Abstract
Objective:
To characterize the relationship between office-based laboratory utilization and Medicare payments for peripheral vascular interventions.
Methods:
Using Centers for Medicare and Medicaid Services Provider Utilization and Payment Data Public Use Files from 2014 to 2017, we identified providers who performed percutaneous transluminal angioplasty (PTA), stent placement (stent), and atherectomy. Procedures were aggregated at the provider and hospital referral region (HRR) level.
Results:
Between 2014 and 2017, 2,641 providers performed 308,247 procedures. Mean payment for OBL stent in 2017 was $4383.39, while mean payment for OBL atherectomy was $13079.63. Change in mean payment varied significantly, from a decrease of $16.97 in HRR 146 to an increase of $43.77 per beneficiary over the study period in HRR 11. Change in the rate of PVI also varied substantially, and moderately correlated with change in payment across HRRs (R2=0.40, p<0.001). The majority of HRRs experienced an increase in rate of PVI within OBLs, which strongly correlated with changes in payments (R2=0.85, p<0.001). Furthermore, 85% of the variance in change in payment was explained by increases in OBL atherectomy (p<0.001).
Conclusion:
A rapid shift into the office setting for PVIs occurred within some HRRs which was highly geographically variable and was strongly correlated with payments. Policymakers should revisit the current payment structure for OBL utilization, and, in particular atherectomy, to better align the policy with the intended goals.
Table of Contents Summary
This study provides evidence for the financial motivation behind increasing rates of office-based laboratory atherectomy in the US, a finding which is in contrast to the relationship found for stenting and additionally calls into question the current payment structure for atherectomy.
Introduction
Significant advances have been made in the treatment of peripheral arterial disease (PAD), with the majority of patients being treated endovascularly for the last decade.1,2 PAD is common and costly, accounting for over $25 billion yearly.3 Increases in the cost of PAD treatment resulted in modifications of physician reimbursement for outpatient and office-based laboratory (OBL) peripheral vascular intervention (PVI) by the Centers for Medicare and Medicaid Services (CMS) in an effort to shift care to the outpatient setting.4 The increased reimbursement for office-based procedures was intended to allow for practices to cover overhead expenses that are meant to be covered by the facility fee for facility based procedures. The relative changes in reimbursement between inpatient and outpatient procedures were intended to increase efficiency and decrease overall costs, however prior studies have shown that no such decrease in overall costs occurred.5
Atherectomy is reimbursed at a higher rate than other modes of PVI to allow providers and facilities to recuperate the increased costs of purchasing and maintaining equipment. Resulting dramatic increases in atherectomy volume have been shown in prior years, both in terms of absolute numbers as well as relative to other modes of PVI, despite a lack of data suggesting clinical superiority.6,7 There has been a substantial increase in the proportion of PAD treatment performed in the outpatient setting.8,9 A relative increase in atherectomy volume, particularly in OBLs, may explain the lack of significant cost savings by CMS following the institution of these reimbursements, however the degree of variation in practice pattern across different healthcare markets and the attributes of these markets have not been thoroughly explored.
In this study, we sought to determine the degree of variation that exists in PVI practice patterns across the US as well as the variation in the utilization of OBLs. Furthermore, we were interested in understanding the change in practice over the study period. We hypothesized that a significant correlation exists between increased utilization of OBLs and payment for PAD treatment across geographic locations.
Methods
Data Source and Population
In this study, we utilized CMS Provider Utilization and Payment Data Public Use Files from 2014 to 2017. This database contains information on procedures and services delivered to Medicare beneficiaries by physicians and other healthcare professionals. Specifically, information is available on utilization, CMS payment, and charges organized by National Provider Identifier (NPI), Healthcare Common Procedure Coding System (HCPCS) code, and location of service. This is based on CMS administrative claims data for beneficiaries enrolled in the fee-for-service program available from the CMS Chronic Condition Data Warehouse. Healthcare providers who performed PVI between January 1, 2014 to December 31, 2017 were identified by their NPI. This study was deemed exempt by the Institutional Review Board at the University of Michigan due to use of public-use data files and deidentified data.
