| Peripheral vascular disease | 4930 | 16.3 |
| Cerebrovascular disease | 2275 | 7.5 |
| Chronic obstructive pulmonary disease | 6605 | 21.9 |
| Chronic kidney disease | 7220 | 23.9 |
| Moderate-severe liver disease | 155 | 0.5 |
| Diabetes | 10,455 | 34.6 |
| Income quartile | ||
| 76th-100th | 7585 | 19.8 |
| 51st-75th | 8390 | 25.8 |
| 26th-50th | 7780 | 27.8 |
| 1st-25th | 5980 | 25.1 |
| Payer status | ||
| Medicare | 26,850 | 89.0 |
| Medicaid | 340 | 1.1 |
| Private | 2370 | 7.9 |
| Other | 545 | 1.8 |
| Hospital region | ||
| New England | 820 | 2.7 |
| Middle Atlantic | 3720 | 12.3 |
| East North Central | 4255 | 14.1 |
| West North Central | 2475 | 8.2 |
| South Atlantic | 6345 | 21 |
| East South Central | 1725 | 5.7 |
| West South Central | 3660 | 12.1 |
| Mountain | 3025 | 10 |
| Pacific | 4150 | 13.8 |
| Hospital volume | ||
| Low volume (1-15 procedures/year) | 3360 | 11.1 |
| Medium volume (16-35 procedures/year) | 7955 | 26.4 |
| High volume (>35 procedures/year) | 18,860 | 62.5 |
| Major adverse event (%) | 1385 | 4.6 |
| In-hospital death | 40 | 0.1 |
| Acute MI | 20 | 0.1 |
| Pericardial effusion/tamponade requiring pericardiocentesis or surgery | 285 | 0.9 |
| Bleeding or transfusion | 885 | 2.9 |
| Stroke (ischemic/hemorrhagic) orTIA | 175 | 0.6 |
| Systemic embolization | 30 | 0.1 |
| Vascular complication | 755 | 2.5 |
| Median length of stay (d) | 1 | |
| Median cost (interdecile range) | $24,500 ($13,900-37,000) |
On a national level, the median unadjusted patient-level hospitalization cost for Watchman device implantation was $24,500 and demonstrated significant variability across admissions in our study cohort (interdecile range, $13,900-37,000). The median hospitalization costs decreased slightly over the study period from $24,600 [IQR, $18,900-30,900] in 2016 to $24,400 [IQR, $18,600-29,800] in 2018 ( p < 0.001). As expected, patient-level costs were significantly ( p < 0.001) greater for patients experiencing a MAE: $28,700 (IQR, $21,700-37,100) compared to those who did not experience a MAE: $24,400 (IQR, $18,600-30,100). Median hospitalization costs were significantly lower in HVH: $24,000 (IQR, $18,700-29,200), compared to LVH: $25,900 (IQR, $20,100-33,400) in our study sample ( p < 0.001). Additionally, there was significant variation ( p < 0.001) in median hospitalization costs based on primary payer: Medicare: $24,600 (IQR, $19,000-30,400), Medicaid: $24,900 (IQR, $19,800-30,900), private pay: $24,900 (IQR, $18,000-30,400), and other pay: $17,100 (IQR, $8800-27,900).
Analysis of random intercept from the mixed model revealed 13.3% (95% CI: 12.1%-14.7%) of total cost variation for Watchman device implantation was due to the center-level differences. Among patients who experienced a MAE, 11.5% (95% CI: 9.6%-13.8%) of interhospital variation in total costs was attributable to center level differences, compared to 13.6% (95% CI: 12.3%-15%) in those who did not experience a MAE. On examining the relationship between annual hospital volume for Watchman device implantation and cost variation, we observed a decline in the proportion of cost variation attributable to the center with 14.7% in LVH (95% CI: 12.6%-17.2%), 14% (95% CI: 11.9%-16.3%) in MVH, compared to 11.4% (95% CI: 9.7%-13.5%) in HVH. For patients primarily insured by Medicare, 13% (95% CI: 11.8%-14.3%) of interhospital variation of total cost was attributable to center-level differences, while this variation was 22.4% (95% CI: 19.6%-26.5%) among non-Medicare (Medicaid, private pay, other) patients ( Figure 2 , Table 2 ). Notably, White race (coef: −0.034, SE: 0.008; p < 0.001) and lower CCI score: 3-4 (coef: – 0.038, standard error: 0.015; p = 0.014) were statistically associated with lower costs.
| Variable | Coefficient estimate | Standard error | P value |
|---|---|---|---|
| Age | −0.001 | 0.0 | 0.109 |
| Female | 9.67 | 0.035 | <0.001 |
| White | −0.034 | 0.008 | <0.001 |
| Income quartile | |||
| 0-25th | −0.007 | 0.008 | 0.342 |
| 26-50th | 0.022 | 0.007 | 0.002 |
| 51-75th | 0.017 | 0.007 | 0.011 |
| 76-100th | Reference | – | – |
| Payer status | |||
| Medicare | 0.339 | 0.018 | <0.001 |
| Medicaid | 0.317 | 0.029 | <0.001 |
| Private | 0.225 | 0.02 | <0.001 |
| Other | Reference | – | – |
| Hospital volume | |||
| Low volume center | 0.042 | 0.02 | 0.04 |
| Medium volume center | 0.083 | 0.017 | <0.001 |
| High volume center | Reference | – | – |
| Major adverse event | 0.096 | 0.012 | <0.001 |
| Comorbidities | |||
| Coronary artery disease | 0.017 | 0.008 | 0.036 |
| Congestive heart failure | 0.052 | 0.006 | <0.001 |
| Peripheral vascular disease | 0.024 | 0.007 | 0.001 |
| Cerebrovascular disease | −0.003 | 0.01 | 0.78 |
| COPD | 0.008 | 0.007 | 0.22 |
| Diabetes | 0.014 | 0.007 | 0.038 |
| Chronic kidney disease | −0.012 | 0.007 | 0.092 |
| Moderate-severe liver disease | 0.109 | 0.034 | 0.001 |
| Charlson comorbidity index score | |||
| CCI ≥ 5 | −0.033 | 0.025 | 0.2 |
| CCI = 3-4 | −0.038 | 0.015 | 0.014 |
| CCI = 1-2 | −0.01 | 0.008 | 0.22 |
| CCI = 0 | Reference | – | – |
| Length of Stay | 0.66 | 0.002 | <0.001 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