Study Variables
PVI of the lower extremity was categorized into 4 distinct groups: A. lower extremity diagnostic angiography (DA) without revascularization using HSPCS codes 75716 and 75710, B. lower extremity angioplasty (percutaneous transluminal angioplasty (PTA)) without stenting or atherectomy using HSPCS codes 37220, 37222, 37224, 37225, and 37228, C. transluminal stent implantation with or without angioplasty (stent) using HSPCS codes 37221, 37223, 37226, 37230, and 37234, D. atherectomy with or without PTA or stent implantation using the HSPCS codes 37227, 37229, 37231 and 37235. The number of PVI services, clinical setting of service, number of Medicare beneficiaries receiving services per procedure, average submitted charge per procedure, and average Medicare payment per procedure were collected for each individual provider’s NPI per year. Clinical setting was defined as the location of care that was submitted on claims and identified as a facility (facility-based) or non-facility (office-based laboratory). Information is available in the dataset regarding city, state, country and zip code for each provider observation. The dataset only contains information for Medicare beneficiaries with Part B Fee-For-Service coverage, and therefore does not represent a complete picture of a physician’s practice. Provider submissions for reimbursement for non-facility-based interventions are essentially limited to office-based laboratories, whereas facility-based interventions includes those performed in hospitals as well as free standing ambulatory surgical centers or other settings where the physician is only requesting reimbursement for the physician fee and a separate request for the facility payment fee is made by the facility and therefore not captured in this dataset.
Statistical Analysis
Provider annual rates of lower extremity PVI for each type of PVI, clinical setting, and number of Medicare beneficiaries were calculated. Zip codes were crosswalked to Dartmouth Atlas hospital referral regions (HRRs) using publicly available crosswalk files.10 A description of the methods for generation of HRRs is available on the Dartmouth Atlas website, but in short, these regions represent a grouping of geographically contiguous zip codes that contain at least one tertiary care center and within which a minimum of 65% of residents receive their medical care.11 Descriptive statistics for each state and HRR as a mean for the providers who submitted claims within each state and HRR for a given year were generated and comparisons were made using chi2 for categorical variables and student’s t-tests for continuous measures. Correlation was assessed using scatter plots and ordinary least squares linear regression. A p-value of less than 0.05 was considered statistically significant for all analyses. All analyses and map figures were performed/generated using STATA version 16 with the maptile package (Statacorp, College Station, TX).
Exclusion of Outliers
After assessment of the scatterplots for the comparison of change in payment per beneficiary and other measures of change across the study period, there was one HRR that stood out as a significant outlier on a variety of measures. Upon further investigation, this HRR was identified as HRR 129. Upon review of the zip codes within this HRR, it became clear that 34474 is the zip code for the Ocala, FL area in which a physician was prosecuted for Medicare fraud involving peripheral vascular intervention billing practices and was excluded from Medicare billing as part of the lawsuit settlement in June of 2016.12 We therefore excluded this HRR from our analyses.
Results
A total of 2641 providers performing a total of 308,247 procedures on a total of 302,206 patients were included. Total Medicare payments for PVI during 2017 were $271,720,506, increasing by 54% over the study period from a total of $175,918,634 in 2014. Average payment for lower extremity intervention during 2017 was $829.82 for outpatient angioplasty, $939.65 for outpatient stent placement, and $6460.12 for outpatient atherectomy. In the office-based laboratory setting, payment during 2017 was $2614.74 for angiogram, $3767.69 for stent placement, and $8084.43 for atherectomy. Trends in overall Medicare professional payments for PVI as well as trends in rates of procedural utilization across the study period are shown in Figure 1.
Figure 1. Trends in Procedural Rates and Payment Over the Study Period.
(A) Overall rates as well as those for each endovascular procedure per 100,000 Medicare beneficiaries for each year included in the study. (B) Total payment overall and for the different types of endovascular procedure for each year included in the study.
Geographic Variation in Overall Payment for PVI
Geographic variation at the HRR level in average physician payment for outpatient PVI per beneficiary is shown in Figure 2A. Significant variation was found, with the lowest HRRs performing no PVI and therefore spending $0 per beneficiary on PVI to as high as $332.07 per beneficiary in HRR 25. Median payment was $5.27 (IQR $1.86-$21.95) per beneficiary across all HRRs.
Figure 2. Geographic Variation in Mean Payment and Change in Mean Payment, by Hospital Referral Region.
Payment is calculated as the mean, in USD, for all qualifying procedures completed within that hospital referral region. (A) Mean payment for all outpatient peripheral vascular interventions within each HRR over the study period. (B) Change in mean payment for all outpatient peripheral vascular interventions within each HRR from 2014 to 2017. HRRs in grey represent regions with missing data. The HRR labeled H represents the region with the highest value for the graph, while the HRR labeled L represents the region with the lowest value for the graph.
Geographic Variation in Change in Payment for PVI
Given the varying prevalence of PAD throughout the US, we sought to understand variation in the change in rates of intervention and payments overtime, which should correct for this varying prevalence as PAD prevalence is unlikely to have changed substantially over the 4 years of the study period. We first evaluated change in payment for PVI over the study period. The geographic variation at the HRR level was substantial, varying from -$16.97 per beneficiary in HRR 146 to +$43.77 per beneficiary in HRR 11. This variation is shown graphically in Figure 2B.
Geographic Variation in Outpatient PVI Utilization
Significant variation in the rate of outpatient PVI utilization was also found across HRRs during the study period, with the lowest utilization in HRR 437 performing 17 per 100,000 beneficiaries compared to the highest utilization in HRR 25 performing 4965 per 100,000 beneficiaries. This is shown geographically in Figure 3A.
Figure 3. Geographic Variation in Rate and Change in Rate of Peripheral Vascular Interventions, by Hospital Referral Region.
Rate is calculated as the number of cases completed by providers within that HRR divided by the total number of Medicare beneficiaries within that region. (A) Rate of all included outpatient peripheral vascular interventions within each HRR over the study period. (B) Change in rate of outpatient peripheral vascular interventions within each HRR from 2014 to 2017. HRRs in grey represent regions with missing data. The HRR labeled H represents the region with the highest value for the graph, while the HRR labeled L represents the region with the lowest value for the graph.
Geographic Variation in Change in Outpatient PVI Utilization
To correlate change in cost with change in utilization, we sought to assess the change in rate of PVI per beneficiary over the study period within each HRR. Again, significant variation was found, with the largest decrease in utilization of PVI per beneficiary of −535 per 100,000 beneficiaries in HRR 186 and the largest increase of 644 per 100,000 beneficiaries in HRR 173. This is shown geographically in Figure 3B.
Correlation Between Change in Payment and Change in PVI Utilization
To investigate the correlation between changes in payment for outpatient PVI per beneficiary and changes in rates of PVI per beneficiary, we first compared the heat maps for overlap. As shown in Figure 4A and 4B, there appeared to be qualitatively moderate overlap between areas of significant change in payment for outpatient PVI per beneficiary and significant change in rate of PVI utilization per beneficiary. This data is shown graphically in Figure 4C to allow for quantitative assessment. Linear regression analysis revealed a moderate and statistically significant correlation with R2 of 0.40, p<0.001.
Figure 4. Correlation Between Change in Mean Payment and Change in Rate of Peripheral Vascular Interventions, by Hospital Referral Region.
Map of change in mean payment is duplicated here as in Figure 1B and map of change in rate of peripheral vascular interventions is duplicated here as in Figure 2B to allow for direct visual comparison. (A) Change in mean payment for all outpatient peripheral vascular interventions within each HRR from 2014 to 2017. (B) Change in rate of outpatient peripheral vascular interventions within each HRR from 2014 to 2017. (C) Scatter-plot with change in mean payment for outpatient peripheral vascular intervention on the y-axis and change in rate of outpatient peripheral vascular interventions on the x-axis. Each data point represents a single HRR. Ordinary least squares fit linear regression line as well as 95% CI is shown along with R2. HRRs in grey represent regions with missing data. The HRR labeled H represents the region with the highest value for the graph, while the HRR labeled L represents the region with the lowest value for the graph.
Correlation Between Change in Payment and Change in OBL PVI Utilization
To test our hypothesis that increasing OBL utilization was correlated with increased payment geographically, we next calculated change in rate of PVI completed in OBL for each HRR and created a heat map for comparison to the map for change in PVI payment per beneficiary as shown in Figure 5A and 5B. On qualitative assessment, there is direct correlation between rate of OBL utilization and change in payment for PVI per beneficiary. HRR 383 had the largest decrease in OBL utilization with a change of -249 per 100,000 Medicare beneficiaries, compared to HRR 173 which had the largest increase in OBL utilization with a change of +707 per 100,000 beneficiaries. Linear regression analysis found a strong and statistically significant correlation with R2 of 0.85 and p<0.001. A scatterplot showing change in payment for PVI as a function of change in rate of OBL utilization is shown in Figure 5C.
Figure 5. Correlation Between Change in Mean Payment and Change in Rate of Office-Based Laboratory Peripheral Vascular Interventions, by Hospital Referral Region.
Map of change in mean payment is duplicated here as in Figure 1B to allow for direct visual comparison. (A) Change in mean payment for all outpatient peripheral vascular interventions within each HRR from 2014 to 2017. (B) Change in rate of peripheral vascular interventions performed within office-based laboratories within each HRR from 2014 to 2017. (C) Scatter-plot with change in mean payment for outpatient peripheral vascular intervention on the y-axis and change in rate of office-based laboratory peripheral vascular interventions on the x-axis. Each data point represents a single HRR. Ordinary least squares fit linear regression line as well as 95% CI is shown along with R2. HRRs in grey represent regions with missing data. The HRR labeled H represents the region with the highest value for the graph, while the HRR labeled L represents the region with the lowest value for the graph.
Correlation Between Change in Payment and Change in OBL Atherectomy Utilization
Atherectomy reimbursement was found to be substantially higher than stent reimbursement in both the OBL and non-OBL settings. To investigate the association of change in rate of OBL atherectomy to change in payment for PVI per beneficiary, we calculated these values for each HRR and plotted them each separately as shown in Figure 6A and 6B. As with our previous qualitative assessment of overall OBL utilization, there was a clear and direct correlation between change in rate of OBL atherectomy and change in payment for outpatient PVI. This was evaluated statistically with linear regression analysis which revealed an R2 of 0.84 and p<0.001. HRR 383 again had the largest decrease in OBL atherectomy rate with a change of -161 per 100,000 Medicare beneficiaries, compared to HRR 173 which again had the largest increase in OBL atherectomy rate with a change of +594 per 100,000 beneficiaries. A scatterplot showing change in payment for outpatient PVI as a function of change in rate of OBL atherectomy is shown in Figure 6C. To investigate whether office-based lab utilization of stenting alone showed a similar relationship, we calculated the change in rate of stenting in the OBL setting per beneficiary and compared this to the calculated change in payment per beneficiary which revealed a direct correlation with an R2 of 0.36 and p<0.001, shown in Supplementary Figure 1.
Figure 6. Correlation Between Change in Mean Payment and Change in Rate of Office-Based Laboratory Atherectomy, by Hospital Referral Region.
Map of change in mean payment is duplicated here as in Figure 1B to allow for direct visual comparison. (A) Change in mean payment for all outpatient peripheral vascular interventions within each HRR from 2014 to 2017. (B) Change in rate of atherectomy performed within office-based laboratories within each HRR from 2014 to 2017. (C) Scatter-plot with change in mean payment for outpatient peripheral vascular intervention on the y-axis and change in rate of office-based laboratory atherectomy on the x-axis. Each data point represents a single HRR. Ordinary least squares fit linear regression line as well as 95% CI is shown along with R2. HRRs in grey represent regions with missing data. The HRR labeled H represents the region with the highest value for the graph, while the HRR labeled L represents the region with the lowest value for the graph.
Discussion
This study used the CMS Provider Utilization Data and Payment Public Use File to characterize the relationship between OBL utilization and Medicare payments for outpatient PVI across geographic regions. We found significant geographic variation in the payment for PVI across HRRs throughout the US and furthermore found significant variation in change in payment for PVI over the four years of this study. We found that there was a statistically significant correlation between change in overall rate of PVI and change in payment over the study period. This correlation was much weaker than the correlation we found between change in OBL utilization and change in payment or change in OBL atherectomy and change in payment. This suggests that physicians who practice in OBLs are much more likely to use atherectomy, and as such receive higher reimbursement. The high correlation between change in rates of OBL atherectomy utilization and change in payment further suggests that the vast majority of the variance seen in changes in payment for PVI geographically can be explained by changes in OBL atherectomy use rather than changes in other PVI procedure modalities.
Our analysis suggests that 85% of the variation in change in professional payment for outpatient PVI across HRRs is explained by variation of the change in rate of OBL atherectomy across HRRs. This correlation accompanies a significant increase in overall spending by Medicare for professional fees for outpatient PVI over the study period. A comparison of stenting rates in the OBL setting to change in professional fee payment per beneficiary, to act as a control for the atherectomy comparison, revealed a significantly lower correlation coefficient with a significant number of providers actually decreasing their utilization of OBL stenting over the study period. Moreover, atherectomy has not been shown to be superior to stenting, a much cheaper alternative, suggesting that decision making regarding PVI technique and specifically the decision to rapidly transition to OBL atherectomy may be, at least partially, financially motivated.
CMS payment policy changes in 2008 came with a significant transition in the treatment of PAD from the inpatient to the outpatient setting, as was the intended purpose of the policy. However, a disproportionate increase in reimbursement for atherectomy and especially office-based atherectomy has accompanied a massive increase in the utilization of this procedure for the treatment of PAD without evidence of clinical benefit. While the exact procedure valuation decision making around PVI is not known, this effect likely represents the unintended consequences of payment policy changes that have not been thoroughly examined prior to implementation. Given the lack of evidence for clinical superiority of atherectomy over stenting, the decisions and policies around reimbursement for peripheral vascular interventions may represent an opportunity to utilize value-based payment models. Reimbursement tied to value may incentivize providers to move away from atherectomy which is significantly more expensive in the current fee-for-service model of payment. However, overutilization of atherectomy is likely complex and correcting it will require a multifaceted approach which may vary across local markets due to differences in patient populations, provider characteristics, provider supply, current utilization patterns, and market characteristics. A partnership between CMS and private payers may leverage the nationwide influence that CMS has over payment policy, especially for PVI where they are the predominant payer nationally, but also the local knowledge that private payers may have about intricacies of particularly geographic regions and healthcare markets which might require alternative strategies.
This work expands on prior work in several ways. We used data regarding changes in rates of utilization over time in order to control for the variation in PAD prevalence across geographic regions. We also aggregated utilization rates at the HRR level in order to examine an association that previously had only been made broadly at the national level immediately after the policy implementation. This study’s utilization of the sequential assessment of correlation between procedures (total PVI rate, OBL PVI rate, and finally OBL atherectomy rate) allows us to determine, with a high degree of specificity that changes in the rates of office-based atherectomy are the primary driver for change in payment for outpatient PVI and show that changes in the overall rate of PVI explain much less of the variation.
These results should be interpreted within the context of several limitations. This data source contains information regarding payment and provider characteristics for claims for outpatient PVI to Medicare for reimbursement, but it does not contain information regarding patient or clinical characteristics or outcomes that might explain the variation in clinical decision making or allow for evaluation of the value for care delivered during these episodes. In this way, we were unable to adjust for any clinical factors that may influence the decision making around procedure type. Adjustment, though, may inadvertently inject its own set of biases as there is no clinical consensus regarding criteria for patient selection for atherectomy vs stent. Recent meta-analyses have found no evidence of superiority of atherectomy over stenting in the treatment of PAD, but the motivations for decision making around procedure type beyond reimbursement cannot be evaluated within the context of this dataset.7 Additionally, these data include patients with Medicare coverage, rather than the broader population of patients undergoing PVI treatment. While PAD is a disease of elderly patients who would be eligible for Medicare, it is possible that patients included here are not representative of the broader population of patients undergoing outpatient PVIs. While there are differences across states with respect to regulations around opening and maintaining OBLs, this study focused on changes within HRRs across time rather than across HRRs. To our knowledge, there were no local coverage determinations or other Medicare policies which would differentially affect Medicare reimbursement geographically during the study period. However, it remains possible that differential regulations were implemented within particular HRRs during the study period that may affect our results. Another important limitation is the lack of patient or clinical information for the procedures available from this data source such as surgical indication. Prior studies have shown that increased OBL and ASC utilization has been associated with overuse of PVI in early claudicants.13 It remains possible that there was a significant change in the relative proportions of indications for atherectomy and other PVI over the period of the study, however we are not aware of any data to suggest this currently. Finally, the data contained in this data set do not provide any information regarding any facility payment, which is billed separate from the physician fee for procedures completed in the hospital, ambulatory surgical center, or other facility setting. There may be significant variation in the facility fee payment by CMS across geographic regions and, while providers likely are significantly less influenced by facility reimbursement than professional fee reimbursement, it is possible that this variation may drive utilization patterns to some degree.
Conclusions
These findings suggest that financial incentives have motivated the increased utilization of office-based atherectomy in the US. The reported motivation behind the implementation of CMS payment policy changes in 2008 was to incentivize providers to transition the care of PAD from predominantly facility-based to the outpatient setting, including OBLs, in an effort to capitalize on the decreased costs associated with OBLs and ultimately result in cost savings overall. These results reveal a transition to the OBL setting, as was intended, but also a transition to the use of more expensive, but clinically equivalent, treatment modalities and likely represent the unintended consequences of inadequately investigated payment policy at the national level.
Supplementary Material
Supplementary Figure 1. Correlation Between Change in Mean Payment and Change in Rate of Office-Based Laboratory Stenting, by Hospital Referral Region. Scatter-plot with change in mean payment for outpatient peripheral vascular intervention on the y-axis and change in rate of office-based laboratory stenting on the x-axis. Each data point represents a single HRR. Ordinary least squares fit linear regression line as well as 95% CI is shown along with R2.
ARTICLE HIGHLIGHTS.
Type of Research:
Retrospective Analysis of Medicare Provider Utilization and Payment Data
Key Findings:
There is a strong and direct correlation geographically between increasing Medicare payment for peripheral vascular interventions and utilization rates for office-based atherectomy in the US, a finding that was not present for other types of peripheral vascular intervention.
Take home Message:
Payment policy changes by CMS to incentivize the transition from inpatient to outpatient treatment for patients with peripheral arterial disease was inadvertently associated with increased utilization of atherectomy, particularly in the office-based laboratory setting, which may explain the lack of cost savings by CMS with implementation of these policies.
Abbreviations:
- PAD
Peripheral Arterial Disease
- CMS
Centers for Medicare and Medicaid Services
- PVI
Peripheral Vascular Intervention
- OBL
Office-Based Laboratory
- NPI
National Provider Index
- HCPCS
Healthcare Common Procedure Coding System
- PTA
Percutaneous Transluminal Angioplasty
- HRR
Hospital Referral Region
Footnotes
Presented at the Academic Surgical Congress, February 5th, 2020 in Orlando, FL, USA.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- 1.Goodney PP, Beck AW, Nagle J, Welch HG, Zwolak RM. National trends in lower extremity bypass surgery, endovascular interventions, and major amputations. J Vasc Surg [Internet] 2009July1[cited2019 Dec 12];50(1):54–60. Available from: https://www.sciencedirect.com/science/article/pii/S0741521409000548 [DOI] [PubMed] [Google Scholar]
- 2.Rowe VL, Lee W, Weaver FA, Etzioni D. Patterns of treatment for peripheral arterial disease in the United States: 1996–2005. J Vasc Surg [Internet] 2009April1 [cited 2019 Dec 12];49(4):910–7. Available from: https://www.sciencedirect.com/science/article/pii/S0741521408019873 [DOI] [PubMed] [Google Scholar]
- 3.Mahoney EM, Wang K, Cohen DJ, Hirsch AT, Alberts MJ, Eagle K, et al. One-Year Costs in Patients With a History of or at Risk for Atherothrombosis in the United States. Circ Cardiovasc Qual Outcomes [Internet] 2008September [cited 2019 Dec 12];1(1):38–45. Available from: https://www.ahajournals.org/doi/10.1161/CIRCOUTCOMES.108.775247 [DOI] [PubMed] [Google Scholar]
- 4.Centers for Medicare & Medicaid Services (CMS), HHS. Medicare program; prospective payment system for long-term care hospitals RY 2008: annual payment rate updates, and policy changes; and hospital direct and indirect graduate medical education policy changes. Final rule. Fed Regist [Internet] 2007May11 [cited 2019 Dec 12];72(91):26869–7029. Available from: http://www.ncbi.nlm.nih.gov/pubmed/17520779 [PubMed] [Google Scholar]
- 5.Levy B, Paulozzi L, Mack KA, Jones CM. Trends in Opioid Analgesic-Prescribing Rates by Specialty, U.S., 2007–2012. Am J Prev Med 2015; [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Mukherjee D, Hashemi H, Contos B. The disproportionate growth of office-based atherectomy. J Vasc Surg [Internet] 2017February1 [cited 2019 Dec 17];65(2):495–500. Available from: https://www.sciencedirect.com/science/article/pii/S0741521416312526 [DOI] [PubMed] [Google Scholar]
- 7.Ambler GK, Radwan R, Hayes PD, Twine CP. Atherectomy for peripheral arterial disease. Cochrane Database Syst Rev [Internet] 2014March17 [cited 2020 Feb 7];2014(3). Available from: http://doi.wiley.com/10.1002/14651858.CD006680.pub2 [DOI] [PubMed] [Google Scholar]
- 8.Smith ME, Sutzko DC, Beck AW, Osborne NH. Provider Trends in Atherectomy Volume between Office-Based Laboratories and Traditional Facilities. Ann Vasc Surg [Internet] 2019July1 [cited 2019 Dec 12];58:83–90. Available from: https://www.sciencedirect.com/science/article/pii/S0890509619300172?via%3Dihub [DOI] [PubMed] [Google Scholar]
- 9.Jones WS, Mi X, Qualls LG, Vemulapalli S, Peterson ED, Patel MR, et al. Trends in Settings for Peripheral Vascular Intervention and the Effect of Changes in the Outpatient Prospective Payment System. J Am Coll Cardiol [Internet] 2015March10 [cited 2019 Dec 17];65(9):920–7. Available from: https://www.sciencedirect.com/science/article/pii/S0735109715001503 [DOI] [PubMed] [Google Scholar]
- 10.Dartmouth Atlas Supplemental Research Data [Internet] [cited 2019 Dec 17]. Available from: https://atlasdata.dartmouth.edu/static/supp_research_data
- 11.Research Methods - Dartmouth Atlas of Health Care [Internet] [cited2019 Dec 17]. Available from: http://archive.dartmouthatlas.org/tools/faq/researchmethods.aspx
- 12.Florida Cardiologist and His Practice Pay Millions and Agree to Three Years of Exclusion to Resolve Alleged False Billings for Unnecessary Procedures and Illegal Kickbacks [Internet]. Department of Justice 2016[cited2020 Feb 3]. p. Press Release Number 16–772. Available from: https://www.justice.gov/opa/pr/florida-cardiologist-and-his-practice-pay-millions-and-agree-three-years-exclusion-resolve
- 13.Hicks CW, Holscher CM, Wang P, Black JH, Abularrage CJ, Makary MA. Overuse of early peripheral vascular interventions for claudication. In: Journal of Vascular Surgery [Internet] Mosby Inc.; 2020. [cited 2020 Dec 28]. p. 121–130.e1. Available from: https://pubmed.ncbi.nlm.nih.gov/31208940/ [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplementary Figure 1. Correlation Between Change in Mean Payment and Change in Rate of Office-Based Laboratory Stenting, by Hospital Referral Region. Scatter-plot with change in mean payment for outpatient peripheral vascular intervention on the y-axis and change in rate of office-based laboratory stenting on the x-axis. Each data point represents a single HRR. Ordinary least squares fit linear regression line as well as 95% CI is shown along with R2.